That many women avoid science is clear. Less clear, however, are explanations for this phenomenon. In the past decade, researchers have looked for causes first by grappling with the question: What's wrong with women that they don't like science? This line of inquiry inspired several studies of cognitive differences between men and women [15, 25, 33] and of differences in their background experiences, including coursework deficits, teacher expectations, and classroom interactions [11, 35]. These individual studies were later summarized by more sophisticated meta-analytic studies, which essentially demonstrated that cognitive differences between the sexes in and of themselves account for little in explaining the variance in achievement scores.(1) Because the results of these studies, though fruitful, did not fully explain the differences between men and women in science participation, some researchers turned their attention from women to science as the subject of inquiry.
This change of focus led to a second question: What's wrong with science that women don't like it? Respondents critiqued the structure of scientific disciplines and the way in which scientific knowledge is produced, raising epistemological arguments that challenged the nature of objectivity [7, 12, 13, 14, 19]. Evelyn Fox Keller  argues that what society defines as "scientific" parallels what society defines as "masculine;" thus, women have felt excluded from scientific endeavors. Ruth Bleier  corroborates the exclusion from science experienced by herself and many other women scientists, stating that science would be very different, if women had not been excluded from it. In a similar vein, Sandra Harding has argued that women need sciences that are "for women in every class, race and culture." .
At the same time that this line of inquiry continues, another related question has surfaced: What is wrong with the teaching of science that women don't like it? This pedagogical question is related to the previous, epistemological one, because our perception of the nature of the discipline of science influences the way we structure our curriculum and teach courses. For example, if we believe science to be a set of objective truths to be learned, we may feel that it is necessary to transmit as much of this body of knowledge as possible through lectures, and that it is unnecessary to integrate controversial material into our courses or to provide time for discussion . An exclusive focus on objectivity alienates some women from science, as is shown in this student's statement: "What's missing in science is a whole sort of human element. It doesn't seem to be infused with any morality. It doesn't even seem to be a world about people anymore" . Indeed, Belenky et al.  found that the Women they interviewed reported that their most valuable learning experiences occurred outside the classroom, possibly because their preferred style of learning, connected knowing, found little room to flourish in academe. According to these authors, connected knowing is a personal, cooperative approach to learning, which values tying theory to experiences and stresses belief rather than doubt. What the academy values highly, however, is something Belenky et al.  call "separate knowing," an adversarial, impersonal approach to objective reasoning. Separate knowing is clearly more valued in most science courses than is connected knowing.
This conflict presents a dilemma. Is it possible to engage women in connected knowing in a discipline and academic culture that value separate knowing? Furthermore, what does connected knowing actually entail? Although Belenky et al.  have provided evidence based on interview data that women prefer this sort of learning, studies of what women actually do to engage in connected learning are very rare.(2) Nevertheless, feminist professors have provided pedagogical ideas about what they do to engage women in connected learning. Some excellent articles have been written on feminist education,(3) but only a few writers have focused on feminist science education . One common theme in most of these articles has been the value of cooperative learning.
One cooperative learning structure that has been used at the college level is Supplemental Instruction (SI). SI encourages thinking outside the college classroom through the use of an SI leader, a peer who has previously been successful in a particular class. Typically in such a program an SI leader meets with groups of students two or three times a week on a voluntary basis to assist them in clarifying their understanding of course concepts. This program has been successful in improving the grades of students attending SI sessions compared to control groups who did not attend SI [for example, 2, 26]. Focusing on achievement as measured by final course grades, however, tells us little about the factors that influence student learning; research on the process of learning rather than on the outcomes is essential.
In order to understand why programs that involve cooperative learning, such as SI, promote achievement, we need to know more about the social interaction process during cognitive development. In Vygotsky's view, modeling and speaking precede learning and thinking - what students can do collaboratively today, they can do independently tomorrow . Thus, social interaction enhances thinking, because individuals can learn to solve problems independently by first solving problems together with competent peers. The purpose of our research, therefore, was to describe the process of social interaction in women learning science through SI.
Background of the Study
This study took place at a private women's college in the Midwest. Several women nursing students were failing health science, one of the junior-level foundation courses for the nursing program at that institution. Health Science was a two-semester course requirement for nursing students that focused primarily on organic disorders and maladaptive syndromes. Prerequisites for the course included anatomy, physiology, psychology, and child psychology. It was a fairly large lecture class (averaging fifty-eight students) and was taught by several instructors who alternated units by teaching their specialty. To improve students' achievement, the nursing department invited two learning center faculty to observe the course in an effort to understand better the intellectual demands placed on students and to design appropriate interventions. After eight visits to the class and numerous discussions with students in the class, the learning center faculty met with the nursing faculty to share their observations. The course required students to memorize knowledge; identify terms, formulae and basic theories; apply theoretical knowledge in practical cases; and generalize to new applications and new knowledge. Unfortunately, during the lectures most students were so caught up scrambling to memorize and identify information that they seemed to miss the connections, applications, and generalizations the professors made. In addition, the nursing program at this women's college was very competitive, which, according to student reports, led to unhealthy anxiety. Thus, the nursing faculty and the learning center instituted Supplemental Instruction (SI) to enable students to think more about concepts and processes. The research reported here encompasses a two-year study of that SI program.
Description of the SI Program
Goals. The goals of this program were to encourage students to develop conceptual understandings; to articulate both understandings and misconceptions in a think-aloud fashion; to connect, relate, and integrate scientific information; to develop confidence and ability in analyzing illnesses; and to learn how to learn science.
Operation. The learning center employed nursing major(s) to lead SI sessions two hours a week outside the scheduled health science class and laboratory times. To prepare for these sessions, the SI leader attended the health science classes, took notes, did the assigned readings and problems, and participated in training meetings. Three of the four SI leaders met with some of the health science professors on a regular basis.
Role of the SI leader. The SI leaders had no power over the students' grades and did not tell the professors who was attending the SI sessions and who was not. Their role was to model the thinking involved in learning health science; to listen to students' questions, comments, concerns, and answers; to ask relevant questions; and to encourage students to work cooperatively in learning scientific concepts.
SI leaders. Four SI leaders (senior nursing students identified as having a thorough understanding of health science) participated in the program and in this research. Two of these women led SI sessions individually for a semester each; the other …