A Spanish continuous voluntary web survey: sample bias, weighting and efficiency

Using micro data from a continuous volunteer web survey (CVWS), the WageIndicator (WI), this paper firstly analyses the type of bias that such a survey method produces. Secondly, following a «modelbased » approach, two alternative data weighting methodologies are implemented. Thirdly, in order to test whether weighting corrects the bias, thus making it possible to obtain conclusions applicable to the whole labour force, the efficiency of the weighting methodologies is evaluated. Since the WageIndicator is a labour market oriented survey, weighting efficiency is evaluated by calculating mean salaries, inequality indices and salary regressions before and after applying weights to WI data, and by comparing the results obtained with those achieved using the Structure of Earnings Survey (SES), a wage survey run by the Spanish National Statistics Institute. It is concluded that, after weighting, estimated statistics and parameters move in the right direction.