Putting the “Co” in Education: Timing, Reasons, and Consequences of College

 Coeducation from 1835 to the Present
Claudia Goldin Harvard University and National Bureau of Economic Research
Lawrence F. Katz Harvard University and National Bureau of Economic Research
The history of coeducation in U.S. higher education is explored through an analysis of a database containing almost all 4-year undergraduate institutions that operated in 1897, 1924, 1934, or 1980. The opening of coeducational in- stitutions was continuous throughout its history, and the switching from single- sex was also fairly constant from 1835 to the 1950s before accelerating in the 1960s and 1970s. Older and private single-sex institutions were slower to become coeducational, and institutions persisting as single-sex into the 1970s had lower enrollment growth than those that switched earlier. Access to coeducational institutions was associated with increased women’s educational attainment.
Women now receive 57 percent of all bachelor of arts degrees in the United States, and 97 percent of them are awarded by coeducational institutions, defined here as institutions that admit both males and fe- males and in which both can take classes together.1 Although almost all undergraduates in the United States today attend a coeducational in- stitution, none could have prior to the appearance of the first such college around 1835. When did institutions of higher education become coeducational, why did they, and what was the impact of coeducation on women’s educational attainment?
Bernie Zipprich provided exceptional research assistance on all parts of the project, and Ryan Sakoda followed in turn. Julia Fifer helped with the coding of the coeducation data set, and Janet Currie and Louis Galambos supplied some of the coeducation dates. Emily Glassberg Sands assisted with the references. Helpful comments on a draft were offered by Stanley Engerman, Julie Reuben, Emily Glassberg Sands, members of the NBER’s Development of the American Economy 2010 Summer Institute, and the Harvard Economic History Workshop. We thank them all. We are grateful to the editor of this Journal and an anonymous referee for further comments. 1 Newcomer (1959) also employs this definition. Whether or not males and females choose to take the same courses and concentrate in the same subjects, and whether or not they are admitted using the same criteria, are different matters.
378 Journal of Human Capital
These subjects are explored through an analysis of a database con- taining information on all institutions of higher education offering 4- year undergraduate degrees that operated in 1897, 1924, or 1934, most of which still exist today.2 Data on all 4-year institutions in 1980 that were founded from 1934 to 1980, or established as 4-year institutions during that period, are added to the database. These data reveal some surprises about the timing of coeducation and the reasons for its increase.3 We find that the founding of coeducational institutions in the public and private sectors occurred at a fairly steady rate from 1835 to 1980. In addition, the rate of switching from single-sex to coeducational status was also relatively continuous from the 1860s through the 1950s. Al- though the rate of switching increased considerably in the 1960s and 1970s, the jump was greatest for Catholic single-sex institutions, espe- cially female-only schools, and in the Northeast. The relative continuity that we find in the evolution of coeducational schools contrasts sharply with the implications of the most commonly found reasons offered for its rise. These reasons include war, ideological change, and national economic downturns.4 Such factors would imply a more episodic evolution of coeducation. The most recent period of switching in the 1960s and 1970s, during which Dartmouth, Princeton, Yale, and a host of elite liberal arts colleges became coeducational, is shown to have increased the fraction of un- dergraduates in coeducational schools by only a small amount. Although the change opened the doors of many of the best institutions in the nation to women, many others had been opened for a long time. We offer a framework to understand the switch from single-sex to coeducational status that emphasizes potential trade-offs between the demand for a coeducational or single-sex environment by current and
2 The 1934 Coeducation College Database contains 769 institutions that existed at some point from 1897 to 1934. Of these, 22 closed by 1934 and 59 of the remaining 748 institutions closed after 1934 (others merged but did not officially close). Therefore, only about 8 percent of the institutions existing in 1934 subsequently closed. 3 The historical literature on coeducation and higher education is sparse. An excellent, but brief, review piece is Rosenberg (1988). See also the essays inMiller-BernalandPoulson (2004). Several articles contain short summaries of the history of coeducation, including Graham (1978). Early writings on coeducation include Thomas (1900) and Woody (1929). On coeducation at Catholic institutions, see Poulson (1995). A thicker literature exists on women’s education. See Newcomer (1959) on women’s colleges and Solomon (1985) on women’s higher education. Some insightful research exists on individual colleges and universities, including Conable (1977) on Cornell University and McGuigan (1970) on the University of Michigan. 4 On the role of the Civil War in reducing the supply of male students, see Newcomer (1959, 12), Graham (1978, 764), and Rosenberg (1988); Solomon (1985, 188) discusses similar factors with regard to World War II. On war and the changing perceptions of women, see Conable (1977), and on the GI Bill in opening women’s colleges to men, see Newcomer (1959), Solomon (1985), and Eisenmann (1997). Newcomer (1959) discusses the Great Depression in spurring coeducation among some schools, and Miller-Bernal (2004) claims that both world wars and the Great Depression created financial hardships for small single-sex institutions and prompted them to switch.
College Coeducation from 1835 to the Present 379
future undergraduates, on the one hand, and expected alumni contri- butions by past and future graduates, on the other. Estimates of hazard models of the time to switching from single-sex to coeducational status suggest the role of alumni influence. We find that older and privately controlled institutions were slower to switch from single sex to coedu- cation. Linear probability models of institutional switchers with controls for both institutional characteristics and time-varying measures of the competitive environment reinforce our findings on continuity in the rate of switching from the 1870s to the 1950s followed by a sharp ac- celeration in the 1960s and 1970s. In a separate analysis we demonstrate that by the 1960s the undergraduate enrollments at institutions that shifted to coeducational status grew faster than those that delayed switching. Greater access to coeducation in higher education, as we demonstrate, had a positive impact on women’s college attainment in the period before the 1930s. Currie and Moretti (2003) present similar findings for the more recent period. Even when women’s colleges existed in an area, an increase in coeducational institutions furthered the college education of women relative to men because women’s colleges were costly and many coeducational institutions were public and less ex- pensive. The paper is structured as follows. We begin with a brief history of coeducation and then move to a discussion of our coeducation database and a description of the evolution of coeducation. We develop a frame- work to understand the shift to coeducation by single-sex institutions and then use our data to detect which factors mattered, thereby testing the validity of the framework. The impact of coeducation on female educational attainment is addressed next. We close with a section on the end of in loco parentis, the spread of sex-blind admissions, and the commencement of real gender equality in higher education.