What do you think?
Rate this book
Talking About Leaving: Why Undergraduates Leave The Sciences
This intriguing book explores the reasons that lead undergraduates of above-average ability to switch from science, mathematics, and engineering majors into nonscience majors. Based on a three-year, seven-campus study, the volume takes up the ongoing national debate about the quality of undergraduate education in these fields, offering explanations for net losses of students to non-science majors. Data show that approximately 40 percent of undergraduate students leave engineering programs, 50 percent leave the physical and biological sciences, and 60 percent leave mathematics. Concern about this waste of talent is heightened because these losses occur among the most highly qualified college entrants and are disproportionately greater among women and students of color, despite a serious national effort to improve their recruitment and retention. The authors' findings, culled from over 600 hours of ethnographic interviews and focus group discussions with undergraduates, explain the intended and unintended consequences of some traditional teaching practices and attitudes. Talking about Leaving is richly illustrated with students' accounts of their own experiences in the sciences. This is a landmark study-an essential source book for all those concerned with changing the ways that we teach science, mathematics, and engineering education, and with opening these fields to a more diverse student body.
444 pages, Paperback
First published November 4, 1996
Ratings & Reviews
Friends & Following
Create a free account to discover what your friends think of this book!
Community Reviews
Displaying 1 - 5 of 5 reviews
August 12, 2024
This book is dated, but it’s a landmark study that’s still worth a read.
I’ll discuss what I see as the good and the bad, but note that my perspective is that of a physics professor. With that comes some appreciation for what social scientists can illuminate for me, but also deep skepticism about their objectivity and quality of data analysis. Seymour and Hewitt clearly have an agenda, and seem pretty happy to quote mine in support of their agenda. There are no statistical tests done in this entire book, no tests for significance, no sign of confidence intervals anywhere. That said, it’s worth hearing what the students think in their own words, as long as you keep in mind that the authors quote mined.
Let’s start near the end, with this key observation:
Teaching is super important, though I'm not sure what halcyon days we're trying to "revive". Maybe when professors read straight from books, in Latin, a language most students could read but not understand a lecture in?
STEM education was bad in the early 90s, when this study was conducted. What’s concerning is that despite massive advances in discipline-based educational research, many faculty still teach in the same old ineffective ways. A big part of the reason for this is that university faculty receive no training in teaching, cognitive psychology, and so on.
The fact that in 2024 we’re still heavily emphasizing multicultural centers and so on while ignoring pedagogy seems like a profound oversight on the part of those who claim to be pushing for equity in STEM. Of course, the focus on pedagogy has benefits for all students, including white men, who actually experience many of the same issues that women and students of color do.
You’d think in an era of declining student enrollment we’d be working on improving our teaching, yeah? But no, we’re really not (in higher education broadly, or at my institution specifically). Instead we found ways to get higher student evaluations that don’t require much work and definitely don’t improve learning, primarily grade inflation.
But maybe we should listen to the students with a less venal interest in grades. “Switchers and non-switchers were virtually unanimous in their view that no set of problems in S.M.E. majors was more in need of urgent, radical improvement than faculty pedagogy” (p. 165).
Other takeaways worth noting:
1. Students are very poor judges of how well they’re doing and how unique their struggles are, and as a result, sociological studies based on interviewing students (like, ahem, this book) can’t give a good objective picture of the degree and extent of difficulties of underrepresented students vs. others. As one student notes, “I remember thinking, ‘This is horrible.’ But it turns out that I was doing pretty good. I had a B when I dropped the class, but I didn’t know that. (Female black engineering switcher”; p. 328).
As a professor, I’ve had this experience many times. A student who is doing perfectly fine in my class drops it without talking to me because they think they’re doing badly. While some get that misperception correct, many do not, and it’s very easy for a student who’s black or female to attribute their “poor” performance to a system that’s rigged against their identity, without understanding how they were actually doing. And then the sociologist who interviews them can collect many such stories and claim it’s a valid theme in their qualitative analysis, but without having any objective evidence that it’s true.
2. Pushing underrepresented students into majors they may not want is counterproductive. “Choosing [majors] to please others (unless this coincides with personal interests and goals) often leads to unhappiness in the major and places them at high risk of leaving it” (p. 386), and as a student says, “there is a danger that you may drop out if you discover that you are unsuited, rather than just changing majors” (p. 383). And “those who encouraged students of color to choose an S.M.E. major were often mistaken about students’ actual interests and abilities, or had limited understanding of what level of ability or preparation was required. Some parents and high school teachers confused students’ expressions of interest in mathematics and science with excellence in those subjects” (p. 325).
One of the problems with this is that when such students do poorly in a college class, it’s the college professors who get accused of running “weed-out” courses, including by this very book, when in fact we may just be setting realistic standards for the first time in the student’s experience.
3. Campuses with larger representation of underrepresented groups show more hostility towards these groups. “The strongest prejudice was actually expressed on campuses where S.M.E. students of color said they felt most accepted” (p. 368) because they could insulate within their own group. But “the larger the ethnic group, the more its members are perceived as an economic threat to the majority, and the greater the degree of racism and hostility which is expressed against them by all students, including those of other ethnicities” (p. 376).
4. Treating people equally can lead to unequal outcomes, due to different cultures (including the socialization of men vs. women). E.g., “faculty are unwittingly discouraging women more than men by behavior which is actually the same for both men and women” (p. 261).
5. It will always be hard to get students to major in STEM, because it’s hard (and has real grades).
Almost a quarter of their entire sample “reported a fall in their level of confidence caused by expectation of high (or easy) grades, and their shock at receiving the lower grades that are traditional of introductory S.M.E. classes” (p. 86). I’ll note that it’s not just intro classes, at least in the physical and mathematical sciences. We have actual grading standards. (Interestingly, at least at my university, it is true for biology that the intro classes have very low grades while the higher-level courses have extremely high average grades.)
This is the problem with treating students as customers. “There’s no point in taking the major unless you’re getting at least a 3.3” (p. 222) says a student, joined by many others with similar sentiments. They want to live in Lake Wobegon, where all the children are above average.
Weaknesses in the book:
1. Taking what their interviewees say at face value, including wild guesses about the internal states of others with no objective evidence.
“I saw less self-questioning from the men. They didn’t sit there saying, ‘Am I stupid? Am I stupid?’ like my women friends and I did” (p. 296). Um, actually, a lot of them did. They just didn’t say it out loud in class (nor did the women, I’m guessing), but they didn’t share it with you. Because men are culturally supposed to be tough and independent, and to not show weakness. This is related to the fact that “more men than women described failure to establish useful working peer groups” (p. 299).
Another example is the student saying “In physics, the only prospect of employment is after grad school… if you want a good job, you need a Ph.D.” (p. 227). This is false—on average, those who get PhDs never recover economically from the opportunity cost. But the researchers took the opinion of an undergrad at face value rather than interrogating it.
2. Subset of the above: taking what interviewees say about grading at face value, even though the interviewees know nothing about the course grades as a whole or how grading works.
Examples are numerous.
This is a common student complaint, because it’s the perspective of a novice. Novices see differences where experts see connections. Novices might see 135 distinct problems where experts see two principles (say, conservation of energy and conservation of momentum). A good professor recognizes this, but there’s no way around it other than the hard work of the novice becoming more expert.
WTF? You didn’t fail, because they curved the grades. That’s not failing. And if they curved the grades so people weren’t failing, that’s not weeding people out. Curved grades, and medians near 50%, are very useful for using the full dynamic range of grading. If nobody in the class is getting below 50%, then your assessment tool is wasting half its range.
But the authors take these parochial student misperceptions as establishing a fact that there is a weed-out system in STEM courses. I know how grades vary across disciplines, and if STEM grades harder than others (which it does), that’s because it has actual standards. We don’t just pass everyone regardless of how much (or little) they learn, like our social science colleagues… so maybe don’t come into our disciplines with normative moral judgments on our grading. (However, there are some cases of real, explicit weed-out mentality. Those are indeed harmful. But they’re much rarer than students seem to think.)
3. Taking what their interviewees say at face value… until it disagrees with a pet theory of the authors.
I mean, maybe, yes, the female students are missing something crucial, or maybe we shouldn’t pick and choose when we believe what the interviewees are telling us based on our own prior beliefs and without any evidence.
4. Cherry picking quotes to support pet theories of the authors.
“Although both men and women cited competition in their major as a factor in their switching decisions, it is our contention that they did so for different, gender-related reasons” (p. 265). Maybe! But you have no evidence beyond a single quote that isn’t all that relevant.
Note that 26% of men cited “morale undermined by competitive S.M.E. culture” as a reason for switching, compared to only 4% of women. And 27% of men cited “discouraged/lost confidence due to low grades in early years” as a reason for switching, compared to only 19% of women. 16% of men cited “lack of peer study group support” as a reason for switching, compared with only 8% of women (pp. 232–233). But it's the self-confidence of women that Seymour & Hewitt describe as “so fragile” (p. 265). It’s unclear how to reconcile this with the actual data.
I’ll discuss what I see as the good and the bad, but note that my perspective is that of a physics professor. With that comes some appreciation for what social scientists can illuminate for me, but also deep skepticism about their objectivity and quality of data analysis. Seymour and Hewitt clearly have an agenda, and seem pretty happy to quote mine in support of their agenda. There are no statistical tests done in this entire book, no tests for significance, no sign of confidence intervals anywhere. That said, it’s worth hearing what the students think in their own words, as long as you keep in mind that the authors quote mined.
Let’s start near the end, with this key observation:
It is also very clear from our data that the most effective way to improve retention among women and students of color, and to build their numbers over the longer-term, is to improve the quality of the learning experience for all students [p. 394].
The strongest [line of practical action], by far, involves a radical revival of the quality of teaching [p. 182].
Teaching is super important, though I'm not sure what halcyon days we're trying to "revive". Maybe when professors read straight from books, in Latin, a language most students could read but not understand a lecture in?
STEM education was bad in the early 90s, when this study was conducted. What’s concerning is that despite massive advances in discipline-based educational research, many faculty still teach in the same old ineffective ways. A big part of the reason for this is that university faculty receive no training in teaching, cognitive psychology, and so on.
The fact that in 2024 we’re still heavily emphasizing multicultural centers and so on while ignoring pedagogy seems like a profound oversight on the part of those who claim to be pushing for equity in STEM. Of course, the focus on pedagogy has benefits for all students, including white men, who actually experience many of the same issues that women and students of color do.
Moving pedagogy from a focus on teaching to a focus on learning, and from selecting talent to nurturing it, will dis-proportionately [sic] increase the persistence rate of able women in S.M.E. majors. It also promises to reduce the loss of able male students [p. 314].
Reports of poor teaching in S.M.E. classes were by far the most common complaint of all switchers and non-switchers. Poor teaching was mentioned by almost every switcher (90.2%), and by far more non-switchers (73.7%) than any other issue [p. 146].
You’d think in an era of declining student enrollment we’d be working on improving our teaching, yeah? But no, we’re really not (in higher education broadly, or at my institution specifically). Instead we found ways to get higher student evaluations that don’t require much work and definitely don’t improve learning, primarily grade inflation.
But maybe we should listen to the students with a less venal interest in grades. “Switchers and non-switchers were virtually unanimous in their view that no set of problems in S.M.E. majors was more in need of urgent, radical improvement than faculty pedagogy” (p. 165).
Other takeaways worth noting:
1. Students are very poor judges of how well they’re doing and how unique their struggles are, and as a result, sociological studies based on interviewing students (like, ahem, this book) can’t give a good objective picture of the degree and extent of difficulties of underrepresented students vs. others. As one student notes, “I remember thinking, ‘This is horrible.’ But it turns out that I was doing pretty good. I had a B when I dropped the class, but I didn’t know that. (Female black engineering switcher”; p. 328).
As a professor, I’ve had this experience many times. A student who is doing perfectly fine in my class drops it without talking to me because they think they’re doing badly. While some get that misperception correct, many do not, and it’s very easy for a student who’s black or female to attribute their “poor” performance to a system that’s rigged against their identity, without understanding how they were actually doing. And then the sociologist who interviews them can collect many such stories and claim it’s a valid theme in their qualitative analysis, but without having any objective evidence that it’s true.
2. Pushing underrepresented students into majors they may not want is counterproductive. “Choosing [majors] to please others (unless this coincides with personal interests and goals) often leads to unhappiness in the major and places them at high risk of leaving it” (p. 386), and as a student says, “there is a danger that you may drop out if you discover that you are unsuited, rather than just changing majors” (p. 383). And “those who encouraged students of color to choose an S.M.E. major were often mistaken about students’ actual interests and abilities, or had limited understanding of what level of ability or preparation was required. Some parents and high school teachers confused students’ expressions of interest in mathematics and science with excellence in those subjects” (p. 325).
One of the problems with this is that when such students do poorly in a college class, it’s the college professors who get accused of running “weed-out” courses, including by this very book, when in fact we may just be setting realistic standards for the first time in the student’s experience.
3. Campuses with larger representation of underrepresented groups show more hostility towards these groups. “The strongest prejudice was actually expressed on campuses where S.M.E. students of color said they felt most accepted” (p. 368) because they could insulate within their own group. But “the larger the ethnic group, the more its members are perceived as an economic threat to the majority, and the greater the degree of racism and hostility which is expressed against them by all students, including those of other ethnicities” (p. 376).
4. Treating people equally can lead to unequal outcomes, due to different cultures (including the socialization of men vs. women). E.g., “faculty are unwittingly discouraging women more than men by behavior which is actually the same for both men and women” (p. 261).
5. It will always be hard to get students to major in STEM, because it’s hard (and has real grades).
Almost a quarter of their entire sample “reported a fall in their level of confidence caused by expectation of high (or easy) grades, and their shock at receiving the lower grades that are traditional of introductory S.M.E. classes” (p. 86). I’ll note that it’s not just intro classes, at least in the physical and mathematical sciences. We have actual grading standards. (Interestingly, at least at my university, it is true for biology that the intro classes have very low grades while the higher-level courses have extremely high average grades.)
Learning not to interpret grades as personal criticism was critical to surviving the discovery that high school had not adequately prepared them for college—intellectually, or in terms of work skills, discipline, or grade expectation [p. 87].Definitely a concern, though it’s sad that in the decades since this study colleges have dealt with this mostly by lowering expectations for students, massively inflating grades, or both. This very much includes Ivy League universities, which mostly have had utterly insane grade inflation, even in STEM.
This is the problem with treating students as customers. “There’s no point in taking the major unless you’re getting at least a 3.3” (p. 222) says a student, joined by many others with similar sentiments. They want to live in Lake Wobegon, where all the children are above average.
Weaknesses in the book:
1. Taking what their interviewees say at face value, including wild guesses about the internal states of others with no objective evidence.
“I saw less self-questioning from the men. They didn’t sit there saying, ‘Am I stupid? Am I stupid?’ like my women friends and I did” (p. 296). Um, actually, a lot of them did. They just didn’t say it out loud in class (nor did the women, I’m guessing), but they didn’t share it with you. Because men are culturally supposed to be tough and independent, and to not show weakness. This is related to the fact that “more men than women described failure to establish useful working peer groups” (p. 299).
Another example is the student saying “In physics, the only prospect of employment is after grad school… if you want a good job, you need a Ph.D.” (p. 227). This is false—on average, those who get PhDs never recover economically from the opportunity cost. But the researchers took the opinion of an undergrad at face value rather than interrogating it.
2. Subset of the above: taking what interviewees say about grading at face value, even though the interviewees know nothing about the course grades as a whole or how grading works.
Examples are numerous.
In the science classes you may end up with an A or B but only get 50 on a test out of 100 points, because they had to curve it up. That always bothered me because it seems like if you’re only… getting 50 percent of the points you’re not really learning everything [p. 157].Listen, bro, I could write a physics test where 90% of competent PhD physicists would get less than 50%. They still know physics. Getting 100% on a test just means it was an easy test.
I had a physics professor who taught stuff that was totally unrelated to what we were being tested on and totally unrelated to what we were reading [p. 155].
This is a common student complaint, because it’s the perspective of a novice. Novices see differences where experts see connections. Novices might see 135 distinct problems where experts see two principles (say, conservation of energy and conservation of momentum). A good professor recognizes this, but there’s no way around it other than the hard work of the novice becoming more expert.
When you have a class average for Chemistry 103 that gives you a B for scores of 62, I say that’s because they’re trying to fail you out. (Female white engineering switcher) [p. 124]Um, actually that’s showing them specifically not trying to fail you out. When you get a B, that’s not a failing grade.
People get discouraged when they are given such hard tests that no one passes them. Then they curve the scores. What’s the point of giving a test in which you know everyone is going to fail?… I guess they want to weed people out. [p. 124]
WTF? You didn’t fail, because they curved the grades. That’s not failing. And if they curved the grades so people weren’t failing, that’s not weeding people out. Curved grades, and medians near 50%, are very useful for using the full dynamic range of grading. If nobody in the class is getting below 50%, then your assessment tool is wasting half its range.
But the authors take these parochial student misperceptions as establishing a fact that there is a weed-out system in STEM courses. I know how grades vary across disciplines, and if STEM grades harder than others (which it does), that’s because it has actual standards. We don’t just pass everyone regardless of how much (or little) they learn, like our social science colleagues… so maybe don’t come into our disciplines with normative moral judgments on our grading. (However, there are some cases of real, explicit weed-out mentality. Those are indeed harmful. But they’re much rarer than students seem to think.)
3. Taking what their interviewees say at face value… until it disagrees with a pet theory of the authors.
One of the more puzzling pieces of our data is women’s resistance to the theory that the rudeness, hostility, taunting, discounting, or other inappropriate behavior experienced from their male peers (or more rarely from faculty) is an important element in women’s alienation, isolation, loss of confidence, and indirectly, to their decisions to leave [p. 315].
I mean, maybe, yes, the female students are missing something crucial, or maybe we shouldn’t pick and choose when we believe what the interviewees are telling us based on our own prior beliefs and without any evidence.
4. Cherry picking quotes to support pet theories of the authors.
“Although both men and women cited competition in their major as a factor in their switching decisions, it is our contention that they did so for different, gender-related reasons” (p. 265). Maybe! But you have no evidence beyond a single quote that isn’t all that relevant.
Note that 26% of men cited “morale undermined by competitive S.M.E. culture” as a reason for switching, compared to only 4% of women. And 27% of men cited “discouraged/lost confidence due to low grades in early years” as a reason for switching, compared to only 19% of women. 16% of men cited “lack of peer study group support” as a reason for switching, compared with only 8% of women (pp. 232–233). But it's the self-confidence of women that Seymour & Hewitt describe as “so fragile” (p. 265). It’s unclear how to reconcile this with the actual data.
August 1, 2013
As the title suggests, this book aims to discover the primary reasons that talented, well qualified undergraduates leave science, mathematics, and engineering majors. It does this by conducting personal interviews and focus groups with students and then providing thorough support of points made in the book using quotes from the student interviews. These quotes offer a unique perspective not often available. The points made in the book were clearly stated however I felt they were often repetitive. To me, the book read like a stretched out journal paper with student anecdotes thrown in between sections. The book seemed to be intended primarily for instructors or department heads. It offers suggestions on how to combat the problem of student attrition from science, math, and engineering majors however some are realizable, some not. A concern I have with the book however is the possibility that picking and choosing sentences out of interviews to back up specific points could have occurred. I would have liked to see some of the more positive parts of interviews concerning majors instead of only focusing on the negatives as this might have provided more insight into students surveyed. Due to the questions asked, it seems almost as though students are encouraged to place the majority of blame on faculty, programs, and other students rather than themselves. Many of the points made are indeed valid but I question whether all of them are truly as big of a cause for concern as the book makes them out to be at all universities. A lot of the concerns brought up in the book were not things I experienced at my undergraduate university (however I graduated from an engineering program) so whether or not these points are always cause for concern is up to individual programs. Overall the book was an interesting read and offers things to watch out for in programs with a high attrition rate however it is not something I would take as an absolute truth.
May 14, 2011
Clear, well put together, and thought provoking. I'm curious how much has changed over the past 15 years.
April 25, 2018
This book was written in 1997 but feels like it could have been written today. Very thoughtful analysis of a large number of focus groups / conversations with students who started in a STEM major in college. Some switched; some stayed.
This book was honestly a bit tough to read at times in light of my own experience as a woman in STEM and my experience dropping (albeit temporarily) my chemistry major (I doubled with math, so always remained in STEM, but I did have some struggles with my chemistry department). Some of the issues mentioned applied to me; some did not. All were compelling. I wonder whether a similar study done today would find different results.
This book was honestly a bit tough to read at times in light of my own experience as a woman in STEM and my experience dropping (albeit temporarily) my chemistry major (I doubled with math, so always remained in STEM, but I did have some struggles with my chemistry department). Some of the issues mentioned applied to me; some did not. All were compelling. I wonder whether a similar study done today would find different results.
July 8, 2016
Anyone in undergraduate science, math and engineering education MUST read this book in its entirety. A priceless resource.
Displaying 1 - 5 of 5 reviews