ABSTRACT
There are two ways, however, to gain construct validity for the people variable of interest. The first alternative is called discriminative validity, that is, the people construct should not be highly correlated to variables that our indicator does not claim to measure. The second alternative is for convergent validity, that is, the correlation of the author measure of interest with the measures of others. In addition, the authors present some ways to operationalize the concepts through indexes that can show the people the measurement variability among different analysis units, in order to compare them or include them within a more complex analysis. The Principal Component Analysis (PCA) technique helps the people define which combination of data best represents the concept the people want to measure. PCA technique gives the people different components, which are the reduction of the dimensionality of a multivariate dataset condensing the dimensions into one single variable.
