In response to recent data from the Department of Education indicating that the number of women earning computer science undergraduate degrees has declined sharply in recent years, a study was conducted to determine gender bias in introductory computer classes. Questionnaires were distributed to students in two classes at Rutgers University and one at Princeton in spring 1995. Sample populations were then constructed of 94 Rutgers students, 30% of whom were female, and 33 Princeton students, 45% of whom were female. An analysis of responses revealed the following: (1) 55% of the Rutgers students, including 43% of the females, and 51% of the Princeton students, including 40% of the females, felt that they were well prepared for the course; (2) women at both schools and the Princeton men strongly disagreed that they could program fluently, but some of the Rutgers men had entered with strong programming skills; (3) at Rutgers, male students were significantly more likely to own a computer than females (91% versus 70%), but at Princeton computer ownership was virtually universal (91%); (4) for both samples, men were significantly more likely to indicate an intent to continue with computer science and men indicated a more positive attitude toward a computer science career; (5) with respect to programming, men rated themselves as more familiar in almost all categories; and (6) the mean grades for men and women were the same for both groups, but women were significantly underrepresented among the highest scorers. It is noted that there seems to be a positive feedback loop operating for men and a negative one operating for women. some changes to introductory courses that might help compensate for varying backgrounds of students and encourage them to try a computer science course include: (1) make the introductory course pass/fail; (2) have multilevel or slower paced introductory courses; (3) have scheduled, supervised labs and smaller classes; (4) involve students in larger collaborative projects; and (5) involve upper level and graduate students as mentors. (Contains 17 references.) (The survey instrument is appended.) (MAB)