Ole to Ole:
A Study on Social Support Reciprocity at St. Olaf College
December 18, 2009
Erin Thompson, Daniel Murphy, Sarah Bosch, Scott Grandt, and fifth author
Abstract.In recent studies of social support among college students the act of providing social support has been overlooked. This study examines the amount of social support college students provide in relation to the amount and sources of social support they receive from peers, family, significant others or established networks. Using survey data from an anonymous random sample survey of students at a small liberal arts college in the Midwest, we tested the hypothesis that the amount of social support a college student receives positively correlates to the amount of social support they provide. Additionally, we explored whether the social support received from peers positively relates to the amount provided to peers, if the social support received from family negatively relates to the amount provided to peers, if the social support received from significant others negatively relates to the amount provided to peers and if the social support received from established networks negatively relates to the amount provided to peers. We found that the amount of social support peers receive from all sources positively relates to the amount of social support they provide to peers.
Introduction
We are interested in examining the amount of social support college students provide in relation to the support they receive, specifically self-disclosure of issues of deep concern. Part of transitioning to adulthood is taking on more responsibilities, including the task of providing social support for families and friends. In congruence with recent studies we examined social support as an interpersonal transaction rather than a one-way relationship.
Literature Review
The term social support has commonly been defined as the giving of emotional comfort, appraisal, and affirmation through a friend or family member (Jou and Fukada 2002). Past research has investigated the relationship between social support and many other phenomena including: chemical substance use, body image perception and outlets of social support such as religion, the internet, group membership, and family dynamics.
As stated above, many recent studies have begun to approach social support as an interpersonal transaction. For example, Gant, Calsyn, and Winter (1999) attempted to find variables that predicted whether two individuals would have a mutual understanding of whether their relationship would be of high or low social support. They found that the openness of both partners, the level of intimacy of the relationship, and the empathy within the relationship had a stronger effect on a couple’s mutual understanding than partners’ personalities.
Another important focus of reciprocal social support relationships among college students is how a student's sense of coherence, one’s ability to see the world as comprehensible, manageable, and meaningful, is affected by different relationships (Darling, Howard, Olmstead and McWey 2007). In a study conducted on a southern university campus, researchers found that females had a greater sense of coherence when friend, love and parent relationships were well maintained. Results for males indicated that maintaining family relationships had the greatest effect on their sense of coherence. Another study found that when students who give more social support felt more supported than those who provide less social support (Crocker and Canevello 2008).
Research centered on the negative aspect of social support reciprocity among college students has focused on how these support relationships are affected by a variety of stressors and on their resulting impacts on students' mental and physical health. In a case study conducted in Japan, researchers found that students who gave less support than what the recipient needed in turn felt they received less support than they needed, leading them to be less physically and mentally healthy (Jou and Fukada 2002). Similarly, another study found individuals who provide more support than they receive tend to feel burdened. However, those who receive more support than they provide tend to feel indebted (Rosen, Mickler and Collins 1987). Thus, a balance between the amount of social support provided and the amount of support received is key to personal well-being.
Methods
This study took a quantitative approach to examine the perceived amount of social support college students provide in relation to the amount and source of social support they receive from peer, family, significant others or established networks. The concept of social support was limited to self-disclosure, which we defined as “an instance in which very personal information is shared with another.”
We drew five hypotheses:
- The amount of social support a St. Olaf student receives positively correlates to the amount of social support they provide.
- The amount of social support a St. Olaf student receives from peers positively relates to the amount of support they provide to peers.
- The amount of social support a St. Olaf student receives from family negatively relates to the amount of support they provide to peers.
- The amount of social support a college student receives from a significant negatively relates to the amount of social support they provide to peers.
- The amount of social support a college student receives from established networks negatively relates to the amount of support they provide to peers.
On October 14th, 2009 we conducted a one-hour focus group with three males and three females to receive feedback about social support on campus. We asked questions regarding how focus group members would define certain concepts, how they would react to certain wordings, and how they would answer some of our survey questions. Using this data, we formulated questions for our survey. The main variables our questions addressed were the amount and sources of social support received, how much support subjects perceive they provide to peers and subjects’ levels of comfort with giving and receiving social support. The independent variables were the amount of social support received and the sources of received social support. The dependent variable was the perceived amount of social support provided.
For the variables concerning sources of social support, we predetermined six categories: peers, family members, significant others, established networks, and other. For the purpose of the study we defined peers as “students of a similar age, on campus, who are not family members and not involved romantically.” Off-campus peers were included in the other category. The category of family members was limited to the immediate family. In regard to significant others, we asked subjects to consider only on-campus relationships, though off-campus relationships could be described in the other category. Established networks where defined as organizations or people whose job it is to listen and offer counseling, were categorized as either on or off-campus. Examples of on-campus networks included counselors, Residence Life Staff and the Pastor’s office. Examples of off-campus networks included off campus pastors or psychiatrists.
Most variables were measured in an interval ratio or rank ordered format asking subjects to choose the number of times in the past week the subject had confided an issue of deep concern in each of our categories (peers, family, on-campus networks, off-campus networks, significant others, and other) and how many times members of each of the categories (peers, family, significant others, and other) had come to the subject to confide something in the past week.
We used multiple questions that covered a variety of dimensions (number of occurrences, number of individuals, and perceived relationships) involved in confiding issues of deep concern. Ideally, this increases reliability, consistency (other researchers can reproduce the study using the same methods), and the internal validity (the design of the research is subject to limited or no errors) of the survey (Neuman 2007). Other variables that were not key to our findings, but were taken into account to reduce spuriousness include: class year, gender, and reciprocity.
We applied for approval from St. Olaf College’s Institutional Review Board to check the ethical nature of the study since it includes human subjects. These procedures protect the rights and well-being of subjects, particularly respect for persons; researchers must recognize that subjects are autonomous and regard them as such and work to protect those whose autonomy might be reduced, beneficence; making sure harm to subject is appropriate and minimized and benefits to subjects is maximized, and justice; making sure that vulnerable populations are not exploited (St. Olaf College Institutional Review Board 2004). Once our study was reviewed and approved by the Institutional Review Board, we were able to conduct the study.
The ethics of working with human subjects were taken into account in creating and conducting this study. Neuman cites students as a “special population” or someone whose participation may be motivated by the possibility of attaining something desired (2007). Participation in the survey was completely voluntary and students were free to stop taking the survey at any time. With careful wording and question choice, we were as conscious of the participants’ comfort level as possible to avoid mental or emotional discomfort. In an effort to increase the student response rate, upon completing the survey, students had the opportunity to enter into a drawing for a gift certificate to the school’s bookstore.
Data were collected via a self-administrated online survey in November 2009 at St. Olaf College. In addition to demographics, we asked where subjects give and receive support, and the level of comfort while receiving and giving. Our study utilized simple random sampling to draw a random sample so that results could be generalized to the St. Olaf population. The target population was the 2813 eligible St. Olaf students; all of those who were currently on campus, did not participate in our focus group, and were not in either research methods class section. Our sample size totaled 703 St. Olaf College students. Of the students invited, 333 participated (47.4%). To ensure that all applicable populations were included, Susan Canon, Director of Institutional Research at St. Olaf College, selected a simple random sample from the target population.
Results
The majority of our data were in ordinal form, due to the structure of our response categories. The general questions concerning how many instances in the past week subjects confided in someone or had someone confide in them had possible answers of 0, 1-3, 4-6, 7-9, and 10+, each answer becoming an ordinal category. For the giving and receiving of social support by peers and family categories, the median and mode were both 1-3, for the category other the median was 1-3 and the mode was 0. For network categories both on- and off-campus and significant other categories, the medians and modes were 0.
We coded the data assigning numerical values to all ordinal categories (1=0, 2=1-3, 3=4-6, etc.) In order to calculate total social support, we created an index by adding all of the coded values together from each of the categories. This made both of the total categories (Total Social Support received and Total Social Support given) scaled data. The means for these categories were 9.72 and 7.03 respectively, with standard deviations at 2.642 and 2.150. (See Table A, Figure A, and Figure B in the Appendix for more details.)
We tested our first hypothesis (the amount of social support received positively correlates to the amount of social support provided) by running a bivariate correlation test. We calculated a Spearman rho correlation coefficient for the relationship between Total Social Support received and Total Social Support given. A strong positive correlation was found (rho (168) = .766, p < .01), indicating a significant relationship between the two variables, supporting our hypothesis (see Table B in Appendix). This level of correlation can be seen in the dot-plot representation of the data in Figure 1.
Figure 1
Table 1. Correlations – Total Social Support without SO |
||||
|
Total Social Support Received without SO |
Total Social Support Given without SO |
||
Spearman's rho |
Total Social Support Received without SO |
Correlation Coefficient |
1.000 |
.728** |
Sig. (1-tailed) |
. |
.000 |
||
N |
264 |
253 |
||
Total Social Support Given without SO |
Correlation Coefficient |
.728** |
1.000 |
|
Sig. (1-tailed) |
.000 |
. |
||
N |
253 |
272 |
||
**. Correlation is significant at the 0.01 level (1-tailed). |
||||
Due to the inclusion of the significant other category, in these first two indexes the number of applicable cases (N) was only 170 (see Appendix). Any case in the category of significant other which indicated the answer not applicable was excluded. In order to increase internal reliability, we ran a second correlation between similar indexes after removing the category significant other. This increased N to 253. The correlation coefficient is still high at .728, and the hypothesis is still supported by the significance level (see Table 1).
Hypotheses two through five involved ordinal data, so we tested each using a cross tabulation test. This test shows the distribution of answers given between two different categories. To test for significance within each cross tabulation (and to test each of our hypotheses), we used Kendall’s tau-b and Kendall’s tau-c because both input categories in each cross tabulation consisted of ordinal data.
Table 2. Cross Tabulation - Social Support from Peers * Social Support to Peers |
|||||||
% within social support from peers |
|||||||
|
Social Support to Peers |
||||||
0 |
1-3 |
4-6 |
7-9 |
10+ |
Total |
||
Social Support from Peers |
0 |
62.8% |
34.9% |
2.3% |
|
|
100.0% |
1-3 |
10.2% |
71.9% |
13.2% |
4.2% |
.6% |
100.0% |
|
4-6 |
|
33.3% |
57.6% |
6.1% |
3.0% |
100.0% |
|
7-9 |
|
|
40.0% |
46.7% |
13.3% |
100.0% |
|
10+ |
|
|
15.8% |
15.8% |
68.4% |
100.0% |
|
Total |
14.2% |
50.6% |
22.6% |
6.8% |
5.8% |
100.0% |
|
We tested our second hypothesis (the amount of social support received from peers positively relates to the amount provided to peers) by running a cross tabulation between the categories social support from peers and social support to peers.
As one can see in Table 2, the most frequently answered category in the to peers section tends to correspond with the similar category in the from peers section. The significance values from both Kendall’s tau-b and Kendall’s tau-c (see Table 3) indicate that our second hypothesis is supported.
Table 3. Symmetric Measures for Peers to Peers Cross Tabulation |
|||||
|
|
Value |
Asymp. Std. Errora |
Approx. Tb |
Approx. Sig. |
Ordinal by Ordinal |
Kendall's tau-b |
.659 |
.032 |
15.308 |
.000 |
Kendall's tau-c |
.537 |
.035 |
15.308 |
.000 |
|
N of Valid Cases |
310 |
|
|
|
|
a. Not assuming the null hypothesis. |
|||||
b. Using the asymptotic standard error assuming the null hypothesis. |
|||||
Our third hypothesis (the amount of social support received from family negatively relates to the amount provided to peers) and fourth hypothesis (the amount of social support received from significant others is negatively related to the amount provided to peers) had similarly positive results. Once again, we ran a cross tabulation using Kendall’s tau-b and Kendall’s tau-c. We rejected both hypotheses as results indicated that there is a positive relationship between each variable (see Table 4 and Table 5).
Table 4. Symmetric Measures for Family to Peers Cross Tabulation |
|||||
|
|
Value |
Asymp. Std. Errora |
Approx. Tb |
Approx. Sig. |
Ordinal by Ordinal |
Kendall's tau-b |
.281 |
.044 |
6.087 |
.000 |
Kendall's tau-c |
.228 |
.037 |
6.087 |
.000 |
|
N of Valid Cases |
309 |
|
|
|
|
a. Not assuming the null hypothesis. |
|||||
b. Using the asymptotic standard error assuming the null hypothesis. |
|||||
Table 5. Symmetric Measures for Significant Other to Peers Cross Tabulation |
|||||
|
|
Value |
Asymp. Std. Errora |
Approx. Tb |
Approx. Sig. |
Ordinal by Ordinal |
Kendall's tau-b |
.200 |
.064 |
3.029 |
.002 |
Kendall's tau-c |
.155 |
.051 |
3.029 |
.002 |
|
N of Valid Cases |
201 |
|
|
|
|
a. Not assuming the null hypothesis. |
|||||
b. Using the asymptotic standard error assuming the null hypothesis. |
|||||
Our fifth hypothesis (the amount of social support received from established networks negatively relates to the amount provided to peers) was divided into two parts: on-campus networks and off-campus networks. Thus, we ran two different cross tabulation tests. Neither test showed any relationship between received from established networks and given peer support (see Tables 6 and 7).
Table 6. Symmetric Measures for On-Campus to Peers Cross Tabulation |
|||||
|
|
Value |
Asymp. Std. Errora |
Approx. Tb |
Approx. Sig. |
Ordinal by Ordinal |
Kendall's tau-b |
.014 |
.057 |
.242 |
.809 |
Kendall's tau-c |
.006 |
.025 |
.242 |
.809 |
|
N of Valid Cases |
297 |
|
|
|
|
a. Not assuming the null hypothesis. |
|||||
b. Using the asymptotic standard error assuming the null hypothesis. |
|||||
Table 7. Symmetric Measures for Off-Campus to Peers Cross Tabulation |
|||||
|
|
Value |
Asymp. Std. Errora |
Approx. Tb |
Approx. Sig. |
Ordinal by Ordinal |
Kendall's tau-b |
.077 |
.063 |
1.171 |
.242 |
Kendall's tau-c |
.025 |
.022 |
1.171 |
.242 |
|
N of Valid Cases |
294 |
|
|
|
|
a. Not assuming the null hypothesis. |
|||||
b. Using the asymptotic standard error assuming the null hypothesis. |
|||||
In order to increase the reliability of our results, we tested for two confounders: gender and class year. No relation was shown between either of the categories and our results.
Beyond our hypotheses, we also tested the relationship between social support received from family and social support given to family, and the relationship between social support received from peers and social support given to family. For both of these situations, a positive relationship was observed. (See Tables C and D in Appendix.)
Discussion
Results indicate that the more social support one receives the more social support one provides (see Table B in Appendix). We attribute this to the internalization of social norms surrounding reciprocal relationships, the need for balance in social support relationships, and the normalization of a highly connected lifestyle.
Recently, researchers have begun to treat social support as an interpersonal transaction dictated by cultural norms and rules (Fant, Calsyn and Winter 1999). Within reciprocal relationships, Nadler-Fisher (Rosen, Mickler and Collins 1987) asserts the self-esteem of both parties can be threatened if the exchange of social support is unbalanced. Therefore, the amount of social support received from peers positively relates with the amount of social support provided to peers because transactions most often occur within reciprocal relationships.
The positive association between social support received from family and social support provided to peers may be due to a need for balance in overallsocial support given and received. The positive association cannot be credited to reciprocity: the data showed a noticeable gap between the amount of social support received from family and the amount of social support individuals provide to family (see Table C in Appendix). Previous research has indicated that many individuals experience feelings of guilt when they receive a much higher amount of social support than they provide (Rosen, Mickler and Collins 1987). It is possible that those who receive higher amounts of support from family provide more support to peers to reconcile feelings of debt in family relationships.
The positive association between social support received from significant others and social support provided to peers, along with the fact that all statistically significant relationships were positive, may be explained partly by the normalization of a highly connected lifestyle. A highly connected lifestyle is one that utilizes a strong network of socially supportive relationships in order to achieve a balance of giving and receiving social support. A study by Crocker and Canevello (2008) suggests that students who provide more social support perceived a greater level of available social support. People accustomed to engaging in social support relationships will be familiar with the cultural norms surrounding interpersonal relationships. A highly connected individual may enter into a reciprocal relationship more readily since there is less risk involved in entering into a new relationship because a larger base of relationships is already established. Furthermore, it is easier to achieve a balance of total social support received and social support provided within a larger network. This could play a part in the positive association between social support received from significant others and the social support provided to peers (see Table 5); these could also be partial explanations why all statistically significant associations were positive.
Our fifth hypothesis, the amount of support from established networks negatively relates to the amount of social support provided to peers, was not statistically significant because too few people reported utilizing established networks of social support (pastor, residence counselor or professional psychologist). This may be because they are not utilized or subjects feared social stigma from seeking help from established networks (see Tables 6 and 7). Though previous research had indicated that gender and class year were confounding variables, our results show that class year and gender were not statistically significant confounding variables.
After testing our hypotheses, we used the data to further examine other relationships and found that some were statistically significant. We found that the amount of social support provided to family positively relates to both the amount of social support received from family and the amount of social support received from peers (see Tables C and D in Appendix). Our literature did not address these connections directly, but we surmise the trend may be a part of maintaining a balance of total social support received and total social support provided within the context of an individual’s total social network.
Conclusion
In this study we examined how the amount of social support college students receive influenced the relationship of the support they provide. Our results indicated that the more social support students receive from peers, family, or significant other, the more support they provide to peers. Studies show that without a balance between support given and received that there was a feeling of guilt or burden to supply the same amount of support. Subjects attempted to balance between the amount of social support given and received both in socially shaped reciprocal relationships and within their greater social networks.
Our study is only applicable to St. Olaf College because generalizations can only be made to the population from which our data was collected. We had a moderate response rate. Some limitations of our study include the issue of self report; it is difficult for us to know for certain that our respondents are telling the truth or are remembering past events accurately. As we have such a predominately Caucasian campus we had a largely homogenous sample. To avoid such limitations in the future, similar research could be done on another campus that is more racially/gender balanced. This could lead to a more accurate generalization about the sample population.
A lot of emphasis regarding social support on our campus is about how and where to receive it, whether it be at the counseling center, academic support center, or from a professor. Our research suggests that the college should encourage friends to be sources of support for one another. Peer to peer social support is important for feelings of coherence and well-being. Ideas for future research include: conducting the same study on a range of campuses in more urban locations, to see if the differences in demographics and characteristics of our population change our results, looking at both receiving and giving social support as indicators of an individual’s social well-being, or studying specific reciprocal relationships more thoroughly. Instead of asking respondents how many people they receive support from and how many they people they give support to, specific couples (platonic or romantic) could be studied.
Works Cited
Crocker, Jennifer, Amy Canevello. 2008. “Creating and Undermining Social Support in Communal Relationships: The Role of Compassionate and Self-Image Goals.” Journal of Personality & Social Psychology 19:3.
Darling, Carol A, Stacy N Howard, Spencer B Olmstead, and Lenore M McWey. 2007.
“College student stress: the influence of interpersonal relationships on sense of coherence.” Stress and Health 23: 215-229.
Jou, Yuh Huey and Hiromi Fukada. 2002. “Stress, Health, and Reciprocity and Sufficiency of Social Support: The Case of University Students in Japan.” The Journal of Social Psychology 142(3): 353-370.
Gant, Judy R., Robert J. Calsyn, and Joel Winter. 1999. “Predicting Agreement Between Providers and Recipients of Social Support.” Journal of Social Behavior & Personality 14: 4.
Newman, Lawrence W. 2007. Basics of Social Research: Qualitative and Quantitative Approaches. Pearson Education, Inc.
Rosen, S., E. Mickler, J.E. Collins. 1987. “Reactions of Would-be Helpers Whose Offer of Help is Spurned.” Journal of Personality and Social Psychology 53:288-297.
St. Olaf College. “Residence Life.” St. Olaf College. Retrieved October 26, 2009 (http://www.stolaf.edu/admissions/campus/reslife.html).
St. Olaf Institutional Review Board. 2004. “The Ethical Foundations of the Federal Regulations Governing Human Subjects Research in the United States.” St. Olaf Institutional Review Board. Retrieved October 26, 2009 (http://www.stolaf.edu /academics/irb/ AboutProtection.htm).
Appendix
Figure A
Figure B
Table A
Descriptive Statistics |
|||||
|
N |
Mean |
Std. Deviation |
Median |
Mode |
Social Support from peers |
311 |
n/a |
n/a |
1-3 times |
1-3 times |
Social Support from family |
310 |
n/a |
n/a |
1-3 times |
1-3 times |
Social Support from on-campus networks |
298 |
n/a |
n/a |
0 times |
0 times |
Social Support from off-campus networks |
295 |
n/a |
n/a |
0 times |
0 times |
Social Support from significant other |
202 |
n/a |
n/a |
0 times |
0 times |
Social Support from other |
272 |
n/a |
n/a |
1-3 times |
0 times |
Social Support to peers |
319 |
n/a |
n/a |
1-3 times |
1-3 times |
Social Support to family |
315 |
n/a |
n/a |
1-3 times |
1-3 times |
Social Support to significant others |
202 |
n/a |
n/a |
0 times |
0 times |
Social Support to other |
272 |
n/a |
n/a |
1-3 times |
0 times |
Total Social Support received |
184 |
9.72 |
2.642 |
9 |
9 |
Total Social Support given |
184 |
7.03 |
2.150 |
7 |
7 |
Total Social Support received without SO |
264 |
8.27 |
2.259 |
8 |
8 |
Total Social Support given without SO |
272 |
5.69 |
1.875 |
5 |
5 |
Table B
Correlations – Total Social Support |
||||
|
Total Social Support Received |
Total Social Support Given |
||
Spearman's rho |
Total Social Support Received |
Correlation Coefficient |
1.000 |
.766** |
Sig. (1-tailed) |
. |
.000 |
||
N |
184 |
170 |
||
Total Social Support Given |
Correlation Coefficient |
.766** |
1.000 |
|
Sig. (1-tailed) |
.000 |
. |
||
N |
170 |
184 |
||
**. Correlation is significant at the 0.01 level (1-tailed). |
||||
Table C
Symmetric Measures for Family to Family Cross Tabulation |
|||||
|
|
Value |
Asymp. Std. Errora |
Approx. Tb |
Approx. Sig. |
Ordinal by Ordinal |
Kendall's tau-b |
.548 |
.038 |
12.499 |
.000 |
Kendall's tau-c |
.417 |
.033 |
12.499 |
.000 |
|
N of Valid Cases |
308 |
|
|
|
|
a. Not assuming the null hypothesis. |
|
|
|
||
b. Using the asymptotic standard error assuming the null hypothesis. |
|
||||
Table D
Symmetric Measures for Peers to Family Cross Tabulation |
|||||
|
|
Value |
Asymp. Std. Errora |
Approx. Tb |
Approx. Sig. |
Ordinal by Ordinal |
Kendall's tau-b |
.297 |
.046 |
6.104 |
.000 |
Kendall's tau-c |
.226 |
.037 |
6.104 |
.000 |
|
N of Valid Cases |
306 |
|
|
|
|
a. Not assuming the null hypothesis. |
|
|
|
||
b. Using the asymptotic standard error assuming the null hypothesis. |
|
||||
