Expanding the Applicant Pool: Exploring the College Decision-Making of Students from Single-Parent Families
Bateman, Mark, Kennedy, Eugene, College and University
This article explores factors which predispose students from single- and two-parent families to pursue postsecondary education and identify implications for enrollment managers and researchers. The results of this study indicate that while parents play the most pivotal role for each group of students, mothers are of primary importance for students from single-parent families.
Enrollment managers and admissions personnel are under continuous pressure to increase the quality, and in many cases the quantity, of matriculants to their institutions. This pressure has led administrators and policymakers to develop programs that draw matriculants from historically untapped areas of the applicant pool of prospective students. For example, the search for African-American students led to policy development at several levels (including state, federal, and institutional), assisting colleges and universities in the recruitment of this segment of the population. However, there continue to be student groups in the applicant pool that need attention if institutions are to fully explore, recruit, and retain potential students. A particularly important group to address are those students from singleparent families. The lack of attention given to these students has left policymakers and administrators with little information to develop effective and efficient strategies to recruit and retain this important and growing segment of the applicant pool.
According to census data, 25 percent of children in the United States are under the age of 18 and growing up in singleparent homes headed by women-a total of 17 million children (Bureau of the Census 1997). Rawlings and Hernandez (1990) traced the dramatic rise in the number of families headed by single-parent females from 1970 and expect the trend to continue. Research has indicated that these children tend to have lower economic attainment in adulthood than those from two-parent families (Krein 1986). Additionally, students from single-parent families have comparatively low levels of educational attainment and are more likely to experience academic failure and leave school prematurely than is true of children from two-parent families (Hauser and Featherman 1976). In fact, single-parent, female-headed families are often included among lists of disadvantageous traits which put students "at-risk" for a host of negative outcomes (Sartain 1989). According to Krein (1986), living in a single-parent family has a direct negative effect on educational level. In addition, Keith and Finley (1988) found parental divorce was associated with lower levels of educational attainment.
Despite the well documented growth of students from single-parent femaleheaded families and inclusion of these students as at-risk, there is a dearth of literature and research on the postsecondary educational plans of these students. Interestingly, although college choice research has identified parents as critical to the decision making process (Hossler and Stage 1992), students from singleparent families have not been studied. The trends regarding students from single-parent families combined with the role parents play in the college choice process provide the basis for research on the development of educational aspirations of students from single-parent families. Such research will aid enrollment managers in developing effective and efficient policies designed to recruit this growing portion of potential students.
The theoretical framework for this study is drawn from two sources, a model of college choice developed by Hossler and Gallagher (1987), and research by Hossler and Stage (1992). Research on college choice has been grounded in economic, sociological, and combined models (Kohn, Manski, and Mundel 1976; Litten 1982; Chapman 1984). This study utilized the Three Stage Model of College Choice by Hossler and Gallagher (1987). This model is classified as a combined model because it includes variables from both econometric and sociological models. The authors identify three stages in the college choice process: (1) predisposition-"a developmental phase in which students determine whether or not they would like to continue their education beyond high school"; (2) search-the time when students seek more information about colleges and universities; and (3) choice-where students evaluate institutions enabling them to select a specific institution. The predisposition stage addresses the formation of educational aspirations and is the focus of this study.
Using the Hossler and Gallagher model of college choice, Hossler and Stage (1992) developed a theoretical model of predisposition which identified the relationships among (a) family and student background characteristics, (b) student high school experience factors, and (c) the postsecondary plans of ninth graders. Variables included in this study are taken from the Hossler and Stage model of predisposition.
The purpose of this study was to compare factors which influence the formation of postsecondary educational plans of students from single-parent, female-headed families with those of students from traditional two-parent (mother and father) families. The following research questions guided this study:
1. What are the factors included in the formation of postsecondary aspirations for students from two-parent families?;
2. What are the factors included in the formation of postsecondary aspirations for students from single-parent families?; and 3. How do factors which are included for students from two-parent families compare with those from single-parent families in the formation of educational aspirations?
To address these issues, structural equation modeling was utilized. Specifically, data from the base year of the National Education Longitudinal Study of 1988 (Ingles, Abraham, Karr, Spencer, and Frankel 1990) were used to generate covariance matrices for eighth grade students from two-parent families and those from single-parent, femaleheaded families. The LISREL VII (Joreskog and Sorbom 1988) computer program was used to fit baseline models (see Figure 1) for each group. Following this, a multisample analysis strategy was used to locate differences in the structural models of the two groups.
Review of Variables
The variables included in this study were identified by Hossler and Stage in the development of their model of predisposition. These variables were extracted primarily from the status attainment and college choice research. As such, the variables included in this study were:
1 Student Ability (Carpenter and Fleishman 1987);
2. Involvement in High School Activities (Astin 1984; Hossler and Gallagher 1987; Hossler and Stage 1992);
3. Parents' Expectations (Sewell and Shaw 1968; Carpenter and Fleishman 1987; Hossler and Stage 1992);
4. Socioeconomic Status (Hause 1969; Jackson 1978; Ekstrom 1985);
5. Parents' Educational Level (Hause 1969; Jackson 1978; Carpenter and Ekstrom 1985; Fleishman 1987)
The Role of Family Structure
Figure 1 presents a modified version of the Hossler and Stage model of predisposition. This model includes indicators of each of the variables discussed above. However, this model assumes the traditional family structure with a mother and father in the household. This model fails to address the possibility that the influence of mothers and fathers may not be the same when one is missing. This issue is particularly relevant to the factor of parents' expectations. As noted in the review of the literature, studies have not examined the possibility that differences exist between mothers' and fathers' expectations in predicting educational aspirations. Separating mothers' and fathers' expectations is an important issue in examining the educational aspirations of students from single-parent families. The question to be asked is, do mothers and fathers have equal impact if both are not in the home? The modification of the Hossler and Stage model attempts to address this question.
Sample The sample for this study was taken from the base year student file of the National Education Longitudinal Study of 1988 [NELS88] (Ingles, Abraham, Karr, Spencer, and Frankel 1990). The NELS88 study consists of nationwide surveys of eighth grade students, their parents, teachers, and administrators. The surveys solicited a variety of information on students' home and school experiences. These data were collected both by questionnaires administered to respondents and by reviews of school records.
For these analyses the sample was restricted to those students who reported that both parents were alive at the time of the survey. This was necessary given that a key aspect of the study involved exploring the relative impact mothers and fathers have on the formation of their children's educational aspirations. This action reduced the base year file from 24,000 to 18,736 students. Of this number, 15,178 students lived with both parents and 2,892 lived with their mothers. These data constitute the samples for the study.
As noted above, this study builds upon the Hossler and Stage (1992) model of predisposition. Accordingly, the background variables considered are (a) father's education (DADED), (b) mother's education (MOMED), and (c) family income (INCOME). The intervening variables are (a) mother's educational expectations for the respondent as perceived by the respondent (MOMEXP), (b) father's educational expectations for the respondent as perceived by the respondent (DADEXP), (c) school grades (GRADES), and (d) school activities (ACTIVITY). Finally, the dependent variable for this study is the student's reported postsecondary educational aspirations (EDASP).
The purpose of this study was to determine if a modified version of the structural model of predisposition to college posited by Hossler and Stage (1992) was adequate for students from singleparent, female-headed families as well as those from traditional two-parent families. To test this hypothesis, the computer package LISREL VII (Joreskog and Sorbom 1988) was utilized to estimate parameters for the path model with latent variables presented in Figure 1. In line with recent recommendations (Byrne 1989), conclusions regarding the adequacy or fit of the model are based on (a) examination of the solution, (b) theoretical predictions, (c) chi-square and other overall fit statistics, and (d) detailed examination of misfit (modification indices and standardized residuals) for purposes of refining the model. It should be noted that although strict reliance on statistical significance levels of chisquare have guided structural equation modeling in the past, researchers now recognize important limitations of this approach. In particular, with large sample sizes, the overall fit chi-square will almost certainly be significant. Although a number of alternative measures of overall fit have been proposed, as Bollen (1989) notes, "None of the fit measures have both calculated values uninfluenced by N (sample size) and means of their sampling distributions unrelated to N." For this study, we present chi-square divided by its degrees of freedom and we emphasize the goodness of fit (GFI), adjusted goodness of fit (AGFI), and root meansquared residual (RMR) indices generated by the LISREL program. Sample size does not enter into the calculation of the GFI and AGFI, but there is evidence that their sampling distributions show a positive shift with larger samples. These indices have a range from 0 to 1.0, with values near 1.0 indicating a good fit. The RMR is a measure of the typical distance between the sample covariance matrix and that implied by the fitted model. Values close to 0 generally indicate an acceptable fit.
This study was also concerned with the possibility of measurement and/or structural differences in the model for the two groups considered. To address this issue a multigroup analysis was undertaken. These analyses tested a series of increasingly restrictive hypotheses regarding group similarities. These hypotheses were as follows:
A. The variance-covariance matrices for the two groups are identical.
B. The pattern or form of the model is the same for the two groups.
C. The structural relations of the model are invariant for the two groups.
D. Factor loadings are invariant for the two groups.
E. Item error variances are the same for the two groups.
Hypothesis A provides an overall test for differences in the variance-covariance matrices for the two groups studied. Should this hypothesis be rejected, this provides support for attempts to identify the sources of the differences. Accordingly, hypotheses B through E provide for increasing levels of similarity between the groups. Hypothesis B argues that the groups have the same number of latent and observed variables and the same location of fixed and free parameters. In other words, the model in Figure 1 holds in both groups but the observed and latent variables may function differently.
Assuming that B holds, hypothesis C argues that the path coefficients among the latent variables and the errors and correlated errors among the endogenous latent variables have the same magnitudes and directions in the two groups. Assuming B and C, hypothesis D argues that the factor loadings for the observed variables on the latent variables are identical for the two groups. And, finally, hypothesis E is the most restrictive of all considered. It argues for total invariance of the models. This is accomplished by adding the constraint that the errors of measurement for the observed variables are identical for the two groups.
These hypotheses are tested sequentially, by examining chi-square divided by its degrees of freedom, the GFI and RMR indices.
It should be noted that this study had several limitations. First, female-headed families were identified as those wherein no male figure was present and the father was alive at the time of the survey. There is some evidence to suggest that the timing of the father's departure is related to social and emotional consequences for children-that is, while initially children may experience multiple role and status disruptions, over time stabilizing adjustments occur (see Heartherington, Cox, and Cox 1978). Future research may find that children's postsecondary plans vary depending upon the timing of the father's departure. Second, this study did not differentiate children from female-headed families with respect to gender, ethnicity, and other demographic characteristics. Future research may find this to be a useful avenue of research.
Table 1 presents Pearson correlations, means, and standard deviations (SD) for the variables considered in this study. These statistics are presented separately for students from two-parent families (mother and father) and those from female-headed families (mother only). With the exception of family income (INCOME), the means for the variables presented in the table are similar for the two groups. As might be expected, the difference in family income indicates that two-parent families are typically better off financially.
The Pearson correlations presented in Table 1 show that, for both groups of students, mother's and father's expectations (MOMEXP) and (DADEXP) yield the strongest correlations with their children's educational aspirations of any variables considered. However, the relationships are much weaker for female-headed families, and for this group, mother's expectations are decidedly more significant than those of the father. Following parent expectations in importance are school grades and parent educational levels. With regard to the latter, in both groups there is a tendency for the correlation between father education (DADED) and aspirations to be stronger than the correlation between mother education (MOMED) and aspirations.
Adequacy of the Model
Figures 2 and 3 present the standardized path coefficients for the structural model. Again, these results are presented separately for the two groups. The fit indices presented, chi-square divided by degrees of freedom (chi-square/df), GFI, AGFI, and RMR, all indicate that the model is a reasonable fit for each group the values for GFI and AGFI are close to 1.0 and RMR is close to 0. The chisquare/df values are large relative to recommendations by some authors (Byrne 1989). However, as noted above, the magnitude of chi-square is misleading given the size of the samples used for this study. To illustrate this point, the covariance matrix for two-parent families was reanalyzed but with sample size specified as 1,000 in the LISREL program. The resulting value for chi-square/df was 1.7.
The total coefficient of determination for structural equations in each group is approximately 0.33. Also, the squared multiple correlation for structural equations predicting educational aspirations is 0.52 for female-headed families and 0.58 for two-parent families. This suggests that the model has slightly more predictive power when both parents are in the home.
With regard to the strength of relations in the data, the results show potential differences across groups. First, when both parents are in the home, the factor loading for father's education on the education latent variable is stronger than the loading for mother's education. However, when the mother is the only parent in the home, the loadings appear to be approximately equal. Second, with regard to the latent variable "parental expectations," when both parents are in the home, the factor loadings for mother's expectations and father's expectations are approximately equal. When the father is absent from the home, the expectations of the mother are far more significant with respect to this latent variable.
Comparison of Models From Single
and Two-Parent Families
The results presented in Figures 2 and 3 suggest that although the model may be acceptable for both groups, there may exist some important within-group differences. Specifically, the standardized coefficients indicate a differential import of father's education and expectations contingent upon whether or not he is in the home. To systematically assess this, the results of multisample analyses described in the text are presented.
Table 2 presents results of the five sequential hypotheses described on page 5. The initial test addressed the hypothesis that the variance-covariance matrices for the two groups were essentially identical. The chi-square/df value of 52.6 and the other goodness of fit indices, particularly for female-headed families, show that this hypothesis should be rejected-the groups differ with respect to the structure of the variance-covariance matrices. Standardized residuals and modification indices (not shown) suggest that the primary sources of misfit pertain to the role of the education and expectations of mothers and fathers. Subsequent tests presented in the Table confirm that the models for the two groups are essentially invariant except as regards the measurement properties of father's and mother's expectations. Specifically, the chisquare/df for model D, which allows for differences in the measurement error of these two variables, is 17.0. Once these error variances are constrained to be equal, this ratio increases to 47.5. Inspection of the standardized solution for each group shows that when both parents are present, father's expectations tend to dominate, but when the father is absent, the educational expectations perceived by the student are largely due to the mother.
The results of this study indicate that the Hossler and Stage model provides a reasonable fit for children from two-parent families as well as those from femaleheaded families. However, the model has slightly more predictive power for children from two-parent families and the importance of mothers and fathers differed for the two samples. Specifically, when the father is not in the home, his impact, relative to that of the mother, is significantly diminished. For example, when both parents are in the home, father's and mother's expectations have equal loadings on the parent expectation factor. But when the father is absent, the factor loading for mother's expectations is significantly higher. Other than these differences, the models for the two groups are remarkably similar, an important finding in the development of policies.
These results lead us to the conclusion that there are more similarities than differences in the ways in which children from two-parent and single-parent families formulate their plans for postsecondary education. Our data also indicate that there are no meaningful differences in the levels of education desired by these two groups. These findings suggest that children from single-parent, female-headed families may not warrant special consideration with regard to predisposition to college.
The results from this study lead us to the following recommendations for researchers and policymakers.
Although this study indicates that students from single- and two-parent families are similar in factors which predispose attendance to a college or university, several questions emerged regarding the college choice process for students from single-parent families, from the development of educational aspirations to enrollment in a college or university. As mentioned previously, these students are reported to have relatively low levels of educational attainment. It is important to determine why a discrepancy exists between aspirations and attainment. The following questions would be useful in guiding further research on students from single-parent families.
Do students from single-parent families have realistic educational aspirations? Future research may seek to examine the basis for students' aspirations and identify the support structures and resources available to these students. Students from families which are unable or unwilling to financially support a college or university education may develop unrealistic aspirations based on lack of understanding of family constraints. This research would require that parents, as well as students, be included in the research. In addition, questions related to student readiness should also be probed. One aspect of readiness is the high school courses a student anticipates taking. The student who expects to enter a college or university while taking vocational courses may have unrealistic aspirations or expectations.
At what point do students from singleparent families alter their decision to pursue postsecondary education? One advantage of the Hossler and Gallagher model is that it presents choice as a process befitting the longitudinal nature of college decision making. The present study examined only the predisposition stage of this model. Future studies should focus on the search and choice phases to determine if these stages highlight changes in the expectations or aspirations of students from single-parent families. The identification of the development period in which students are likely to alter their plans would lead to targeted intervention strategies.
Do gender and ethnic differences exist between students from single-parent families? This study was an important first step in exploring the college decision making process of students from singleparent families. However, we did not address the possibility that differences may be present between or within ethnic groups. This line of research is particularly important considering the differential rates of single-parent homes among ethnic groups. For example, 63 percent of all African-American families are headed by a single-parent (Bureau of the Census 1996).
Although the findings from this study suggest that students from single- and two-parent families are similar in the predisposition stage of college choice, and thus influenced by the same enrollment strategies, we believe the study also provides empirical leads for policy development. These recommendations are targeted at the mothers of students from singleparent families and highlight the importance of financial aid.
Unlike previous research, this study indicates that mothers', and not fathers', expectations are the central and most important determinant of college decision making. Thus, the mothers of students from single-parent families should be targeted by policymakers at a variety of levels (particularly state and institutional) with information regarding postsecondary attendance for their son or daughter. This information should relate to both pursuing education after high school and providing guidance on individual types of institutions available for their child.
Not surprisingly, the respondents from female-headed, single-parent families are less well off financially than those from two-parent families, indicating a potential need for financial assistance to be addressed during recruitment and provided throughout a student's educational experience. Additionally, mothers of potential students may need assistance in planning for the costs of higher education. Such assistance should address the need to plan for postsecondary education, an understanding of alternative methods for financing a college or university education (i.e. pre-paid tuition plans), and relative costs of institutions. Thus, when recruiting students from single-parent, female-headed families, financial considerations should be addressed early.
Astin, A. 1984. Student involvement: A developmental theory for higher education. Journal of College Student Personnel. 25(4): 297-308.
Bollen, K. A. 1989. Structural Equations With Latent Variables. New York: John Wiley & Sons.
Bureau of the Census. 1997. Statistical Abstract of the United States. Washington DC: U.S. Department of Commerce, Bureau of the Census.
Byrne, B. M. 1989. A Primer of LISREL: Basic Applications and Programming for Confirmatory Factor Analytic Model. New York: Springer-Verlag.
Carpenter, P. G. and J. A. Fleishman. 1987. Linking intentions and behavior: Australian students' college plans and college attendance. American Educational Research Journal. 24(1): 79-105.
Chapman, R. C. 1984. Toward a Theory of College Selection. Unpublished Manuscript. Alberta, Canada: University of Alberta.
Christensen, S., J. Melder, and B. Weisbrod. 1975. Factors affecting college attendance. Journal of Human Resources. 10(1): 77-94.
Davidson, N. 1990. Life without fathers: America's greatest social catastrophe. Policy Review. 51 (Winter): 40-44.
Ekstrom, R. B. 1985. A Descriptive Study of Public High School Guidance: Report to the Commission for the Study of Precollege Guidance Counseling. Princeton, NJ: Educational Testing Service.
Hause, J. C. 1969. Ability and schooling as determinants of lifetime earnings, or if you're so smart, why aren't you rich. American Economic Review. 59: 289-298.
Hauser, R. M. and D. Featherman. 1976. Equality of schooling: Trends and prospects. Sociology of Education. 49: 99-120.
Heartherington, E. M., M. Cox, and R. Cox. 1978. The aftermath of divorce.In J. H. Stevens, Jr., and M. Matthews (eds.) Mother-Child, Father-Child Relations. Washington, D.C.: National Association for the Education of Young Children.
Hossler, D. and K. S. Gallagher. 1987. Studying student college choice. A three phase model and the implications for policymakers. College & University. 2(3): 207-221.
Hossler, D. and F. Stage. 1992. Family and high school experience influences on the postsecondary educational plans of ninth-grade students. American Educational Research Journal. 29(2): 425-451.
Ingles, S. J., S. Y. Abraham, R. Karr, B. D. Spencer, and M. R. Frankel. 1990. National Education Longitudinal Study of 1988: Base Year Student Component File User's Manual. Washington, D.C.: U.S. Department of Education.
Jackson, G. A. 1978. Financial aid and student enrollment. Journal of Higher Education. 49: 548-574.
-----. 1982. Public efficiency and private choice in higher education. Educational Evaluation and Policy Analysis. 4(2): 237-247.
-----.1986. MISSA, the Fall of Saigon, and College Choice, 1972 to 1980. Paper presented at the annual meeting of the Association for the Study of Higher Education. San Diego, CA.
Joreskog, K. G. and D. Sorbom. 1988. LISREL 7: A Guide to the Program and Applications. Chicago: SPSS.
Keith, V. M. and B. Finley. 1988. The impact of parental divorce on children's educational attainment, marital timing, and likelihood of divorce. Journal of Marriage and Family. 50(3): 797-809.
Kohn, M. G., C. F. Manski, and D. Mundel. 1976. An empirical investigation of factors influencing college going behaviors. Annuals of Economic and Social Measurement. 5(4): 391419.
Krein, S. F.1986. Growing up in a singleparent family: The effects on education and earnings of young men. Family Relations. 35(1): 161-168. Litten, L. M. 1982. Different strokes in the applicant pool: Some refinements in a model of student college choice. Journal of Higher Education: 53(4): 383-402.
Rawlings, S .W. and D. Hernandez. 1990. Household and Family Characteristics: March 1997 and 1989. (Contract No. 447). Washington, DC: Bureau of Census.
Sartain, H.W. 1989. Nonachieving Students at Risk: School, Family, and Community Intervention. Washington D.C.: National Education Association.
Sewell, W. H. and V. P. Shaw. 1968. Social class, parental encouragement, and educational aspirations. American Journal of Sociology. 73: 559-572.
Mark Bateman is Associate Dean for Academic Affairs and Associate Professor in the School of Education at Baylor University. His research interests are college choice and education policy.
Eugene Kennedy is an Associate Professor in the Department of Educational Leadership and Research at Louisiana State University. His interests include educational evaluation, research methods, and school reform.…
Questia, a part of Gale, Cengage Learning. www.questia.com
Publication information: Article title: Expanding the Applicant Pool: Exploring the College Decision-Making of Students from Single-Parent Families. Contributors: Bateman, Mark - Author, Kennedy, Eugene - Author. Journal title: College and University. Volume: 75. Issue: 1 Publication date: Summer 1999. Page number: 2+. © American Association of Collegiate Registrars and Admissions Officers Fall 1999. Provided by ProQuest LLC. All Rights Reserved.
This material is protected by copyright and, with the exception of fair use, may not be further copied, distributed or transmitted in any form or by any means.