The objective of this study was to investigate differential tendencies to avoid guessing as a function of three variables: (1) lingual-cultural-group; (2) gender; and (3) examination year. The Psychometric Entrance Test (PET) for universities in Israel was used, which is administered in Hebrew, Arabic, English, French, Spanish, and Russian. The PET is a battery of five subtests, and encompasses about 200 test items. Three of the five subtests were used in this study: figural reasoning, mathematical reasoning, and English. Data for 12,440 male and 10,532 female examinees were analyzed. The tendency to avoid guessing was measured by the proportion of two types of unanswered items: unreached items, and omitted items. A factor analysis using VARIMAX rotation indicated a strong two-factor structure, in which all indices based on omitted items loaded on the first factor and all indices based on unreached items loaded on he second factor. An analysis of covariance with the corrected-for-guessing scores as a covariate indicated that all three effects and paired interactions were significant. The examination year appeared to have the strongest effect on the proportion of omitted items, while the language-version seemed to affect the proportion of unreached items most strongly. The gender effect was not found to be as strong as were the other two main effects. Three data tables are included. (RLC)