An Application of Item Response Theory to Detect Distortions in Non-cognitive Questionnaires

In non-cognitive studies, there often arises a question whether the collected data can be trusted or not, and, if so, how much? How can we examine the credibility of the data, especially in cases of sensitive questions? In this study we apply the Item Response theory to identify fakings of respondents in the scale of teachers’ self-efficacy, which is included in teachers questionnaire TALIS-2013. The result of the analysis show, that the data contain more than 30% fakings, 21% of which may be characterized as positive faking. These are the respondents' attempts to improve their score. From the results of the analysis scale of teachers’ self-efficacy scale and respondent's faking attempts we colnclude that the scale is easily falsifiable, which is a serious threat to the validity of measurements and the quality interpretation.