The first study hypothesized that child molesters would display lower self-esteem than either nonsexual offenders or community participants and that child molesters' attitudes toward children would be negatively correlated with self-esteem, and attitudes toward women would be positively correlated with self-esteem. A total of 24 nonfamilial child molesters and 19 property offenders were recruited from 3 Canadian Federal penitentiaries. All the child molesters had been convicted of sexually molesting a female child under 14 years old. The instruments used to test the hypotheses were the Social Self-Esteem Inventory and an adaptation of the Semantic Differential. The Marlowe-Crowne Social Desirability Scale was included to correct (if necessary) for the tendency on the part of the participants to present themselves in a positive light. The only significant results in the first study were that the child molesters had lower self-esteem than did the subjects in the other two groups. Given the failure of the first study to find significant differences between groups in their attitudes toward children and adults, as well as the failure to find a relationship between the child molesters' attitudes toward children and their self-esteem, the second study used the same subjects to examine the issue of cognitions and self-esteem in a different way. The second study repeated the use of the Social Self-Esteem Inventory but replaced the Semantic Differential with Bumby's Molest Scale, which assessed attitudes toward sex between adults and children. As hypothesized, the findings show that child molesters' evidenced a greater tendency to distort their perceptions about sex between adults and children; this tendency was significantly related to their lack of self-confidence. The second study failed to support the hypotheses that suggested that child molesters would have lower self-esteem than the other two groups and that they would have a greater tendency to use sex when coping with distress. Results are discussed in terms of the differences between the findings and earlier studies and their implications for future research. 2 tables and 33 references