The aim of this study was to examine whether personality traits predict sample depletion, which occurs in the studies with the two stages of data collection. Methodologically, the problem was studied in terms of the five-factor theory of personality (McCrae, Costa, 2013). The study was conducted in two stages. At Stage 1, 1030 students participated in a psychological study in classrooms, where they fulfilled the main battery of the tests. At Stage 2 the participants completed a questionnaire off-site, and 277 participants of the main sample emailed a file with the completed test to the researcher. Agreeableness, conscientiousness, and introversion predicted that the participants will take part at Stage 2. The logistic regression analysis estimated the contribution of the reflective characteristic adaptations – meta-traits and attitudes to traits – to these effects. Meta-traits are the individuals’ judgments on how their personality traits are perceived by the significant others, in our case – by parents. Attitudes toward traits are the valent (positive/ negative) evaluations of these traits by the individuals without the direct reference to their personality. Reflective characteristic adaptations incrementally predicted to the participation at Stage 2, while the role of the traits themselves was eliminated. The depletion of the sample was predicted by various meta-traits and attitudes toward various traits. The ambivalent role of attitudes toward traits and meta-traits has been further discussed in terms of the reflective-impulsive model by F. Strack and R. Deutsch. A model was empirically supported, in which the personality traits (extraversion, agreeableness, conscientiousness) related with depletion of the sample indirectly, via reflective characteristic adaptations. Relatively small values of the coefficients of determination presumes that the role of personality characteristics in sample depletion is rather small, but yet substantial.