ICT skills measurement in social surveys: Can we trust self-reports?

Abstract

Self-reports are the most common way of measuring information and communications technology (ICT) skills in social surveys. Studies comparing self-reported computer skills with objective assessments have found evidence of significant overreporting of skills, but were conducted only among non-representative groups of individuals. This paper fills an important gap by analysing the degree to which ICT skills are overreported in the working-age population of Poland, and the potential causes of this behaviour. We compare answers to Eurostat questions on ICT usage with direct assessments of the corresponding tasks. The results suggest that those individuals who are most likely to possess ICT skills are also most likely to overreport having these skills. The propensity to overreport decreases with age and increases with years of education and numeracy level; women are less likely than men to overestimate their skills. The positive relationship between the probability of a group to overreport their own skills and their expected levels of skills suggests that social desirability bias may explain this phenomenon.