Skip to main content
NCBI home page
NIHPA Author Manuscripts logoLink to NIHPA Author Manuscripts
. Author manuscript; available in PMC: 2020 Jan 1.
Published in final edited form as: Health Psychol. 2018 Nov 5;38(1):84–93. doi: 10.1037/hea0000694

Retrospective memories of parental care and health from mid to late life

William J Chopik 1, Robin S Edelstein 2
PMCID: PMC6309211  NIHMSID: NIHMS993904  PMID: 30394760

Abstract

Objective:

Perceptions of early caregiving experiences are hypothesized to be influential across the lifespan. However, previous research testing this hypothesis focuses primarily on young adults using mostly cross-sectional designs. The current study examined associations between memories of early caregiving experiences and trajectories of depressive symptoms and physical health in two large samples of middle-aged and older adults.

Method:

Sample 1 consisted of participants from the National Survey of Midlife Development in the United States (n = 7,108) followed for 18 years (from 1995/1996 until 2013/2014). Sample 2 consisted of participants from the Health and Retirement Study (n = 15,234) followed over a 6-year period. Memories of caregiver support prior to age 18 were retrospectively assessed in middle and older adulthood; associations between these memories and changes in self-rated health, chronic health conditions, and depressive symptoms over time were examined.

Results:

Memories of higher parental affection in early childhood were associated with better self-rated health and lower depressive symptoms over intervals of approximately six and eighteen years, in both samples; the results for chronic health conditions was more mixed. These associations persisted over time and were not moderated by time.

Conclusion:

Associations between perceptions of caregiver support and health persisted over time, underscoring the importance of memories of close relationships for health and depressive symptoms across the lifespan, even into late life. Findings are discussed in relation to models that link perceptions of early life experiences to later life outcomes.

Keywords: perceptions of early experiences, parental affection, lifespan development, middle-age, older adulthood

Are memories of a rosy past relationship with our parents associated with depression and physical health over time? Memories of positive relationships likely serve many functions—they instill us with a sense of gratitude (Algoe, 2012), make us feel good (Gable, Gonzaga, & Strachman, 2006), and give us hope for our current and future relationships (Abeyta, Routledge, & Juhl, 2015). There are also theoretical frameworks that suggest that memories of our past experiences, especially those with our parents, contribute to working models and representations about close relationships and the world more generally (Collins & Allard, 2004; Fraley, 2007; Pietromonaco & Barrett, 2000). Memories of past experiences, whether they serve to regulate emotions or guide future interpersonal behavior, are thus likely implicated in individuals’ health behavior and how they regulate stress (regardless of the accuracy of those memories; Cohen & Pressman, 2006; Pietromonaco, Uchino, & Dunkel Schetter, 2013; Reuben et al., 2016). Indeed, perceptions of positive early caregiving environments are associated with adaptive development and adjustment in adolescence and early adulthood (Enns, Cox, & Clara, 2002; Luecken, 2000; Newcomb & Bentler, 1988; Oriña et al., 2011; Richman & Flaherty, 1986; Russek & Schwartz, 1997; Shaw, 2006; Shaw & Krause, 2002; Shaw, Krause, Chatters, Connell, & Ingersoll-Dayton, 2004; Wills & Cleary, 1996). However, very few studies have examined associations between perceptions of early relationships and health and well-being beyond adolescence or young adulthood; even fewer include longitudinal follow-up assessments of health and well-being. In this report, we present data from two panel studies to examine whether perceptions of early caregiving experiences are associated with health and well-being over an eighteen- and six-year period, respectively, among middle-aged and older adults.

According to Collins (1996), in the context of interpersonal situations, people are often guided by memories of prior relationships and situations. The activation of memories of relationships then triggers cognitive and emotional responses to interpersonal situations, which in turn guide our behavior in those situations. Interpersonal behavior is thought to affect stress response systems and health behavior, thus affecting downstream health outcomes (Pietromonaco, DeBuse, & Powers, 2013; Pietromonaco, Uchino, et al., 2013). Aside from our behavior in interpersonal situations, the positive affect that results from positive memories of past relationships may also provide some health benefits, primarily through stress reduction or the initiation of positive health behavior (Hill, Allemand, & Roberts, 2013; Sin, Moskowitz, & Whooley, 2015). However, it is unclear whether memories of our past relationships are associated with current health and well-being in middle and older adulthood, which would be a natural extension of process models linking the positive benefits of interpersonal behavior and emotion regulation for health.

The preponderance of evidence from the few studies examining the association between perceptions of early caregiving and health and well-being suggests that perceptions of greater parental affection in early childhood are associated with a variety of positive outcomes in young adulthood, including higher quality work and personal relationships, lower substance abuse, lower depression, and fewer health problems (e.g., Shaw, 2006; Shaw et al., 2004). A related issue is quantifying the influence of perceived affection that is received from mothers versus fathers. Much of the existing work on parental care relies exclusively on maternal warmth and rarely examines fathers’ roles in development (Featherstone, 2004; Lamb, 2000; Sarkadi, Kristiansson, Oberklaid, & Bremberg, 2008). Little is known the few prospective studies that do exist about the relative magnitudes of the contributions of perceptions of maternal versus paternal affection on depression and physical health, particularly in samples beyond adolescence (Sarkadi et al., 2008). Are perceptions of maternal and paternal affection both associated with health and well-being? Or does one type of affection matter less when you control for the other type? Or do perceptions of maternal or paternal affection diminish in influence over time? Considering the mechanisms that link positive parental memories to less depression and better physical health, it seems logical that more positive memories of paternal care would also be related to better health. Unfortunately, our uncertainty about these questions stems from some of the methodological decisions made in past research. For example, some studies only measure perceptions of maternal affection (Davis, Rudy, Su-Russell, & Zhang, 2018; Jackson‐Newsom, Buchanan, & McDonald, 2008); other studies combine perceptions of maternal and paternal affection into one summary score (Shaw et al., 2004). As such, large studies examining the associations between perceptions of maternal and paternal affection on health among middle aged and older adults are needed.

The Current Study

In the current study, we examined associations between early caregiving perceptions and trajectories of depressive symptoms and physical health across middle and older adulthood in two nationally representative samples. Parental affection was assessed using retrospective reports of participants’ relationship with their mothers and fathers. Drawing on life course theories that suggest working models of early experiences set individuals on particular trajectories of health and well-being over time (Collins & Feeney, 2004; Elder & Giele, 2009; Fraley, 2007; Pietromonaco, Uchino, et al., 2013), we predicted that retrospective reports of paternal affection in childhood would be positively associated with better physical health and fewer depressive symptoms both over time and at different points in the adult life span. We did not make explicit hypotheses about the relative magnitude of the associations between perceived maternal versus paternal affection and health, but we expected that they would each demonstrate unique associations with health. We tested these hypotheses in two large samples of middle-aged and older adults followed over intervals of approximately eighteen and six years, respectively.

Method

Participants

Two publicly available data sets were used to test the association between perceptions of caregiver affection and longitudinal changes in depressive symptoms and physical health. The use of existing data sets was determined to be exempt from ethical oversight by the institutional review board at Michigan State University.

Sample 1 consisted of participants from the National Survey of Midlife Development in the United States (MIDUS; Brim, Ryff, & Kessler, 2004). The first wave of the MIDUS study (MIDUS 1; 1995–1996) sampled 7,108 English-speaking adults in the United States (Mage = 46.38, SD = 13.00; 51.1% Female; 90.7% White, 5.2% Black/African American, 4.1% other race/ethnicities, MdnEducation = 1–2 years of college). In the second wave of data collection (MIDUS 2; 2004–2005), approximately 70% of the original sample (n = 4,963) were successfully contacted for follow-up assessments. In the third wave of data collection (MIDUS 3; 2013–2014), approximately 66% of the sample from the second wave (n = 3,294) were successfully retained for follow-up assessments.

Sample 2 consisted of participants from the Health and Retirement Study (HRS), a nationally representative prospective panel study that has surveyed more than 22,000 Americans aged 50 and above every two years since 1992 (Sonnega et al., 2014). In 2006, a random 50 percent of HRS respondents were given a self-report questionnaire that asked questions about parental affection. In 2008, the remaining 50 percent of HRS respondents completed the self-report questionnaire. Thus, two distinct cohorts were formed that had measures of parental affection collected once (at different waves), but depressive symptoms and physical health were assessed every two years (i.e., Cohort 1: Assessed in 2008 (Wave 1), 2010 (Wave 2), 2012 (Wave 3), and 2014 (Wave 4); Cohort 2: Assessed in 2010 (Wave 1), 2012 (Wave 2), and 2014 (Wave 3)). The cohorts were combined into one sample for the present analyses to increase statistical power and precision (as done in Chopik, Kim, & Smith, 2015); in line with findings from past studies, interaction terms with cohort showed that cohort source did not moderate any of the associations reported below. The final combined sample comprised 15,234 older adults (Mage = 67.73, SD = 11.00; 59.2% Female; 73.5% White, 14.2% Black/African American, 9.5% Hispanic, and 2.7% other race/ethnicities, MeanYearsOfEducation = 12.72, SD = 3.06).

Measures

Perceptions of parental affection.

In MIDUS 1 (1994–1995), perceptions of parental affection were assessed with 14 items that asked about respondents’ relationships with their mothers (7 items) and fathers (7 items) before the age of 18. In HRS, perceptions of parental affection were assessed in 2008/2010 with 5 items asking about relationships with mothers (4 items) and fathers (1 item) before the age of 18. The items in each study focused on perceptions of the availability of emotional support from parents as well as the quality of their relationship; there was a substantial degree of overlap between the items used in the two samples.

In MIDUS 1, six items each for mothers and fathers were asked on a 4-point Likert-scale ranging from 1(a lot) to 4(not at all): “How much did she/he understand your problems and worries?,” “How much could you confide in her/him about things that were bothering you?,” “How much love and affection did she/he give you?,” “How much time and attention did she/he give you when you needed it?,” “How much effort did she/he put into watching over you and making sure you had a good upbringing?,” and “How much did she/he teach you about life?”. A seventh item, “How would you rate your relationship with your mother/father during the years you were growing up?” was rated on a 5-point Likert scale ranging from 1(excellent) to 5(poor). MIDUS personnel reverse-scored and multiplied this last item by a .75 factorial to maintain continuity with the other items. These seven items for both mothers and fathers were recoded and averaged such that higher scores reflect more maternal (α = .91) and paternal (α = .93) affection.1

In HRS, perceptions of maternal affection were assessed with variants of the same items used in MIDUS. Three items were rated on a 4-point Likert scale ranging from 1(a lot) to 4(not at all): “How much time and attention did your mother give you when you needed it?,” “How much effort did your mother put into watching over you and making sure you had a good upbringing?,” and “How much did your mother teach you about life?”. A fourth item, “I had a good relationship with my mother before age 18,” was rated on a 5-point Likert scale ranging from 1(strongly disagree) to 5(strongly agree). To create continuity among items, they were recoded, standardized, and averaged (α = .86) such that higher scores reflect more maternal affection. Perceptions of paternal affection was assessed with a single item, “I had a good relationship with my father before age 18,” rated on a 5-point Likert scale ranging from 1(strongly disagree) to 5(strongly agree). Items measuring time/attention, effort, and teaching about life were unavailable for paternal relationships, which is a limitation of the current study.

Overall health.

In both MIDUS and HRS, self-rated health was assessed with the same single item at both waves: “Would you say your health is excellent, very good, good, fair, or poor?” Participants rated their health on a scale ranging from 1(poor) to 5(excellent). Self-rated health is a strong predictor of mortality (Idler & Benyamini, 1997; Schnittker & Bacak, 2014).

Chronic conditions.

In both studies, participants received a list of self-reported chronic health problems and conditions. Participants’ self-reports of chronic health conditions are reliably associated with physician reports of these conditions (Fisher, Faul, Weir, & Wallace, 2005).

In MIDUS, respondents reported whether they had experienced or been treated for any of 27 chronic conditions (e.g., thyroid disease, urinary or bladder problems, gall bladder trouble, high blood pressure, diabetes or high blood sugar, stroke, ulcer, hernia/rupture) in the past 12 months. The number of conditions was summed so that higher values reflected more health problems. Additional chronic conditions were assessed at waves 2 and 3, but we limited our investigation to the 27 common chronic conditions across the study window.

In HRS, an index of seven major chronic conditions was computed for each participant. Each participant self-reported eight physician diagnosed conditions: (1) high blood pressure, (2) diabetes, (3) cancer or a malignant tumor of any kind (excluding minor skin cancer), (4) lung disease, (5) coronary heart disease including heart attacks, angina, and congestive heart failure, (6) arthritis or rheumatism, and (7) stroke. The number of chronic conditions was summed so that higher values reflected more health problems.

Depressive Symptoms.

In MIDUS, depressive symptoms were measured at both waves with the World Health Organization’s World Mental Health Composite International Diagnostic Interview-Short Form (CIDI-SF; Kessler, Andrews, Mroczek, Ustun, & Wittchen, 1998). The CIDI-SF is a checklist of seven depressive symptoms (losing interest in most things, feeling lower energy, changes in appetite, trouble falling asleep, trouble concentrating, feeling down, and thoughts about death). Participants self-reported whether or not they experienced each of seven symptoms in the last 2 weeks. The number of depressive symptoms was summed, with higher levels indicating higher levels of depression. Internal consistency for the CIDI-SF checklist was somewhat low (αMIDUS1 = .52), which is a limitation of the current study. The use of a more reliable depressive symptoms measure in HRS (see below) provides some additional confidence that we are assessing the parental memories—depression association appropriately. Nevertheless, future research should employ more reliable depression measures in population studies.

In HRS, depressive symptoms were measured at all waves with a modified eight-item version of the Center for Epidemiological Studies Depression Scale (CES-D; Radloff, 1977). Participants responded to interview questions about whether or not (i.e., yes or no) they experienced each of eight symptoms in the past week (feeling depressed, felt everything they did was an effort, restless sleep, happiness, lonely, enjoyed life, sad, felt unmotivated). The number of dichotomous depressive symptoms was summed, with higher levels indicating higher levels of depression (αWave1 = .82).

Analytic Approach

To examine longitudinal changes in health, chronic conditions, and depression, and whether these changes were moderated by perceptions of parental affection in childhood, six multilevel models were constructed (one for each outcome—health, chronic conditions, and depression—for both samples). Multilevel modeling allows for the analysis of the entire sample, whereas traditional regression approaches utilize listwise deletion of subjects who do not have complete data on all measures. Thus, if an individual had missing data on paternal affection perceptions (e.g., if they only provided information on their relationship with their mother), the coefficient is only estimated among people with data on this variable. Health, chronic conditions, and depression at each wave were treated as within-subjects variables that varied over time. The linear coefficient of time was modeled on these within-subject observations. Perceptions of maternal and paternal affection at wave 1 were entered as time-invariant predictors of health, chronic conditions, and depressive symptoms. Further, interactions between maternal/paternal affection perceptions and time were included to test whether the association between perceptions of affection and health became stronger, weaker, or stayed the same over time. Participant gender (−1 = male, 1 =female), age (at Wave 1), race (−1 = white, 1 = non-white), and education (at Wave 1) were included as covariates based on previous research examining associations between caregiver affection and health (Shaw, 2006). Because variation in retrospective memories might be partially attributable to variation in health and depression, we controlled for baseline standing in self-rated health, chronic health conditions, and depression in each analysis. When examining variation in a particular dependent variable (e.g., depression), an individual’s standing at baseline on that variable (e.g., depression) need not be controlled for, as baseline standing is already accounted for in the intercept and linear slope estimates (e.g., an individual’s baseline standing on the other two dependent variables (self-rated health, chronic health conditions) were controlled for in the depression analyses, see Tables 2a and 2b). All continuous variables were centered on the sample mean.

Table 2a.

Growth Models Predicting Health, Chronic Conditions, and Depression in MIDUS Participants

95% CI (b)
Health b SE β df t p LB UB
Intercept 3.35 .02 3.37 6680.72 180.63 < .001 3.31 3.38
Time −.10 .01 −.10 11869.25 −11.19 < .001 −.12 −.08
Maternal Affection .05 .02 .04 6107.47 2.99 .003 .02 .09
Paternal Affection .01 .01 .01 6063.55 .96 .34 −.01 .04
Maternal Affection × Time −.001 .02 −.001 11937.55 −.09 .93 −.03 .03
Paternal Affection × Time .003 .01 .002 11983.49 .24 .81 −.02 .03
Gender .06 .01 .06 5828.19 5.57 < .001 .04 .07
Age −.01 .001 −.11 6085.99 −11.00 < .001 −.01 −.01
Education .08 .004 .20 5841.19 20.30 < .001 .07 .09
Race −.12 .02 −.12 6536.99 −6.48 < .001 −.15 −.08
Conditions T1 −.13 .004 −.31 6048.40 −29.31 < .001 −.14 −.12
Depression T1 −.03 .01 −.05 5902.17 −5.14 < .001 −.04 −.02
95% CI (b)
Conditions b SE β df t p LB UB
Intercept 2.64 .05 2.57 6679.16 55.37 < .001 2.54 2.73
Time .17 .02 .17 10461.53 7.18 < .001 .12 .21
Maternal Affection −.15 .04 −.10 6010.79 −3.32 .001 −.23 −.06
Paternal Affection −.07 .04 −.06 5982.36 −1.99 .047 −.15 −.001
Maternal Affection × Time .04 .04 .03 10593.19 1.06 .29 −.04 .12
Paternal Affection × Time .03 .03 .02 10598.56 .83 .41 −.04 .09
Gender .12 .03 .23 5616.02 9.20 < .001 .18 .28
Age .03 .002 .39 5849.55 15.31 < .001 .03 .04
Education −.02 .01 −.06 5601.46 −2.20 .03 −.04 −.003
Race .10 .05 .10 6467.59 2.22 .03 .01 .20
Health T1 −.82 .03 −.79 5802.14 −30.08 < .001 −.88 −.77
Depression T1 .24 .01 .44 5690.07 16.93 < .001 .21 .26
95% CI (b)
Depression b SE β df t p LB UB
Intercept .58 .03 .57 66.08.62 18.50 < .001 .52 .65
Time −.13 .02 −.13 12706.35 −7.34 < .001 −.17 −.10
Maternal Affection −.18 .03 −.13 5838.12 −6.39 < .001 −.24 −.13
Paternal Affection −.07 .02 −.06 5790.70 −2.97 .003 −.12 −.02
Maternal Affection × Time .03 .03 .02 12739.73 .97 .33 −.03 .09
Paternal Affection × Time .04 .03 .03 12764.24 1.66 .10 −.01 .09
Gender .08 .02 .08 5692.74 5.04 < .001 .05 .12
Age −.02 .001 −.23 5976.94 −13.75 < .001 −.02 −.02
Education −.01 .01 −.03 5691.99 −1.63 .10 −.02 .002
Race −.01 .03 −.01 6536.43 −.29 .77 −.07 .05
Health T1 −.11 .02 −.11 5880.99 −5.82 < .001 −.15 −.07
Conditions T1 .14 .01 .33 5936.45 17.88 < .001 .12 .15
Open in a new tab

Note. Gender: −1 = male, 1 = female. Race: −1 = white, 1 = non-white

Table 2b.

Growth Models Predicting Health, Chronic Conditions, and Depression in HRS Participants

95% CI (b)
Health b SE β df t p LB UB
Intercept 3.02 .01 3.04 15293.19 403.91 < .001 3.00 3.03
Time −.03 .002 −.03 41874.07 −15.18 < .001 −.03 −.03
Maternal Affection .04 .01 .03 15375.39 4.27 < .001 .02 .05
Paternal Affection .01 .01 .02 15295.63 2.67 .008 .004 .03
Maternal Affection × Time −.005 .003 −.004 41854.04 −1.85 .06 −.01 .0003
Paternal Affection × Time .002 .002 .002 41867.28 .84 .38 −.002 .01
Gender .04 .01 .04 14531.96 6.11 < .001 .03 .05
Age .0002 .001 .002 14915.24 .33 .74 −.001 .001
Education .05 .002 .16 14583.90 24.84 < .001 .05 .06
Race −.12 .01 −.12 14485.29 −15.43 < .001 −.13 −.10
Conditions T1 −.28 .005 −.37 14775.95 −56.16 < .001 −.29 −.27
Depression T1 −.14 .003 −.27 14691.29 −.42.32 < .001 −.15 −.14
95% CI (b)
Conditions b SE β df t p LB UB
Intercept 2.21 .01 2.16 13785.99 188.54 < .001 2.18 2.23
Time .09 .001 .09 32026.26 73.78 < .001 .09 .10
Maternal Affection −.01 .01 −.01 13831.52 −.48 .63 −.03 .02
Paternal Affection −.01 .01 −.01 13809.04 −.94 .35 −.03 .01
Maternal Affection × Time −.004 .002 −.004 32023.93 −2.44 .02 −.01 −.001
Paternal Affection × Time .001 .001 .001 32025.34 .84 .40 −.001 .003
Gender −.03 .01 −.03 13559.03 −3.36 .001 −.05 −.01
Age .04 .001 .43 13685.96 42.64 < .001 .04 .04
Education .00002 .004 .0001 13580.87 .01 .99 −.01 .01
Race −.06 .01 −.06 13545.05 −5.18 < .001 −.09 −.04
Health T1 −.51 .01 −.55 13618.16 −50.10 < .001 −.54 −.49
Depression T1 .04 .01 .07 13597.01 6.38 < .001 .02 .05
95% CI (b)
Depression b SE β df t p LB UB
Intercept 1.45 .01 1.42 15578.97 101.52 < .001 1.43 1.48
Time .01 .004 .01 41470.32 3.48 < .001 .01 .02
Maternal Affection −.17 .02 −.14 15639.79 −10.42 < .001 −.20 −.13
Paternal Affection −.07 .01 −.08 15541.14 −6.16 < .001 −.09 −.04
Maternal Affection × Time .003 .01 .003 41461.38 .55 .58 −.01 .01
Paternal Affection × Time .0004 .004 .001 41461.67 .11 .91 −.01 .01
Gender .15 .01 .15 14755.57 12.33 < .001 .12 .17
Age −.01 .001 −.08 15256.21 −6.45 < .001 −.01 −.01
Education −.06 .004 −.18 14900.00 −14.11 < .001 −.07 −.05
Race .08 .01 .08 14692.10 5.42 < .001 .05 .11
Health T1 −.58 .01 −.62 14994.68 −45.38 < .001 −.60 −.55
Conditions T1 .09 .01  .11 15027.03 8.30 < .001 .07 .11
Open in a new tab

Note. Gender: −1 = male, 1 = female. Race: −1 = white, 1 = non-white

Results

Preliminary Analyses

Attrition analyses.

A number of individuals from each sample did not complete any follow-up assessments (N = 2,144 MIDUS participants (30.2%); N = 1,166 HRS participants (7.7%)). Below we review how these individuals differed from those who provided complete data.

MIDUS.

Compared to individuals who have only one wave of data, those with at least two waves were more highly educated (d = .35), healthier (d = .28), had fewer chronic conditions (d = .12), and were less depressed (d = .10). Those who had longitudinal data and those who did not were otherwise comparable on other study variables (ds < .04).

HRS.

Compared to individuals who have only one wave of data, those with at least two waves were more highly educated (d = .68), younger (d = .25), healthier (d = .57), received less paternal affection when they were younger (d = .09), had fewer chronic conditions (d = .58), and were less depressed (d = .27). Those who had longitudinal data and those who did not were otherwise comparable on other study variables (ds < .03).

Bivariate correlations.

Descriptive statistics and correlations among primary study variables for the two samples are presented in Tables 1a (MIDUS) and 1b (HRS). In the MIDUS sample, men perceived higher levels of maternal and paternal affection and reported better health (but only at wave 1), fewer chronic conditions, and fewer depressive symptoms compared to women. In the HRS sample, men perceived higher levels of maternal affection and reported fewer depressive symptoms across waves. In both samples, age was associated with perceiving greater maternal and paternal affection, such that older adults reported receiving more affection from their mothers and fathers during childhood. Older adults also reported worse physical health, more chronic conditions, and fewer depressive symptoms. Participants with higher levels of education had better health, fewer chronic conditions, and fewer depressive symptoms. There was also a correlation in which more educated participants perceived more maternal affection in both samples (and more paternal affection in MIDUS), but these associations were very small in all cases. Non-white participants perceived greater paternal affection (in MIDUS only) and reported worse health than white participants. Non-white participants also reported more depressive symptoms in the HRS sample.

Table 1a.

Correlations among primary study variables for MIDUS participants

n Mean (SD) 1 2 3 4 5 6 7 8 9 10 11 12 13 14
1. Gender 7027 −− (−−)
2. Age 7049 46.38 (13.00) .02
3. Education 7095 6.77 (2.49) −.10* −.10*
4. Race 6176 −− (−−) .03 −.10* −.04*
5. Maternal Affection 6255 3.14 (.69) −.13* .04* .03 .02
6. Paternal Affection 5988 2.74 (.80) −.03 .07* .03 .04* .47*
7. Health W1 7097 3.53 (.99) −.03 −.16* .25* −07* .09* .06*
8. Cond W1 6308 2.41 (2.51) .12* .18* −.13* .02 −.14* −.09* −.41*
9. DepW1 7108 .79 (1.93) .09* −.11* −.07* .02 −.15* −.12* −.16* .26*
10. Health W2 4962 3.54 (1.02) −.02 −.18* .26* −.05* .10* .08* .53* −.32* −.13*
11. Cond W2 4041 2.42 (2.53) .11* −.11* −.06* −.01 −.12* −.10* −.11* .15* .30* −.16*
12. Dep W2 4963 .53 (1.64) .12* 20* −.15* .06* −.10* −.06* −.34* .53* .18* −.41* .21*
13. Health W3 3293 3.43 (1.04) −.02 −.08* .23* −.08* .08* .07* .44* −.30* −.13* .55* −.16* −.36*
14. Cond W3 2676 2.64 (2.48) .16* .15* −.14* .03 −.11* −.07* −.33* .53* .18* −.37* .59* .18* −.45*
15. Dep W3 3294 .47 (1.57) .10* −.08* −.06* .002 −.13* −.10* −.14* .20* .25* −.14* .32* .20* −.18* .23*
Open in a new tab

Note. Cond: Conditions. Dep: Depression

*

p < .01 Gender: −1 = male, 1 = female. Race: −1 = white, 1 = non-white. Education: 1 = no school/some grade school, 2 = eighth grade/junior high school, 3 = some high school, 4 = GED, 5 = graduated from high school, 6 = 1 to 2 years of college, no degree yet, 7 = 3 or more years of college, no degree yet, 8 = grad, 2 year college, vocational school, or associates degree, 9 = grad, 4 or 5 year college or bachelors degree, 10 = some graduate school, 11 = masters degree+

Table 1b.

Correlations among primary study variables for HRS participants

n Mean (SD) 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17
1. Gen 15234 −− (−−)
2. Age 15234 67.73 (11.00) −.04*
3. Educ 15195 12.72 (3.06) −.04* −.13*
4. Race 15231 −− (−−) .05* −.19* −.27*
5. Mat. Affection 15176 0.00 (.87) −.07* .10* .05* .01
6. Pat. Affection 14068 3.92 (1.25) −.01 .12* −.01 −.004 .43*
7. Hea W1 15224 3.16 (1.09) −.002 −.09* .29* −.18* .08* .05*
8. Cond W1 15213 1.95 (1.37) −.03* .36* −.15* −.02* −.01 .01 −.44*
9. Dep W1 14793 1.40 (1.94) .09* −.03* −.21* .14* −.14* −.10* −.41* .19*
10. Hea W2 13893 3.17 (1.07) .003 −.11* .28* −.16* .08* .05* .68* −.41* −.37*
11. Cond W2 13870 2.09 (1.39) −.02* .35* −.16* −.02 −.01 .01 −.43* .94* .18* −.44*
12. Dep W2 13315 1.38 (1.95) .10* −.04* −.21* .14* −.14* −.11* −.37* .19* .58* −.42* .20*
13. Hea W3 12696 3.11 (1.06) .01 −.10* .28* −.16* .08* .07* .65* −.40* −.34* .68* −.41* −.37*
14. Cond W3 12684 2.22 (1.40) −.02* .33* −.17* −.001 −.02 −.01 −.42* .90* .18* −.43* .95* .20* −.44*
15. Dep W3 12212 1.40 (1.95) .10* −.02* −.21* .13* −.14* −.10* −.37* .19* .55* −.38* .20* .60* −.43* .22*
16. HeaW4 5157 3.06 (1.05) −.01 −.10* .26* −.14* .07* .08* .59* −.36* −.30* .63* −.37* −.35* .68* −.39* −.38*
17. Cond W4 5150 2.40 (1.38) −.05* .28* −.15* .01 −.01 −.02 −.41* .84* .19* −.42* .90* .21* −.43* .95* .22* −.42*
18. Dep W4 4878 1.38 (1.94) .10* .02 −.20* .12* −.15* −.12* −.37* .19* .52* −.37* .20* .58* −.38* .21* .62* −.40* .23*
Open in a new tab

Note. Gen: Gender. Educ: Education. Hea: Health. Cond: Conditions. Dep: Depression

*

p < .01 Gender: −1 = male, 1 = female. Race: −1 = white, 1 = non−white. Education: Years of education

Perceptions of Caregiver Affection and Changes in Depression and Physical Health over Time

The results from these multilevel models are presented in Tables 2a (for MIDUS) and 2b (for HRS). In both samples, perceiving higher levels of maternal affection in early childhood were each associated with better physical health and fewer depressive symptoms in both samples. Perceiving higher levels of paternal support was also associated with better physical health (in the HRS sample only) and fewer depressive symptoms (in both samples). The results for chronic health conditions were less straightforward—perceptions of maternal and paternal affection were both associated with fewer chronic conditions in the MIDUS sample, but not the HRS sample. The standardized coefficient of perceived maternal affection was usually larger than the standardized coefficient of paternal affection on each outcome. Worth noting, in nearly all cases, the interactions between perceived parental affection and time were not significant. The only exception was a significant perceived maternal affection × time interaction predicting chronic conditions in HRS. Decomposing this interaction revealed that the impact of perceived maternal affection on chronic conditions increased in magnitude over time, albeit this association was small. Specifically, those perceiving lower (−1SD) levels of maternal affection in childhood increased in the number of chronic conditions over time (b = .098, SE = .002, t = 49.83, p < .001) compared to those perceiving higher (+1SD) levels of maternal affection in childhood (b = .090, SE = .002, t = 47.48, p < .001). These results suggest that the positive associations between parental affection on depressive symptoms and physical health persist over time and across the lifespan.2,3

Discussion

Higher perceived parental affection was associated with better self-rated health, fewer chronic conditions, and fewer depressive symptoms. These findings were largely invariant over time, suggesting that the associations between memories of early experiences on our health and well-being persist over time, across middle and older adulthood. Perceived maternal and paternal affection were each associated with fewer depressive symptoms and better physical health, although the associations for paternal affection were smaller than those for maternal affection. The associations for parental affection perceptions were generally stronger for depressive symptoms than for physical health.

The reconstruction of parental memories is likely the product of many different forces—societal expectations, adult children’s own caregiving experiences (with their children), and their current relationships with their parents. For example, that memories of maternal affection had larger associations with health than memories of paternal affection may reflect the broader cultural circumstances into which participants from both samples were born. Because mothers were most likely primary caregivers during their childhood and the existence of cultural expectations that mothers serve primary caregiving roles (Bretherton, 1992), participants’ evaluations of relationship quality in childhood may be more heavily weighted toward perceptions of their mothers. Indeed, when engaging with caregiving experiences of their own children, individuals often reflect and focus on reconstructing memories of their relationships with the own mothers in the past (Baruch & Barnett, 1983). Worth noting, memories of parent-child relationship quality are also more positive when individuals report more positive relationships with their mothers at the time of assessment (Dalton III, Frick-Horbury, & Kitzmann, 2006). Parent-child relationships generally improve in middle age and older adulthood, especially after difficult life transitions (Kaufman & Uhlenberg, 1998). Thus, the positive retrospective memories of maternal caregiving might also reflect the relationship improvements individuals experience with their mothers in mid- and late life. However, memories of affectionate paternal care were also associated with fewer depressive symptoms and better physical health, over-and-above the contribution of maternal memories. With shifting cultural norms about the roles of fathers in caregiving settings, it is possible that the associations of parental caregiving memories with health may be different in more recent birth cohorts, as individuals may reflect and focus more on relationships with their fathers (Bogossian et al., 2017; Cabrera & Tamis-LeMonda, 2014; Lamb, 2000; Sarkadi et al., 2008). Although many of these explanations are speculation, it is possible that participants are actively reflecting on the quality of their relationships with both their mothers and fathers, but for reasons unclear, maternal memories may be more influential when examining associations with health.

Examining exactly how positive caregiving experiences influence health in later life can provide insight into the interpersonal and health decisions that individuals make across the lifespan. For example, concurrent support in adulthood (from family members, spouses, and friends) is related to health and mortality (House, Landis, & Umberson, 1988). However, the developmental antecedents of close relationship quality in adulthood (particularly in older adulthood) are rarely examined longitudinally. Indeed, adults who find it easy to depend on other adults retrospectively report more positive relationships with their parents during childhood (Collins & Read, 1990). The association between perceptions of positive caregiving experiences and current relationship functioning suggests that memories of past relationships may guide interpersonal behavior in adult relationships (Brumbaugh & Fraley, 2007; Pietromonaco & Barrett, 2000). Further, perceptions of different types of caregiver support (e.g., instrumental or emotional) earlier in development may be related to health in different ways, as they are in adult relationships (Wilson et al., 1999). Long-term longitudinal studies that include multiple assessment waves are needed to appropriately test mediating factors between early caregiving experiences and health across the lifespan (Cole & Maxwell, 2003).

Limitations

There are limitations to the current study that are worth acknowledging. First, our samples consisted of middle aged and older adults. The extent to which perceptions of parental caregiving during childhood are associated with depression and physical health in young adulthood is currently unknown. Although longitudinal data are useful in characterizing the association between perceptions of childhood caregiving and health, there may be a number of unmeasured variables that might explain the associations that we found in the current study. We attempted to rule out any health/depression-related explanations for the findings by adjusting for baseline health and depression in each model. Nevertheless, several possibilities for alternative explanations still exist. For example, because memories of parental care at least partially depend on current relationship quality (Dalton III et al., 2006), and relationship quality across a variety of different relationships is associated with health (Chopik, 2017), current relationship quality might explain variation in both retrospective memories and health. Likewise, individual differences in cognitive ability may also affect both retrospective memories and health (e.g., Vuolo, Ferraro, Morton, & Yang, 2014). Future research can more directly rule out alternative explanations of the link between perceptions of childhood caregiving, depressive symptoms, and physical health in middle and late life. To properly address the issues of third variable explanations of the associations presented in this report, multiple assessments of each of these variables over longer intervals of time are necessary.

In both samples, we found associations between perceptions of childhood caregiving and a summary index of chronic conditions. Unfortunately, chronic conditions were rare in both samples, precluding our ability to examine individual chronic health conditions. Future research can investigate similar questions in samples that have more variability with respect to chronic health conditions. Knowing the exact health conditions affected by perceptions of caregiving can potentially lead to interventions that target specific health pathways to reduce chronic illness. Finally, many of the associations in the current study are relatively small in magnitude. With large sample sizes, smaller associations can be captured and more precise effect sizes can be estimated. However, there are likely large individual differences in the associations between perceptions of caregiving and health. Future research can examine moderators of the associations between perceptions of caregiving and health to reveal why perceptions might be more strongly associated with health for some individuals compared to others.

In the current study, memories of early parenting experiences were assessed in middle and older adulthood, generally at least thirty and forty years after they occurred. Do these self-reported memories accurately reflect those early parenting experiences? Much of the research examining this question has found that, although there is occasionally some overlap between perceptions of early caregiving support and actual support experienced in childhood, the association between the two is often small (Booth‐LaForce & Roisman, 2014; Reuben et al., 2016; Roisman & Haydon, 2011; Yarrow, Campbell, & Burton, 1970). This suggests that memories of parental support in childhood rarely resemble the actual support received when early experiences are measured retrospectively through evaluative judgments of recalled experiences (e.g., were they positive or negative?; as in the current studies). Although it appears that retrospective reports have little fidelity with respect to actual experiences, it could be the case that evaluative judgments about memories still exert a considerable influence over interpersonal behavior. For example, the process through which working models affect cognition, emotions, and interpersonal behavior is rooted in the accessibility of past relationship memories and their valence (Pietromonaco & Barrett, 2000). Further, Reuben et al. (2016) compared the predictive ability of retrospectively and objectively reported adverse childhood experiences on various indices of health at age 38. Retrospective memories of childhood experiences indeed predicted life outcomes, controlling for whether these experiences objectively occurred or not. Worth noting, objective occurrence of childhood experiences still predicted life outcomes, but retrospective memories were stronger predictors of self-rated health, cognitive impairment, mental health, and quality of social relationships. Thus, retrospective accounts and perceptions of childhood experiences may be, in some cases, more closely related to interpersonal behavior, health, and well-being in adulthood than the objective events that happen to individuals. However, additional research that combines objectively assessed childhood events and retrospective evaluations of childhood events to examine important life outcomes is needed.

Conclusion

In conclusion, the current study provides evidence that perceptions of early caregiving experiences are associated with an individual’s health and well-being, and these associations are largely invariant over time. Memories of positive caregiving experiences were associated with better health, fewer chronic illnesses, and fewer depressive symptoms longitudinally in both middle and older adulthood. Future research can address the methodological challenges involved in linking early caregiving perceptions to health and well-being across the lifespan, as well as the mechanisms that link these processes.

Acknowledgments

This research was supported by a grant from the National Institute of Aging awarded to the first author (1 R03 AG054705-01A1). The MIDUS study was supported by a grant from the National Institute on Aging (P01-AG020166). The Health and Retirement Study is sponsored by the National Institute on Aging (NIA U01AG09740) and is conducted by the University of Michigan.

Footnotes

1

Embedded within MIDUS are also items assessing parental discipline (e.g., “How strict was she/he with her rules for you.”) and generosity (e.g., “How generous and helpful was she/he to people outside the family?”). Although the focus of the current report is on parental affection, we examined how these questions were associated with depressive symptoms and physical health. After controlling for maternal and paternal warmth and the covariates found in Table 2a, the coefficients for maternal/paternal discipline and generosity were non-significant across all the outcomes (ps > .08).

2

Worth noting, HRS samples individuals at the household level, introducing a degree of non-independence in the data. To account for this, we re-ran the three HRS models by adding an additional level (household) in the multi-level modeling analyses (using STATA’s xtmixed command). These analyses led to nearly identical estimates (and identical levels and direction of significance) to the results presented in Table 2b, suggesting that the non-independence arising from household sampling had little influence in biasing the individual estimates (in this scenario with this data set).

3

In a supplementary analysis, we tested two-way perceived maternal support × perceived paternal support interactions predicting each of the outcomes. These interactions test whether unique combinations of maternal and paternal support perceptions are associated with better health. For five of the outcomes, this two-way interaction was not significant (bs < |.02|, ps > .10). The one exception was a significant maternal support × paternal support interaction predicting chronic conditions in the MIDUS sample. Decomposing this interaction revealed that the participants reporting the fewest chronic conditions were those with more positive memories of maternal and paternal caregiving from their childhood. Among those perceiving lower (−1SD) levels of paternal affection in childhood, the coefficient of perceived maternal affection was negative and significant (b = −.214, SE = .056, t = 3.84, p < .001). Among those perceiving higher (+1SD) levels of paternal affection in childhood, the coefficient of perceived maternal affection was negative, significant, and slightly stronger (b = −.431, SE = .068, t = 6.31, p < .001).

Contributor Information

William J. Chopik, Department of Psychology, Michigan State University, East Lansing.

Robin S. Edelstein, Department of Psychology, University of Michigan, Ann Arbor.

References

  1. Abeyta AA, Routledge C, & Juhl J (2015). Looking back to move forward: Nostalgia as a psychological resource for promoting relationship goals and overcoming relationship challenges. Journal of Personality and Social Psychology, 109(6), 1029–1044. [DOI] [PubMed] [Google Scholar]
  2. Algoe SB (2012). Find, remind, and bind: The functions of gratitude in everyday relationships. Social and Personality Psychology Compass, 6(6), 455–469. [Google Scholar]
  3. Baruch G, & Barnett RC (1983). Adult Daughters’ Relationships with Their Mothers. Journal of Marriage and Family, 45(3), 601–606. [Google Scholar]
  4. Bogossian A, King G, Lach LM, Currie M, Nicholas D, McNeill T, & Saini M (2017). (Unpacking) father involvement in the context of childhood neurodisability research: a scoping review. Disability and rehabilitation, 1–15. [DOI] [PubMed] [Google Scholar]
  5. Booth‐LaForce C, & Roisman GI (2014). I. INTRODUCTION. Monogr Soc Res Child Dev, 79(3), 1–14. [DOI] [PubMed] [Google Scholar]
  6. Bretherton I (1992). The origins of attachment theory: John Bowlby and Mary Ainsworth. Dev Psychol, 28(5), 759. [Google Scholar]
  7. Brim OG, Ryff CD, & Kessler RC (2004). How healthy are we? A national study of well-being at midlife Chicago, IL: University of Chicago Press. [Google Scholar]
  8. Brumbaugh CC, & Fraley RC (2007). Transference of attachment patterns: How important relationships influence feelings toward novel people. Personal Relationships, 14(4), 513–530. [Google Scholar]
  9. Cabrera NJ, & Tamis-LeMonda CS (2014). Handbook of father involvement: Multidisciplinary perspectives: Routledge. [Google Scholar]
  10. Chopik WJ (2017). Associations among relational values, support, health, and well-being across the adult lifespan. Personal Relationships, 24(2), 408–422. [Google Scholar]
  11. Chopik WJ, Kim ES, & Smith J (2015). Changes in optimism are associated with changes in health among older adults. Social Psychological and Personality Science, 6, 814–822. [DOI] [PMC free article] [PubMed] [Google Scholar]
  12. Cohen S, & Pressman SD (2006). Positive affect and health. Current directions in psychological science, 15(3), 122–125. [Google Scholar]
  13. Cole DA, & Maxwell SE (2003). Testing mediational models With longitudinal data: Questions and tips in the use of Structural Equation Modeling. Journal of Abnormal Psychology, 112(4), 558–577. [DOI] [PubMed] [Google Scholar]
  14. Collins NL (1996). Working models of attachment: Implications for explanation, emotion, and behavior. Journal of Personality and Social Psychology, 71(4), 810–832. [DOI] [PubMed] [Google Scholar]
  15. Collins NL, & Allard LM (2004). Cognitive representations of attachment: The content and function of working models. Social Cognition, 75–101.
  16. Collins NL, & Feeney BC (2004). Working models of attachment shape perceptions of social support: Evidence from experimental and observational studies. Journal of Personality and Social Psychology, 87(3), 363–383. [DOI] [PubMed] [Google Scholar]
  17. Collins NL, & Read SJ (1990). Adult attachment, working models, and relationship quality in dating couples. Journal of Personality and Social Psychology, 58(4), 644–663. [DOI] [PubMed] [Google Scholar]
  18. Dalton WT III, Frick-Horbury D, & Kitzmann KM (2006). Young Adults’ Retrospective Reports of Parenting by Mothers and Fathers: Associations With Current Relationship Quality. The Journal of General Psychology, 133(1), 5–18. [DOI] [PubMed] [Google Scholar]
  19. Davis AN, Rudy D, Su-Russell C, & Zhang C (2018). Chinese and European American Undergraduates’ Perceptions of Maternal Warmth and Negativity as Predictors of Self-Esteem and Life Satisfaction. Cross-Cultural Research, 52(2), 192–212. [Google Scholar]
  20. Elder GH Jr., & Giele JZ (2009). The craft of life course research New York, NY, US: Guilford Press. [Google Scholar]
  21. Enns MW, Cox BJ, & Clara I (2002). Parental bonding and adult psychopathology: Results from the US national comorbidity survey. Psychol Med, 32(6), 997–1008. [DOI] [PubMed] [Google Scholar]
  22. Featherstone B (2004). Fathers matter: a research review. Children & Society, 18(4), 312–319. [Google Scholar]
  23. Fisher GG, Faul JD, Weir DR, & Wallace RB (2005). Documentation of chronic disease measures in the Heath and Retirement Study (HRS/AHEAD) Survey Research Center University of Michigan, Ann Arbor, MI. [Google Scholar]
  24. Fraley RC (2007). A connectionist approach to the organization and continuity of working models of attachment. Journal of Personality, 75(6), 1157–1180. [DOI] [PubMed] [Google Scholar]
  25. Gable SL, Gonzaga GC, & Strachman A (2006). Will you be there for me when things go right? Supportive responses to positive event disclosures. Journal of Personality and Social Psychology, 91(5), 904–917. [DOI] [PubMed] [Google Scholar]
  26. Hill PL, Allemand M, & Roberts BW (2013). Examining the pathways between gratitude and self-rated physical health across adulthood. Personality and Individual Differences, 54(1), 92–96. [DOI] [PMC free article] [PubMed] [Google Scholar]
  27. House J, Landis K, & Umberson D (1988). Social relationships and health. Science, 241(4865), 540–545. [DOI] [PubMed] [Google Scholar]
  28. Idler EL, & Benyamini Y (1997). Self-rated health and mortality: A review of twenty-seven community studies. Journal of health and social behavior, 38(1), 21–37. [PubMed] [Google Scholar]
  29. Jackson‐Newsom J, Buchanan CM, & McDonald RM (2008). Parenting and Perceived Maternal Warmth in European American and African American Adolescents. Journal of Marriage and Family, 70(1), 62–75. [Google Scholar]
  30. Kaufman G, & Uhlenberg P (1998). Effects of life course transitions on the quality of relationships between adult children and their parents. Journal of marriage and the family, 924–938.
  31. Kessler RC, Andrews G, Mroczek D, Ustun B, & Wittchen H-U (1998). The World Health Organization Composite International Diagnostic Interview short-form (CIDI-SF). International Journal of Methods in Psychiatric Research, 7(4), 171–185. [Google Scholar]
  32. Lamb ME (2000). The history of research on father involvement: An overview. Marriage & Family Review, 29(2–3), 23–42. [Google Scholar]
  33. Luecken LJ (2000). Attachment and loss experiences during childhood are associated with adult hostility, depression, and social support. Journal of Psychosomatic Research, 49, 85–91. [DOI] [PubMed] [Google Scholar]
  34. Newcomb MD, & Bentler PM (1988). Impact of adolescent drug use and social support on problems of young adults: a longitudinal study. J. Abnorm. Psychol, 97, 64–75. [DOI] [PubMed] [Google Scholar]
  35. Oriña MM, Collins WA, Simpson JA, Salvatore JE, Haydon KC, & Kim JS (2011). Developmental and dyadic perspectives on commitment in adult romantic relationships. Psychological science, 22(7), 908–915. [DOI] [PMC free article] [PubMed] [Google Scholar]
  36. Pietromonaco PR, & Barrett LF (2000). The internal working models concept: What do we really know about the self in relation to others? Review of General Psychology, 4(2), 155–175. [Google Scholar]
  37. Pietromonaco PR, DeBuse CJ, & Powers SI (2013). Does attachment get under the skin? Adult romantic attachment and cortisol responses to stress. Current directions in psychological science, 22(1), 63–68. [DOI] [PMC free article] [PubMed] [Google Scholar]
  38. Pietromonaco PR, Uchino B, & Dunkel Schetter C. (2013). Close relationship processes and health: Implications of attachment theory for health and disease. Health Psychology, 32(5), 499–513. [DOI] [PMC free article] [PubMed] [Google Scholar]
  39. Radloff LS (1977). The CES-D Scale: A self-report depression scale for research in the general population. Applied Psychological Measurement, 1(3), 385–401. [Google Scholar]
  40. Reuben A, Moffitt TE, Caspi A, Belsky DW, Harrington H, Schroeder F, … Danese A (2016). Lest we forget: comparing retrospective and prospective assessments of adverse childhood experiences in the prediction of adult health. J Child Psychol Psychiatry, 57(10), 1103–1112. [DOI] [PMC free article] [PubMed] [Google Scholar]
  41. Richman JA, & Flaherty JA (1986). Childhood relationships, adult coping resources and depression. Social Science and Medicine, 23, 709–716. [DOI] [PubMed] [Google Scholar]
  42. Roisman GI, & Haydon KC (2011). Earned-security in retrospect: Emerging insights from longitudinal, experimental, and taxometric investigations. In Cicchetti D & Roisman GI (Eds.), The origins and organization of adaptation and maladaptation: Minnesota symposia on child psychology (Vol. 36, pp. 109–154). New York NY: Wiley. [Google Scholar]
  43. Russek LG, & Schwartz GE (1997). Perceptions of parental caring predict health status midlife: A 35-year follow-up of the Harvard mastery of stress study. Psychosom Med, 59, 144–149. [DOI] [PubMed] [Google Scholar]
  44. Sarkadi A, Kristiansson R, Oberklaid F, & Bremberg S (2008). Fathers’ involvement and children’s developmental outcomes: a systematic review of longitudinal studies. Acta Pædiatrica, 97(2), 153–158. [DOI] [PubMed] [Google Scholar]
  45. Schnittker J, & Bacak V (2014). The increasing predictive validity of self-rated health. PLoS ONE, 9(1), e84933. [DOI] [PMC free article] [PubMed] [Google Scholar]
  46. Shaw BA (2006). Lack of emotional support from parents early in life and alcohol abuse later in life. International Journal of Aging and Human Development, 63(1), 49–72. [DOI] [PubMed] [Google Scholar]
  47. Shaw BA, & Krause N (2002). Exposure to physical violence during childhood, aging, and health. Journal of Aging and Health, 14, 467–494. [DOI] [PubMed] [Google Scholar]
  48. Shaw BA, Krause N, Chatters LM, Connell CM, & Ingersoll-Dayton B (2004). Emotional support from parents early in life, aging, and health. Psychology and Aging, 19(1), 4–12. [DOI] [PubMed] [Google Scholar]
  49. Sin NL, Moskowitz JT, & Whooley MA (2015). Positive Affect and Health Behaviors across Five Years in Patients with Coronary Heart Disease: The Heart and Soul Study. Psychosom Med, 77(9), 1058–1066. [DOI] [PMC free article] [PubMed] [Google Scholar]
  50. Sonnega A, Faul JD, Ofstedal MB, Langa KM, Phillips JWR, & Weir DR (2014). Cohort profile: The Health and Retirement Study (HRS). International Journal of Epidemiology, 43(2), 576–585. [DOI] [PMC free article] [PubMed] [Google Scholar]
  51. Vuolo M, Ferraro KF, Morton PM, & Yang T-Y (2014). Why Do Older People Change Their Ratings of Childhood Health? Demography, 51(6), 1999–2023. [DOI] [PMC free article] [PubMed] [Google Scholar]
  52. Wills TA, & Cleary SD (1996). How are social support effects mediated? A test with parental support and adolescent substance use. Journal of Personality and Social Psychology, 71, 937–952. [DOI] [PubMed] [Google Scholar]
  53. Wilson DK, Kliewer W, Bayer L, Jones D, Welleford A, Heiney M, & Sica DA (1999). The influence of gender and emotional versus instrumental support on cardiovascular reactivity in African-American adolescents. Annals of Behavioral Medicine, 21(3), 235–243. [DOI] [PubMed] [Google Scholar]
  54. Yarrow MR, Campbell JD, & Burton RV (1970). Recollections of childhood. A study of the retrospective method. Monogr Soc Res Child Dev, 35(5), 1–83. [PubMed] [Google Scholar]

RESOURCES