Psychological Health Benefits of Companion Animals Following a Social Loss

Dawn C. Carr, PhD; Miles G. Taylor, PhD; Nancy R. Gee, PhD; Natalie Sachs-Ericsson, PhD


Gerontologist. 2020;60(3):428-438. 

In This Article


Risks related to major social losses increase as we age, and we tend to rely more heavily on close friends and family, especially our spouses (Antonucci et al., 2014). The loss of primary sources of social and emotional support can have significant negative consequences to well-being. The greatest concerns relate to growing evidence surrounding social isolation and loneliness. However, individuals often report emotional benefits from pets, which could have heightened relevance during a time of social loss. The current longitudinal study examined the effect of having a CA when middle-aged and older adults are faced with major social loss. Specifically, we examined whether CAs helped to maintain psychological health (i.e., avoid increases in depressive symptoms/loneliness) after experiencing a social loss (i.e., loss of spouse through divorce or death). Previous research has produced inconsistent findings, likely related to methodological challenges in HAI research (Friedmann & Gee, 2018; Gee et al., 2017). Primary among these is selection bias. Using longitudinal data from the HRS, the present study addressed selection bias by using a propensity score method allowing us to compare multiple pet/loss groups.

We examined four pet/loss groups: no-pet/no-loss; no-pet/loss; pet/loss; pet/no-loss. We expected that regardless of CA ownership, a major social loss would be associated with negative psychological consequences, but that CA owners would experience relatively fewer negative consequences. Our hypotheses were both supported. For depressive symptoms, individuals who experienced a loss had significantly greater increases in overall symptoms than those without a loss, but having a pet buffered these effects (p < .05). For loneliness, the no-pet/loss group experienced greater negative effects than the no-loss control groups, and the difference between the two loss groups (pet/loss vs. no-pet/loss) was statistically significant at the p < .01 level. It is important to note a key limitation that should be considered in interpreting these findings. The timing of the data collection for loneliness (see earlier the description about data structure for this outcome measure) had a wider range of error in relation to the loss compared with that shown by depression. Future research should examine effects of CAs on immediate and long-term loneliness in the wake of negative life events.

Overall, our findings suggest CAs are beneficial with respect to the attenuation of depressive symptoms and loneliness symptoms among middle-aged and older adults experiencing a major social loss. Our results, based on a longitudinal national sample and adjusting for important selection factors, are important in guiding future research on HAI, particularly in light of the growing body of research on social support and loneliness in later life. Our findings indicate that although pets may not be able to fully prevent individuals from experiencing psychological health decline after a social loss, they do seem to minimize the negative consequences. There are very few interventions that have been designed to help off-set the negative consequences of social losses for overall well-being, particularly loneliness, and our study can inform ways that individuals may be able to maintain well-being outside of interventions involving human interactions alone.

Several limitations of our study should be considered when interpreting our study findings. First, although this is the first longitudinal study of its kind to test the potential effects of CAs for middle-aged and older adults when faced with a major social loss, we rely on a relatively simplistic measure of CA ownership that does not allow us to understand which factors associated with CA ownership offer the greatest benefits, such as the number or type of pets. Second, our sample is based on an experimental survey in the HRS, resulting in a small sample from which to identify individuals experiencing social losses. As a result, it is not possible to conduct more refined analyses. In the future, using a larger sample, we hope to explore whether these effects are differentially beneficial for certain social groups (e.g., men vs. women). In addition, our sample size limitations prevented us from examining whether the benefits of CA ownership might hold for a range of different social losses. We were able to conduct our analysis on only those who experienced widowhood (i.e., excluding those who become divorced). The results for widowhood alone are very similar to those presented. A notable difference, however, is that the effect sizes and level of significance are more robust with respect to loneliness compared with the results presented in the present article. Our sample sizes were quite small for this analysis, and given the complex, two-stage analytic technique used here, we chose not to show the detailed results (available on request). Descriptive analysis indicated that the widowhood group is significantly older (p < .001) and more likely to be female (p < .01) than those who experienced divorce, which may make CA ownership more beneficial for buffering loneliness for widowhood. Although the propensity score models adjust for age, gender, and other potential confounding observable factors, in future analyses with larger study samples, it would be important to examine moderators of the benefits of CA ownership at the time of a major social loss in later life like gender and age for different loss types. Future studies should also examine the effect of child and parental losses, which was not directly examined in this study. Finally, for some of the individuals included in our study, the outcome measure was collected in 2014 for which we do not have data about pet ownership. Thus, it is plausible that some individuals identified as CA owners in the 2012 module may no longer have a CA by the time the social loss was reported in 2014.

Despite these limitations, this study is novel, offering a sophisticated approach to assessing the effect of CAs on the well-being of middle-aged and older adults. For gerontologists, our findings suggest a need for evaluation research on how pets might serve in therapeutic contexts. Moreover, the specific mechanism by which CA convey such benefits is an important area of additional focus. Our findings also have consequences for social policies. For instance, it may be beneficial to include CA in the treatment for individuals residing in residential facilities for middle-aged and older adults, or reducing barriers to pet ownership in such settings.