The article discusses theoretical and methodological aspects of cognitive dissonance as a social behavior factor. It is concluded that the modern sociological methodology does not provide researchers with any generally accepted techniques for measuring cognitive dissonance in surveys. It is proved that sociologically relevant manifestations of cognitive dissonance should be carried out according to the model of the contradiction of people’s knowledge about their behavior and the situation in which they found themselves with their self-esteem. Self-justification acts as a model for overcoming cognitive dissonance. There are three criteria for the use of approaches to measuring cognitive dissonance: 1) the measurement method should measure exactly the latent variable that can be interpreted as cognitive dissonance; 2) the method should also be simple and economical to implement; 3) the method should have a fairly broad applicability to the research topic. It is proposed to use the technique of measuring cognitive dissonance in mass surveys, which is a “scale of difference” – a set of identical statements applied to the respondents themselves and external instances (for example, other people). The differences in ratings in these identical statements, according to the self-justification hypothesis, should reflect the magnitude of cognitive dissonance. The hypothesis was confirmed in an experiment that induced cognitive dissonance according to the Festinger–Carlsmith model of “forced consent”, and in a survey that demonstrated a correlation between the magnitude of the difference and the intensity of smoking. The technique has passed primary validation, it shows the presence of cognitive dissonance where it should be observed, and to some extent reflects its magnitude. However, the relatively low significance of the differences in the first study and the weakness of correlation in the second one necessitate further research, including in the direction of development and refinement of the technique.