Educational Performance and Persistence of Bereaved College Students
Anecdotal evidence and clinical impressions suggest that bereaved college students are at risk for academic difficulties and dropout (Balk, 2001; Rickman, 1996; Toth, Stockton, & Browne, 2000; Zinner, 1985). However, no investigations have directly addressed the effects of a significant death loss on the educational performance or persistence of bereaved undergraduates. Before universities can initiate intervention programs targeted at fostering the academic success of bereaved college students, determining whether or not this population is at risk is necessary. As retention is a by-product of both academic and institutional success, the identification of at-risk populations is of utmost importance in the world of higher education.
Tinto (1975, 1993) described the process of dropout as a longitudinal one in which students cyclically evaluate their commitment to their academic goals and to the institution they are attending. According to Tinto (1975), the outcome of this evaluation (e.g., dropout vs. persist) hinges upon the level of both academic and social integration experienced by students. Academic and social integration are largely determined by the "interactions between the individual and the academic and social systems of the college during the person's experiences in those systems" (p. 94). In [End Page 225] essence, the more engaged a student is within both the academic and social realms of the university system, the higher the likelihood that he/she will persist. According to Tinto (1975; 1993) grade performance is the single best predictor of academic integration, while peer-group associations are the most directly related to social integration.
Tinto (1975) acknowledged that "very frequently, events in the social system external to the college can affect integration within the more limited social and academic systems of the college" (p. 97). Although he recognized that these events can lead to dismissal or withdrawal, he went on to emphasize his belief that the effects of such events are best understood with regard to their influence upon the student's ever-changing commitments to his/her academic goals and to the institution.
The experience of a death loss is a clear example of an event in the social system often external to the college that affects not only the academic and social integration of college students, but also their subsequent commitments to their academic goals and to the institution. The impact of a death loss on academic and social integration acts as a catalyst for students to re-evaluate their commitments and may lead to decreased educational performance, academic probation, academic dismissal, or voluntary withdrawal.
At any one point in time, 22-30% of college undergraduates are likely to have experienced a death loss in the previous 12 month period, while 35-48% are likely to be in their first 24 months of grieving (Balk, 2001). These numbers may appear counterintuitive and perhaps somewhat shocking as the college experience is often viewed as a time of youth, growth, and promise (LaGrand, 1985); however, they highlight the salience of bereavement as a critical issue for college students (Balk, 2001).
Despite the surprising lack of research on bereaved college students as a distinct group of grievers (Balk et al., 1998; LaGrand, 1981), there are some indications from the literature that grief affects the academic experience and peer-group associations of bereaved college students. The idea that bereaved colleges students are at risk with regard to academic difficulties is suggested by clinical experience and observations carried out on college campuses (Balk, 2001; LaGrand, 1981; Rickman, 1996; Toth et al., 2000). Two small qualitative empirical investigations were located that addressed the idea that bereaved college students experience difficulties with concentration and studying (Balk & Vesta, 1998; Silverman, 1987). Research focused on early and middle adolescents (i.e., middle and high school students), although somewhat mixed (Fleming & Balmer, 1996), does indicate academic difficulties as an aspect of the grief experience (Balk, 1983 as cited in Fleming & Balmer, 1996; Gray, 1987 as cited in Fleming & Balmer; Harris, 1991; Martinson & Campos, 1991). The current lack of quantitative investigations focused on actual educational performance is problematic and the present investigation seeks to address this disparity.
In relation to peer association, the literature on bereaved college students suggests that the university campus is often an isolating and lonely place for grieving undergraduates (Balk, 2001; Balk, Tyson-Rawson, & Colletti-Wetzel, 1993; LaGrand, 1985; Stephenson, 1985; Toth et al., 2000). Silverman (1987) found that female college students grieving father deaths received little social support from their peers, often being told "you'll get over it" (p. 393). Balk (1997) found that although most bereaved students described talking about the death as helpful, most of the people they interacted with on campus appeared [End Page 226] uncomfortable when the topic was brought up. These findings are consistent with research describing the experience of younger bereaved adolescents. For example, parentally bereaved high school students reported discomfort in interpersonal interactions (Harris, 1991; Servaty & Hayslip, 2001), and those experiencing the death of a sibling reported negative alterations in their peer relationships (Balk, 1990; Tyson-Rawson, 1996).
When viewed through the lens of Tinto's (1975) model, these academic and peer group related difficulties represent threats to academic and social integration, which may lead to bereaved college students being at greater risk for academic problems, including dropping out. The purpose of the present investigation was to empirically test the following hypotheses:
1. Bereaved college students will earn lower GPAs during the semester of their death loss than a group of matched peers.
2. Bereaved college students will complete fewer credit hours during the semester of their loss than a group of matched peers.
3. Bereaved college students will be more likely to be classified with a problematic academic standing (e.g., academic probation, dropped out) during the semester of their death loss than a group of matched peers.
In addition to these hypotheses (all focused on the semester of the death loss), differences between bereaved students and their matched peers with regard to GPA, credit hours completed, and academic standing were explored in the semester post death loss.
In order to empirically explore the educational performance and persistence of bereaved college students, we studied 227 students who had experienced a death loss sometime between the summer/fall 2001 and the end of the spring 2004 semester. These students were identified through their interaction with the Office of the Dean of Students (ODOS). On average, 60-65% of bereaved students had self-presented to the ODOS, 15-20% had notified staff members who then contacted the office, and 15-20% had family members who called the ODOS. The primary purpose for the contact with ODOS was to request that a letter be sent notifying instructors of the death loss. The setting of the study was a large, public, land grant university in the midwest. The institution is predominately residential with one of the highest percentages in the U.S. of students living in campus organized housing units (e.g., residential halls, Greek houses).
Lists containing the bereaved student IDs were forwarded to the Office of the Registrar following the completion of each semester. Data files were created containing the GPAs, credit hours completed, and academic standing for students during the semester within which their death loss occurred and for the semester following the death loss. Just under half of the bereaved students had experienced the death of a parent (n = 106; 46.7%), while about one fifth indicated experiencing the death of a grandparent (n = 50; 22.1%). Although data on the date of death were not collected, it is likely that the timing of the death losses was closely connected with when contact was made with the Office of the Dean of Students; most students/families requested letters be sent to instructors prior to the funeral services. The time between interaction with the ODOS and the end of the death loss semester ranged from 0 days to approximately 6 months (summer death losses were grouped with fall death losses) with a mean of 63 days (SD = 38 days). The time between interaction with the ODOS and the end of the semester [End Page 227] did not correlate with GPA (r = .03, p .05) or credit hours completed (r = .02, p .05).
Comparative academic data for the corresponding semesters (i.e., death loss semester and post death loss semester) were generated by the Registrar's Office for 227 control participants who were matched as closely as possible to the bereaved students based on sex, age, race, entering SAT score, semester of study (e.g., first semester sophomore, second semester senior) and school of study. All pairs were exactly matched on sex and race, and only five were not matched on school of study. Of those not matched on age (n = 57 pairs), the average age difference was 2.4 years. Of those not matched on semester in school (n = 75 pairs), the average difference was approximately 2 semesters. It was possible for only 133 pairs to be matched on entering SAT scores, and the average difference in scores for these pairs was 56.6 points. With regard to the overall sample, males and females were about equally represented (52.9% females), the average age was 20.7 years (SD = 3.9), and the majority of students were White (78.4%). Other races/ethnicities represented in the sample were Black (5.3%), Hispanic-American (5.3%), Asian-American (2.6%), and Native American (.4%). In addition, 24 (5.3%) individuals were identified as international students. Common majors were liberal arts (28.2%), management (10%), science (9.8%), and consumer and family sciences (9.1%). There were significant differences between the distribution of students in university majors and those in the present sample who interacted with the ODOS in association with a death loss. For example, in the College of Liberal Arts, with an undergraduate population of 6,185, 59 bereaved students interacted with ODOS, while in the College of Engineering with a comparable population (n = 6,434), only 17 bereaved students were seen by the ODOS, Pearson χ2 (1, N = 12,595) = 23.92, p < .001).
To enhance the homogeneity of the sample through a focus on more traditionally aged undergraduate students, individuals who were either over the age of 25 years and/or were enrolled at the graduate level (n = 36) were excluded from the primary analyses. In addition, the total n for analyses varied as a result of missing data.
Two planned 2 (group: bereaved vs. non-bereaved) x 2 (sex) x 4 (year in school) ANOVAs were calculated for GPA and credit hours completed (during the semester of the death loss). For GPA, there were main effects for group membership, F(1, 384) = 5.19, p < .05, and year in school, F(3, 384) = 5.14, p < .01. More specifically, when collapsed across sex and year in school, the bereaved students had significantly lower GPAs (M = 2.62, SD = .07) than their peers in the matched group (M = 2.84, SD = .07). When collapsed across group and sex, second year students (M = 2.43, SD = .10) had significantly lower GPAs than juniors (M = 2.87, SD = .10) and seniors (M = 2.93, SD = .11). A secondary analysis was performed for GPA using group (bereaved vs. non-bereaved) and race (White vs. non-White) as independent variables. The results indicated main effects for group, F(1, 428) = 6.74, p < .01, and race, F(1, 428) = 5.98, p < .05, but no interaction effect, F(1, 428) = 3.24, p .05. When collapsed across race, bereaved students (M = 2.51, SD = .08) had lower GPAs than the matched group (M = 2.80, SD = .08), and when collapsed across group, non-White students (M = 2.52, SD = .10) had lower GPAs than their White peers (M = 2.80, SD = .05).
For credit hours completed, there was a [End Page 228] main effect for sex, F(1, 385) = 4.49, p < .05, and a group by sex by year of study interaction, F(3, 385) = 3.42, p < .05. Regardless of group membership or year in school, female students (M = 14.7, SD = .17) completed more credit hours than males (M = 14.18, SD = .18). However, these findings must be viewed in light of the three-way interaction. Bereaved females, in their junior year, completed more credit hours than all other juniors, while bereaved males in their junior year took fewer credit hours than junior males and females in the matched group (see Table 1).
A two-way contingency table analysis indicated that bereaved college students were, statistically speaking, no more or less likely to exhibit satisfactory, honors, or problematic academic standing during the semester within which they experienced their death loss than those in the matched group, Pearson χ2 (2, N = 413) = 3.70, p = .16. However, the frequencies were in the anticipated directions, with bereaved students less likely to be recognized with honors and more likely to be identified with a problematic academic standing (e.g., probation, dropped, withdrawn). Frequencies are reported in Table 2. [End Page 229]
Two follow-up ANOVAs were performed again using group, gender, and year in school as independent variables and GPA and credit hours completed during the semester following death loss (see Table 3). For GPA, there were main effects for sex, F(1, 258) = 7.24, p < .01, and year in school, F(3, 258) = 2.85, p < .05. More specifically, females, regardless of group membership or year in school, had higher GPAs than males. In addition, when data were collapsed across group membership and sex, juniors had higher GPAs than first year students. For credit hours completed, there was a main effect for class, F(3, 260) = 3.26, p < .05, with juniors completing more credit hours than first year students and seniors. A two-way contingency table analysis using post death loss semester academic standing, indicated that bereaved college students had academic standings similar to the matched group, Pearson χ2 (2, N = 312) = 1.04, p = .60. It is important to note that the drop in n for the post death loss semester is primarily due to the graduation of seniors, but is also associated with students who dropped out, withdrew, or transferred.
The current findings provide the first empirical support for clinical impressions that bereaved college students are at risk for decreased academic performance. Results indicated that bereaved students earned significantly lower GPAs, during the semester of their death loss, than did a matched group of students. According to Tinto (1975, 1993), grade performance is the single best predictor of academic integration, a factor closely associated with retention/attrition. Although the [End Page 230] frequency of bereaved college students identified with a problematic academic standing (during the death loss semester) was not statistically greater than the frequency in the matched group, the fact that this finding was in the expected direction suggests the need for further study. In addition, further study is needed to explore the potential effects of bereavement, gender, and year in school on the completion of credit hours. Qualitative data gathering may discover the true nature of the current interaction regarding credit hours completed during the death loss semester.
In contrast to the significant difference in GPA for the semester of the death loss, there was a lack of significant difference in GPA between the bereaved and matched groups during the semester following the death loss. This finding highlights the importance of identifying and serving bereaved students during the semester of the death loss. The results suggest that bereaved students experience academic difficulties during the death loss semester, but those who are retained on campus tend to subsequently perform at a similar level to their matched peers.
Significant differences in the representation of students seeking services by academic area of study may indicate either differential sensitivity regarding help-seeking on the part of students or differences in awareness of available support resources on the part of students, faculty, advisors, or other university personnel within specific academic units. Research supports the idea that individuals with similar interests and personality traits are drawn to similar career paths (Boone, van Olffen, & Roijakkers, 2004; Lowman, 1991). Those in the liberal arts may be more likely to seek support such as that offered by the ODOS than are those studying engineering. It could also be that this finding is unique to the present sample in that certain academic units on this specific campus may be more knowledgeable regarding available services. Further research could explore the nature of this disparity in student representation.
Overall, the present research adds credence to arguments that colleges and universities should intervene on behalf of bereaved students (Balk, 2001; Rickman, 1996; Wrenn, 1999). The findings suggest, however, that primary emphasis would best be placed on the semester of the death loss. Because resident assistants and academic advisors have substantial contact with students in the course of their regular responsibilities, they are quite likely to be the first college personnel to learn about the death losses experienced by students. They are, therefore, uniquely poised to be of assistance (Zinner, 1985). However, neither of these groups is likely to have specialized training related to grief and bereavement. In-service trainings focused on the common symptoms of grief, the importance of normalizing individuals' experiences, and information regarding the bereavement-related services offered on campus would be particularly valuable (LaGrand, 1985). While working with bereaved students, resistant assistants and academic advisors can stress the potential negative effect that bereavement can have on academic performance and retention. They must encourage students to immediately access the resources available to them through offices such as the dean of students' office (e.g., letters sent to professors, assistance with acquiring incompletes) and the campus counseling center (e.g., individual and group counseling services).
Balk (2001) and Wrenn (1999) both argued the need for an identifiable place on campus where bereavement support and information can be readily received. Such a [End Page 231] center/office could provide specialized services to bereaved students, while also working to increase the awareness of grief and bereavement issues across campus (Balk, 2001; Wrenn). Although a counseling center may be an option, Balk argued that bereaved students are unlikely to view grief as a mental health issue and may perceive seeking help after a loss as a sign of weakness. However conceptualized, one goal of such a center/office must be improving the identification and referral of bereaved students during the semester of the death loss. It is not enough for such a center to exist. All college personnel must be made aware through workshops/training of the bereavement-related services available on campus and of the importance of encouraging students to seek support immediately. In addition, innovative approaches to reaching and educating bereaved students and their peers with regard to the services available are likely required. For example, outreach to resident halls and Greek organizations, public service announcements on campus radio and TV, and editorials/coverage in the campus newspaper could be effective strategies.
A number of the needs of grieving students identified by Wrenn (1999) pertain to academics. More specifically, students need instructors "to allow late work, a make-up exam or an incomplete for a class" (p. 134). Although instructors may be hesitant to provide such accommodations over a long period of time, the present findings suggest that they are most critical during the semester of the death loss. An action that would be of particular value to bereaved students would be the establishment of academic "bereavement leave" policies. Despite the fact that bereavement leave policies are commonplace for university faculty/staff, the authors know of no such policies for students. Such a policy would provide consistency and protection for bereaved students. Rather than allowing instructors to make idiosyncratic decisions with regard to the offering/denying of student accommodations, faculty would be required to adhere to a university-wide policy that would provide students with, for example, the 3–5 bereavement days usually offered in the workplace. Student affairs professionals can advocate for the creation of such policies.
The primary limitation of the current investigation is the composition of the bereaved student sample. The sample was composed solely of bereaved students who interacted with the ODOS. As indicated earlier, these students either directly or indirectly contacted the ODOS requesting that letters be sent to instructors informing them of the students' death loss. It is difficult to determine the full nature of the bias that this sampling procedure produced. These students were clearly aware of the system and knew to seek at least one bereavement-related service offered by the ODOS. These students may have been more conscientious about their academic performance than bereaved students who did not seek services. It could be argued that these bereaved students, regardless of the bias, are those with whom campus professionals are most likely to interact. However, the present findings may not generalize to bereaved students who do not seek bereavement-related services.
The difficulties associated with recruiting participants for death- and dying-related research are well documented (Balk, 1995; Cook, 2001). In fact, much of the existing research has been done with samples of convenience (i.e., acquired through hospices, funeral homes; Cook). Ethical concerns must be balanced with the potential contribution of knowledge. The present approach to sampling used archival data, and the potential harm to participants was clearly minimized. [End Page 232] Bereaved college students are not an easily recognizable or identifiable population. An alternative approach would be to recruit participants through advertising, which would result in a biased sample as well. Mass surveying could be a method that would reduce bias; however, it is an inefficient and more intrusive approach.
Other limitations include sample size, lack of information regarding death losses of the matched group, and limited information on the specifics of the death losses experienced. A larger sample size would increase power and also allow for more detailed analyses (i.e., breakdown of problematic standing). It is possible that students in the matched group experienced death losses during the semesters of interest, but were not identified as bereaved because they had not interacted with the ODOS. Information regarding cause of death, closeness to the deceased, current social support, etc. would have allowed for the investigation of possible confounding variables.
Despite these limitations, the present study represents a significant contribution to the literature because it is the first empirical attempt to quantify the possible academic impact of bereavement on the lives of college students. The findings indicate that bereaved college students are at risk for academic difficulties that may result in attrition. Bereaved college students warrant practical and empirical attention by university personnel such as administrators, academic advisors, counselors, faculty members, and resident assistants. These individuals have the opportunity to acknowledge the experience of, offer assistance to, and work to more fully understand this potentially vulnerable population.
Heather L. Servaty-Seib is Assistant Professor of Educational Studies
Lou Ann Hamilton is a Counselor in the Office of the Dean of Students