An attempt was made to determine whether student honesty is affected by a take-home test, referred to as an unstructured evaluation. A group of 27 students in a history class at a university was given a teacher-made take-home test as a mid-term examination. A second group of 27 students was given the same examination in class. Unknown to the students, the tests were identical. One week later, a surprise objective test was given to both groups in class on the same material. A questionnaire accompanied this test, asking students how much effort they had spent preparing for the original test and how much they would have spent had they prepared for the other type of test. The group given the in-class test performed significantly better on the surprise test, although the take-home test group had higher scores on the original test. Members of the take-home test group said that they would have studied more had they prepared for an in-class test. It is concluded that student dishonesty accounts both for the poorer scores on the second test, indicative of less preparation, and the higher scores on the original test. Some students in the take-home test group probably did cheat and did not adequately learn the subject matter. (SLD)