Grandparenting and Physical Activity

Abstract

Objective:

The aim of this study was to examine a continuum of grandparenting intensity and its association with physical activity using three perspectives: grandparents are active, the more constraints perspective, and the selection bias perspective.

Method:

We use 2014 data from the Health and Retirement Study (HRS), a nationally-representative panel study of the US population over the age of 50 and their spouses (n = 17,851).

Results:

We found that greater grandparenting intensity was inversely associated with physical activity, providing support for both the more constraints perspective and the selection bias perspective.

Discussion:

We discuss the implications of inequality in which the most advantaged with physical activity are those who were either not grandparents, or grandparents who provided less care.

Physical activity in later life is important for the health and well-being of older adults (Lieberman et al., 2021; Rowe & Kahn, 2015). Many older adults benefit from an active lifestyle, and this is associated with lower levels of disease and longer life expectancy (Lieberman et al., 2021). Being a grandparent is one way that older adults may get exercise and engage in activities. In this paper we examine three perspectives to understand the association between grandparenting intensity and physical activity.

Having grandchildren may keep one connected to family and increase lifespan (Carr & Utz, 2020). Engaging in multiple roles, such as grandparenting, may promote overall health (Rozario et al., 2004), including sustaining mental and physical activity in older adults (Hayslip et al., 2021; Sneed & Schulz, 2019). Grandparents who live with their grandchildren or spend time taking care of them are possibly more physically active because of the work involved with raising a child. Many grandparents take care of their grandchildren by babysitting, playing, and going on outings. In addition, they may engage in care work such as feeding, diapering, and bathing young grandchildren. As the children get older, grandparents may engage in activities such as driving and attending school functions. The larger social networks associated with grandparents may also lead to increased physical activities (Quirke et al., 2019).

Alternatively, taking care of grandchildren can take away from discretionary activities that keep one active. With more responsibilities in taking care of grandchildren, less time can be spent pursuing interests and staying active. Additionally, selection into spending a lot of time with grandchildren, including being a primary caretaker is more common among those who are socially and economically disadvantaged. Those who are socially and economically disadvantaged have worse health on average and do not have as much leisure time (Kegler et al., 2022). To our knowledge, no one has asked how grandparenting intensity is associated with physical activity. Much of the research on aging does not use theory (Alley et al., 2010); accordingly, we apply three theoretical perspectives to examine this relationship. This research is exploratory in that it theoretically examines associations between grandparenting intensity and physical activity.

Among those ages 65 and older in the US, about 71% have grandchildren (Westrick-Payne, 2023). The level of involvement of grandparenting varies from not having any grandchildren to having full custody. In 2019, an estimated 7 million grandparents live with their grandchildren in their home, and of these, about 2.3 million are living with their grandchildren under 18-years old (American Community Survey, 2019). By examining the grandparenting and physical activity relationship we can better understand how older adults spend their time. In this paper we measure grandparenting intensity by the household arrangement and number of hours providing care.

Background

A Continuum of Care

Researchers often examine two types of family arrangements when grandparents live with grandchildren: that of coresidence families (in which the three generations of grandparent, grandchild, and the grandchild’s parent live together) or the skipped-generation household (in which the grandparent raises the child without the parent present) (Choi, 2020). With coresidence or custodial care, the grandparent has more frequent interaction with grandchildren and may be more active than nonresidential grandparents. We first pose that grandparents may have an advantage in staying physically active, as compared to those who are not grandparents. Additionally, level of grandparent involvement may also matter. Intensive grandparenting may lead to even greater physical activity. Figure 1 is a heuristic device which shows our conceptualization of the continuum of care in grandparenting activity in which one can range from not having a grandchild to the highest level of intensity (that of being in a skipped generation household in which there is a child and grandparent, but no parent). This conceptualization is modeled after the work of Choi (2020) and is operationalized in the measures section. We build upon Choi’s (2020) work that shows how grandparenting affects mortality. While Choi controls for a dichotomous measure of vigorous activity, our research examines a robust measure of physical activity as the dependent variable (Latham & Williams, 2015).

Figure 1.

Heuristic device for potential continuum of care for grandparenting.

Grandparenting and Physical Activity

There is a classic debate between role strain theory versus role enhancement theory (Chan et al., 2023). In role strain theory, multiple roles lead to strain, thus more grandparenting intensity means greater strain to the individual; whereas in role enhancement more grandparenting activity is associated with greater well-being as the multiple roles lead to a broader sense of purpose and larger networks (Rozario et al., 2004). The evidence shows that both theories are supported depending on multiple factors such as geographical location in the world, gender, and resources (Chan et al., 2023). While inspired by this tension, we chose a more nuanced approach to explaining physical activity. As such, we offer three perspectives for explaining the relationship: grandparents are active, the more constraints perspective, and the selection bias perspective.

Grandparents are Active Perspective

This perspective stresses the importance of individuals remaining active, vital, and engaged (Rowe & Kahn, 2015). In this perspective, grandparenting is active and rewarding, even when grandparenting is concurrent with paid employment. The multiple roles can be beneficial for health, supporting a role accumulation perspective (Meyer, 2014; Rozario et al., 2004). Grandparents may take care of grandchildren by babysitting, giving baths, driving them, attending school functions, shopping for them and so on. These activities increase reasons to engage in more physical activity. In remaining physically active, older adults have higher levels of overall health (Danielsbacka et al., 2022; Rowe & Kahn. 2015). Grandparenting may decrease frailty, so long as one is not a “skipped generation” grandparent (Chen et al., 2015). In the HRS data, Choi (2020) found that light nonresidential grandparenting is associated with better health behaviors (such as more physical activity), that ultimately leads to lower mortality, as compared to those who do not grandparent. Research from Europe and Asia similarly shows grandparenting is beneficial for health (Choi & Zhang, 2018; Di Gessa et al., 2016; Xu et al., 2022; Zhou et al., 2017). Research also shows that grandparenting may be especially beneficial when grandparents live separately from the children (Danielsbacka et al., 2022). According to the grandparents are active perspective, as compared to non-grandparents, we expect the following:

• Hypothesis 1. Greater levels of grandparenting intensity are associated with greater physical activity.

More Constraints Perspective

Alternatively, the more intensive the grandparenting, the less free time individuals have to pursue their own interests. In the more constraints perspective, taking care of grandchildren means less time for pursuing activities that could be beneficial to one’s health. In this view, taking care of children is potentially burdensome. Grandparenting in coresidence or skipped generation in particular may be onerous because of the time constraints of caring for a child versus self-care. A role constraint perspective means that more time with committed activity at home interrupts grandparents’ freedom. With difficult problems in the family such as drug abuse, incarceration, and/or death of an adult child, grandparents can be further burdened (Hayslip et al., 2021). With greater intensity, grandparents may be more home-bound, with their free time consumed by childcare as compared to those with little to no grandparenting duties (Arpino & Bordone, 2017; Danielsbacka et al., 2022). If the grandparents are working themselves, they are even further limited in time (Meyer, 2014). As such, it is to the detriment of their regular health routines and levels of physical activity. Instead of personal time for healthy behavior and self-care (such as exercise), their time becomes a form of “foregone” care in which leisure time is spent caretaking. In the more constraints perspective, as compared to non-grandparents, grandparenting intensity will be associated with less physical activity.

• Hypothesis 2. Greater levels of grandparenting intensity are associated with less physical activity.

Selection Bias Perspective

A selection bias may occur in which those most disadvantaged are the most likely to have the highest levels of grandparenting intensity. This group may have become active in their grandchildrens’ lives out of necessity. Sometimes it is due to a cultural norm of living in an intergenerational household. However, usually when the grandparent(s) are the guardian, it involves situations with the parents such as drug use, mental health problems, divorce, young age of the mother, financial need, incarceration among the parent(s), or premature death (Carr & Utz, 2020). The parents may live with the grandparents as well (co-residence) and those parents may cycle in and out of the household. It is more common among those who are socially and economically disadvantaged (Hayslip et al., 2021; Luo et al., 2012). For instance, most grandchildren in skipped generation households receive some form of public assistance (Drentea, 2019). With lower economic resources, activities may be home-based and involve less physical activity. Lower socioeconomic neighborhoods may be less safe, with fewer community resources, and grandparents may not venture of their house as often (Kegler et al., 2022; Longoria, 2009). They also experience more health problems, in part due to their lower socioeconomic status and additional stressors (Danielsbacka et al., 2022). Disadvantaged grandparents are not equally distributed among all races and ethnicities; coresidence and skipped generation grandparenting is more common among Black and Hispanic families, (Luo et al., 2012), with one study showing that 29% of these children are Black (Drentea, 2019). Alternatively, greater socioeconomic status allows for more resources and flexibility when faced with problems associated with increased responsibility for grandchildren, thus lacking resources is more stressful (Shorey & Ng, 2022). Greater economic resources also can be associated with more fun outings such as going to parks, shopping, and entertainment, all of which are associated with physical activity. As compared to non-grandparents, the selection bias perspective would expect:

• Hypothesis 3. Greater levels of grandparenting intensity are associated with less physical activity and this will be explained by SES, mobility, and health variables.

Gender and Race Effects

We focus on examining overall patterns of physical activity and grandparenting intensity; however, there will be differences by race and gender. Regarding gender, the expectations are not clear. Women are more likely to experience multi-generational and skipped generation households, and childcare work falls mostly on women (Meyer, 2014). Women are much more likely to do the physical work of caring for young children involving feeding, dressing, and bathing young children. They are more likely to drive around grandchildren and take care of their day-to-day needs (Meyer, 2014). In one study from Europe, regular care of grandchildren was associated with fewer social activities for grandmothers, but not for grandfathers (Arpino & Bordone, 2017). For men, greater intensity of grandparenting may be attributed to more time with family members (older men’s primary social networks), which is positively associated with more physical activity (Quirke et al., 2019). However, grandfathers may still have more freedom to do activities (Arpino & Bordone, 2017). For men in retirement, Szinovacz and Davey (2006) found that freedom from grandparent duties was associated with greater well-being; whereas, for women, grandparenting while still working is protective against depressive symptomatology. As such, we will expect differences between men and women, but do not state a direction.

Physical activity and grandparenting patterns may also vary by race and ethnicity. There are cultural norms of multi-generational households found among Latine, Black, and Asian families. For instance, filial piety is the expectation among some Asian adult children that they will care for, and possibly live with their parents in older age (Demir & Drentea, 2016). In Black and Hispanic families, grandparents have traditionally taken care of their grandchildren and have played an active role in their children’s lives (Smith & Whitley, 2014; Sneed & Schulz, 2019). Most Hispanic older adults provide extensive caregiving to their grandchildren as part of their responsibilities (Drentea, 2019). Black and Hispanic grandparents are more likely to experience a multi-generational or skipped generational household and therefore not have as much time to engage in physical activity on their own (Chen et al., 2015; Choi, 2020), but may be much more active in the day-to-day caretaking needed to care for children. White grandparents may have more economic resources to engage in more leisure physical activity. Additionally, they may live in environments more conducive to physical activity (Longoria, 2009). As such, we expect differences among different race/ethnicities that affect both grandparenting activity and physical activity. We expect that the effect of grandparenting intensity on physical activity will vary by race/ethnicity and present these results as post hoc.

Data and Methods

We use the 2014 wave of the Health and Retirement Study (HRS), a nationally-representative panel study that collects information on the economic, health, marital, and family status of the US population over the age of 50 and their spouses who were not institutionalized at their first interview (Heeringa & Connor, 1995). Launched in 1992, the HRS collects data on a biennial schedule. Additional individual-level data on adult demographics, health, and socioeconomic status (SES) comes from the RAND HRS dataset, a cleaned version of the data with consistent naming of variables across waves and imputations for key variables often missing in the raw data (Bugliari et al., 2020). Using a multistage clustered probability frame, the HRS oversamples Blacks (1.86:1), Hispanics (1.72:1), and Floridians (1.72:1). The initial sample size in 2014 was 18,747. The 2014 response rate was 87.1% (HRS Staff, 2017).

Analytic Sample

Eligible respondents included respondents over the age of 50 and their partners or spouses of any age. To be included respondents needed to be categorized into one of five grandparenting categories. Grandparenting categories were created using variables from respondent- and family-level datasets, including the number of reported grandchildren, residency status of grandchildren (resident in the respondent’s household vs. non-resident in the respondent’s household), number of children in the household, and intensity of care provided to grandchildren. A total of 18,393 respondents had complete information on these variables. Less than 1% (355 respondents) could not be assigned a grandparent category and were excluded. An additional 142 were missing on physical activity and excluded. After dropping all other missing data, our final sample had 17,851 respondents. We use OLS regression because the dependent variable, physical activity, is a continuous variable that ranges from 0 to 17.6 and is close to a normal distribution. To test our three hypotheses, we use a progressive adjustment model. We first examine in Model 1 grandparenting intensity (with sociodemographic controls), in the second model we add race/ethnicity, in the third model we add SES, and in the fourth model we add in the health variables. Interactions with gender and race are examined post hoc.

Measures

Dependent Variable.

Following (Latham & Williams, 2015) physical activity was defined using three variables. Respondents were asked about the frequency in which they participate in light, moderate, and vigorous activity in daily life. Respondents were asked “How often do you take part in sports or activities that are vigorous/moderately energetic/mildly energetic?” Among the vigorous activities are running, jogging, swimming, cycling, aerobics, gym workouts, tennis, and digging with a spade or shovel. Activities such as gardening, cleaning the car, walking at a moderate pace, dancing, and floor exercises are moderate physical activities. Household chores such as vacuuming, laundry, and repairs are considered light activities. The variable was coded “never” (coded 0), “1–3 times per month” (coded 1), “one time per week” (coded 2), “more than one time per week” (coded 4) “every day” (coded 5). The index was then weighted each type of activity by intensity (1.2 for light, 1.4 for moderate, 1.8 for vigorous). The total intensity-weighted physical activity score was derived by summing the three weighted scores, with a final range from 0 to 17.6. Higher values reflected greater physical activity.

Grandparenting

We followed HRS researchers (Chen et al., 2015; Choi, 2020) coding of grandparenting. We created 5 types of grandparenting by measuring both the household caregiving arrangement and level of intensity (measured by number of hours). The categories of non-grandparent, light-care non-coresident grandparents (99 or fewer hours), medium care non-coresident grandparents who provide care (100 hr and over), and multigenerational grandparenting, and skipped-generation grandparenting were defined using respondentlevel and household roster information (coding scheme Supplemental Figure 4 available online).

Non-grandparent represents those in the HRS who do not have any grandchildren–neither themselves nor from a spouse or partner. This includes biological, step- or adopted children. All others were grandparents.

Grandparents were further classified based on (1) the number of hours of care provided over the previous 2 years, (2) whether a grandchild was present in the household; and (3) the presence of adult children in the household.

Light-care non-coresident grandparents were grandparents who reported (a) providing between 0 and 99 hr of care for their grandchildren (and great grandchildren) in the last 2 years and (b) living in a separate household from their grandchildren.

Non-coresidency status was determined using roster information, in which HRS respondents were asked to define the relationship of other household members to themselves and whether they were a resident of the household. If a grandparent indicated that no other resident of the household was related to them as a grandchild, the respondent was classified as a non-coresident grandparent. If a grandparent indicated that at least one other household resident was related to them as a grandchild, they were classified as a coresident grandparent.

Medium-care non-coresident grandparents were grandparents who reported providing more than 100 hr of care in the last 2 years for their grandchild and living in a household with no resident grandchildren or great grandchildren.

Multigenerational household grandparents indicated living in a household with at least one resident grandchild and residing in a household with at least one adult child present.

Skipped generation grandparents indicated that at least one other member of the household was related to them as a grandchild and living in a household without the presence of a resident adult child.

Sociodemographics

All controls came from the 2014 wave, consisting of demographic characteristics, socioeconomic status (SES), and health conditions. Age was constructed as a continuous variable, and then was centered. Gender was coded (1 = female, 0 = male). Self-reported marital status was coded using a four-category measure: married/cohabiting = reference, separated/divorced, widowed, and never married. Race and ethnicity groups included non-Hispanic white (reference group), non-Hispanic black, Hispanic, and other minorities (only about 3%). Participants were asked to identify both their race and ethnicity in interviews. If they reported multiple races, participants were asked to choose the one they consider to be their primary race.

SES included three indicators: educational attainment, household income, and net household wealth. Respondents were asked how many years of education they received. Household income reflects the total income of respondent and spouse for the last calendar year. Education was measured with a continuous measure of schooling years, and the average was 12.70 years. Both household income and household assets were transformed using the natural logarithm.

Heath status consisted of subjective and functional health. Subjective health was assessed with self-rated health (poor, fair, good, very good, excellent (1–5). The mobility index indicated the limitations in completing tasks such as walking several blocks, walking one block, walking across the room, climbing several flights of stairs, and climbing one flight of stairs. Mobility limitations ranged from 0–5.

Results

Table 1 reports weighted data for the means and standard deviations of the variables in the analysis. The average level of physical activity is 7.9 on an index of physical activity ranging from 0 to 17.6 and higher values indicating greater physical activity. Twenty-eight percent of the sample was not a grandparent. Among grandparents, 45% provide light non-coresident care, 19% provide medium non-coresident care, 5% provide care in a multi-generational household, and approximately 2% live in a skipped-generation household and provide full care.

Table 1.

Weighted Descriptive Statistics for the RANDHRS Analysis Sample at 2014 (N = 17,851).

Variable	Min	Max	Mean	Std. Dev.
Dependent variable
Physical activity	0	17.6	7.921	4.311
Grandparenting
Non-grandparent(reference)	0	1	.284
Light-care non-coresident	0	1	.454
Medium-care non-coresident	0	1	.193
Multigenerational household	0	1	.047
Skipped generation	0	1	.022
Sociodemographic controls
Women (men = reference)	0	1	.559
Age (centered)	−49.907	36.093	−2.084	10.168
White non-hispanic (reference)	0	1	.784
Black non-Hispanic	0	1	.094
Hispanic	0	1	.084
Other	0	1	.038
Marital status
Married	0	1	0.687
Separate/divorced	0	1	.134
Widowed	0	1	.135
Never married	0	1	.044
SES
Education	0	17	13.307	2.914
Household inc (log)	−2.12	16.232	10.855	1.119
Household assets (log)	0	16.535	10.166	5.198
Health conditions
Mobility limitations	0	5	1.035	1.442
Self-rated health	1	5	3.210	1.057

% for dichotomous variables.

In terms of sociodemographics, women were 56% of the sample and the average age was 68 years. The sample was 78% white, 9% Black, 8% Hispanic, and 4% were in “other.” Most were married (69%), 13% separated or divorced, 13% widowed, and 4% never married. The average level of education was slightly over 12 years (a high school diploma).

Next, we examined the bivariate relationship between physical activity and grandparenting intensity. Figure 2 shows that those who were not grandparents had the highest level of physical activity overall, with a significantly higher mean level of physical activity relative to other groups in each pairwise comparison (p < .001) (see horizontal bar across the top of the figure). ANOVA results show some of the group means differing from each other (p < .05). Using the post-hoc Tukey test, results indicate that physical activity levels are significantly lower in individuals providing light, multigenerational, or skipped generational care compared to non-grandparents, (p < .05) whereas those providing medium non-residential care show no significant difference in activity levels from non-grandparents. Individuals providing medium non-residential care exhibit higher levels of physical activity compared to those engaged in light non-residential care (p < .05). Physical activity levels do not show statistically significant differences when comparing individuals in multi-generational or skipped generational care settings to those providing light care without co-residency. Individuals in multigenerational or skipped generational care settings exhibit significantly lower physical activity levels than those providing medium non-coresidential care. There is no statistically significant difference in physical activity levels between skipped generational care and multigenerational care. The findings suggest that medium non-residential caregivers and non-grandparents have similar physical activity levels, both exhibiting higher activity than those in light, multigenerational, or skipped generational care. Overall, the patterns suggest that those who likely have the most responsibility have the least activity (the skipped and multigenerational care). We also examined variables in the analysis to see how each group on the continuum of grandparenting varied. Patterns show that those in multi-generational and skipped households are more likely to be female, less likely to be married, and much less likely to be white. Those who do not engage in care are younger, white, and have higher levels of education (see Supplemental Table 6).

Figure 2.

Physical activity by type of grandparenting (no controls) (N = 17,851).

Table 2 reports the results of a weighted OLS regression model regressing physical activity level on grandparenting types, sociodemographics, race/ethnicity, socioeconomic status (SES), health, and mobility. The analysis has four models, progressively including sets of variables to observe their associations with physical activity levels. In model 1, those who are not grandparents are the most physically active, supporting the more constraints perspective. As expected, females and those who are older are significantly less active. Being separated/divorced, widowed, or never married was associated with less physical activity than those who were married. Model 3 adds race/ethnicity. Non-Hispanic blacks, Hispanics, and non-Hispanic other are all less active than non-Hispanic whites. After adding these covariates to the model, medium non-coresident is no longer statistically significant, indicating that race and ethnicity account for lowered physical activity among non-residential medium care. Model 4 adds SES and health. Greater education, income, and assets are significantly associated with physical activity. Not surprisingly, a higher count of mobility impairments is significantly associated with less physical activities, and higher self-rated health is associated with greater physical activity. In this final model, non-coresident light care is no longer associated with less physical activity, indicating that SES and health are explaining the negative effect of non-coresident medium care, which lends support to the selection bias perspective. Similarly, residing in a skipped generation household was no longer significantly associated with physical activity, suggesting that it is the poor health and lower SES of the skipped generation grandparents driving lower levels of physical activity. Only grandparents residing in a multi-generation household was significantly associated with less physical activity as compared to those non-grandparents.

Table 2.

Ordinary Least Squares Regression for Weighted Physical Activities by Grandparenting Types, Race, Socioeconomic Status, and Controls (N = 17,851).

	Model 1	Model 2	Model 3	Model 4
Grandparenting (ref. non-grandparents)
Non-coresident light care	−.90*** (.13)	−.80*** (.12)	−.31** (.11)	−.15 (.10)
Non-coresident medium care	−.46* (.18)	−.40* (.18)	−.07 (.17)	.03 (.15)
Multigenerational household	−1.79*** (.16)	−1.51*** (.17)	−.80*** (.19)	−.42** (.15)
Skipped generation household	−1.74*** (.27)	−1.48*** (.26)	−.77** (.27)	−.11 (.23)
Sociodemographic features
Female	−.36*** (.08)	−.36*** (.07)	−.33*** (.08)	−.17* (.07)
Age (centered)	−.08*** (0)	−.09*** (0)	−.08*** (0)	−.05*** (0)
Marital status
Separate/divorced	−.62*** (.13)	−.53*** (.13)	.31* (.14)	.31* (.12)
Widowed	−.81*** (.11)	−.72*** (.11)	.11 (.13)	.11 (.12)
Never married	−1.02*** (.18)	−.99*** (.18)	.04 (.20)	.05 (.18)
Race/Ethnicity (ref. = white)
Non-hispanic black		−1.18*** (.13)	−.53*** (.12)	−.34** (.11)
Hispanic		−.83*** (.14)	.49** (.14)	.39** (.13)
Non-hispanic other		−1.04** (.28)	−.60* (.29)	−.37 (.26)
Socioeconomic status
Years of schooling			.21*** (.02)	.11*** (.02)
Household income (Ln)			.47*** (.05)	.14*** (.04)
Household assets (Ln)			.09*** (.01)	.05*** (.04)
Health and mobility impairments
Count of mobility limitations				−.98*** (.02)
Self-rated health				.70*** (.03)
R-squared	.08	.09	.14	.31

Standard errors are in parentheses.

^***p < .001

^**p < .01

^*p < .05.

Interaction Effects

While not the focus of this paper, we include results from interaction terms in a post hoc test. In the final step we tested interactions between grandparenting intensity and gender as well as race/ethnicity for physical activities (Table 3). The first model shows the final model for regression analysis (i.e., Model 4 in Table 2). The interaction term between grandparenting types and gender was assessed in the second model and the interaction between grandparenting types and race in the third model. Significant interactions indicated there were gender and race differences in the relationship between grandparenting and physical activities. Table 3 only presents the significant associations in Model 2 and Model 3. For gender, Model 2 indicated that the association between medium care non-coresidential grandparenting and physical activity differed by sex, such that the association was weaker among females (b = −0.45, p < .05). However, the question for medium care actually asks whether respondents or their spouse provided 100 or more hours of care to grandchildren, thus these results may be somewhat conflated. For race, results from Model three indicate that the association between grandparenting type and physical activity varied by race/ethnicity, such that the association with light care non-coresidential grandparenting was statistically significant among Black (b = .87, p < .05) and Hispanic respondents (b = .81, p < .05) (white = reference category). A similar pattern emerged for both multigenerational households and skipped generation households, and these grandparenting types were associated with increased physical activity relative to white individuals who were non-grandparents.

Table 3.

Ordinary Least Squares Regression for Regressions for Physical Activities by Grandparenting Types, Race and Gender (N = 17,851).

	Model 1	Model 2	Model 3
Grandparenting types (ref. = non-grandparents)
Non-coresident light care	−.15 (.10)	−.11 (.16)	−.30* (.12)
Non-coresident medium care	.03 (.15)	.28 (.21)	−.05 (.16)
Multigenerational household	−.42** (.15)	−.53* (.25)	−.86*** (.20)
Skipped generation household	−.11 (.23)	.32 (.40)	−.37 (.30)
Female	−.17* (.07)	−.04 (.16)	−.17* (.07)
Grandparenting types × female
Medium care non-coresidential × female		−.45* (−.22)
Race (ref. White)
Non-hispanic black	−.34** (.11)		−.96** (.27)
Hispanic	.39** (.13)		−.22 (.33)
Non-hispanic other	−.37 (.26)		−.86* (.40)
Grandparenting types × race
Non-coresident light care × Black			.87** (.32)
Non-coresident light care × hispanic			.81* (.38)
Multigenerational household × Black			1.36** (.43)
Multigenerational household × hispanic			1.18* (.47)
Skipped generation household × Black			.90* (.41)
Skipped generation household × hispanic			1.60** (.59)

Standard errors in parentheses.

Only significant interactions shown. Controlling for age, marital status, SES, mobility, and health impairments.

^***p < .001

^**p < .01

^*p < .05.

In Figure 3, Black and Hispanic grandparents tend to engage in more physical activities than Whites. Skipped generation households vary by race showing Hispanics in skipped generation households have the highest levels of physical activity, on the other hand, Other races in skipped generation households report the lowest activity levels, indicating a complex interplay of factors influencing physical activity within grandparenting roles. See below for all groups.

Figure 3.

Marginal effects of grandparenting type by race.

Sensitivity Analysis

Many respondents had a physical activity score of 0. Thus, we tested dividing the physical activity variable into two categories: Those who were completely sedentary (0), and those who were not. We experimented with the level of sedentariness; whereby logistic regression models were fit with the same covariates. Overall, conclusions were similar except that grandparents of multigenerational households were not significantly associated with sedentariness (Tables 4 and 5 not shown, available with supplementary analysis).

Discussion

We find that with greater intensity of grandparenting, there is less physical activity. There is no support for hypothesis one stating that greater levels of grandparenting intensity are associated with greater physical activity. We did find partial support for both hypothesis two stating that greater levels of grandparenting intensity are associated with less physical activity. This could be because there are more constraints on time. The findings also support a role strain perspective. Finally, we found partial support for hypothesis three stating that greater levels of grandparenting intensity are associated with less physical activity and this will be explained by SES and mobility and health variables.

The strongest support is for hypothesis three because not only is greater grandparenting intensity associated with less physical activity, it is also in part due to SES, mobility issues, and self-rated health. Only multigenerational household remains significantly associated with lower physical activity after controls. Once we controlled for one’s health, the effect of skipped generation households was explained, and the effect of multi-generation households was diminished. Therefore, it is the health conditions of those grandparents that was associated with the limited amount of physical activity. As such, our research corroborates research that finds no association between grandparenting and physical activity (Ates, 2017). Disadvantaged and/or those grandparents with health conditions are less mobile, may live in unsafe neighborhoods, or may be busy with mundane activities such as driving children and being more homebound in general (Hayslip et al., 2021; Longoria, 2009).

Rather, our story may be one of privilege and freedom to pursue activities. With less grandparenting intensity, older adults have more discretionary time to choose what they want to do. With higher socioeconomic status, older adults also have more autonomy and resources and as such could engage in more leisure activities (Drentea, 2019). These findings speak to Link and Garuía’s (2021) assertion that we must also study those who are privileged, so that we do not engage in diversions in which we only study those who have difficult life circumstances and low social power. In this case, those with greater physical activity are not grandparents may be privileged in terms of health.

Unlike some of the research coming from Europe and Asia, we did not find that greater intensity is associated with greater physical activity and ultimately better health (Choi & Zhang, 2018; Di Gessa et al., 2016; Xu et al., 2022; Zhou et al., 2017). Rather, we found evidence to show that high levels of grandparenting intensity is associated with disadvantage. Further, our research shows that the effects of grandparenting intensity are complicated and situational depending on family household type, family roles, and cultural differences (Chan et al., 2023). It is perhaps not surprising that with our U.S. results, those who are less obligated to family ties in terms of caring for grandchildren have higher levels of physical activity. Our individualistic society does not elevate and support the role of grandparenting in ways that more collectivistic cultures may (Hossain et al., 2020). As such, the social structure and norms of intergenerational families may facilitate healthy activities at individual, familial, and societal levels.

In the U.S. context, in our study, greater grandparenting intensity is associated with fewer physical activities one may want to do. Grandparents who spend more hours caring for grandchildren concurrently have less time in their own individual pursuits. It can take away time spent doing other activities that are conducive to health and well-being (Arpino & Bordone, 2017). These findings suggest a profile of those who are privileged with more free time for leisure and physical pursuits, actively creating their own lives of hobbies and exercise. They also support Choi’s (2020) work that light grandparenting is associated with better health behaviors. Less family responsibility may lead to more time for oneself. Future research could investigate how fewer to no grandchildren to care for leads to more freedom and better health. With the growing population of those who will be single in later life with few if any family ties, it may be interesting to examine under which conditions these single older adults thrive, in comparison to those who have obligations to families and grandchildren (Klinenberg, 2013; Marsh, 2023).

Regarding the interaction effects, our findings show that in non coresidential intensive care, grandmothers engage is less physical activity compared to men who are not grandparents. Grandfathers in general may engage in more activities such as sports and recreation (Bates & Goodsell, 2013). Whereas grandmothers on the other hand are often working themselves, while handling much of the day-to-day difficult work of childcare, and thus have less time for fun activities (Arpino & Bordone, 2017; Meyer, 2014; Moore & Rosenthal, 2015). For race, the effect of intensity on physical activity differs by race and ethnicity. Blacks and Hispanics are more active than Whites in general. Caring for grandchildren may bring about a sense of accomplishment for older adults and increase connections with family members (Rowe & Kahn, 2015; Rozario et al., 2004), especially for those who are facing social role compression due to no longer working. The race and ethnicity effects show complex relationships with intensity and health outcomes, as has been shown in other articles (cf. Chan et al., 2023; Chen et al., 2015; Choi, 2020).

As with any study, there are several limitations to this paper. First, this is a cross-sectional study and it is establishing correlations, and not causation between physical activity and intensity of grandparenting. Though our theoretical focus placed grandparenting intensity as a determinant of physical activity, physical activity may be associated with intensity of grandparenting. Our contribution was exploratory and was the first step in establishing how the three theoretical perspectives represent the relationship. While our research aim was appropriately assessed using cross-section data to show concurrent associations, future research should explore how these three theoretical perspectives endure over time.

Another limitation is that Figure 1 shows that progressive levels of grandparenting may not lead to a perfect lockstep increase in intensity. For instance, in the light-care non-coresidential group, some are likely providing 0 hr of care, while others may be providing up to 99 hr-we do not have the exact number of hours. Also, while it is likely that many grandparents in multigenerational households provide more grandchild care than non-resident grandparents, there may also be households where the parents are caring for both the grandparent and the grandchild. Thus while useful for visualizing grandparenting, it is used as a heuristic device. We were guided by Choi’s (2020) operationalization as an established framework for understanding the levels of grandparenting intensity. Another consideration is that because our sample starts at 50-years old, it does not include the grandparents who are very young-some presumably as young as in their 30s. One would expect much higher levels of disadvantage in these families. Finally, the household income measure does not include income from other household members, thus we were not able to capture income per capita.

Our findings show that grandparents are active when they have less grandparenting intensity, and that those without grandchildren engage in higher levels of physical activity. The HRS measures light, moderate, and physical activity using three questions. Each question contains specific examples of the types of physical activity, which serve to guide participants as to what type of activity constitutes light, moderate, and physical activity. The selection of aid-to-recall items included in the question is intended to shape how respondents interpret and respond to the question. Many of the examples of activities provided for vigorous and moderate activity are leisure-like activities. Thus, we are tapping into leisure physical activities as well. A review of physical activity and parenthood found parents to engage in less physical activity, with commonly reported barriers being lack of time or social support, fatigue, and other role obligations (Bellows-Riecken & Rhodes, 2008). However, as noted by these authors, all but one of the studies included in the review were based on self-reported measures. An accelerometerbased assessment of physical activity by mothers of children found that having at least one child aged 6 months to 5 years old in the same household was associated with fewer sedentary minutes and extended periods sedentary behavior (Do et al., 2022). It is thus important to interpret our findings in light of the focus on leisure-like activities in the HRS assessment of physical activity, which may miss more nuanced changes in physical activity among grandparents caring for children. Also, when grandparents are in skipped generation households, they may be handling other special needs of the children who have possibly experienced trauma. Research in which grandparents explain exactly what they do physically would elucidate the findings (cf. Hayslip et al., 2021).

In sum, our research underscores the disadvantaged experience of intensely caring for grandchildren. We show that awareness of how those with more grandparenting intensity could be supported by focusing on their physical activity, and this may provide insights into targeted interventions to improve quality of life. Service providers who work with older adults may ask about clients’ grandparenting responsibilities and suggest activities that can be done by those who experience greater grandparenting intensity. For instance, small changes such as parking farther in a parking lot or walking around during a grandchild’s activity could help grandparents build in walking. Simple suggestions for creating ways to play with younger grandchildren could increase activity and leisure could be suggested. At a community-based level, having safe neighborhoods, newer sidewalks that are flat, and neighborhood activities could all help grandparents get outside and increase physical activity with their grandchildren. Encouraging groups can be supportive to grandparents who may feel more comfortable knowing that there are many adults watching the children. Public places that are enclosed such as playgrounds and libraries are helpful because it makes it easier to contain younger children in a safe area.

Additionally, examining those who are privileged in later life, either as grandparents or those who are not, may elucidate understanding health disparities. By understanding what it is about their lives that lead to relative advantage in activities, we can better explicate under which conditions grandparenting is beneficial for physical activity (Link & García, 2021). Additionally, providing support to those who are providing a lot of care for their grandchildren, such as in the multi- and skipped-generation households would be valuable.

Supplementary Material

Supplemental material for this article is available online.