"Do Female BYU Students Have Lower Educational Ambition Than Their Male Counterparts? Results from a Recent Survey"

Stephen Cranney

SquareTwo, Vol. 3 No. 3 (Fall 2010)

 

5 Comments

Homepage

 
   

 

            Despite the educational emphasis of the LDS Church, differential graduation rates from BYU-Provo suggest that there is a gender-based component affecting women’s level of education among members of the church. The following table of enrollment versus graduation rates for female versus male students at BYU illustrates the issue:

Table 1: Enrollment versus Graduation Rates, Disaggregated by Sex, BYU, 2002-2008 [1]
Year Females as % of All Enrollees (Males) Females as % of All Graduates (Males) Difference Between Enrollment and Graduation Rates for Females
2002     54% (46%)     52% (48%)     -2.0
2003     56.1 (43.9)     51 (49)     -5.1
2004     56.8 (43.2)     51 (49)     -5.8
2005     56.7 (43.3)     50 (50)     -6.7
2006     54.5 (45.5)     50 (50)     -4.5
2007     56.3 (43.7)     50 (50)     -6.3
2008     54.2 (45.8)     50 (50)     -4.2

            Table 1 shows that, while the graduating class has about equal representation from both genders, the initial freshman classes show disproportionality towards females, implying a level of educational attrition. There appears to be something causing females at BYU to drop out of college at a higher rate than men. While theoretically the LDS Church’s culture of maternal child-rearing could be identified as the causal mechanism, no studies that I am aware of have provided explicit evidence that connects the two. This is perhaps a fruitful area for future research.

            As part of a larger study assisted by the Brigham Young University Department of Church History and Doctrine, I administered an anonymous survey to several large general education religion classes and received 379 responses. As these surveys come from required classes, they provide a representative cross-section of the BYU student body. The survey contained questions that, among other things, measured educational ambition. Each survey contained questions that had several measures of religiosity and educational ambition, as well as background questions (gender, year-in-school, etc.).

            I utilize three different measures for ambition: GPA, expected future salary, and the decision to attend graduate school. The GPA measure operates under the assumption that ambitious students work harder in school and therefore do better in school ceteris paribus. On this measure, in order to assure a GPA more representative of their entire scholastic career, the question was not asked of those who had only attended one or less complete semesters of school. The decision to attend graduate school measures an educational level of ambition, while the expected future salary measures financial ambition. Actual question wordings are included in the appendix. 

Table 2: OLS Regressions, Gender on Three Measures of Educational Ambition, With Control Variables Included
  (1) OLS (2) OLS (3) OLS
Measure of Educational Ambition (Dependent Variable) Cumulative GPA Expected Future Salary Not Planning to Attend Graduate School
Constant 3.55***
 (.44)
93591.04***
(14005.28)
.23***
(.037)
Male Gender -0.55
(.15)
167223.4***
(46823.7)
-.236***
(.037)
Observations 366 374 374
Adjusted R-squared -.0024 .032 .089

*=significant at the .05 level
**=significant at the .01 level
Notes: All regressions include an intercept. Heteroskedasticity-robust standard errors appear in parentheses below the estimated coefficients.

           These results demonstrate that there exists a real difference between the graduate school plans of females and those of males within the LDS context. Specifically, expected future salary and plans to attend graduate school seem to be robustly correlated with gender, with males significantly more likely to plan to attend graduate school and to expect a higher salary. Interestingly, there does not seem to be a correlation between GPA and gender, thus implying that even though women are not as ambitious in terms of expectations of future formal education, women do not lack in ambition during their undergraduate education compared to men. The R-squared values caution us against reading too much into these results, as gender accounts for only about 3% of the variation on salary and 9% of the variation on the decision to attend graduate school or not.

           Overall, these regressions support the plausible conclusion drawn from the graduation rates figures: something within the LDS community disincentivizes female educational ambition. Despite the fact that women's GPAs do not differ significantly from men's, there is still a noticeable attrition among female students at BYU when we compare female graduate rates to female enrollment. Furthermore, female BYU students are significantly less likely than their male counterparts to plan to attend graduate school, and they also expect to be paid a lower salary after school. While the exact causal mechanisms remained unexamined, this is a fruitful area for future research.

Appendix: Survey Questions Related to Educational Ambition

What is your current overall BYU GPA? If this is your first semester at BYU, please leave blank.

Approximately how much do you see yourself personally making in 30 years?

What type of graduate or professional school do you plan on attending? Check all that apply. (Options: None, Law school, other, medical/dental school, Masters degree, PhD, Business school).

           

NOTES:

[1] Brigham Young University Y Facts. Graduates by gender. http://yfacts.byu.edu/viewarticle.aspx?id=218. Brigham Young University Y Facts. Freshman enrollment by gender past eight years. http://yfacts.byu.edu/viewarticle.aspx?id=83. [Back to manuscript]

 

Full Citation for This Article: Cranney, Stephen. (2010) "Do Female BYU Students Have Lower Educational Ambition Than Their Male Counterparts? Results from a Recent Survey," SquareTwo, Vol. 3 No. 3 (Fall), http://squaretwo.org/Sq2ArticleCranneyEducation.html, accessed [give access date].

Would you like to comment on this article? Thoughtful, faithful comments of at least 300 words are welcome. Please submit to SquareTwo.

COMMENTS: 5 Comments

1) Robert Couch, Salem Oregon, 9 March 2011

Thank you for an interesting article.

Do you have any sense if this is a BYU-specific tendency, or does it simply reflect a broader cultural tendency?  

Liang Zhang has shown with older data from 1997 that women were in general less likely to enroll in graduate school (“Advance to Graduate Education: The Effect of College Quality and Undergraduate Majors,” in The Review of Higher Education, v. 28 no. 3, 2005, pp. 313-338). It would be very interesting to know if this result still holds in samples after 1997, since this would help in establishing some sense as to the extent this result is a BYU-specific or Mormon specific effect.

____________________________________________

2) Stephen Cranney responds to Robert Couch, 9 March 2011:

Given BYU's cultural iconic status within the LDS community, it would seem reasonable to generalize these tendencies to the LDS culture in general. However, the risk of confounding factors is too great to take a firm stance on this, and a degree of uncertainty must be admitted. For example, family goals undoubtedly play a large role in the decision to come to BYU for many of its students; this could risk a selection bias as BYU attracts more people for whom family goals play a more exclusive of their future plans. Consequently, the data could just be showing that BYU attracts a disproportionate number of young women who see pursuing a career and family as mutually exclusive. Without further, more specific data, we are prevented from drawing any sure conclusions on the matter. 

Whatever the case, the data do strongly indicate that something Mormon-specific is going on here; as the Population Reference Bureau now records a growing higher-education gap between female enrollment (at the high end) and male enrollment, (https://www.prb.org/Articles/2007/CrossoverinFemaleMaleCollegeEnrollmentRates.aspx). As this survey was not a time-series, we cannot see how or if Mormon women are following this trend. What is established, however, is that in terms of educational intention they are still lagging behind males which, if enrollment is taken as a mirror indicator of intention, is opposite the trend found in the US in general.

__________________________________________________________________

3) Roy Zwahlen

Great article on a fascinating subject.

I am curious about your reference to the Population Reference Bureau study in the comment section. Are there any numbers on the enrollement and/or graduation rate comparison nationally? I know the study references percentage of the total population enrolled in college by female/male (which is difficult to compare to your enrollment/graduation rate comparison). I think your numbers say something about the difference between men and women at BYU but I am not sure what it says about what effect the "Mormon" factor has since we have no national numbers to compare it to. While your conclusions may likely be true, it is hard to have certainty without that comparison.

Another "Mormon" cultural factor which may make LDS/NonLDS comparison difficult may be the LDS emphasis on education and the male providing for the family (which arguably is stronger in LDS culture when compared to NonLDS American culture generally). Returned missionaries at BYU may be more motivated to graduate from college than their nonLDS counterparts, have a better GPA, have more ambition for graduate school, and possibly even greater amibition for a higher salary (although I think there is a broad assumption across the spectrum that if you go to graduate school you will command a higher salary which means that these two factors may be dependant or exactly the same when and individual selects both). It seems all of your measurements may be affected by this factor.

In my mind, I think data that would reinforce (or possibly change) your conclusions would be the comparison of enrollment/graduation rates between LDS women at BYU and NonLDS women in America. This would help (I suspect) affirm your conclusions.

_______________________________________________________________

4) Ian Fillmore

In his article, Stephen Cranney publishes survey results that relate to possible explanations for why females comprise a larger share of Freshmen enrollees (roughly 55%) than graduates (roughly 50%).  The author wants to test whether women at BYU are less educationally ambitious than their male peers.  Unfortunately, the main contribution of the paper—the survey results in Table 2—contains two blatant errors and one suspicious looking result.  In light of these problems, it is doubtful whether any meaningful conclusions can be drawn from the results as presented.  I list the two errors below along with two additional concerns.

Cranney presents three regressions based on a survey of BYU students.  The regressions are simple—amounting to no more than a comparison of means between men and women.  According to column 1 of Table 2, the women in the sample had a mean GPA of 3.55, while the men had a mean GPA of 3.00.  Cranney claims that the difference between men and women (-0.55) is not statistically significant.  But the standard error on that difference (0.15) implies a t-statistic of 3.67 which is significant at the 1% level.

Second, in column 3 of Table 2, Cranney reports that 23% of women in the sample reported no intention of attending graduates school.  Among the men, the fraction is much lower.  In fact, it is so low that it drops below zero (-0.6%).  Fractions can’t be negative.

Column 2 of Table 2 says that women reported an average expected income of $93,591.  The men reported an average income of $260,814.  Obviously, these incomes are implausibly high.  While this is not an error like those mentioned above, I would have expected some discussion about why the responses were so unrealistic.  Are students uninformed about real world salaries?  Did a few respondents accidentally add an extra zero and significantly skew the results?

I’m also a bit suspicious about Table 1 which motivated the entire exercise.  Freshman admissions are not the only way that students arrive at BYU.  Perhaps transfer students skew male which would also lower female representation among graduates.  One could imagine other explanations as well.  Thus, Table 1 alone does not provide convincing evidence that BYU women have lower attrition rates than their male peers.

_______________________________________________________________

5) Stephen Cranney, in response to Ian Fillmore

Thanks to Ian Fillmore for his helpful critique of my article "Do Female BYU Students Have Lower Educational Ambition Than Their Male Counterparts? Results from a Recent Survey." He does indeed point out serious errors in the numbers reported in the table.  

The paper in question was written nearly four years ago when I was an undergraduate. In an attempt to figure out exactly what happened I have attempted to locate my original statistical coding, but have been unsuccessful. However, I have been able to retrieve the original data file that was sent to me. Here I re-do my results in an attempt to clarify the relationships discussed later and in order to speculate about how the errors crept in. Here I articulate my coding in order to assure that my results are replicable.

In this version all three models in Table 3 are generated by simply using “regression [dependent variable] male, r” in the statistical program Stata. The new results for Table 2 are below:

Table 2: OLS Regressions, Gender on Three Measures of Educational Ambition, With Control Variables Included

 

Cumulative GPA

Expected Future Salary

Not Planning to Attend Graduate School

Male

-0.0625

170712.1***

-0.226***

 

(0.145)

(46517.9)

(0.0363)

 

 

 

 

Constant

3.552***

93536.1***

0.281***

 

(0.142)

(13943.9)

(0.0322)

Observations

371

379

379

Adjusted R2

-0.002

0.034

0.087

Standard errors in parentheses
+ p < .10, * p < .05, ** p < .01, *** p < .001
Notes: All regressions include an intercept. Heteroskedasticity-robust standard errors appear in parentheses below the estimated coefficients.

As can be seen by comparing these results, the constant for cumulative GPA is the same across both studies. Here I have a few more observations than were reported in the original study; again, without the specific coding used to clean the data I can’t say for sure why there is a slight difference, but the difference is indeed slight. The adjusted R squareds are also very similar to their counterparts in the original paper. For expected future salary, the numbers are again slightly different (167223.4 for the male coefficient in the original article versus 170712.1 in the updated version, and 93591.04 for the constant in the original one versus 93536.1 in the updated one), but once again are very similar. In his critique Dr. Fillmore states that he believes these numbers are “implausibly high,” and wonders if some people accidentally added some extra zeros. I investigated this possibility even further. There were a few outliers, but nothing radically out of the ordinary. Specifically, one person expected to be making $6 million a year (a male), one $4 million (a male), one $3 million (a male), one $2 million (a male), 1 $1.7 million (a female), one $1.8 million (a female), and 10 $1 million (9 males, 1 female). Once these outliers are taken out of the equation, the average expected salary for a male is $76,615.02 and for females it is $71,672.94. It is interesting to note that taking the outlier out reduces the gender differential (although they are still significantly different). However, it is hard to know if this is the result of the male respondents not taking the survey seriously, or a case of inflated male ego about their potential to make money.

Finally, on the not planning to attend graduate school measure, my results are once again very similar. However, he makes the valid critique that in the original results men have a negative percentage.

This, combined with the male coefficient in the cumulative GPA model, are two major errors in the paper.

Given that the analyses were rather simple, and that the rest of the cells seem to be in order, I am inclined to attribute these two cell-specific errors to my own sloppiness while inputting the tables by hand. When I first submitted the paper, I remember (vaguely, it has been four years) simply dumping the raw tables generated by Stata in the original paper, so it is likely that these errors crept in as a result of my own sloppiness during the preparation for publication phase.

On a more conceptual note, Mr. Fillmore also points out that “Freshman admissions are not the only way that students arrive at BYU. Perhaps transfer students skew male which would also lower female representation among graduates. One could imagine other explanations as well. Thus, Table 1 alone does not provide convincing evidence that BYU women have lower attrition rates than their male peers.” I see no reason why transfer students would skew male, but without more detailed data on college entry it is true that one can’t just assume a simplistic freshman to senior college career. I don’t think that makes the table completely irrelevant, but he is right that it does not provide “convincing evidence.”

While it is worth pointing out that the newly updated figures do not change the final take away from the paper, Dr. Fillmore was right to be concerned about the validity of the main points given the errors in the table. The original data are available upon request for anyone who wants to replicate my results by using the syntax above.

Ian Fillmore is to be commended for taking the time to point out my errors, and I thank him.