Student researchers

David Brinker, Jr.—Graduate: PhD student
Marisa Greenberg—Graduate: MA student
Lauren E. Herbers—Graduate: PhD student
Eric Meczkowski—Graduate: MA student
This paper was based on a project as part of the “COMM 506” course.


Dr. S. Shyam Sundar


The term “sandwich generation” describes middle-aged adults who have both living elderly parents and dependent children (Pierret, 2006). Pierret describes this generation as being “squeezed between the simultaneous demands of caring for their aging parents and supporting their dependent children” (p. 3). Fingerman, Miller, Birditt, and Zarit (2009) argue family social support can manifest itself in a variety of forms: financial, practical, advice, information, guidance, emotional, and companionship. These types of support are malleable and contingent on constraints, resources, and the family structure (Fingerman, Van der Drift, Dotterer, Birditt, & Zarit, 2011). Research has examined support given in a variety of contexts (e.g., married adult children, Gerstel & Sarkisian, 2007), and from different vantage points (e.g., discrepancies in support giving, Kim, Zarit, Eggebeen, Birditt, & Fingerman, 2011), yet there is no research on the evaluation of the support received by the youngest generation, with knowledge that other family members need support as well. This study examined young adults’ perceptions of support received from their middle-aged parents and their grandparents’ support needs. Furthermore, we were interested in the impact that communicative closeness might have on parental support received and parental support evaluations.


H1: Grandparent support needs should be inversely related to young adult Parental support received.
H2: Young adults will report a positive relationship between parental support received and parental support evaluation.
H3a: Individuals that report higher levels of communicative closeness will report higher parental support received.
H3b: Individuals that report higher levels of communicative closeness will report higher parental support evaluation.
RQ1: For young adults, controlling for communicative closeness, do grandparent needs moderate the relationship between parental support received and parental support evaluation?
RQ2: Does communicative closeness mediate the relationship between parental support received and parental support evaluation?


We initially recruited 376 college-enrolled young adults to fill out an online questionnaire. From this sample, 272 reported having living grandparents that were not considered their guardian. Data was collected using an online questionnaire developed and administered using Qualtrics software. Respondents were asked to identify both their legal guardians and any living grandparents and then asked questions based on our four measures: grandparent support needs, parental support received, parental support evaluation and communicative closeness. The average completion time for the questionnaire was 11 minutes.

To answer the first research question, we conducted a two-way ANOVA to test for a significant interaction between Parental Support Received and a dichotomous variable Grandparent Support Needs Two-Level, where the low and high values represents those whose reported grandparent needs scores fell below and above the median. When accounting for parent gender, separate analyses show the influence of Maternal Support Received on Aggregate Parental Support Evaluation is moderated by Grandparent Needs Two-Level (F = 21.94, p < .0001; b = .108, SE .11, p = .23), while Paternal Support Received is not. A dichotomous (high/low) Communicative Closeness control variable was introduced for the significant maternal support interaction model. The data show a significant 3-way interaction term (F = 13.62, p < .0001; b = .529, SE 0.24, p = .031). Analyzed separately, those reporting low (below-median) Communicative Closeness return a significant interaction between ‘maternal support received’ and ‘Grandparent Support Needs Two-Level’ (F= 9.01, p < .0001; b = -.393, SE 0.17, p = .023) (see figure 1). However, among those reporting “high closeness,” the relationship between Maternal Support Received and Aggregate Parental Support Evaluation is not moderated by grandparent support needs (see figure 2).




The data show the same pattern of interaction in the relationship between ‘Categorical Financial Support Received’ and ‘Grandparent Support Needs Two-Level’ on ‘Financial Support Evaluation,’ where the interaction term is significant only for those reporting “low closeness” (b = -.14, SE .049, p < .01) (see figure 3) but not for those reporting “high closeness” (see figure 4).



The second research question asked, “Does communicative closeness mediate the relationship between parental support received and the recipient’s evaluative satisfaction with that support?”  In order to determine whether Communicative Closeness mediates the relationship between Parental Support Received and Aggregate Support Evaluation,’ Hayes & Preachers’ (2012) macro “MEDIATE” for indirect-effects estimation using bootstrapping was employed. There were two major findings from these analyses. The first is that when Aggregate Parental Support Received is analyzed as the predictor, with Communicative Closeness as the mediator and Aggregate Parental Support Evaluation as the criterion variable, the results indicate a significant indirect effect through Communicative Closeness, with a point indirect estimate of .3183 and 95% BCa (bias-corrected and accelerated) bootstrap confidence interval (CI) of .2118, .4437.  Because the direct effect (.3337, p < .05) remains significant with the introduction of the mediator, ‘Communicative Closeness’ appears to be one mediating variable of the relationship, in addition to other unmeasured mediators and/or the true direct effect. The other major finding from analyses is that Maternal Emotional Support appears insignificant as a predictor variable in the total effects model (b = -.0928, p =.4921); however, the indirect path is significant (b = .2059, p = .0157), as is the mediation with a point estimate of .2059 and a 95% BCa CI of .0550, p < .3382 (see figure 5).  The coefficient sign reversal is indicative of a suppression effect (Tzelgov & Henik, 1991), because the combined direct and indirect effects (i.e., total effect) are net insignificant. The influence of ‘Parental Financial Support on ‘Financial Support Evaluation’ was partially mediated by ‘Communicative Closeness,’ with a point estimate of .0732 and a 95% BCa CI of .0429, .1097 (see figure 6).




All four formal hypotheses were tested using Pearson Product-Moment Correlation (Pearson’s “r”). The results show sufficient evidence for rejection of the null for each of the four hypothesized relationships.


The data show that grandparents’ support needs moderate respondents’ evaluations of the emotional and financial support they receive from their parents. Furthermore, perceptions of communicative closeness partially mediate respondents’ parental support evaluations, suggesting an important communicative component that affects the relationship between received parental support and the perceptions of that support. A proposed explanation for these results is that communicative closeness serves as a proxy for support. The fungibility of communicative closeness allows for an “either/or” relationship with support given; meaning, if a young adult is communicatively close with their parents or parent, then actual support received from the parents matters less when the young adult is evaluating the support. In contrast, when the young adult does not report having communicative closeness with their parents, then actual support received from one’s parent works as a substitute and is the key ingredient for how a young adult evaluate the support given. This investigation emphasizes that young adults continue to need support, but that they are aware of the other constraints that affect the support provided from their parents. Results further confirm contingency theory and its ability to predict support in these multi-generational contexts, more specifically that there is a downward stream of support, as proposed by Fingerman et al. (2010).

For more details regarding the study contact

Dr. S. Shyam Sundar by e-mail at or by telephone at (814) 865-2173

>> Return to abstract listings page
>> Return to the main research page

Media Effects Research Lab at College of Communications, Penn State University