Skip to main content


<?xml version="1.0"?>

<!DOCTYPE TEI.2 SYSTEM "base.dtd">




<title>Questionnaire design and attitude measurement</title></titleStmt>

<publicationStmt><distributor>BASE and Oxford Text Archive</distributor>


<availability><p>The British Academic Spoken English (BASE) corpus was developed at the

Universities of Warwick and Reading, under the directorship of Hilary Nesi

(Centre for English Language Teacher Education, Warwick) and Paul Thompson

(Department of Applied Linguistics, Reading), with funding from BALEAP,

EURALEX, the British Academy and the Arts and Humanities Research Board. The

original recordings are held at the Universities of Warwick and Reading, and

at the Oxford Text Archive and may be consulted by bona fide researchers

upon written application to any of the holding bodies.

The BASE corpus is freely available to researchers who agree to the

following conditions:</p>

<p>1. The recordings and transcriptions should not be modified in any


<p>2. The recordings and transcriptions should be used for research purposes

only; they should not be reproduced in teaching materials</p>

<p>3. The recordings and transcriptions should not be reproduced in full for

a wider audience/readership, although researchers are free to quote short

passages of text (up to 200 running words from any given speech event)</p>

<p>4. The corpus developers should be informed of all presentations or

publications arising from analysis of the corpus</p><p>

Researchers should acknowledge their use of the corpus using the following

form of words:

The recordings and transcriptions used in this study come from the British

Academic Spoken English (BASE) corpus, which was developed at the

Universities of Warwick and Reading under the directorship of Hilary Nesi

(Warwick) and Paul Thompson (Reading). Corpus development was assisted by

funding from the Universities of Warwick and Reading, BALEAP, EURALEX, the

British Academy and the Arts and Humanities Research Board. </p></availability>




<recording dur="01:32:12" n="14413">


<respStmt><name>BASE team</name>



<langUsage><language id="en">English</language>

<language id="it">Italian</language>



<person id="nm0976" role="main speaker" n="n" sex="m"><p>nm0976, main speaker, non-student, male</p></person>

<person id="sm0977" role="participant" n="s" sex="m"><p>sm0977, participant, student, male</p></person>

<person id="sm0978" role="participant" n="s" sex="m"><p>sm0978, participant, student, male</p></person>

<person id="sm0979" role="participant" n="s" sex="m"><p>sm0979, participant, student, male</p></person>

<person id="sm0980" role="participant" n="s" sex="m"><p>sm0980, participant, student, male</p></person>

<person id="sm0981" role="participant" n="s" sex="m"><p>sm0981, participant, student, male</p></person>

<person id="sm0982" role="participant" n="s" sex="m"><p>sm0982, participant, student, male</p></person>

<person id="sm0983" role="participant" n="s" sex="m"><p>sm0983, participant, student, male</p></person>

<person id="sf0984" role="participant" n="s" sex="f"><p>sf0984, participant, student, female</p></person>

<person id="sm0985" role="participant" n="s" sex="m"><p>sm0985, participant, student, male</p></person>

<person id="su0986" role="participant" n="s" sex="u"><p>su0986, participant, student, unknown sex</p></person>

<person id="su0987" role="participant" n="s" sex="u"><p>su0987, participant, student, unknown sex</p></person>

<personGrp id="ss" role="audience" size="s"><p>ss, audience, small group </p></personGrp>

<personGrp id="sl" role="all" size="s"><p>sl, all, small group</p></personGrp>

<personGrp role="speakers" size="14"><p>number of speakers: 14</p></personGrp>





<item n="speechevent">Lecture</item>

<item n="acaddept">Agricultural and Food Economics</item>

<item n="acaddiv">ss</item>

<item n="partlevel">PG</item>

<item n="module">Marketing Research Methods</item>




<u who="nm0976"> what i'd like to do this morning is to get into the sort of practical details of market research <pause dur="0.6"/> if you remember what the <pause dur="0.2"/> the programme looks like <pause dur="0.3"/> over the next few weeks <pause dur="1.4"/> in the next few weeks starting next week you're going to get quite involved with the <pause dur="0.4"/> # the detail of of # <pause dur="1.9"/> methodology <pause dur="0.2"/> so next week <pause dur="0.4"/> Professor <gap reason="name" extent="1 word"/> is going to be talking about conjunct analysis which is a <pause dur="0.4"/> quite a modern market research methodology to <pause dur="0.5"/> # <pause dur="0.3"/> basically <pause dur="1.0"/> find out what consumers like about new products <pause dur="0.3"/> # <pause dur="0.2"/> potential new products <pause dur="0.8"/> but before <pause dur="0.2"/> we do all that we've deliberately dealt with <pause dur="0.3"/> last week the kinds of data that you'll be <pause dur="0.4"/> coming across in this # <pause dur="1.4"/> # <pause dur="0.6"/> course and the sorts of data that market researchers deal with <pause dur="1.1"/> and this week we want to deal with the practical aspects of collecting that data <pause dur="1.4"/> and there are in a sense there are always <pause dur="0.4"/> we always <pause dur="0.5"/> deal with this in two ways <pause dur="0.5"/> and the first way <pause dur="0.4"/> or or <trunc>th</trunc> <trunc>th</trunc> there are <pause dur="0.2"/> in a sense two themes to this <pause dur="0.4"/> there's a qualitative stream of market

research <pause dur="0.3"/> and there's a quantitative stream of market research <pause dur="1.4"/> i'm going to deal with the <pause dur="0.2"/> basically the quantitative <pause dur="0.2"/> stream of data collection first <pause dur="0.3"/> and then <pause dur="0.3"/> look briefly at the <trunc>k</trunc> <pause dur="0.3"/> the qualitative stream but just to draw the distinction between the two <pause dur="1.9"/> qualitative research is essentially <pause dur="0.2"/> exploratory <pause dur="1.2"/> as i <pause dur="0.2"/> i think i mentioned in week one you're trying to find out ideas <pause dur="0.2"/> concepts <pause dur="0.4"/> # <pause dur="0.2"/> you might be testing people's reactions to certain things <pause dur="0.6"/> but in a qualitative way and what that <pause dur="0.3"/> effectively means is that you will never have <pause dur="0.4"/> large enough numbers to do statistical tests <pause dur="0.8"/> and this is why we're going to try and get this out of the way <pause dur="0.3"/> there are <pause dur="0.2"/> ways of dealing with you <gap reason="name" extent="1 word"/> last week talked to you about attitude measurement <pause dur="0.4"/> and some of that verges on qualitative but <pause dur="0.5"/> what she was referring to was essentially building up enough information about people's attitudes to be able to <pause dur="0.8"/> render it useful for statistical analysis of one sort or another <pause dur="0.2"/> and attitudes <pause dur="0.6"/> as you will see

fall into today's session as well attitude measurement <pause dur="1.0"/> so qualitative research <pause dur="0.2"/> is about <pause dur="0.7"/> basically exploratory work <pause dur="0.8"/> today's the first part of today's session is more about <pause dur="0.4"/> quantitative <pause dur="0.2"/> work and <pause dur="0.2"/> collecting enough data to be able to do something <pause dur="0.4"/> statistical with it <pause dur="2.1"/> now qualitative research in some ways boils down to a <trunc>f</trunc> very few basic techniques and <pause dur="0.3"/> at the end of the morning i'm going to just show you a video which <pause dur="0.4"/> shows you how focus groups work <pause dur="0.3"/> because focus groups <pause dur="0.5"/> are the # the main <pause dur="0.7"/> vehicle that are used for qualitative research <pause dur="1.8"/> it's often said that qualitative research should come before quantitative research <pause dur="0.5"/> and for the sake of variety i'm doing it the other way round it doesn't really matter <pause dur="0.3"/> because <pause dur="0.8"/> qualitative research can stand on its own <pause dur="1.8"/> you know as you'll see in a second in order to design a decent questionnaire <pause dur="0.3"/> you need to have some idea of <pause dur="0.2"/> where people stand on the issue you're looking at <pause dur="0.6"/> it's dangerous of course to <pause dur="0.5"/> # <pause dur="2.1"/> impose your views as a researcher <pause dur="0.8"/> so

what you what what we'll be looking at are ways of <pause dur="0.4"/> in qualitative work <pause dur="0.2"/> ways of <pause dur="0.5"/> prompting the sort of information we need to design questionnaires <pause dur="1.7"/> but we'll start off with <pause dur="0.8"/> some thoughts and some <pause dur="0.5"/> methods for <pause dur="0.4"/> dealing with quantitative data <pause dur="2.1"/> and just to give you a flavour for what's <pause dur="0.9"/> the sorts of problems and the sorts of issues that we're <pause dur="0.3"/> going to be looking at later on <pause dur="1.9"/><kinesic desc="changes slide" iterated="n"/> these are common market research problems <pause dur="0.3"/> analysing product usage <pause dur="1.2"/> very straightforward in a way just saying okay how <trunc>m</trunc> how much <pause dur="0.6"/> of <trunc>th</trunc> <pause dur="0.3"/> this product do people use why do they use it <pause dur="0.5"/> how much more might they use and so on so a fairly straightforward <pause dur="0.4"/> # quantitative <pause dur="1.0"/> assessment of of product usage or sales <pause dur="2.3"/> market researchers always talk about usage because it's <pause dur="0.3"/> more than just sales sales is just a number so much was sold <pause dur="0.3"/> usage is far more concerned with <pause dur="0.3"/> the way people <pause dur="0.5"/> react to the product the way people use the product what people do with it and so on <pause dur="1.7"/> you're going to get in <pause dur="0.4"/> the next couple of weeks quite a lot on

market segmentation <pause dur="1.0"/> so quantitative research to describe market segments <pause dur="0.6"/> and if you're trying to split a mass market into lots of <pause dur="0.4"/> submarkets or segments <pause dur="0.5"/> you need quantitative research usually <pause dur="0.5"/> because <pause dur="0.2"/> it's quantitative research that says <pause dur="0.2"/> this many people are likely to be <pause dur="0.3"/> in this segment if you're trying to <pause dur="0.3"/> target market <pause dur="0.6"/> if you're trying to <pause dur="0.2"/> focus on a particular segment <pause dur="0.4"/> then you will need to know <pause dur="0.4"/> # basically how many people are in that segment or whether it's worth your while <pause dur="0.3"/> going for it <pause dur="1.3"/> so describing market segments market segmentation is usually <pause dur="0.4"/> quantitative and you're going to look at <pause dur="0.5"/> cluster and factor analysis and other things <pause dur="0.7"/> to <pause dur="0.6"/> # <pause dur="0.2"/> describe market segments <pause dur="0.6"/> assessing attitudes as <gap reason="name" extent="1 word"/> spoke to you last <pause dur="0.2"/> week <pause dur="0.4"/> # <pause dur="0.7"/> attitudes of purchasers are assessed mostly by quantitative research <pause dur="1.6"/> describing product image <pause dur="0.4"/> is # on the border <pause dur="0.5"/> again <pause dur="0.2"/> quantitative research can be used you can collect data from large numbers of people <pause dur="0.5"/> # to <trunc>descr</trunc> and and see <pause dur="0.3"/> what they think

of product image but there's a there's a qualitative <pause dur="0.6"/> aspect to that as well <pause dur="1.7"/> identify market opportunities and redefine and reposition products <pause dur="0.6"/> <gap reason="name" extent="1 word"/> talked to you about scaling well you're going to look later on at <trunc>ma</trunc> # at multidimensional scaling <pause dur="0.6"/> which <pause dur="0.3"/> comes up with <pause dur="0.3"/> # ideas on those two those last two <pause dur="1.4"/> # <pause dur="0.4"/> items there <pause dur="3.1"/> now a brief recap on last week <pause dur="0.5"/> <gap reason="name" extent="1 word"/> talked about the types of data well <pause dur="0.4"/> there are different <pause dur="0.9"/><kinesic desc="changes slide" iterated="n"/> types of data categories of data <pause dur="0.3"/> and these <pause dur="0.4"/> have a big impact <pause dur="0.5"/> on <pause dur="0.4"/> # <pause dur="0.3"/> how you collect <pause dur="0.2"/> the information the the the sorts of <pause dur="0.7"/> # vehicles you can use for that <pause dur="0.4"/> so <pause dur="0.2"/> numbers as labels is nominal data <pause dur="0.2"/> ordinal data where you put it in order <pause dur="0.3"/> interval data where <pause dur="0.3"/> there's some <pause dur="2.5"/> implication of the distances between each <pause dur="0.2"/> point let's say on a scale <pause dur="0.5"/> and ratio data where there's some absolute zero point and <pause dur="0.5"/> the <pause dur="0.2"/> differences between the the points means something <pause dur="2.1"/> now <pause dur="0.2"/> how are you going to <pause dur="0.2"/> get hold of people <pause dur="1.3"/> we're looking at questionnaire design which is a sort of general <pause dur="0.3"/> # description of of <pause dur="0.3"/>

of variety of different <pause dur="0.4"/> ways of collecting information <pause dur="0.5"/> but <pause dur="2.2"/><kinesic desc="changes slide" iterated="n"/> the method of contact is obviously important how are you going to get hold of these people <pause dur="0.5"/> and these are <pause dur="0.3"/> the four <pause dur="0.6"/> sort of most common ways not <pause dur="0.4"/> necessarily in order but if you're thinking of <pause dur="0.3"/> how market researchers collect their information <pause dur="0.4"/> those are the ways they do it <pause dur="0.3"/> number four now <pause dur="0.5"/> is <pause dur="0.3"/> being used computers are being used to support <pause dur="0.5"/> # <pause dur="1.7"/> market researchers <pause dur="0.5"/> a great deal more and <pause dur="0.4"/> the whole business of <pause dur="1.0"/> both selling things over the telephone and doing market research over the telephone <pause dur="0.5"/> has become <pause dur="0.2"/> a very important issue in market research <pause dur="0.4"/> and you'll see reference to <trunc>th</trunc> <pause dur="0.2"/> to # <pause dur="0.9"/> terms like C-A-T-I <pause dur="2.8"/><kinesic desc="writes on board" iterated="y" dur="3"/> computer assisted telephone interviewing <pause dur="0.4"/> i <trunc>me</trunc> i think it probably goes without saying now that when you're phoned up <pause dur="0.8"/> and somebody wants to conduct a market research interview with you <pause dur="0.3"/> # they're probably sitting in front of <pause dur="0.8"/> a P-C and we'll we'll <pause dur="0.8"/> look at some of the <pause dur="0.4"/> implications of that but <pause dur="0.3"/> one of the main ones of

course is that <pause dur="0.4"/> the <pause dur="0.2"/> data entry <pause dur="0.5"/> occurs at the same time as the <pause dur="1.2"/> asking of the questions so there's huge savings in terms of that and indeed some of the analysis <pause dur="0.3"/> can go on <pause dur="0.2"/> more or less as you're speaking <pause dur="1.2"/> things like <pause dur="0.6"/> you know in questionnaires you'll need to <pause dur="0.4"/> need to skip <pause dur="0.6"/> from one section to another well the <pause dur="0.3"/> computer does that automatically <pause dur="0.6"/> next week you're going to hear about a technique called adaptive conjoint analysis <pause dur="0.4"/> and this is an analysis method that <pause dur="0.5"/> <trunc>i</trunc> <pause dur="0.3"/> as it suggests sort of adapts to the person who's being interviewed <pause dur="0.4"/> and starts to <pause dur="0.3"/> # <pause dur="0.3"/> react or ask different questions depending on the <pause dur="0.5"/> # the person <pause dur="1.9"/> telephone interviewing <pause dur="1.0"/> # is <pause dur="0.5"/> increasing in in # <pause dur="0.2"/> its coverage its its importance <pause dur="1.2"/> but <pause dur="0.8"/> in some ways <pause dur="0.2"/> it's postal questionnaires that we <pause dur="0.3"/> we want to <pause dur="0.2"/> concentrate on today because <pause dur="0.3"/> it's postal questionnaires that in a sense have to be the most accurate <pause dur="1.1"/> because posnal <trunc>k</trunc> postal questionnaires <pause dur="0.5"/> are <pause dur="0.3"/> the ones where the respondent doesn't have any help at all <pause dur="1.4"/> # <pause dur="1.6"/> there may be follow-ups and you may

follow up by telephone and so on <pause dur="0.4"/> but <pause dur="0.4"/> # <pause dur="0.2"/> it's postal questionnaires which need <pause dur="0.3"/> to be the most accurate if you like <pause dur="1.9"/> personal interviewing <pause dur="3.3"/> in some ways is very good very high levels of response <pause dur="0.6"/> because although you might have <pause dur="1.2"/> told <pause dur="0.4"/> somebody on Broad Street who is trying to hassle you to answer a few questions to go away <pause dur="0.3"/> but at at that <pause dur="0.2"/> the response rate for <pause dur="0.4"/> # personal interviews is actually far higher <pause dur="0.5"/> than these other methods <pause dur="0.9"/> people find it a lot more difficult to turn away somebody who's <pause dur="0.6"/> sort of standing there in front of them <pause dur="1.3"/> the problem with personal interviews of course is that <pause dur="0.4"/> the interviewer is there <pause dur="0.5"/> and the interviewer <pause dur="0.2"/> themselves can <pause dur="0.4"/> bias <pause dur="0.3"/> the results and i think that's just <pause dur="0.5"/> it's a it's a <trunc>l</trunc> it's a lesson <pause dur="0.4"/> in research in general <pause dur="1.1"/> that <pause dur="1.3"/> interviewer bias of course is to be avoided <pause dur="0.9"/> but if you've got somebody in person and and <pause dur="0.7"/> <trunc>i</trunc> it's very interesting if you think about this perhaps <pause dur="0.2"/> having done this course you might think well <pause dur="0.5"/> you know i i tend to submit to these interviewers just

to see how they do it and whether they're any good at it <pause dur="0.7"/> and <pause dur="0.6"/> you quite often find personal interviews where the interviewer is <pause dur="0.3"/> hassling you to answer the question <pause dur="0.7"/> and so you you're you're you're sort of thinking about your and they say <pause dur="0.3"/> well <pause dur="0.2"/> how about or do you mean <pause dur="0.3"/> and this kind of thing and and this is where distortions can come in <pause dur="1.0"/> so <pause dur="1.7"/> personal interviews <pause dur="0.5"/> are <pause dur="0.5"/> good high response rates <pause dur="0.4"/> but there is the problem of bias and of course they're very expensive you're <trunc>em</trunc> <pause dur="0.2"/> employing <pause dur="0.6"/> real people <pause dur="0.5"/> to ask these questions <pause dur="2.5"/> telephone interviewing less expensive but a less good response rate and again some problems of <pause dur="0.6"/> of bias <pause dur="0.4"/> there's a there's a there's a problem whenever you're a person who's asking another person questions <pause dur="0.5"/> there's always a problem of bias because <pause dur="0.2"/> you want people to <pause dur="0.5"/> # <pause dur="0.3"/> expand on their answers and you want people to <pause dur="0.5"/> # sort of chat about what they're interested in <pause dur="0.3"/> and therefore you have to interact with them <pause dur="0.4"/> and that interaction is what can cause

the biases <pause dur="1.4"/> # <pause dur="0.3"/> the alternative is to have a very strict interviewing schedule and a very strict questionnaire <pause dur="0.5"/> and you do get this particularly on the telephone where <pause dur="0.8"/> you know you get this sort of automaton who's actually a person but they're <pause dur="0.4"/> they're <pause dur="0.5"/> it's a very stilted kind of interview and some would say that the quality of the data <pause dur="0.6"/> that is <pause dur="0.9"/> collected as a result is not that high <pause dur="1.4"/> so we're concentrating on postal questionnaires but <trunc>as</trunc> accepting that you need good <pause dur="0.6"/> a good <pause dur="0.3"/> data collection <pause dur="0.5"/> # <pause dur="1.1"/> device for all of those things <pause dur="2.5"/><kinesic desc="changes slide" iterated="n"/> well sampling you're probably all familiar with sampling but <pause dur="0.3"/> you've got to make these decisions <pause dur="1.1"/> you've got to identify the population of people that you're <pause dur="0.6"/> # <pause dur="0.3"/> researching population of people you're talking to <pause dur="0.4"/> you need to draw up a sample frame that is <pause dur="0.2"/> some device which says <pause dur="0.2"/> here are <pause dur="0.3"/> # the rules that we're going to use in sampling these people <pause dur="3.8"/> the sampling unit are you going to sample households <pause dur="0.5"/> or individuals <pause dur="0.4"/> males females and so on so picking

the <pause dur="0.2"/> the people <pause dur="2.7"/> the sampling method <pause dur="0.8"/> well <pause dur="0.6"/> statisticians would love everything to be probability samples preferably random <pause dur="0.8"/> # <pause dur="0.3"/> in practice <pause dur="0.4"/> i think most market researchers are much more pragmatic than that <pause dur="0.4"/> so <pause dur="0.2"/> that pragmatism means that they <pause dur="0.3"/> start <pause dur="0.5"/> looking at non-probability samples <pause dur="0.7"/> quota samples saying right we'll have certain numbers of people in each of these categories <pause dur="0.4"/> # <pause dur="1.1"/> when you get on to things like consumer panels <pause dur="0.5"/> they're probably sampled more <pause dur="0.4"/> # <pause dur="0.6"/> accurately if you like or more scientifically <pause dur="0.3"/> the only problem with them is that you've got the same <pause dur="0.4"/> people every time that's a good thing but it's also a bad thing because they <pause dur="0.6"/> they become their characteristics <pause dur="0.5"/> then get <pause dur="0.4"/> # <pause dur="1.4"/> # in a sense put into the the the research process and they they <pause dur="0.4"/> # may be atypical by <pause dur="1.0"/> cooperating with you <pause dur="1.1"/> sample size you've got to decide <pause dur="0.5"/> sample plan how are you actually going to do it <pause dur="3.9"/><kinesic desc="changes slide" iterated="n"/> now i'm going to show you lots of examples of different types <pause dur="0.2"/> of <pause dur="0.3"/> questions that you can ask <pause dur="1.6"/> here are some very

general <pause dur="0.3"/> design issues though <pause dur="1.6"/> questions need to be precise as you'll you'll see in a moment <pause dur="1.0"/> they need to be <pause dur="0.5"/> well-ordered <pause dur="1.4"/> incidentally <trunc>s</trunc> sorry i should have said this earlier <pause dur="0.8"/> the <pause dur="2.1"/> assessment for this course <pause dur="0.4"/> # <pause dur="0.2"/> we will i think announce next week formally but what it's going to be <pause dur="0.6"/> is a <pause dur="1.1"/> case study <pause dur="0.3"/> basically and you will be asked to <pause dur="0.5"/> evaluate comment appraise the case study <pause dur="0.4"/> and it'll be a case study describing a sort of typical market research process <pause dur="0.9"/> # but it will also include data <pause dur="1.2"/> so there will be data that you can analyse to support your case <pause dur="0.3"/> and you will be able to analyse it basically in whatever way you want to <pause dur="0.6"/> and that'll be up to you <pause dur="0.7"/> i i say that now because <pause dur="0.4"/> you won't <pause dur="0.9"/> if somebody was asking earlier about will we have to do a questionnaire and that they've probably been talking to people who did it last year where <pause dur="0.4"/> we <pause dur="0.2"/> everybody <pause dur="0.2"/> every single individual ran a questionnaire and it got <pause dur="0.7"/> it actually got rather out of hand it was <pause dur="0.2"/> extremely difficult to mark because

people producing <pause dur="0.3"/> huge volumes of things so <pause dur="0.2"/> this is this session is just to introduce you to how <pause dur="0.6"/> this sort of data is collected but you won't be doing this as part of the assessment <pause dur="1.6"/> so questions have got to be precise and i'm going to show you some examples of <pause dur="0.2"/> good and bad questions in a second <pause dur="1.0"/> you need to decide <pause dur="0.5"/> very <pause dur="0.5"/> carefully on the ordering <pause dur="0.8"/> i think there's no excuse these days for <pause dur="0.4"/> in a sense for getting this wrong and certainly <pause dur="0.2"/> presentation is really important <pause dur="0.7"/> and so we'll talk a little bit about presentation and how you order <pause dur="0.3"/> the questions to make sure that you get <pause dur="0.3"/> there's a there are different schools of thought <pause dur="0.3"/> to make sure that you get <pause dur="0.4"/> an optimum response <pause dur="2.9"/> in preparing yourself for this you need to <pause dur="0.2"/> have some idea of what you think people are going to <pause dur="0.4"/> respond or how you think people are going to respond <pause dur="0.9"/> so the sort of categories the sort of ways people might respond <pause dur="0.3"/> the different possible answers to multiple choice questions for example <pause dur="1.6"/> you need <pause dur="0.3"/> instructions <pause dur="0.3"/> for interviewers <pause dur="1.2"/> and as i say this can range from <pause dur="0.2"/> very <pause dur="0.3"/> explicit <pause dur="0.3"/> instructions to as to how somebody should ask particular questions <pause dur="0.4"/> or it could just be <pause dur="0.3"/> # <pause dur="1.7"/>

so some questionnaires as you'll see have very <pause dur="0.5"/> # clear instructions for interviewers some have <pause dur="0.3"/> instructions that are really <pause dur="0.5"/> may well be over the top may well be too detailed <pause dur="1.4"/> and then data processing requirements <pause dur="0.6"/><vocal desc="cough" iterated="n"/> and you'll see some examples of how <pause dur="0.5"/> # we <pause dur="0.3"/> mark up questionnaires in order to <pause dur="0.5"/> make them amenable to <pause dur="0.5"/> data processing <pause dur="3.0"/> so how are we going to develop this thing <pause dur="1.4"/><kinesic desc="changes slide" iterated="n"/> well <pause dur="0.7"/> you need <pause dur="0.7"/> some kind of <pause dur="0.7"/> # <pause dur="0.8"/> statement of objectives the sort of resources you have available <pause dur="0.2"/> and the constraints to <pause dur="0.2"/> the development of the <pause dur="0.3"/> the survey <pause dur="1.5"/> so <pause dur="0.5"/> you've got to set very clear objectives as to what <pause dur="0.4"/> your <pause dur="0.5"/> # questionnaire is designed to achieve <pause dur="1.2"/> you need to say something about <pause dur="0.4"/> how <pause dur="0.6"/> you're going to collect the data the sorts of <pause dur="0.7"/> types of question that you're going to have the <pause dur="0.2"/> the way you word them <pause dur="0.4"/> the flow of the questionnaire and so on <pause dur="0.4"/>

obtaining approval is important <pause dur="1.8"/> in the university we have # <pause dur="1.0"/> a <pause dur="1.7"/> body known as the Ethics Committee <pause dur="0.8"/> and technically speaking if you <pause dur="0.4"/> go out <pause dur="0.5"/> well if you go outside the university <pause dur="0.6"/> to <pause dur="0.5"/> # <pause dur="1.6"/> research anything you need to get the approval of the Ethics Committee <pause dur="0.7"/> and the Ethics Committee is <pause dur="0.3"/> in some ways is quite a good idea the university and indeed <pause dur="0.5"/> any market research body <pause dur="0.4"/> doesn't want <pause dur="0.3"/> the <pause dur="0.3"/> its name to be pulled down by the market research process of course <pause dur="0.7"/> # <pause dur="0.2"/> and so we have to if we're going out and indeed if <pause dur="0.3"/> students are doing projects we have to <pause dur="0.2"/> get the implicit approval of the Ethics Committee <pause dur="0.5"/> sometimes that can come from the head of department <pause dur="1.0"/> but # <pause dur="1.4"/> two or three years ago just <pause dur="1.2"/> for your information the # <pause dur="0.4"/> this group <pause dur="0.4"/> was actually a group of undergraduate students <pause dur="0.9"/> decided to do a <pause dur="0.5"/> a market research project which was the assessment for the course <pause dur="0.6"/> # <pause dur="0.4"/> and they were given a free choice as to what <pause dur="0.2"/> subject they wanted to <pause dur="0.7"/> # <pause dur="1.1"/> ask people about and <pause dur="1.2"/> # the <pause dur="0.5"/> explicit <pause dur="0.4"/> instruction was that the people

they researched should only be members of the course <pause dur="0.7"/> and this group came and said <pause dur="0.3"/> we want to do <pause dur="1.5"/> a kind of sex survey <pause dur="1.2"/> and what this was it <trunc>w</trunc> it was fairly innocent although <pause dur="0.5"/> # i did say <pause dur="0.5"/> you know this must strictly be kept within the <pause dur="1.2"/> group and it it sort of <pause dur="1.3"/> i won't go into the details but it it was asking various fairly personal questions <pause dur="0.6"/> and <pause dur="0.2"/> the <pause dur="1.0"/> first i heard well the next thing i heard of it was <pause dur="0.5"/> somebody called up from who'd been <pause dur="0.2"/> accosted by one of these students <pause dur="0.4"/> somewhere down in town <pause dur="0.6"/> and had been asked these questions and <pause dur="0.5"/> we got into a little bit of trouble about it and <pause dur="0.8"/> we hadn't <trunc>cle</trunc> i hadn't cleared it with the committee because it wasn't <pause dur="0.6"/> # <pause dur="0.3"/> i didn't believe it was going out so anyway <pause dur="0.2"/> there are there are that's a a very obvious problem but you do need to obtain approval <pause dur="0.2"/> there is a market research society code of practice <pause dur="0.3"/> on <pause dur="0.6"/> on asking <pause dur="0.7"/> # <pause dur="0.5"/> questions on on <pause dur="0.5"/> how to do research <pause dur="2.8"/> you need to pretest the questionnaire this is really important those of you some of you

will be doing this for <pause dur="0.5"/> # <pause dur="0.8"/> your <pause dur="1.1"/> dissertations some of you i know are collecting <pause dur="0.2"/> primary data <pause dur="0.6"/> you need to pretest the thing because <pause dur="0.4"/> you're the researcher you're very close to the subject you know what you're talking about <pause dur="0.3"/> but you've got to check that other people do as well and if you want a <pause dur="0.2"/> statistically valid sample of a hundred people or two-hundred people <pause dur="0.5"/> you've got to make sure you're collecting the data properly <pause dur="0.6"/> and it's <pause dur="0.2"/> these pretests or pilots <pause dur="0.5"/> that will tell you whether it's <pause dur="0.2"/> going to work or not <pause dur="0.7"/> so make sure you do pilots and and <pause dur="0.4"/> you know this this can be up to <trunc>s</trunc> <pause dur="0.3"/> sort of half a dozen <pause dur="0.4"/> different <pause dur="0.6"/> # people that you question and you will soon find <pause dur="0.5"/> if you've got any doubts about the length of the questionnaire or the style of particular questions or whether the sort of <pause dur="0.3"/> issues you're asking them about are valid <pause dur="0.4"/> you'll soon find out <pause dur="0.7"/> # from that <pause dur="0.9"/> so <pause dur="0.3"/> # <pause dur="0.5"/> piloting is really important <pause dur="0.6"/> prepare the final questionnaire and then implement the survey <pause dur="1.6"/><kinesic desc="changes slide" iterated="n"/> this is the sort of <pause dur="0.2"/> layout

that you might use <pause dur="0.3"/> for <pause dur="0.5"/> a questionnaire it's there are no <pause dur="0.2"/> right and wrong answers here really <pause dur="0.6"/> but <pause dur="0.2"/> # <pause dur="3.5"/> this is a layout <pause dur="4.5"/> every questionnaire of course has to carry an identification number i think i <pause dur="1.6"/> in the first week i was talking about using <pause dur="0.3"/> # a spreadsheet <pause dur="0.5"/> basically a a matrix to enter the data <pause dur="0.9"/> a lot of people <pause dur="0.3"/> use Excel for that <pause dur="1.0"/> which is fine <pause dur="0.4"/> # <pause dur="0.5"/> the bad thing about Excel i think i mentioned is that it will let you sort the thing <pause dur="0.7"/> in a rather rudimentary way and if you're trying to sort <pause dur="0.5"/> your data <pause dur="0.5"/> # and you <pause dur="0.7"/> if you know in Excel <vocal desc="cough" iterated="n"/><pause dur="1.0"/> you can select <pause dur="0.5"/> three or four columns with the data in it and forget to select the others and if you then sort it <pause dur="0.3"/> all the data will be <pause dur="1.2"/> out you you will <pause dur="0.3"/> it will be non-consistent <pause dur="0.3"/> things like Access <pause dur="0.4"/> and S-P-S-S don't let you do that <pause dur="0.4"/> if you're going to delete something or sort something it the whole row would be sorted so <pause dur="0.3"/> Excel is quite a good database but be careful because you do have to <pause dur="0.5"/> # <pause dur="0.3"/> you do have to watch that <pause dur="0.4"/> the <trunc>ob</trunc> <pause dur="0.2"/> obviously an I-D number is one way of

sorting things so <pause dur="0.3"/> you sort by the order in which you collected the data or something <pause dur="1.8"/> screeners <pause dur="0.2"/> are qualifying questions and this quite often <pause dur="0.3"/> # this is <pause dur="1.0"/> quite often <pause dur="0.4"/> important <pause dur="0.2"/> because sometimes <pause dur="0.4"/> you'll be wanting to collect data on people <pause dur="0.4"/> who <pause dur="0.3"/> # <pause dur="0.4"/> have particular characteristics and you won't know <pause dur="0.5"/> until you <pause dur="0.5"/> # ask them whether they have those characteristics one classic example is <pause dur="0.5"/> and this this happens with student projects and other work we've done is <pause dur="1.8"/> quite often with food products you want to talk to vegetarians because <pause dur="0.4"/> you think that they might <pause dur="0.5"/> # <pause dur="0.3"/> have well they obviously have different <pause dur="0.2"/> physically they have different <pause dur="0.5"/> purchasing habits but they might have different attitudes to food <pause dur="0.8"/> but <pause dur="0.3"/> you can't well i can't spot a vegetarian just by looking at him or her <pause dur="0.6"/><vocal desc="laughter" n="su0986" iterated="y" dur="1"/> # so you need to ask <trunc>s</trunc> screening questions to find out <pause dur="0.4"/> and quite often <pause dur="0.3"/> you'll # you may have come across this you'll get stopped in the street and someone will say <pause dur="0.3"/> you know <pause dur="0.5"/> have you been to the cinema in the last six months

if you say no they say right thanks very much <pause dur="0.4"/> goodbye that's a it's a screening question <pause dur="0.9"/><vocal desc="laughter" n="ss" dur="2" iterated="y"/> okay so <trunc>y</trunc> <pause dur="0.3"/> those qualifying questions are really important <pause dur="1.8"/> # <pause dur="1.1"/> i've got a a questionnaire to show you in a second but we'll just go through this <pause dur="1.8"/><vocal desc="cough" iterated="n"/> you need some introductory remarks and again this may well be a <pause dur="0.2"/> # <pause dur="0.2"/> a comment that the <pause dur="0.9"/> the personal interviewer <pause dur="0.6"/> makes but if it's if you're if this is a mail questionnaire you'll need some remarks to explain what you're doing <pause dur="2.0"/> it's no good just having you know first question <pause dur="0.4"/> how old are you or something # you need some explanation of the work you're doing and why <pause dur="0.5"/> and all market researchers do this <pause dur="0.4"/> there is one <pause dur="1.8"/> i suppose one exception to that there's a a <pause dur="1.1"/> technique if you like called an omnibus survey and this is where several firms pay <pause dur="0.4"/> to <pause dur="0.3"/> include questions about their own interests on a big survey and you may <pause dur="0.4"/> may well have come across these

and these tend to be <pause dur="0.3"/> very mixed and so <pause dur="0.6"/> lots of introductory remarks about the survey are not necessarily relevant <pause dur="2.3"/> transition questions these are ones that sort of take you from <pause dur="0.2"/> one # part of the questionnaire to another if you're asking people about <pause dur="0.3"/> food consumption and you're <pause dur="0.2"/> you're starting off on pricing issues for example <pause dur="0.3"/> then you might have a question that <pause dur="0.3"/> # links price with the amount they buy or something and then you go on to <pause dur="0.4"/> deal more <pause dur="0.5"/> closely with the amount they buy and demand and so on <pause dur="1.4"/> difficult questions <pause dur="0.7"/><vocal desc="cough" iterated="n"/><pause dur="0.4"/> and this is <pause dur="1.0"/> the textbook <pause dur="0.3"/> # idea <trunc>r</trunc> really this depends on <pause dur="0.6"/> # <pause dur="1.1"/> the practice on the ground if you like and what you're asking but <pause dur="0.4"/> i don't think you'd put difficult questions right at the top or very sensitive questions right at the start of the questionnaire <pause dur="1.0"/>

you need to distinguish questions <pause dur="0.3"/> very clearly from instructions <pause dur="0.7"/> and you need to think about <pause dur="0.2"/> classification data about the respondent <pause dur="0.9"/> virtually all market research studies as you'll see <pause dur="0.5"/> will and and particularly when you get into market segmentation you need information about the people <pause dur="0.7"/> you're <pause dur="0.3"/> asking you're not <pause dur="0.3"/> you're very unlikely <pause dur="0.4"/> to be <pause dur="0.6"/> researching something related to some kind of mass market <pause dur="0.5"/> but you say <pause dur="0.2"/> okay well it doesn't matter who these people are we just want a hundred people and we'll ask them some questions <pause dur="0.8"/> because <pause dur="0.8"/> by far ninety-nine per cent of the time you'll be saying well we need to know <pause dur="0.3"/> whether older people think this and younger people think that <pause dur="0.4"/> or people in the north think <pause dur="0.2"/> one thing and people in the south think another thing <pause dur="0.9"/> so those <pause dur="0.5"/> the the classification data or the demographic data is is extremely important <pause dur="0.5"/> and for for

clustering and <sic corr="factorizing">factronizing</sic> various other things <pause dur="0.3"/> as you'll find out <pause dur="0.4"/> this <pause dur="0.2"/> data is is vital <pause dur="2.7"/> there are various schools of thought on this # <pause dur="0.4"/> you know there used to be big debate about how you could ask people their age and and <pause dur="0.7"/> there was <pause dur="0.4"/> sort of # <pause dur="2.0"/> various schools of thought like you you didn't ask people their age that would put them off completely <pause dur="0.5"/> so you just had a <pause dur="0.6"/> # <pause dur="1.5"/> you know you had a blank box and you evaluated their age yourselves <pause dur="0.2"/> but now the <pause dur="0.6"/> it's current <pause dur="0.2"/> idea seems to be you just ask people how old are you <pause dur="0.8"/> # <pause dur="0.8"/> same thing about income <pause dur="0.4"/> that is more sensitive it's very difficult to ask people about incomes <pause dur="0.2"/> so as you know you start <pause dur="0.4"/> talking to people about <pause dur="0.5"/> # <pause dur="0.3"/> either you have a range of incomes and ask people to fill that in <pause dur="0.5"/> or you talk about the occupation of the main income earner <pause dur="0.7"/> so if you ask about occupation you get <trunc>r</trunc> rather more of an idea of <pause dur="0.9"/> what people <pause dur="1.1"/><kinesic desc="changes slide" iterated="n"/> earn <pause dur="2.4"/> so allied to the <pause dur="0.2"/> the types of

data <pause dur="0.4"/> we were talking about earlier <pause dur="0.8"/> this <pause dur="0.5"/> really helps classify the types of question <pause dur="0.5"/> that <pause dur="0.4"/> you ask <pause dur="0.5"/> and they fall into various categories <pause dur="0.4"/> there are direct factual <pause dur="0.2"/> questions <pause dur="1.7"/> how much do you earn would be one of them <pause dur="1.1"/> you know are you male or female <pause dur="0.7"/> how tall are you this kind of thing it's all very factual <pause dur="0.4"/> # <pause dur="0.5"/> if you think about it in data analysis terms <pause dur="0.6"/> a lot of it is <pause dur="1.2"/> quite quantitative by which i mean <pause dur="0.4"/> it's it's useful in further <trunc>ana</trunc> it's it's particularly useful in further analysis so if you're collecting things on quantities or prices <pause dur="0.7"/> then <pause dur="0.3"/> you know as economists you'll know that that's data that can be used <pause dur="0.5"/> # for other things <pause dur="0.4"/> # personal details <pause dur="0.3"/> if you're <pause dur="0.6"/> just collecting information on <pause dur="0.3"/> you know whether somebody's going to vote yes in an election or vote no in an election <pause dur="0.4"/> it's very simple data it's a dummy variable but it's it's <pause dur="0.3"/> it can be useful <pause dur="1.4"/> however if you're <trunc>ope</trunc> asking open-ended questions <pause dur="1.1"/> like <pause dur="0.8"/> what do you feel about the new building <pause dur="1.8"/> # <pause dur="0.4"/> right <pause dur="0.5"/>

people could write an essay on that <pause dur="0.8"/> # <pause dur="0.4"/> would be very useful because they'd give you loads and loads of information but it would be <pause dur="0.4"/> extremely difficult <pause dur="0.4"/> to analyse <pause dur="0.5"/> so you've got to <pause dur="0.2"/> strike a balance <pause dur="0.2"/> there is the use of the data and <pause dur="0.4"/> how easy it will be to actually use it <pause dur="0.5"/> multiple choice questions and dichotomous questions dichotomous being yes or no <pause dur="0.5"/> just look at the different characteristics in those <pause dur="2.5"/><kinesic desc="changes slide" iterated="n"/><vocal desc="cough" iterated="n"/> so open-ended questions are great as first questions <pause dur="1.2"/> they've still got to establish <pause dur="0.3"/> if you think of <trunc>th</trunc> the questionnaire remember it's postal <pause dur="0.6"/> that's what we're looking at at the moment <pause dur="0.6"/> you think of it as <pause dur="0.3"/> being you <trunc>s</trunc> you as the researcher speaking to people <pause dur="1.4"/> then you need something that establishes some rapport with people <pause dur="1.5"/> and you'll see in a second different ways of doing that <pause dur="1.3"/>

so <pause dur="0.3"/> they're they're good first questions for that reason <pause dur="0.8"/> there's very little influence on the response what do you think of the new building <pause dur="0.6"/> is a very open question you can say exactly what you like <pause dur="0.6"/> you can write as much as you like you could give a one word answer or you could <pause dur="0.4"/> write an essay on it <pause dur="1.5"/> so there's very little influence on the <trunc>res</trunc> you're not saying what do you think of the new building <pause dur="0.4"/> you know good bad or indifferent <pause dur="0.8"/> you're not <pause dur="0.2"/> you're not <pause dur="0.4"/> imposing on people <pause dur="0.8"/> and so you can get <pause dur="0.3"/> insights <pause dur="0.2"/> however <pause dur="0.9"/> there can be <pause dur="0.2"/> interviewer bias certainly if it's a <pause dur="0.4"/> a personal interview <pause dur="0.5"/> but the <trunc>pr</trunc> main problem with open-ended questions is <pause dur="0.7"/> we're looking at <pause dur="0.4"/> quantitative methods we want to <pause dur="0.4"/> quantify the results we therefore need <pause dur="0.2"/> some form of computerized analysis <pause dur="0.4"/> and coding the things coding <pause dur="0.4"/> open-ended questions <pause dur="0.5"/> is <pause dur="0.3"/> the main problem <pause dur="0.3"/> that is <pause dur="0.5"/> putting a number really to the question <pause dur="0.2"/> to the sorry to the answer that you've been given <pause dur="0.8"/> you can precode open-ended questions so you'd have a <pause dur="0.3"/> a list of

things that you think people are going to respond but they always say <pause dur="0.2"/> different things <pause dur="0.7"/> so you you you need to <pause dur="0.6"/> # bear that in mind you can <pause dur="0.4"/> it's usually better to do it afterwards <pause dur="3.5"/><kinesic desc="changes slide" iterated="n"/> multiple choice questions <pause dur="0.3"/> dead easy they reduce <pause dur="0.3"/> interviewer bias <pause dur="0.5"/> very easy for people to <pause dur="0.2"/> very easy and fast for people to answer <pause dur="0.8"/> very easy for data processing <pause dur="2.0"/> but the argument goes that they are <pause dur="0.9"/> rather difficult to design the thing about multiple choice questions <pause dur="0.5"/> is that you are forcing people into certain <pause dur="0.6"/> answers <pause dur="0.7"/> if you do this is a good reason for piloting if you have a multiple choice question <pause dur="0.5"/> and <pause dur="0.4"/> # <pause dur="1.4"/> you pilot it you may find that people are not <pause dur="0.5"/> they don't put the issue that you're asking them about into <pause dur="0.2"/> that particular set of categories that you've imposed <pause dur="0.3"/> so that's where your pilots and qualitative research will help <pause dur="1.0"/> let me just show you an example of this <pause dur="21.2"/><kinesic desc="changes slide" iterated="n"/> so this is a <pause dur="1.5"/> questionnaire was actually done out of this course a couple of years ago <pause dur="0.3"/> and it was to try to

ask <pause dur="0.6"/> # <pause dur="0.3"/> customers of a farm shop <pause dur="0.2"/> what they thought of the <pause dur="0.7"/> # outlet they were visiting <pause dur="0.6"/> and we had a screening question because <pause dur="0.4"/> # <pause dur="0.3"/> it's not actually on there but we had a screening process because <pause dur="0.2"/> we wanted to interview customers at the farm shop and non-customers so there was a <pause dur="0.7"/> # <pause dur="0.6"/> a screener for that but <pause dur="0.2"/> these <pause dur="0.3"/> are <pause dur="0.4"/> basically multiple choice questions of various sorts <pause dur="2.1"/> here's one thinking about your visits to the shop do you visit at specific times of the year <pause dur="0.3"/> and there were <pause dur="0.4"/> different options <pause dur="0.2"/> and it's multiple choice now if you're <pause dur="0.2"/> asking do you visit at specific times of the year <pause dur="1.4"/> summer autumn winter spring there are no other options <pause dur="0.5"/> # <pause dur="0.2"/> oh except for all the year round <pause dur="1.2"/> so <pause dur="0.8"/> you know that's a a a relatively simple example <pause dur="0.9"/> # <pause dur="2.4"/> we said thinking about your visits to the shop do you <pause dur="0.2"/> more often than not visit at specific times of the day <pause dur="0.4"/> morning lunchtime afternoon <pause dur="0.4"/> now <pause dur="0.4"/> for all we knew they might come in the middle of the night <pause dur="0.6"/> so we had <pause dur="0.4"/> a category called other and this is this is the sort of <pause dur="0.7"/>

<trunc>p</trunc> a # a pilot would find out if people did unusual things or <pause dur="0.3"/> did things that were not <pause dur="0.3"/> listed on here <pause dur="1.2"/> # <pause dur="0.3"/> what form of transport did you use <pause dur="1.3"/> car bus taxi foot bicycle <pause dur="0.4"/> or other <pause dur="0.3"/> so other in multiple choice questions other is obviously a catch-all <pause dur="0.4"/> # if you find everyone's ticking other <pause dur="0.3"/> then you've designed the thing badly <pause dur="0.8"/> because you haven't <pause dur="0.4"/> obviously identified the right <pause dur="0.8"/> answers that people are going to give and this again is where piloting and so on will <pause dur="0.6"/> # help <pause dur="4.1"/> the other thing is about bias <pause dur="0.8"/> and it's said that <pause dur="0.4"/> if <pause dur="0.4"/> you offer people options <pause dur="1.9"/> # <pause dur="0.2"/> the order you put them in does make a difference <pause dur="0.2"/> now this is a <pause dur="0.3"/> this is all sort of fairly <pause dur="0.2"/> straightforward data i don't know if we can find <pause dur="1.2"/><kinesic desc="changes slide" iterated="n"/> something else <pause dur="1.8"/> okay how did you hear about the existence of the shop <pause dur="0.4"/> well on the radio from friends <trunc>newsp</trunc> now if they're trying to think about oh how did i <pause dur="0.3"/> hear about this <pause dur="0.7"/> you need to be careful about biasing the survey <pause dur="0.3"/> because of the order these are in and people might oh radio yeah <pause dur="0.4"/>

i'll put radio down <pause dur="1.1"/> okay and there's there's there's various ways of getting round this and one is simply to <pause dur="0.4"/> randomize the order and have different <pause dur="0.2"/> questionnaires with different orders <pause dur="0.6"/> where do you do your main food shopping for the week <pause dur="0.7"/> again <pause dur="0.2"/> the the order may well be important <pause dur="8.4"/><kinesic desc="changes slide" iterated="n"/> dichotomous questions this is yes or no male or female whatever <pause dur="1.1"/> # very easy <pause dur="1.9"/> coding is very simple and it offers you the possibility of <pause dur="0.6"/> # <pause dur="0.4"/> using dummy variables in this <pause dur="3.1"/> so <pause dur="0.4"/> you can isolate how males and how females think about things <pause dur="0.4"/> but <pause dur="0.5"/> with some issues the it may not be seen in this dichotomous way it may not be <pause dur="0.6"/> # an issue of <pause dur="0.4"/> # either or <pause dur="1.0"/> and the wording becomes very important <pause dur="0.5"/> in order to to get the right sort of response <pause dur="0.4"/> just <pause dur="1.1"/><kinesic desc="changes slide" iterated="n"/> one thing i just forgot <pause dur="0.7"/> while we're talking about multiple choice questions <pause dur="0.8"/> the analysis and coding of multiple choice questions is really crucial <pause dur="1.2"/> because <pause dur="0.3"/> if you're thinking about <pause dur="2.1"/> the <pause dur="0.7"/> variables <pause dur="1.7"/> remember <pause dur="0.6"/> you're putting this into a spreadsheet or into some sort of analysis

package <pause dur="0.3"/> and you've got variables and the variable has got to <pause dur="0.4"/> identify <pause dur="0.3"/> each of the bits of data that you're dealing with <pause dur="2.4"/> and if i just take one of these <pause dur="1.3"/> says where do you do your main food shopping for the week <pause dur="0.2"/> or month that's a very deliberate question because <pause dur="0.4"/> we felt <pause dur="0.3"/> that there should only be one answer to that <pause dur="0.3"/> where do you do your main food shopping <pause dur="0.4"/> but if we'd said <pause dur="0.2"/> where do you do your food shopping <pause dur="1.8"/> then these people could quite easily have ticked more than one answer <pause dur="0.7"/> and indeed they did even though we put the word main in there <pause dur="1.2"/> so you then come up with a problem <pause dur="0.5"/> because <pause dur="1.0"/> the question <pause dur="0.7"/> arises what is the variable here <pause dur="2.3"/> if we had one variable <pause dur="0.3"/> to cover <pause dur="0.4"/> main food shopping so we'd call it main food shop or something <pause dur="0.3"/> as the variable name <pause dur="0.6"/> then you'd enter a number <pause dur="1.0"/> dependent on which one of these they ticked <pause dur="0.5"/> but if they ticked two you're in real problems because <pause dur="0.5"/> you can't <pause dur="0.6"/> if if you've just got the one variable about main food shopping and then say i do my main <trunc>shop</trunc>

food shopping in Asda and Waitrose <pause dur="1.7"/> then you can't interpret that <pause dur="0.5"/> and this is a key thing about multiple choice questions is that you <pause dur="0.5"/> need to decide whether it's tick one or tick more than one <pause dur="0.7"/> and if it's tick more than one <pause dur="0.4"/> then each of these has to become a variable <pause dur="2.0"/> and if you're doing this later on it's really important to to get that difference right <pause dur="0.7"/> you you sometimes see people do <pause dur="0.3"/> what <pause dur="0.7"/> what are really rather silly things like <pause dur="0.8"/> if if it's a a tick once so this is one variable called main food shopping <pause dur="1.3"/> you then you you see people <pause dur="0.3"/> trying to <pause dur="1.3"/> create a number so right we'll say number seven <pause dur="0.3"/> we'll call it number seven if they tick Tesco and Asda <pause dur="1.6"/> right and then we'll call it number eight if they tick Asda and Sainsbury but then what if they tick three <pause dur="0.3"/> and so on it just gets out of hand so if <pause dur="0.8"/> there's a possibility in a multiple choice answer of them ticking more than one box you need to <pause dur="0.3"/> create <pause dur="0.5"/> five <pause dur="0.2"/> variables <pause dur="0.3"/> and we've got some rather horrendous <pause dur="1.2"/> things here <pause dur="1.7"/> please

indicate <pause dur="0.5"/> where you shop for the bulk <pause dur="0.3"/> your needs for the following products <pause dur="0.2"/> there's an awful lot of data <pause dur="0.3"/> that can be collected in a table like that <pause dur="0.7"/> and again <pause dur="0.6"/> as long as people only tick one we're we're all right but of course each of these is a variable <pause dur="1.0"/> they start ticking more than them then each cell <pause dur="0.4"/> becomes a variable <pause dur="0.6"/> and you get <pause dur="0.6"/> you can get very <pause dur="0.6"/> confused and get an awful lot of variables <pause dur="5.8"/><kinesic desc="changes slide" iterated="n"/> now there are lots of problems obviously in asking people questions <pause dur="4.2"/> they relate to things like sensitivity <pause dur="0.5"/> is the issue a tricky one <pause dur="3.3"/> they relate to the complexity of the questions <pause dur="1.7"/> they relate to the applicability of the questions can people answer them <pause dur="0.7"/> they relate to length <pause dur="1.0"/> so <pause dur="0.2"/> the length of the questions themselves and the <trunc>c</trunc> the length of the questionnaire <pause dur="0.4"/> is <pause dur="0.3"/> an issue <pause dur="1.1"/> and then <pause dur="0.7"/> as you'll see <pause dur="0.3"/> asking people hypothetical questions may be something that they're not very good at <pause dur="0.4"/> dealing with <pause dur="0.5"/> and <pause dur="0.4"/> asking people <pause dur="0.2"/> leading questions <pause dur="0.4"/> as you'll see can bias your results <pause dur="2.7"/><kinesic desc="changes slide" iterated="n"/> to help avoid this there's a little list of things <pause dur="0.5"/> that you need to avoid <pause dur="1.0"/> <unclear>follow</unclear> you want simple words you want clear words avoid things like leading people biasing people <pause dur="0.7"/> implicit assumptions like assuming somebody knows something <pause dur="0.6"/> # which they may not <pause dur="1.6"/>

unnecessary estimation <pause dur="0.6"/> double-barrelled questions and so on <pause dur="1.5"/> you'll see some examples of this now <pause dur="0.4"/> so <pause dur="4.9"/><kinesic desc="changes slide" iterated="n"/> you need <pause dur="0.3"/> to be specific <pause dur="0.2"/> that still on <pause dur="3.0"/><event desc="noise from computer" iterated="n"/> <unclear>what</unclear> was that <pause dur="2.2"/> would you here's here's a question <reading>would you support increased taxes to pay for educational programmes for your children</reading> <pause dur="1.9"/> now <pause dur="0.4"/> the trouble with this <pause dur="0.3"/> topic if you like is it's it's more of an art than a science and it's very easy to say here are some precise rules about how you do it <pause dur="1.0"/> # <pause dur="1.0"/> but <pause dur="1.6"/> it is an art and and that looks a perfectly good question <pause dur="1.3"/> but <pause dur="1.4"/> and and in some ways because it's an art you can go too far you could make this extremely precise or you could leave it relatively vague like this <pause dur="0.3"/> but some would say that's a bit open-ended <pause dur="0.5"/> increased taxes well how much <pause dur="1.4"/> you could be a bit more specific <pause dur="0.4"/> i mean that also says increased taxes that could be somebody else's <pause dur="3.6"/><event desc="noise from computer" iterated="n"/> anybody explain <shift feature="voice" new="laugh"/> where that's coming from <shift feature="voice" new="normal"/><vocal desc="laughter" n="ss" iterated="y" dur="2"/> </u><u who="ss" trans="latching"> <gap reason="inaudible, multiple speakers" extent="4 secs"/></u> <pause dur="0.5"/> <u who="nm0976" trans="pause"> what this thing </u><pause dur="0.3"/> <u who="ss" trans="pause"> <gap reason="inaudible, multiple speakers" extent="4 secs"/></u><u who="nm0976" trans="overlap"> oh you think that's my <pause dur="0.4"/> well that's </u><u who="ss" trans="overlap"> <gap reason="inaudible, multiple speakers" extent="6 secs"/></u><u who="nm0976" trans="overlap"> gosh you can do far

more exciting things than that you can have a <vocal desc="laughter" n="ss" dur="1" iterated="y"/> gunshot going off or something <vocal desc="laughter" n="sl" dur="1" iterated="y"/> <pause dur="0.3"/> # <pause dur="0.5"/> well sorry about that but <pause dur="0.5"/> you'll have to put with it <pause dur="0.6"/> these <pause dur="0.4"/> # P-Cs are too <pause dur="1.3"/> efficient <pause dur="0.5"/> # <pause dur="1.1"/> so there's a little bit more <pause dur="1.1"/> it's a little bit more specific would you support an increase in your taxes to pay for education programmes for your children <pause dur="2.2"/> <event desc="noise from computer" iterated="n"/><vocal desc="laughter" n="sl" dur="1" iterated="y"/> i'm sorry i'm going to have to turn that off <pause dur="0.7"/> can't cope with that <pause dur="0.9"/> what is it it's in </u><u who="ss" trans="overlap"> <gap reason="inaudible, multiple speakers" extent="4 secs"/></u> <u who="nm0976" trans="overlap"> eh </u><u who="ss" trans="overlap"> <gap reason="inaudible, multiple speakers" extent="18 secs"/></u><u who="nm0976" trans="overlap"> where is it</u><u who="ss" trans="overlap"> <gap reason="inaudible, multiple speakers" extent="6 secs"/><vocal desc="laughter" n="ss" dur="1" iterated="y"/> </u><u who="nm0976" trans="overlap"> says no sound look </u><u who="ss" trans="overlap"> <gap reason="inaudible, multiple speakers" extent="2 secs"/></u> <pause dur="0.5"/> <u who="sm0977" trans="pause"> turn the volume off </u><pause dur="0.5"/><u who="nm0976" trans="pause"> we could probably do that in a minute </u> <u who="sm0978" trans="overlap"> <gap reason="inaudible" extent="3 secs"/></u> <u who="nm0976" trans="overlap"> actually you can do volume here look <pause dur="3.1"/> should do <pause dur="1.4"/> sorry about that <pause dur="2.6"/><kinesic desc="changes slide" iterated="n"/> there we go <pause dur="1.0"/> # <pause dur="0.5"/> so <pause dur="1.4"/> be <trunc>s</trunc> be relatively specific <pause dur="1.2"/> you need to know your respondents' ability to answer the sort of questions you're asking <pause dur="0.5"/> and the context <pause dur="2.6"/> that <pause dur="0.2"/> is a very <pause dur="0.5"/> is a question that's very dear to my heart but it might not mean a lot to you i don't know <pause dur="0.9"/> the Dearing Committee was the body that a year ago <pause dur="0.3"/> reviewed higher education for the Higher Education Funding Council of

England <pause dur="0.3"/> now i know that but <pause dur="0.2"/> you might not <pause dur="0.3"/> and your respondents might not <pause dur="4.1"/><kinesic desc="changes slide" iterated="n"/> you need to be specific about time and this is something that <pause dur="0.4"/> quite often <pause dur="0.2"/> you're asking people questions about things they've done well you <pause dur="0.7"/> very very often asking people questions about things they've done <pause dur="1.3"/> so you need to be specific <pause dur="2.5"/> again there's a fine line between being specific and being either <pause dur="0.4"/> you know overlong or being patronizing <pause dur="0.7"/> but how often did you exercise <pause dur="0.3"/> in the past week <pause dur="0.2"/> start with today's date and count back seven days some would say well i know what a week is thank you <pause dur="0.4"/><vocal desc="laughter" n="sm0979" iterated="y" dur="1"/> and i can count that but <pause dur="1.1"/> get the idea <pause dur="1.7"/> that is a totally <pause dur="0.4"/> inappropriate question <pause dur="0.9"/> because you're asking people <pause dur="0.3"/> a question which really relies on you know enormous memory and <pause dur="0.4"/> you know # <pause dur="0.2"/> how would they <pause dur="0.2"/> work that out but you see this sort of thing <pause dur="3.4"/><kinesic desc="changes slide" iterated="n"/> again on this fine line between being <pause dur="0.2"/> overspecific and being # being too long and being too brief <pause dur="0.6"/> here's one place of residence well you could be a bit <pause dur="1.0"/> more specific about that

what's the name of the city where you currently live <pause dur="0.7"/> # <pause dur="0.3"/> these open-ended things <trunc>accim</trunc> accidents among children are <pause dur="0.4"/> you might want to be <pause dur="0.4"/> more specific <pause dur="0.3"/> and # indicating agreement with a statement or something is <pause dur="0.5"/> # a better way of doing that <pause dur="2.2"/><kinesic desc="changes slide" iterated="n"/> here's another one for you <pause dur="1.2"/> in your view does the <pause dur="0.2"/> U-O-R provide a service worthy of the fees you pay <pause dur="1.5"/> well of course that relates to the Union of Railworkers <pause dur="0.5"/> but you may have thought it was relating to the University of Reading <pause dur="0.2"/> you need to be <pause dur="0.8"/> specific <pause dur="2.5"/> and this is something you must <pause dur="0.2"/> have views on <pause dur="6.8"/><kinesic desc="changes slide" iterated="n"/> right i know what this means because <pause dur="0.2"/> i've sort of been involved in <pause dur="0.9"/> educational research and stuff <pause dur="0.4"/> but # you might prefer <pause dur="0.4"/> should each module tutor conduct a review of how students feel the course has gone each term <pause dur="0.4"/> it's the <trunc>sa</trunc> it's exactly the same question but <pause dur="0.3"/> you've got to phrase it in a way <pause dur="0.5"/> that people understand if you were asking <pause dur="0.4"/> this if this questionnaire was aimed at a load of educational specialists <pause dur="0.2"/> no problem <pause dur="3.9"/><kinesic desc="changes slide" iterated="n"/> this is

a classic sort of loaded question many prominent people have publicly admitted they've sought help for problems relating to alcohol abuse <pause dur="0.5"/> what about you <pause dur="1.5"/><vocal desc="laughter" iterated="y" dur="1"/> okay <pause dur="0.4"/> it's a very <pause dur="1.6"/> i'll give you <shift feature="voice" new="laugh"/>another example<shift feature="voice" new="normal"/> <pause dur="0.7"/> this is a very famous one <pause dur="1.9"/> # it's out of a <pause dur="0.2"/> ninety-fifties article called <pause dur="0.2"/> asking the embarrassing question <pause dur="1.2"/> or the loaded question <pause dur="1.9"/> and just just to give you the context one virtually can guarantee meaningless responses by <trunc>direst</trunc> <pause dur="0.2"/> directly <pause dur="0.3"/> asking questions such as <reading>have you ever defaulted on a credit account</reading> <pause dur="0.4"/> <reading>do you smoke pot at least once a week</reading> <pause dur="0.5"/><vocal desc="laughter" n="ss" iterated="y" dur="1"/> or <reading>have you ever been involved in an unreported <pause dur="0.2"/> car accident</reading> <pause dur="1.2"/> <unclear>right</unclear> they're stupid questions people aren't going to <pause dur="0.7"/> # <pause dur="1.6"/> respond to them <pause dur="0.3"/> but Alan Barton it says here <reading>Alan Barton made the point best in ninety-fifty-eight when he <pause dur="0.7"/> posed the following <pause dur="0.3"/> parody on ways to ask the question did you kill your wife</reading> <pause dur="1.7"/><vocal desc="laughter" n="ss" dur="2" iterated="y"/> and the first approach that you might take to asking that question is the casual approach <pause dur="1.4"/> which is <pause dur="1.0"/> <reading>do you happen

to have murdered your wife</reading> <pause dur="2.0"/><vocal desc="laughter" n="ss" iterated="y" dur="1"/> the second approach is multiple choice <pause dur="1.8"/><vocal desc="laughter" n="ss" iterated="y" dur="2"/> <reading>would you please read off the number on this card which <pause dur="0.2"/> corresponds to what became of your wife <pause dur="0.6"/><vocal desc="laughter" n="ss" iterated="y" dur="1"/> you hand the card to the respondents and the card has one <pause dur="0.3"/> natural death <pause dur="0.8"/><vocal desc="laughter" n="ss" iterated="y" dur="1"/> two i killed her <pause dur="0.5"/><vocal desc="laughter" n="ss" iterated="y" dur="1"/> and three other</reading> <pause dur="0.7"/><vocal desc="laughter" n="ss" iterated="y" dur="2"/> whatever other might be <pause dur="1.6"/> you could <pause dur="0.3"/> have the sort of # <pause dur="0.2"/> peer group <pause dur="0.2"/> approach <pause dur="0.6"/> where you say to you you ask you say to people <pause dur="0.4"/> or you say to the respondent <pause dur="0.7"/> <reading>as you know many people have been killing their wives recently <pause dur="0.7"/><vocal desc="laughter" n="ss" iterated="y" dur="1"/> do you happen to have killed yours</reading> <pause dur="1.7"/><vocal desc="laughter" n="ss" dur="1" iterated="y"/> # <pause dur="2.2"/> and another other people approach would see <pause dur="0.2"/> <reading>do you know any people who've murdered their wives and how about yourself</reading> <pause dur="1.7"/><vocal desc="laughter" n="ss" iterated="y" dur="3"/> # <pause dur="0.5"/> and then finally it says <reading>stare firmly into the respondent's eyes ask in simple clear cut language <pause dur="0.3"/> such as that to which the respondent is accustomed and with an air of assuming that everyone's done everything <pause dur="0.5"/> do you ever kill your wife</reading> <pause dur="1.1"/><vocal desc="laughter" n="ss" iterated="y" dur="1"/> so <pause dur="0.5"/><vocal desc="laughter" iterated="y" dur="1"/> <pause dur="0.4"/> there are different <shift feature="voice" new="laugh"/>ways of<shift feature="voice" new="normal"/> asking sensitive questions <pause dur="1.4"/><kinesic desc="changes slide" iterated="n"/> so on controversial subjects for example if you're talking to farmers about their <pause dur="0.3"/> attitudes to <trunc>convers</trunc> conservation <pause dur="0.8"/> it's not particularly <pause dur="0.6"/>

# <pause dur="0.7"/> approachable to say how many kilometres of hedges have you taken out in the past five years <pause dur="0.7"/> so you might do this sort of thing it's the same idea as <pause dur="0.3"/> did you kill your wife <pause dur="0.8"/> <reading>please describe your policy for hedgerow removal</reading> <pause dur="4.2"/><kinesic desc="changes slide" iterated="n"/> you've got to be as clear as possible again without going over the top <pause dur="0.5"/> daily newspapers read regularly <pause dur="0.4"/> is a bit vague <pause dur="0.5"/> you might want to <pause dur="0.5"/> specify <pause dur="0.6"/> read or glanced through most days of the week <pause dur="0.5"/> you've got to adapt this question to how people <pause dur="0.3"/> deal with this kind of issue so <pause dur="0.3"/> <trunc>m</trunc> maybe people you think maybe people just have a <trunc>c</trunc> a quick glance at the paper and you'd call that reading <pause dur="0.3"/> fine <pause dur="1.4"/> you could get even more specific do you read regularly any daily newspaper at least three out of <pause dur="0.2"/> and so on <pause dur="1.8"/> but you've got to be clear <pause dur="3.8"/><kinesic desc="changes slide" iterated="n"/> if you're asking somebody about their <pause dur="0.8"/> employment <pause dur="0.3"/> their job <pause dur="2.6"/> just saying is yours an interesting job or a routine job <pause dur="0.6"/> might well <pause dur="0.6"/> # <pause dur="1.6"/> you might well find people a bit <pause dur="0.6"/> unhappy about answering that so you could <pause dur="0.3"/> ask it another way is yours a job in which you do a lot of hard thinking or a job which <pause dur="0.5"/> once you've learned it you always know

how to do it <pause dur="0.8"/> it's a sort of flavour to how you ask this sort of question you might argue that the answer you get from <pause dur="0.4"/> from those two would be rather different but <pause dur="0.4"/> you get the idea you've got to ask the question in a way that people will <pause dur="1.6"/> be prepared to answer it <pause dur="3.0"/><kinesic desc="changes slide" iterated="n"/> this is used <pause dur="0.3"/> consistently in advertising <pause dur="0.8"/> prestige buyers in advertising or prestige in advertising is used as a <trunc>m</trunc> <pause dur="0.3"/> a major selling point obviously <pause dur="1.0"/> so doctors are always seen advertising headache remedies <pause dur="1.9"/> # <pause dur="0.4"/> sports stars are always seen promoting sports drinks and so on <pause dur="0.8"/> but <pause dur="0.5"/> it works against you if you're <trunc>s</trunc> trying to get <pause dur="0.4"/> unbiased responses from people so doctors say that increased fibre is good for your health do you agree well <pause dur="1.7"/><vocal desc="laugh" iterated="n"/><pause dur="1.1"/> this week is probably not the week to talk about this after <pause dur="0.8"/> the Harold Shipman case <pause dur="0.2"/> # <pause dur="0.4"/> doctors say that a quick injection is good for you do you agree <pause dur="1.6"/> <vocal desc="laughter" n="ss" iterated="y" dur="2"/> # <pause dur="0.6"/> have you <shift feature="voice" new="laugh"/>increased the fibre content of your diet<shift feature="voice" new="normal"/> over the last few months <pause dur="1.2"/> would be another way of putting it <pause dur="0.7"/> and if you

wanted to be really vague about it <pause dur="0.2"/> have you changed your shopping habits over the last few months due to health concerns <pause dur="0.4"/> what products have you been buying more of if somebody says <pause dur="0.3"/> i've been buying loads more <pause dur="0.4"/> products with fibre in them then you've really got them you've really got a <pause dur="0.3"/> a valid answer there <pause dur="0.2"/> so <pause dur="0.4"/> the first one would certainly <pause dur="0.4"/> bias people's responses and and the others might be <pause dur="0.3"/> better <pause dur="1.9"/><kinesic desc="changes slide" iterated="n"/> now this is another <pause dur="0.4"/> one of the most famous <pause dur="0.8"/> areas in which <pause dur="0.2"/> well in Britain anyway in which <pause dur="0.6"/> sort of market research gets <pause dur="0.2"/> condescending and patronizing is in tax returns <pause dur="0.5"/> the Inland Revenue <pause dur="0.5"/> these days you actually have to sort of do it yourself as well it's self-assessment <pause dur="0.5"/> but they send you forms and these forms <pause dur="0.4"/> are <pause dur="1.7"/> sort of <shift feature="voice" new="laugh"/>aimed at<shift feature="voice" new="normal"/><pause dur="2.5"/> sort of five year olds i think because they they <pause dur="0.6"/> speak to you basically <pause dur="0.4"/> they're they're not done if anyone i don't know if anyone's filled in a tax return <pause dur="0.4"/> but they're not <pause dur="1.2"/> they're not # <pause dur="1.7"/> just <pause dur="1.0"/> full of information they're not <trunc>f</trunc> just far too long but <pause dur="0.6"/> part of the

length is due to the fact that they're trying to explain things to you to the sort of nth degree <pause dur="0.5"/> and so you get things like this this is a <pause dur="0.4"/> a <pause dur="0.4"/> a <trunc>th</trunc> the the sort of spirit of it <pause dur="0.7"/> you could say <pause dur="0.4"/> instead of machinery purchased before grant ninety-seven <pause dur="0.4"/> please indicate below the approximate value of any major items of machinery <pause dur="0.9"/> before grant what the tax return approach is <pause dur="0.8"/> says <pause dur="0.3"/> in the question below i'm asking you to think back over the past <pause dur="0.2"/> the tax return talks to you as if it's a person or a friend of yours it's dreadful <pause dur="0.4"/> it says now i would like you to turn the page <pause dur="0.9"/> well if i've got to the bottom of the page i probably would have worked that out for myself <pause dur="0.5"/> in the question below i'm asking you to think back <pause dur="0.3"/> and so on <pause dur="1.1"/> # <pause dur="0.2"/> so you <trunc>s</trunc> you just see the differences <pause dur="2.2"/><kinesic desc="changes slide" iterated="n"/> this is the double-barrelled thing did you drink coffee with lunch and tea yesterday <pause dur="0.4"/> do you mean with lunch and tea or do you mean with lunch or tea <pause dur="0.4"/> or do you mean did you drink coffee with lunch and did you

drink coffee with tea <pause dur="3.2"/> at tea <pause dur="1.4"/><kinesic desc="changes slide" iterated="n"/> # avoid embarrassing questions obviously <pause dur="0.3"/><vocal desc="laughter" n="ss" iterated="y" dur="2"/> or at least find a way of asking them <pause dur="1.2"/> so a sort of summary <pause dur="0.2"/> of the way you can <pause dur="0.5"/> design these questions <pause dur="1.5"/> they've got to be simple you've got this in front of you they've got to be simple <pause dur="0.2"/> direct familiar <pause dur="0.4"/> words used should be clear specific avoid ambiguity <pause dur="0.7"/> cover one point with one question so don't say are you satisfied with the cost and convenience of this product you've got to <pause dur="0.4"/> ask that separately <pause dur="1.3"/> avoid <pause dur="0.2"/> leading or loaded questions and <pause dur="0.2"/> <trunc>emo</trunc> emotional questions find ways of asking them <pause dur="0.4"/> if you can't ask them in the questionnaire then <pause dur="0.3"/> you maybe have to resort to qualitative research <pause dur="0.6"/> which we'll look at in a minute some other ways of collecting that data <pause dur="1.1"/> <unclear>short</unclear> questions that are applicable to respondents avoid very complex instructions <pause dur="0.3"/> and various sorts of bias <pause dur="0.7"/> check and recheck the length of it <pause dur="0.8"/> people are naturally going to be put off by anything very long it depends who they are what you're asking them

about <pause dur="0.8"/><kinesic desc="changes slide" iterated="n"/> and make sure it reads well and it's presented well as i say i don't think these days there's any excuse for poor presentation of <pause dur="0.5"/> questionnaires because we've all <pause dur="0.4"/> more or less got the tools to do it <pause dur="0.2"/> properly <pause dur="2.0"/> so i think it's important it does affect response rate <pause dur="0.9"/> if you it depends of course what you're asking but sort of consumer research if you get a response rate from a <pause dur="0.2"/> postal questionnaire of ten per cent you're relatively happy <pause dur="1.7"/> doesn't sound a lot but <pause dur="0.4"/> you're relatively happy with that <pause dur="0.2"/> if it's up to sort of twenty or twenty-five that's pretty good <pause dur="2.0"/>i think we'll take a short break there are there any questions <pause dur="0.7"/> take a short break and come back to how you'd code this and i'll give you some other examples before we look at <pause dur="0.5"/><kinesic desc="changes slide" iterated="n"/> qualitative research <pause dur="12.2"/>

right just to finish off the discussion of questionnaires we <pause dur="1.8"/> have collected the data let's assume that <pause dur="3.0"/> when i first <pause dur="0.4"/><vocal desc="laugh" iterated="n"/> <pause dur="0.4"/> when i first arrived here <pause dur="0.2"/> some years ago <pause dur="0.6"/> there was a unit called the data processing unit in our department <pause dur="0.7"/> and they still operated <pause dur="0.5"/> or there was one at that time there was one P-C in fact it was a Macintosh in the department <pause dur="1.6"/> and i think it had <pause dur="1.0"/> it had a hundred-and-twenty-eight K of <pause dur="0.6"/> RAM or something <pause dur="1.1"/> # <pause dur="1.0"/> anyway <pause dur="1.2"/> there was the data processing unit and they they still operated with punched cards <pause dur="0.6"/> and quite a lot of the terminology of <pause dur="0.2"/> market research comes from <pause dur="0.5"/> that <pause dur="0.7"/> that # <pause dur="0.6"/> period and you still got talk about people punching data in <pause dur="0.6"/> which is <pause dur="0.3"/> entering data in <pause dur="0.7"/> common parlance but these punch cards were <pause dur="0.3"/> exactly that little cards with holes in them <pause dur="0.9"/> # <pause dur="0.3"/> and i remember seeing them do do a sort on this data so you had <pause dur="0.3"/> a hundred cards each of which represented a respondent <pause dur="0.5"/> and the sorting process if you wanted to <pause dur="0.4"/> take out the males or the females # <pause dur="0.2"/> was more or

less a <trunc>s</trunc> <pause dur="0.3"/> a simplification but it <pause dur="0.3"/> it more or less involved sticking a sort of needle through <pause dur="0.5"/> the the hole that represented male <pause dur="0.2"/> and all the female cards dropped out and you had the <pause dur="0.6"/> the male cards were hanging on this little pole <pause dur="0.7"/> # <pause dur="1.1"/> nowadays we're a bit more sophisticated than that as you know <pause dur="0.9"/> but it is important to get the questionnaire properly coded it's very important that <pause dur="0.4"/> the <pause dur="0.3"/> as i i mentioned with respect to multiple choice questions <pause dur="0.4"/> # it's very important that <pause dur="0.2"/> the data is entered correctly <pause dur="1.5"/> # <pause dur="0.2"/> and in a way that can be <pause dur="0.7"/> relatively easily understood <pause dur="0.5"/> and when we looked at S-P-S-S in the first week <pause dur="0.7"/> i said things like making sure the variable names are sensible <pause dur="0.3"/> making sure the descriptions of those variables are then even more sensible so you can <pause dur="0.4"/> remember and understand what you're analysing is <trunc>v</trunc> very important <pause dur="0.7"/> var one to var twenty is not actually very useful for anybody <pause dur="1.6"/> but multiple choice and factual questions basically <pause dur="0.2"/> code themselves as you've seen <pause dur="1.5"/> # <pause dur="8.4"/><kinesic desc="changes slide" iterated="n"/> you've got little

coding <pause dur="0.7"/> numbers in there and that's quite standard on a questionnaire <pause dur="0.3"/> and they they code themselves and you use those numbers <pause dur="0.3"/> in this case <pause dur="0.4"/> # <pause dur="1.2"/> again if it's if it's a single answer in this case that will just be a number representing <pause dur="0.3"/> that answer it's simply a label it's nominal data <pause dur="4.4"/><kinesic desc="changes slide" iterated="n"/> factual questions code themselves <pause dur="0.2"/> how many kilometres of hedges have you removed <pause dur="0.6"/> they just give you a number <pause dur="0.6"/> open-ended <pause dur="0.3"/> as i said they need a coding frame in advance or <pause dur="0.3"/> ex-post codings after you've got all the <pause dur="0.3"/> results in you sort things into categories <pause dur="0.5"/> with open-ended questions and indeed with qualitative data there are <pause dur="0.2"/> bits of software that can help you with this there's a thing called Nudist you may have come across <pause dur="0.4"/> which is for classifying qualitative data <pause dur="0.8"/> largely for classifying <pause dur="0.5"/> data which is in the form of <pause dur="0.6"/> # transcriptions of recordings interviews focus groups or whatever <pause dur="1.2"/> and what Nudist basically does is pick out <pause dur="0.6"/> key words and organise them in a sort of hierarchy <pause dur="0.8"/> to try to

establish what are the things that are most important to people <pause dur="1.9"/>

<kinesic desc="changes slide" iterated="n"/> dichotomous use basically done in variables ones or zeros <pause dur="1.4"/> # <pause dur="0.9"/> so there's the questionnaire <pause dur="0.4"/> you'll see <pause dur="0.8"/><kinesic desc="changes slide" dur="3" iterated="y"/> just remind you of the things that we <pause dur="0.5"/> were asking <pause dur="0.3"/> so you've got <pause dur="0.8"/> the questionnaire number the interviewer the date the time <pause dur="0.5"/> # how far have you travelled what form of transport did you use <pause dur="0.3"/> what that looks like <pause dur="2.3"/> is that when you collect <pause dur="0.6"/> remember each <pause dur="0.3"/> row is a <pause dur="0.4"/> an observation each row is a person <pause dur="0.4"/> so you've got code number name date time <pause dur="0.3"/> and then you have to start getting into these <pause dur="0.2"/> variable names which <pause dur="0.5"/> i think it it ought to be the next <pause dur="0.7"/> innovation in computing i <shift feature="voice" new="laugh"/>think <pause dur="0.4"/> is that they ought to be able<shift feature="voice" new="normal"/> to let variable names go over eight letters and S-P-S-S <pause dur="0.5"/> the next versions i think should work on that <pause dur="0.2"/> because <pause dur="0.2"/> how far and transport and <pause dur="0.8"/> M-U-M-U-eighteen <pause dur="2.5"/> which must be number of <pause dur="0.5"/> can't even remember </u><u who="ss" trans="latching"> <gap reason="inaudible, multiple speakers" extent="3 secs"/></u><u who="nm0976" trans="latching"> number of people under eighteen with yeah okay so <pause dur="0.7"/> mm <pause dur="0.5"/> it was a few years ago but you do

forget and and you need to try and keep those as sensible as possible <pause dur="0.5"/> but as i mentioned <pause dur="0.4"/> it's <pause dur="0.8"/> variables across the top <pause dur="0.4"/> observations down the side <pause dur="0.5"/> # and then you'll be <pause dur="1.1"/> # <pause dur="1.8"/> safe in in analysing this <pause dur="0.3"/> see some ninety-nines in there that's missing data <pause dur="0.8"/> as long as ninety-nine is not a valid observation you can use something like ninety-nine <pause dur="0.4"/> for missing data <pause dur="0.3"/> S-P-S-S incidentally <pause dur="0.4"/> prefers you to use a <pause dur="0.4"/> # full stop <pause dur="0.3"/> for missing data <pause dur="1.0"/> you could always get Excel to change the ninety-nines to full stops <pause dur="0.6"/> again i think this is probably harking back to the old punch card days when i think minus-ninety-nine was the thing that they used <pause dur="0.4"/> as something that would be unlikely to occur anywhere else <pause dur="0.6"/> and so all the variables are across there <pause dur="1.2"/> there was one <pause dur="1.4"/> # <pause dur="0.3"/> open-ended question and and you see they've <pause dur="1.2"/> the researchers have noted down <pause dur="0.2"/> the <pause dur="0.3"/> responses basically to the open-ended question but that needs some sort of post <pause dur="0.5"/> # analysis <pause dur="7.8"/><kinesic desc="changes slide" iterated="n"/> i'm not going to <trunc>c</trunc> this is <pause dur="0.3"/> part of this presentation but

<gap reason="name" extent="1 word"/> was basically talking to you about this last time <pause dur="0.4"/> scaling being used to estimate <pause dur="0.3"/> the extent of somebody's predisposition to act or their attitude <pause dur="0.7"/> # <pause dur="0.3"/> and <pause dur="0.6"/> i think you covered that <pause dur="2.2"/> are there any <pause dur="0.3"/> questions on how you <pause dur="0.5"/> ask questions <pause dur="5.6"/> right we'll move on and in some ways this is moving back <pause dur="0.2"/> because <pause dur="0.5"/> # <pause dur="2.5"/> what i want to talk about now is <pause dur="0.5"/> qualitative <pause dur="0.4"/> getting getting hold of qualitative information <pause dur="4.1"/> there are <pause dur="0.3"/> a number of different ways these aren't in your notes but i will add them to the web site <pause dur="1.1"/> there are a number of ways <pause dur="0.5"/> does anyone get frustrated with this paperclip by the way </u><pause dur="0.7"/><u who="ss" trans="pause"> mm </u><pause dur="0.7"/> <u who="sm0985" trans="pause"> <gap reason="inaudible" extent="1 sec"/> </u><u who="nm0976" trans="overlap"> it needs to be turned off and <pause dur="1.3"/> i never know how to do it so i'll just get it out of the way <pause dur="10.6"/><event desc="closes computer program" iterated="n"/> # <pause dur="2.4"/> qualitative research as i say <pause dur="0.3"/> is quite often used as a precursor to quantitative research <pause dur="2.9"/><kinesic desc="changes slide" iterated="n"/><event desc="noise from computer" iterated="n"/> partly to reassure yourself that you're asking the right questions <pause dur="0.8"/> so you use <pause dur="0.2"/> qualitative research in a in an exploratory way <pause dur="0.7"/> and <pause dur="0.3"/> it helps to sort of frame the <pause dur="0.2"/> types of questions you're going to ask <pause dur="6.9"/> there

are essentially three <pause dur="0.2"/> broad methods <pause dur="0.5"/> of <pause dur="0.6"/> or broad schools of thought with respect to qualitative research <pause dur="0.4"/> and by far the most important method <pause dur="0.4"/> and of course these are sort of <pause dur="0.6"/> # themes as well is the whole <pause dur="0.3"/><kinesic desc="writes on board" iterated="y" dur="3"/> idea surrounding <pause dur="0.3"/> focus groups <pause dur="0.3"/>and focus groups are by far these days by far the most <pause dur="0.5"/> # <pause dur="1.1"/> used <pause dur="0.9"/> form of qualitative research <pause dur="0.6"/> but you also <pause dur="2.2"/> hear of with <trunc>y</trunc> <trunc>the</trunc> there are also <pause dur="0.8"/><kinesic desc="writes on board" iterated="y" dur="6"/> there's also the possibility of conducting depth interviews <pause dur="3.7"/> which i'll say a little bit about in a moment and then there's a a <pause dur="1.1"/> area of research called motivational research <pause dur="4.6"/><kinesic desc="writes on board" iterated="y" dur="8"/> which in some ways can be used <pause dur="0.5"/> as part of these two <pause dur="2.9"/> so we'll just deal with <pause dur="0.5"/> we'll deal with these first get them out of the way and then talk about focus groups <pause dur="2.9"/> just to repeat the <pause dur="0.5"/> objective of carrying out <pause dur="0.5"/> # <pause dur="1.2"/> qualitative research <pause dur="0.4"/> is <pause dur="0.2"/> really to generate ideas <pause dur="0.8"/> things like <pause dur="0.5"/> background information on a problem or a topic <pause dur="1.7"/> identifying concepts so concept identification i will as i said <pause dur="0.5"/> put some notes up that refer to this <pause dur="1.7"/> identifying relevant behaviour patterns and themes in the way people's <trunc>b</trunc> people behave <pause dur="1.3"/> establishing priorities how people <pause dur="0.5"/>

rate certain things or how important things are to people in a qualitative way <pause dur="1.0"/> preliminary screening focus groups and qualitative research are used a lot <pause dur="0.3"/> in <pause dur="0.2"/> screening product ideas <pause dur="3.1"/> new product development as you'll hear from <gap reason="name" extent="2 words"/> in due course is a <pause dur="0.7"/><vocal desc="laugh" iterated="n"/><pause dur="0.7"/> is a sort of inexact science in that they use very detailed scientific methods <pause dur="0.7"/> # but <pause dur="0.3"/> it's said in the food industry that eighty per cent of new product introductions fail <pause dur="0.9"/> # now not all of those will reach the <pause dur="0.2"/> the retailer <pause dur="0.4"/> Marks and Spencer's are <pause dur="0.5"/> famous for example of actually not test marketing <pause dur="0.8"/> Marks and Spencer's basically test market by putting <pause dur="0.2"/> products in store <pause dur="1.0"/> and so <pause dur="0.2"/> new product ideas like <pause dur="0.4"/> # <pause dur="0.7"/> M and S are famous for ready meals and so on they they basically put them in store see how they <pause dur="0.3"/> go and if they're no good they pull them again <pause dur="0.7"/> so there's as it it's a relatively <pause dur="0.4"/> # <pause dur="0.5"/> inexact science <pause dur="0.9"/> but a lot of products fail <pause dur="2.0"/> focus groups however are used sorry <pause dur="0.6"/> # qualitative research <pause dur="0.3"/> however is used to try

to <pause dur="0.4"/> # ensure that that doesn't happen <pause dur="0.3"/> so screening of products <pause dur="0.3"/> and concept testing is very important <pause dur="0.4"/> # area <pause dur="1.0"/> of this <pause dur="0.7"/> also post research investigations in other words to sort of back up the results of a survey <pause dur="0.4"/> you might well conduct some further qualitative research <pause dur="1.2"/> piloting as i mentioned <pause dur="0.3"/> my pilot <pause dur="0.4"/> questionnaires during the qualitative phase <pause dur="1.1"/> and as i also said <pause dur="0.8"/> # talking about personal issues and personal questions <pause dur="0.4"/> they will be easier to do with some form of qualitative research <pause dur="0.6"/> now <pause dur="0.6"/> depth interviews just to get them out of the way <pause dur="0.2"/> depth interviews are exactly as they sound <pause dur="0.5"/> # <pause dur="1.7"/> relatively long <pause dur="0.4"/> relatively <pause dur="0.4"/> # <pause dur="1.2"/> well very personal in other words it's just one person plus the interviewer <pause dur="0.7"/> trying to uncover people's often people's motivations people's attitudes <pause dur="0.4"/> and so on <pause dur="0.4"/> highly expensive of course 'cause it's very intensive you need a <pause dur="0.4"/> highly trained interviewer to do it to do it <pause dur="0.3"/> and extremely difficult to evaluate <pause dur="0.4"/> the consequences because <pause dur="0.4"/> each respondent basically has a <pause dur="0.4"/> a write-up of a rather <pause dur="0.4"/> conventionally rather long discussion <pause dur="0.6"/> so depth interviews <pause dur="0.3"/> i think to

some extent are <pause dur="0.4"/> waning in in # <pause dur="1.0"/> popularity amongst market researchers <pause dur="0.8"/> focus groups are not <pause dur="0.6"/> the Blair government <pause dur="0.5"/> # <pause dur="0.4"/> has popularized focus group use in the <shift feature="voice" new="laugh"/> U-K <shift feature="voice" new="normal"/><pause dur="0.6"/> in the sense that they were very <pause dur="0.6"/> # <pause dur="0.4"/> open about having used focus groups to try to find out what sort of image they ought to put across <pause dur="0.3"/> so things like image <pause dur="0.6"/> image of a political party is something that lends itself very well <pause dur="0.4"/> to a focus group discussion <pause dur="2.2"/> a focus group and i'll show you two examples in a moment <pause dur="0.6"/> is <pause dur="2.8"/> usually a relatively small number of people <pause dur="0.5"/> between <pause dur="0.2"/> six and ten <pause dur="1.0"/> eight being the <pause dur="0.3"/> probably the optimum in most cases <pause dur="7.2"/> participants are brought together <pause dur="0.4"/> under the direction of a <pause dur="0.3"/> group leader of some sort <pause dur="0.9"/> # <pause dur="1.7"/> a trained person who is usually <pause dur="0.4"/> referred to as the moderator <pause dur="0.5"/> of the focus group <pause dur="2.4"/> in fact to call them the leader is probably wrong <pause dur="0.7"/> they are the moderator they're trying to moderate a discussion between <pause dur="0.8"/> eight people <pause dur="1.0"/> and <pause dur="0.2"/> moderation means that they shouldn't really be <pause dur="0.4"/> too involved in it ideally <pause dur="0.6"/> a good

focus group would <pause dur="1.1"/> last for an hour and a half and the moderator wouldn't <trunc>n</trunc> need to intervene at all <pause dur="0.8"/> that doesn't <pause dur="0.4"/> often happen because there needs to be some introduction there needs to be some direction <pause dur="0.4"/> as you'll see in a second <pause dur="1.7"/> the group <pause dur="0.5"/> the <trunc>m</trunc> the moderator <pause dur="0.2"/> should be just that they shouldn't intervene in the discussion they're not <pause dur="0.3"/> there to be part of the discussion <pause dur="0.8"/> quite often <pause dur="0.6"/> other members of a research team want to be involved in the focus group they'll say oh i'll sit in <pause dur="0.6"/> but that <pause dur="0.3"/> can distort the discussion as well it <trunc>s</trunc> <pause dur="0.4"/> it should be very genuinely a <pause dur="0.3"/> an independent discussion between eight <pause dur="0.3"/> people who are <pause dur="0.4"/> # <pause dur="0.8"/> genuine respondents <pause dur="4.9"/> the leader has various sorry the moderator has various <pause dur="0.4"/> # functions <pause dur="1.1"/> the moderator has got to <pause dur="0.2"/> make sure the discussion follows the route <pause dur="0.5"/> that it's supposed to <pause dur="0.3"/> but <pause dur="1.0"/> they need to do this in a relatively subtle way <pause dur="0.6"/> we did some work as you'll see <pause dur="0.8"/> looking at <pause dur="0.8"/> the attitudes of consumers to British meat products <pause dur="1.6"/> # now <pause dur="0.7"/> we were actually

looking at the attitudes of Belgian consumers to British meat products <pause dur="0.6"/> # for the Meat and Livestock Commission <pause dur="0.3"/> who <pause dur="0.4"/> at the time of the B-S-E <pause dur="0.4"/> the height of the B-S-E affair were facing <pause dur="0.4"/> an export ban on beef and therefore were looking at <pause dur="0.5"/> the <pause dur="0.4"/> # prospects for other <pause dur="0.9"/> products like <pause dur="0.5"/> # lamb and pork so we were looking particularly at those products <pause dur="1.2"/> # <pause dur="2.2"/> we didn't <pause dur="0.7"/> we or the the research team did not go straight into a focus group and say right we're going to talk about British meat <pause dur="1.7"/> because what we wanted to do was to get people's confidence and to build up <pause dur="0.4"/> their <pause dur="0.3"/> # or build up rapport with them so # okay they knew it was about food consumption <pause dur="0.6"/> # <pause dur="0.2"/> what you ideally want in a focus group <pause dur="0.3"/> is for people to lead you <pause dur="0.7"/> so if you're the moderator and there's a group of people <pause dur="0.5"/> chatting away about something and somebody says <pause dur="0.5"/> oh i'm not very keen on British meat <pause dur="1.4"/> that is your cue to bring them in and to <trunc>s</trunc> start the <pause dur="0.2"/> the subject going on rather than saying right it's <pause dur="0.4"/> you know it's half past twelve now we're going to talk about British meat <pause dur="0.4"/> you want

to let people <pause dur="0.5"/> # <pause dur="0.8"/> guide you in some ways and the and the skill of the moderator's in recognizing those signs and making sure that <pause dur="0.4"/> # they're picked up on and the the discussion is <pause dur="0.5"/> # <pause dur="0.7"/> or follows a <pause dur="0.2"/> a <pause dur="0.6"/> particular course <pause dur="2.4"/> focus groups are popular for lots of <pause dur="0.7"/> perfectly valid reasons <pause dur="1.3"/> in some ways they're relatively cheap <pause dur="2.2"/> certainly compared with individual interviews it's relatively cheap to get eight people together <pause dur="0.3"/> even so <pause dur="0.5"/> the going rate at the moment for <pause dur="0.7"/> participation in focus group is about twenty-five quid <pause dur="0.5"/> we're trying to recruit some people at the moment and <pause dur="0.3"/> twenty-five quid <pause dur="0.4"/> is # <pause dur="0.3"/> seen as <pause dur="0.8"/> a reasonable rate to pay people for a <pause dur="0.3"/> couple of hours of their time <pause dur="0.3"/> if you're doing this for a professional agency you probably also <pause dur="0.5"/> pay travel expenses and so on <pause dur="0.4"/> you've got to <pause dur="0.4"/> hire a <pause dur="0.4"/> facility <pause dur="1.1"/> there's one in Reading called Sight and Sound <pause dur="0.5"/> typical facility would be <pause dur="0.6"/> probably about the a room about this size but <pause dur="0.3"/> from here to the wall would be split off <pause dur="0.4"/> and there'd be a big <pause dur="0.5"/> # single

way mirror <pause dur="0.4"/> in the wall <pause dur="0.5"/> and <pause dur="0.2"/> the research group can sit behind the mirror <pause dur="0.4"/> and observe what's going on in the <pause dur="0.3"/> in the room sometimes video is used <trunc>m</trunc> more often than not <pause dur="0.4"/> video is used either in addition to that or instead of it <pause dur="0.9"/> so focus groups are <pause dur="0.4"/> tape recorded they're videoed <pause dur="0.4"/> and <pause dur="0.5"/> # this this facility in Reading is extremely good and you can <pause dur="0.3"/> you can watch as a member of the research team <pause dur="0.6"/> market research code of ethics <pause dur="0.5"/> says that people have to be informed <pause dur="0.2"/> that the research team is sitting behind the mirror <pause dur="0.9"/> # which i think is fair enough and <pause dur="0.7"/> as as the <pause dur="0.6"/> moderator says this they say oh <pause dur="0.4"/> you know you ought to know that behind that wall there's a few people there <trunc>obser</trunc> <trunc>a</trunc> and <pause dur="0.7"/> focus group goes completely quiet and oh my God you know can't <pause dur="0.3"/> you know can't put up with that and within five minutes they've completely forgotten about it <pause dur="0.4"/> as is as they've forgotten about the fact they're being videoed <pause dur="0.9"/> # <pause dur="0.7"/> and then after two hours you sort of introduce yourselves to them and like

oh yeah i forgot about you <pause dur="0.9"/> # <pause dur="1.8"/> focus groups <pause dur="1.1"/> have <pause dur="0.2"/> these important advantages <pause dur="0.3"/> the main advantage of a focus group over an individual <pause dur="0.2"/> interview is that <pause dur="0.5"/> there is <pause dur="0.5"/> interaction <pause dur="0.9"/> people stimulate ideas off each other <pause dur="1.5"/> and that is the main <pause dur="0.8"/> # <pause dur="0.5"/> the main advantage people are <pause dur="0.3"/> interacting with each other it's very dynamic <pause dur="0.3"/> they can <pause dur="0.5"/> # <pause dur="0.6"/> prompt each other they can decide <pause dur="1.0"/> # <pause dur="0.7"/> sorry they can they can # <pause dur="0.2"/> <unclear><trunc>grou</trunc></unclear> or the group dynamic means that the <pause dur="1.1"/> # information you gain tends to be richer <pause dur="2.7"/> it fosters creativity <pause dur="1.1"/> and they start thinking provided it's well moderated <pause dur="0.2"/> there's nothing worse than a focus group that is <pause dur="0.4"/> # <pause dur="1.6"/> badly <trunc>mor</trunc> moderated i want to show you <pause dur="0.3"/> two examples one that i consider is reasonable and one that's <pause dur="0.5"/> # not so good <pause dur="0.8"/> # <pause dur="1.4"/> but <pause dur="0.2"/> yeah the <trunc>m</trunc> the moderator is there to guide it <pause dur="0.2"/> moderator's also there to bring people out who are not <pause dur="0.7"/> participating <pause dur="0.3"/> or <pause dur="0.4"/> to <pause dur="0.4"/> shut somebody up who is participating too much <pause dur="0.4"/> and it's a question of <pause dur="0.3"/> trying to get the views of everybody <pause dur="1.4"/> good focus group <pause dur="0.3"/> spontaneity candour

can produce really good results <pause dur="0.5"/> after the focus group you need to write the thing up it needs a <pause dur="0.4"/> either a full transcription or some kind of summary of it <pause dur="0.4"/> # <pause dur="1.3"/> the <pause dur="0.2"/> the people who did this work in Belgium were <pause dur="0.3"/> they in fact conducted ten focus groups five in the French speaking part <pause dur="0.5"/> of Belgium five in the English speaking part <pause dur="0.2"/> sorry in the Flemish speaking part <pause dur="0.6"/> and <pause dur="0.4"/> # <pause dur="2.5"/> the write-up of ten focus groups is <pause dur="0.3"/> not a <shift feature="voice" new="laugh"/> trivial task<shift feature="voice" new="normal"/> these things some of them lasted three hours <pause dur="0.4"/> so <pause dur="0.4"/> it's it's <pause dur="0.2"/> it's a fairly large <pause dur="0.3"/> # undertaking <pause dur="0.3"/> but <pause dur="0.2"/> the information you can get is very rich <pause dur="0.4"/> what you can't say is that we conducted ten focus groups <pause dur="0.3"/> there were eight people in each one and therefore <pause dur="0.3"/> # we've got eighty respondents and that's a valid sample and you start doing statistical analysis that is not the idea <pause dur="0.6"/> it's exploratory only it's <pause dur="0.2"/> giving ideas and the output that they this group produced <pause dur="0.5"/> for the Meat and Livestock Commission were <pause dur="0.4"/> just basic ideas on how <pause dur="0.3"/> you might market <pause dur="0.3"/> British <pause dur="0.7"/> # <pause dur="0.9"/> lamb and pork in

Belgium <pause dur="1.4"/> i'll show you first of all what <pause dur="0.2"/> i think is not the best example <pause dur="0.4"/> # of how to conduct a focus group <pause dur="0.4"/> can you <pause dur="0.3"/> <gap reason="name" extent="1 word"/> can you pull that blind right down and could you pull <pause dur="0.4"/> those blinds down as well please </u><pause dur="55.7"/><event desc="moves blinds" n="ss" iterated="y" dur="1:05"/> <event desc="prepares video" n="nm0976" iterated="y" dur="1:35"/> <u who="nm0976" trans="pause"> right now i can't see what is happening so <vocal desc="laughter" n="sl" iterated="y" dur="2"/></u> <pause dur="35.8"/> <kinesic desc="starts video" n="nm0976" iterated="n"/><kinesic desc="video plays" n="nm0976" iterated="y" dur="3:38"/> <u who="nm0976" trans="pause"> that was inspired timing they start talking about S-P-S-S <pause dur="0.4"/> this was # <pause dur="0.3"/> work that was conducted actually for Microsoft <pause dur="1.3"/> trying to find out i think this <trunc>w</trunc> one is in Birmingham you probably <pause dur="0.2"/> judge by the accents in a minute <pause dur="0.7"/> # <pause dur="0.6"/> and it was trying to find out what people <pause dur="0.2"/> thought of Microsoft software and it was a couple of years ago so some of it's a bit <pause dur="0.8"/> a bit dated <pause dur="10.5"/> you'll find <pause dur="0.3"/> i think that the room looks a bit cramped it's not <pause dur="0.2"/> an ideal location <pause dur="0.6"/> and i don't know if i've picked out the right bit but you'll <pause dur="0.4"/> you'll find that the moderator <pause dur="0.2"/> i think tends to be very <pause dur="0.3"/> intrusive and <pause dur="0.3"/> he certainly at points leads the respondents on <pause dur="0.5"/> and it's more like a question and answer session really than a discussion <pause dur="8.8"/>

this is the moderator standing at the front bit like a sort of school teacher <pause dur="0.2"/> writing things up on a <pause dur="0.7"/> # <pause dur="0.6"/> flipchart <pause dur="16.2"/> but you find that <pause dur="0.2"/> each time <pause dur="0.3"/> one of them talks they're talking to the moderator which is not the idea at all they should be talking to each other <pause dur="12.6"/> moderator has a rather tedious voice as well which doesn't i think help <pause dur="12.4"/> but that's a sort of typical scene <pause dur="0.4"/> little rules simple things like you don't <pause dur="1.9"/> it's better not to serve alcohol <pause dur="1.5"/> because after two or three hours particularly with groups of men <pause dur="1.0"/> # you tend to get slightly <pause dur="1.2"/> excited <pause dur="0.2"/> respondents and it doesn't necessarily <trunc>ne</trunc> <pause dur="0.2"/> lead to good answers <pause dur="0.8"/> focus groups with men are <pause dur="0.5"/> decidedly more difficult than focus groups with women particularly on food <pause dur="0.6"/> # <pause dur="0.6"/> what we found <pause dur="0.6"/> with this Belgium work for example we were looking at <pause dur="0.5"/> among other things we were looking at ready meals <pause dur="1.5"/> and <pause dur="0.4"/> we were asking for people's impressions <pause dur="0.2"/> of <pause dur="0.3"/> packages of <pause dur="0.2"/> ready meal <pause dur="0.3"/> pork and and lamb <pause dur="0.8"/> and # <pause dur="2.2"/> women <pause dur="1.8"/> i'm i'm now doing what i told you not to do i'm

generalizing over ten focus groups when you <pause dur="0.7"/> you're not supposed to do that but <pause dur="0.4"/> in general the Belgian women <pause dur="1.0"/> # <pause dur="0.6"/> talked about the product and talked about <pause dur="0.3"/> you know what they would do with it how they would cook it and the sort of image and so on the ideas that <pause dur="0.5"/> # it <pause dur="0.3"/> gave them <pause dur="0.4"/> the men <pause dur="0.8"/> the Belgian men <pause dur="0.2"/> and we had two groups of men were <pause dur="0.2"/> obsessed with the package and technical details of the <shift feature="voice" new="laugh"/>package<shift feature="voice" new="normal"/> and <pause dur="0.4"/> oh yeah this is <pause dur="0.2"/> gas atmospheric packaging and you know and it <pause dur="0.4"/> it rises here because there's <pause dur="0.3"/> the product is <pause dur="0.3"/> you know two or three days old and it's giving off some blah blah blah <pause dur="0.4"/> and and we got <pause dur="0.5"/> quite distorted results because <pause dur="0.2"/> the males didn't appear to want to talk about the <pause dur="0.4"/> product itself and the image which is what we were looking for <pause dur="1.5"/> okay that's one example you see they're very static the people are very static the moderator is standing there <pause dur="0.4"/> and and making notes <pause dur="0.2"/> now this is <pause dur="0.7"/> an alternative approach now </u><pause dur="6.1"/><event desc="stops video" iterated="n" n="nm0976"/><u who="sm0980" trans="pause"> would that have been because of <pause dur="0.7"/> the people there didn't know each other # # they were not comfortable speaking because </u><u who="nm0976" trans="overlap"> i think </u><u who="sm0980" trans="overlap"> although some people <gap reason="inaudible" extent="4 secs"/></u> <u who="nm0976" trans="overlap"> yeah </u><u who="sm0980" trans="overlap"> <gap reason="inaudible" extent="4 secs"/></u><pause dur="0.4"/> <u who="nm0976" trans="pause"> you you would aim to include people who don't know each other

in the sort of random sampling <pause dur="0.5"/> kind of ideal <pause dur="0.5"/> # <pause dur="0.7"/> what <pause dur="1.6"/> quite often when we've done this on student projects and things <pause dur="0.3"/> and and it <pause dur="0.3"/> moderation certainly takes <pause dur="0.4"/> a bit of learning and so you're throwing somebody in at the deep end if you ask them to moderate a focus group <pause dur="0.6"/> without having sort of practised it or not done it before <pause dur="0.7"/> and <pause dur="0.4"/> people are most nervous about silence <pause dur="1.3"/> and <pause dur="0.2"/> you're right to start with every focus group is a bit stilted <pause dur="0.3"/> and you certainly start by sort of going round the room saying who are you and maybe asking a particular question of people <pause dur="0.6"/> # <pause dur="0.3"/> but <pause dur="0.5"/> more often than not people get talking <pause dur="0.7"/> # <pause dur="0.2"/> and that that's what you want </u><pause dur="0.2"/> <u who="sm0981" trans="pause"> i suppose in the first guy was speak on like in the experience which you had in Belgium <pause dur="0.6"/> is that speaker speaking about <pause dur="0.2"/> this chemical reaction that's what has this <pause dur="0.3"/> thing to swell <gap reason="inaudible" extent="4 secs"/></u><u who="nm0976" trans="overlap"> <trunc>i</trunc> <pause dur="0.8"/> yeah no it does happen <pause dur="0.5"/> yeah </u><u who="sm0981" trans="overlap"> <gap reason="inaudible" extent="1 sec"/> your comments then become <gap reason="inaudible" extent="3 secs"/></u><u who="nm0976" trans="overlap"> sure sure <pause dur="0.8"/> but it's the role of the moderator to try and bring people in <pause dur="0.2"/> and to <pause dur="0.3"/> to make sure it's a genuine discussion</u><pause dur="0.4"/> <u who="sm0982" trans="pause"> <gap reason="inaudible" extent="2 secs"/> </u> <pause dur="0.4"/> <u who="nm0976" trans="pause"> pardon <vocal desc="laughter" n="ss" iterated="y" dur="2"/> what did he say <pause dur="0.2"/> were people taking <gap reason="inaudible" extent="1 word"/></u><u who="sm0982" trans="overlap"> <gap reason="inaudible" extent="5 secs"/></u><pause dur="0.9"/> <u who="nm0976" trans="pause"> # i don't think in that one they were drinking alcohol it's certainly <pause dur="0.7"/><vocal desc="laughter" n="ss" iterated="y" dur="1"/> # </u><u who="sm0983" trans="latching"> <gap reason="inaudible" extent="2 secs"/> # how how do you identify the respondents # <pause dur="0.3"/> especially given that it's small like a specialistic area</u><pause dur="0.8"/> <u who="nm0976" trans="pause">

it's i mean the process that you use if you're working <pause dur="0.4"/> there there is a there's a commercial <pause dur="0.4"/> # side to this if you working for a commercial agency then they would use a commercial recruiting <pause dur="0.6"/> firm <pause dur="0.7"/> # <pause dur="0.2"/> and how these firms work the one we used in Belgium had fifteen-thousand on their panel <pause dur="0.6"/> and they were people <pause dur="0.5"/> whose demographic characteristics and sort of attitudes had been measured <pause dur="0.7"/> so that when we said we want people of a certain type and <pause dur="0.5"/> you know we want people who definitely eat <pause dur="0.6"/> lamb and pork <pause dur="0.3"/> then they could find these people for us <pause dur="0.4"/> otherwise it's sort of tramping the streets and trying to <pause dur="0.4"/> trying to get people it depends <pause dur="1.2"/> quite often <pause dur="0.2"/> we've used # if we haven't used a dedicated facility we've used like training rooms in supermarkets or something so you've recruited somebody from the supermarket <pause dur="0.3"/> or sorry recruited people from the supermarket but again <pause dur="0.4"/> it's it's a fair question and you're trying to get people with certain characteristics and <pause dur="0.5"/> recruitment is actually far more difficult than running the thing <pause dur="1.1"/>

# <pause dur="2.1"/> here's a different example <pause dur="4.7"/><event desc="changes video" iterated="n"/> hope </u><pause dur="22.3"/><kinesic desc="starts video" iterated="n"/><kinesic desc="video plays" dur="31" iterated="y"/> <u who="nm0976" trans="pause"> this is something i want to mention in a minute the guy is handing out things to try to <pause dur="0.7"/> # <pause dur="2.4"/> he's actually handing out things and people are about to do <pause dur="0.7"/> have a little activity basically to sort of <pause dur="0.5"/> # <pause dur="0.3"/> reinforce the focus group let me just go back a bit and show you <pause dur="1.1"/> something else <pause dur="6.9"/><event desc="rewinds video" dur="53" iterated="y"/> i just wanted to emphasize the difference of approach for the <pause dur="0.3"/> moderator <pause dur="10.9"/> incidentally those of you who're familiar with sort of <pause dur="0.4"/> participatory research approaches <pause dur="0.5"/> this is this is very much related to that and and <pause dur="0.6"/> participatory approaches to <pause dur="0.5"/> # <pause dur="0.8"/> looking into particular issues is is <pause dur="0.2"/> very much uses these kinds of techniques sometimes adapted but <pause dur="0.5"/> focus groups or group <pause dur="0.4"/> activities are very <pause dur="0.5"/><kinesic desc="starts video" iterated="n"/><kinesic desc="video plays" dur="32" iterated="y"/> much to the fore now <pause dur="0.7"/> this guy <pause dur="0.7"/> insists on sitting on the floor basically # <pause dur="0.8"/> and his theory is that you should be <pause dur="0.6"/> at the same height or lower <pause dur="0.2"/> than your <pause dur="0.4"/> respondents now i just caught him when he was actually handing something out but <pause dur="0.6"/> the the Microsoft guy was sort of standing there like a

school teacher this guy <pause dur="0.5"/> is <pause dur="0.4"/> very much trying to interact with people <pause dur="1.7"/> just you just have to watch him for a little while to see his sort of style <pause dur="6.7"/> he's trying to bring someone else in <pause dur="8.6"/> i think this is probably early on he's sort of going around the the room a bit <pause dur="13.0"/> it was noticeable that he didn't sit on the floor with the male group <pause dur="4.5"/> he basically was <pause dur="0.5"/> chatting these ladies up i think # <pause dur="0.7"/> but they <pause dur="0.2"/> they certainly <vocal desc="laughter" n="su0987" iterated="y" dur="1"/> responded and the results were good <pause dur="12.1"/> very it's a very good facility this very very sort of comfortable and <pause dur="0.5"/> as you can see quite well <pause dur="0.6"/> appointed <unclear>at that</unclear> i'm not <pause dur="0.5"/> quite finding what i wanted to show you </u><pause dur="20.6"/><event desc="fast-forwards video" iterated="y" n="nm0976" dur="6"/><kinesic desc="starts video" iterated="n" n="nm0976"/><kinesic desc="video plays" iterated="y" dur="6:38"/> <u who="nm0976" trans="pause"> you'll see on it we might not find it but you'll see on occasions that this guy <pause dur="0.6"/> if <pause dur="0.2"/> later on in the group where people start talking <pause dur="0.6"/> at the moment he's going round the room but people start talking he sort of stops them and and he's very <pause dur="0.5"/> he's very # <pause dur="0.2"/> sort of interactive but very physical with people but it <trunc>s</trunc> actually seemed to work very well <pause dur="0.7"/> # <pause dur="0.6"/> extremely experienced <pause dur="0.6"/> market researcher <pause dur="0.5"/> now the thing well i'll just leave it going you can <pause dur="0.3"/> keep an eye on it <pause dur="0.4"/> # <pause dur="0.8"/> the final thing that i wanted

to <pause dur="0.2"/> mention was that <pause dur="1.7"/> a lot of <pause dur="1.1"/> qualitative research is impressionistic it is exploratory <pause dur="0.7"/> and <pause dur="0.6"/> there's a <pause dur="0.2"/> a a range of things under the title of motivational research <pause dur="0.3"/> that can be used to enhance <pause dur="0.7"/> group discussions of one sort or another and he was handing out cards <pause dur="0.3"/> where <pause dur="0.3"/> i think he got people into pairs to discuss a particular topic so it's sort of breaking things up and and <pause dur="0.4"/> there are various ways of doing that <pause dur="0.9"/> # <pause dur="3.2"/> so subgroup <pause dur="0.3"/> work or or # <pause dur="0.4"/> small group work is one way <pause dur="1.4"/> but motivational research was a sort of school of thought in the <pause dur="0.3"/> the U-S in the nineteen-fifties which <pause dur="1.0"/> took <pause dur="1.0"/> i mean there are there are lots also of sort of experimental ways of doing market research where you set up <pause dur="0.5"/> # <pause dur="1.4"/> you know # simulated shopping experiments and this kind of thing and experimental work is quite <pause dur="0.5"/> common but <pause dur="0.6"/> in the U-S in the in the fifties they really got into <trunc>ma</trunc> motivational research trying to uncover people's sort of hidden motivations <pause dur="0.5"/> there's lots of # <pause dur="0.3"/> references <pause dur="0.3"/> on the web site for that <pause dur="0.5"/> but <pause dur="0.2"/> just

one or two techniques that are used in <pause dur="0.5"/> focus groups and <kinesic desc="puts on transparency" iterated="n"/> other qualitative research to try to <pause dur="0.5"/> # enhance <pause dur="0.4"/> the <pause dur="1.5"/> the # experience <pause dur="6.7"/> things like <pause dur="0.6"/> simply <pause dur="1.8"/> oh still leave them going behind there <pause dur="0.4"/> is just sentence completion test this is <pause dur="0.5"/> i think it has two things it collects data <pause dur="0.3"/> it enables us to collect data <pause dur="0.3"/> but it also <pause dur="0.3"/> # <pause dur="0.7"/> gives people something to do and it breaks up the discussion and <pause dur="0.3"/> and if they've <pause dur="0.2"/> they've filled in this sort of thing <pause dur="0.3"/> they can then <pause dur="0.2"/> <trunc>s</trunc> <pause dur="0.4"/> sort of carry on with the discussion <pause dur="3.6"/><kinesic desc="changes transparency" dur="2" iterated="y"/> there are techniques like this <pause dur="2.9"/> and on the face of it it looks a bit <pause dur="0.2"/> sort of banal but in fact <pause dur="0.5"/> this sort of thing can produce extremely <pause dur="0.3"/> useful results so <pause dur="1.8"/> if you fill in what <pause dur="0.3"/> <gap reason="name" extent="1 word"/> says and when i've done this with <pause dur="0.4"/> groups of students they usually <pause dur="0.4"/> well there are usually two <trunc>th</trunc> <pause dur="0.2"/> two types of responses one <pause dur="0.4"/> oh <pause dur="0.2"/> stick it into the NatWest <pause dur="0.4"/> bank <pause dur="0.2"/> you know they have a good interest rate and the other is <pause dur="0.5"/> go and blow it at the union on a Saturday night <pause dur="0.8"/> # but <pause dur="0.4"/> you get some impression as to how

your respondents are thinking and how they're feeling about things <pause dur="1.5"/> <kinesic desc="changes transparency" iterated="y" dur="4"/> personification or or # <trunc>s</trunc> story completion is another one <pause dur="1.0"/> now you may think this is a bit <pause dur="0.3"/> strange but in fact this will uncover lots of <pause dur="0.6"/> different <pause dur="0.6"/> # feelings that <pause dur="0.2"/> that people have about food it says <reading>once upon a time <pause dur="0.5"/> <gap reason="inaudible" extent="1 word"/> woman decided to invite some friends round to their house for dinner <pause dur="0.5"/> so and so went off to buy the food at</reading> this will <pause dur="0.3"/> say something about who's <pause dur="0.4"/> # who's the main food buyer and and whether there's any <pause dur="0.6"/> # implication about people coming for dinner <reading>so-and-so went off to buy the food <pause dur="0.3"/> they decided to serve <pause dur="0.3"/></reading> what <pause dur="0.3"/> <reading>finishing off with</reading> what because blah blah blah <pause dur="0.3"/> <reading>the dinner went very well <pause dur="0.3"/> but a number of the guests began to feel unwell several had to go home early it was because of</reading> <pause dur="0.5"/> now when i've done this <pause dur="0.2"/> in the last <pause dur="0.4"/> two or three years <pause dur="0.8"/> food safety has sort of <pause dur="0.5"/> become this big issue <pause dur="0.2"/> so loads of people say they decided to serve beef <pause dur="0.4"/> and it was food poisoning probably caused by B-S-E

well <pause dur="0.3"/> <vocal desc="laughter" n="ss" dur="2" iterated="y"/> it's not strictly accurate i think <pause dur="0.3"/> but <pause dur="0.7"/> quite often people put <pause dur="0.2"/> you know it's eggs which must have been caused by <trunc>salm</trunc> problem caused by salmonella and so on but <trunc>i</trunc> <pause dur="0.3"/> it <pause dur="1.8"/> in a qualitative way in an exploratory way it reveals a lot about the way people are thinking <pause dur="1.3"/> and of course this sort of thing is used for <pause dur="0.5"/> advertising all the time i'll just finish off with an example of <pause dur="0.3"/> exploratory <trunc>re</trunc> research <pause dur="0.8"/><kinesic desc="changes transparency" dur="13" iterated="y"/> and this is <pause dur="0.5"/>this is not personal experience it's <trunc>p</trunc> pinched out of a <pause dur="0.5"/> textbook but it this is exploratory research which <pause dur="0.5"/> can <pause dur="0.6"/> as i say it's used for advertising a lot so you you can sort of turn people's answers into <pause dur="1.2"/> elements of your <pause dur="0.3"/> # <pause dur="1.7"/> advertising campaign <pause dur="1.7"/> let's just see what this chap's <pause dur="0.2"/><event desc="turns off overhead projector" iterated="n"/> doing if we've got anywhere else <pause dur="1.3"/> # <pause dur="4.8"/> okay see what i mean he gets rather physical and stares into their eyes a lot <pause dur="14.2"/> right there are three people talking at the moment at some point he's got to <pause dur="0.3"/> calm this down a little bit and <pause dur="0.7"/> there you go <pause dur="0.8"/> he focuses in on one person <pause dur="4.2"/> if any of you are going to do

focus groups as part of your dissertations or anything then <pause dur="5.6"/> he's recording the discussion <pause dur="0.4"/> he's hearing what they're saying </u><pause dur="0.7"/> <u who="sm0982" trans="pause"> <gap reason="inaudible" extent="1 sec"/></u><pause dur="1.6"/> <u who="nm0976" trans="pause"> well <pause dur="0.3"/> i i thought <trunc>w</trunc> what i was going to say was when three people are talking it's very difficult to establish what each one of them is saying <pause dur="0.6"/> and and that's another problem <pause dur="0.8"/> # <pause dur="2.9"/> but he's sort of got it calmed down again now as you can see</u><u who="sm0983" trans="overlap"> # <pause dur="1.1"/> <gap reason="inaudible" extent="1 sec"/><pause dur="1.4"/> <gap reason="inaudible" extent="5 secs"/></u> <u who="nm0976" trans="overlap"> this <pause dur="0.9"/> this was just</u> <u who="sm0983" trans="overlap"> <gap reason="inaudible" extent="4 secs"/></u> <u who="nm0976" trans="overlap"> this was a static <pause dur="0.4"/> camera it wasn't anybody writing anything down <pause dur="0.5"/> we were we were in another room <pause dur="0.3"/> it was relayed to us on a T-V screen <pause dur="0.3"/> so there was a static camera obviously up this end <pause dur="0.7"/> and <pause dur="0.6"/> it was actually quite a long way away from them they seemed to <pause dur="0.4"/> sort of get over any fears of <pause dur="0.2"/> being recorded <pause dur="0.6"/> # <pause dur="1.4"/> but yes i mean the transcription is after this and you use this as the basis of <pause dur="0.3"/> your <pause dur="0.5"/> transcription <pause dur="4.5"/> so just to finish off ideas for <pause dur="4.8"/> just just <pause dur="0.2"/> give you an idea of how qualitative research can <pause dur="2.1"/><event desc="stops video" iterated="n"/><kinesic desc="turns on overhead projector showing transparency" iterated="n"/> help with # advertising so <pause dur="0.2"/> this was a <pause dur="0.5"/> an American study of <pause dur="0.6"/> # using open-ended questions as a way of sort of prompting people and making them think about things <pause dur="0.5"/> people who drive

a convertible <pause dur="0.4"/> are <pause dur="0.5"/> factory workers usually drive most of the new cars are when i drive <gap reason="inaudible" extent="1 sec"/><pause dur="0.3"/> now <pause dur="0.2"/> if you <pause dur="1.4"/> if you are <trunc>adverti</trunc> think about it you're you're thinking of the sort of segments you're appealing to you need to <pause dur="0.4"/> you want to think about <pause dur="0.2"/> how people regard certain categories of <pause dur="1.1"/> of <pause dur="0.9"/> the public and what they drive and blah blah blah <pause dur="0.5"/> this might have been for <pause dur="0.2"/> some sort of market segmentation exercise or it might have been to try and <pause dur="0.3"/> establish people's responses to a particular <pause dur="0.4"/> vehicle <pause dur="0.5"/> but they were finding <pause dur="1.8"/> different responses from men and women <pause dur="2.8"/> so where the sentence stem was when you first get a car <pause dur="0.8"/> # this is <pause dur="0.7"/> horribly out of date and i'm sure it wouldn't happen these days <pause dur="0.8"/> women's responses you can't wait till you drive you would go for a ride you would take rides in it of course you would put gas and go places so it's all about what you'd do with the car <pause dur="0.9"/> # men <pause dur="0.2"/> you'd take good care of it you'd give it a good polish <pause dur="0.5"/> you check the engine <pause dur="0.5"/><gap reason="inaudible due to background noise" extent="2 secs"/> now <pause dur="1.0"/><vocal desc="laughter" n="sl" iterated="y" dur="1"/> <pause dur="1.0"/> # <pause dur="0.4"/> as i say i'm

sure this was a few years back and it probably doesn't apply at all now <pause dur="0.3"/> but it actually <pause dur="0.4"/> resonates with what i was <shift feature="voice" new="laugh"/>saying about<shift feature="voice" new="normal"/> our <pause dur="0.6"/> Belgian males and females <pause dur="1.1"/> # a car of your own <pause dur="1.7"/> pleasant convenient say women it's fine to have it's nice to have <pause dur="0.8"/> # <pause dur="0.4"/> men say i take good care of it it's a good thing it's absolutely a necessity <pause dur="0.5"/> so you get <pause dur="0.4"/> if you were an advertiser this would be really important and <pause dur="0.3"/> let's face it whatever you think about this and however out of date it is <pause dur="0.6"/> # <pause dur="0.9"/> <shift feature="voice" new="laugh"/>gender <pause dur="0.2"/> is extremely important<shift feature="voice" new="normal"/> in advertising <pause dur="0.6"/> you could you can watch almost any <pause dur="0.6"/> # <pause dur="0.3"/> particularly T-V advert because they're the most visual and the most active <pause dur="0.4"/> you can watch almost any one and if you start thinking about <pause dur="0.3"/> what segment are they aiming at <pause dur="0.6"/> and what is the sort of message they're putting across <pause dur="0.2"/> you'll soon find that <pause dur="0.6"/> things like this become very important <pause dur="0.6"/> and <pause dur="1.0"/> you know if <pause dur="0.5"/> it's certainly in the U-K <pause dur="0.7"/> # <pause dur="2.2"/> and i think i mean cars is always a good example of this <pause dur="0.4"/> but <pause dur="0.5"/> you can <trunc>i</trunc> i'll i'll leave you to

draw your own conclusion but you can look at <pause dur="0.5"/> car ads for things like the Renault Clio <pause dur="0.9"/> and Citroen Saxa <pause dur="1.0"/> and you could <pause dur="0.3"/> almost certainly identify a gender bias in that sort of advertising <pause dur="0.5"/> and same for different models so <pause dur="0.2"/> just keep an eye on that because <pause dur="0.4"/> # <pause dur="0.8"/> i can guarantee <pause dur="0.3"/> that <pause dur="0.2"/> almost every T-V advert is backed up by qualitative research of some sort where <pause dur="0.4"/> the researchers have tried to establish <pause dur="0.4"/> the sort of basic motivations that people have <pause dur="0.3"/> for dealing with that <pause dur="0.5"/> vehicle or </u><u who="sf0984" trans="overlap"> i was going to say that there's direct correlation between that and advertising cars is the new Fiat <pause dur="0.2"/> we've got the woman talking about the men the man's point of view # <pause dur="0.6"/> <gap reason="inaudible" extent="1 sec"/> the man talking about the woman's point of view <gap reason="inaudible" extent="1 word"/> have you seen that one yet </u><u who="nm0976" trans="latching"> right is that <pause dur="0.8"/> i mean the <trunc>fi</trunc> Fiat have have done a lot of this they had that <distinct lang="it"> spirito di punto</distinct> thing for the Fiat Punto </u><u who="sf0984" trans="overlap"> mm </u><pause dur="0.5"/> <u who="nm0976" trans="pause"> there's <trunc>alwa</trunc> there's often a male female thing in <pause dur="0.2"/> car i mean <pause dur="0.9"/> if you want sort of classic marketing texts you read something like <pause dur="0.7"/> # if there's a <pause dur="0.2"/> book called the hidden persuaders i think i've got it on a <pause dur="1.3"/> on a reference list somewhere <pause dur="0.6"/> by a guy called Vance

Packard which again is about nineteen-fifties America about how <pause dur="0.3"/> car <pause dur="0.3"/> manufacturers persuade people to buy things and and <pause dur="1.3"/> i mean at that time if you think of American cars in the fifties and the sort of style and size and <pause dur="0.6"/> and shape of these things they were <pause dur="0.2"/> they were making very <pause dur="0.2"/> definite sort of psychological claims on people <pause dur="1.0"/> # <pause dur="0.4"/> and where you <pause dur="0.5"/> you know particularly sports <shift feature="voice" new="laugh"/>cars <shift feature="voice" new="normal"/><pause dur="0.5"/> and the sort of <pause dur="0.3"/> connotations that go with owning and driving a sports car <pause dur="0.5"/> which i'll leave to your imagination <pause dur="0.7"/><vocal desc="laughter" n="ss" iterated="y" dur="1"/> but # <pause dur="1.2"/> you know it's a lot of it would be based on this sort of work definitely <pause dur="2.6"/> okay so it's conjoint analysis next week that's a way of <pause dur="0.6"/> designing new products basically <pause dur="0.5"/> common market research problem <pause dur="1.3"/> thanks <pause dur="3.3"/> can you pull the blinds up <pause dur="0.2"/> thanks