Please read our student and staff community guidance on COVID-19
Skip to main content Skip to navigation

Weighting

Once the different variables in the index have been normalised, they have to be ‘added together’ to generate the index. At this point, there are two alternatives.

First, the index can be constructed by taking a simple average of the variables. This is straightforward, but of course imposes the assumption that all variables entering the index are equally important.

Second, the index can be constructed by taking a weighted average of the variables, where the weights are positive and add up to one. There are then two ways of assigning the weights.

First, the weights can be subjective, i.e. chosen by the researcher according to his/her subjective belief about which variables are more important. This is the approach taken by Foreign Policy and A.T.Kearney in the construction of their Globalisation Index. In fact, in the latest version of their index,  they double-weight  FDI flows, as they believe  that this is a particularly important indicator of globalisation.

Second, the weights can be chosen to maximise the informativeness of the index: statistically optimal weights. We take the second approach, as we feel this avoids any subjective bias on the part of the researcher as to which weights are important.

Briefly, statistically optimal weighting (or principal component weighting, as it is more properly known) works as follows (if you want to know more, see the Appendix!). Take for example, the four variables that make up the economic globalisation index. In 2000, these were available for 119 countries. Thus, in 2000, we are using 119x4 pieces of information. When we aggregate these four variables to make the index, we end up with only one piece of information about each country: thus we are ‘throwing away’ 119x3 pieces of information. Statistically optimal weighting ensures that when we do this, we retain as much as possible of the original information about the countries.

The statistically optimal weights used in the construction of our index are: 

Table 3

Sub-Index
Variables
Weight

Economic Globalisation

Trade

0.418

Fdi

0.092

Portfolio

0.220

Income

0.270

 

 

 

 

 

 

 

Social Globalisation

People

0.331

Foreign Stock

0.266

Foreign Flow

0.629

Worker Remittances

0.079

Tourists

0.026

Ideas

0.669

Phone calls

0.004

Internet users

0.303

Films

0.061

Books and newspapers

0.577

Mail

0.054

 

 

 

Political Globalisation

Embassies

0.378

UN Peace Missions

0.357

International Organisations

0.266