More or less help? A longitudinal investigation of positive and negative consequences of divorce for informal helping

ABSTRACT

As divorce rates have risen, scholars have expressed concern that a breakdown of traditional family bonds might negatively influence community life. This study examines the impact of divorce on one form of community involvement, namely informal community helping, and whether this impact depends on household income, having adult children and being full-time employed. We hypothesized that informal community helping can both increase and decrease after divorce and that the impact of divorce is smaller for people with higher household income, adult children or a full-time job. Utilizing longitudinal data from the first four waves (1986–2002) of the Americans’ Changing Lives panel study (N = 6,185), this study employed fixed-effects regression models. These demonstrated that people did not change their informal community helping after divorce. The impact of divorce did not depend on household income or full-time employment, but people with adult children increased their informal community helping after divorce less than average.

KEYWORDS:

• Informal helping
• divorce
• panel fixed effects regression
• parent-child relationships
• household income
• employment status

Introduction

For more than a century, scholars have expressed concern about a presumed decline of community life (e.g. Bellah et al., 1985; Putnam, 2000). Community life here is defined as prosocial behavior among groups of individuals connected through a shared interest, identity or space. A trend often linked to the decline of communities is the rising number of divorces, as people who are divorced have been said to be less involved in community life than married people (Hofmeister & Edgell, 2015; Lee & Brudney, 2012; Putnam, 2000; Rossi, 2001; Rotolo & Wilson, 2004).

A relevant part of community involvement is informal helping; that is, helping people outside one’s own home without the involvement of external organizations or institutions (Einolf et al., 2016). Examples of informal help are providing transport, doing housework and caring for others’ children or grandchildren. Informal help is often provided to neighbors, friends or relatives living in a different household (Gundelach et al., 2010; Li & Ferraro, 2005; Wilson & Musick, 1997). Moreover, informal help is mainly an outgrowth of existing social relations between community members and has a low threshold for participation. This results in a widespread prevalence of informal help, regardless of affluence or other resources (Williams, 2005). A major argument for why divorced people might engage less in such help than married people is that divorcees have a smaller social network than married people (Lee & Brudney, 2012; Musick & Wilson, 2008; Putnam, 2000). Because informal helping relies on one’s direct social network, it may be more impacted by divorce than other forms of community involvement, such as formal volunteering and charitable giving. Furthermore, people generally reciprocate informal help and expect others to do so as well (Amato, 1990; Manatschal & Freitag, 2014). Not providing informal help may thus result in receiving less support, which may be especially harmful for divorced people who require support to cope with this transition. This makes informal helping a particularly interesting form of community involvement to study to understand the impact of divorce.

Despite the overall expectation that divorce reduces informal helping, studies have reported mixed findings on the topic (Lee & Brudney, 2012; Mutchler et al., 2003; Perks & Haan, 2011; Plagnol & Huppert, 2010). Some have indicated that non-married people provided less informal help than married people (Cramm & Nieboer, 2015; Helms & McKenzie, 2014; Li & Ferraro, 2005; Lim & Laurence, 2015), while others reported the opposite (Plagnol & Huppert, 2010), or found no differences in informal helping by marital status (Erlinghagen, 2010; Lee & Brudney, 2012; Van Tienen et al., 2011; Wang et al., 2017). However, these studies only examined differences in informal helping by marital status; they did not explicitly examine the impact of divorce on informal helping. Moreover, most of these prior studies lack a clear focus on the theoretical link between marital status and informal helping (for an exception see Lee & Brudney, 2012). This makes it difficult to put the mixed findings into perspective.

To better understand the relationship between divorce and informal helping, this study proposes two competing hypotheses. The first is based on the aforementioned premise that divorce reduces the size of the social network and on marriage premium literature. This latter strand of literature presupposes a decline in well-being after divorce (Booth & Amato, 1991; Dupre & Meadows, 2007; Johnson & Wu, 2002; Williams & Umberson, 2004), and subsequently a decline in informal helping. The second and competing hypothesis proposes an increase in informal helping after divorce. Based on social integration literature, it is expected that although social network size may decline, an increase in close social relations occurs after divorce, which subsequently results in a rise in informal helping.

Moreover, since these mechanisms could cancel each other out, the present study examines three groups for whom either positive or negative consequences might be more prominent: people with a high household income, in full-time employment and having adult children. Mapping the impact of these moderators can shed light on the theoretical mechanisms underlying the association between divorce and informal helping. Furthermore, understanding why people from certain social groups increase or reduce their informal helping after divorce may help us to understand in what types of communities rising divorce rates are more or less problematic for community life. Hence, this study aims to answer the following research question: To what extent is divorce related to changes in informal helping, and to what extent is this relationship moderated by household income, being full-time employed and having adult children?

To answer this question, we used longitudinal information from the Americans’ Changing Lives study (House, 2018), collected from 1986 to 2002 and thus covering 15 years of adults’ life course. The panel design of the study provided an advantage over prior research, which only compared married individuals to non-married individuals (Lee & Brudney, 2012; Mutchler et al., 2003; Plagnol & Huppert, 2010; Van Tienen et al., 2011) and results likely were affected by selection bias. For example, sociable people may be more likely to remain married but also provide more informal help. As a result, previously reported associations of not being married may not be indicative of positive or negative consequences of divorce, but instead point to a third factor that explains both divorce and informal helping. By using panel data and estimating fixed effects models, we controlled for such stable, unobserved (third) factors.

Theoretical framework

Negative consequences of divorce

Various studies suggested that never married and previously married people are less involved in community life than married persons (Lee & Brudney, 2012; Putnam, 2000; Rossi, 2001; Rotolo & Wilson, 2004). This implies that divorce reduces community involvement, including informal helping. Two reasons have been suggested for this decline. First, divorce reduces the size of a person’s social network (Eckhard, 2021; Terhell et al., 2004; Wrzus et al., 2013). Divorcees often lose contact with their former spouse’s family and friends (de Bel, 2020; Terhell et al., 2004). Moreover, divorce often induces moving to another house, possibly in a different neighborhood or city (Mikolai & Kulu, 2018). This diminishes one’s proximate social network, at least in the short term. Since larger social networks are expected to increase the likelihood of knowing someone who needs help, they bring more opportunities for helping (Lee & Brudney, 2012). Altogether, since divorce generally reduces the size of a person’s social network, it is likely that informal helping is reduced as well.

A second reason why community involvement, such as informal helping, may decline after divorce relates to a person’s physical and mental well-being (Musick & Wilson, 2008). Divorce is stressful and may have severe negative impacts on a person’s well-being (Booth & Amato, 1991; Dupre & Meadows, 2007; Johnson & Wu, 2002; Williams & Umberson, 2004). A major explanation is the loss of economic resources (Gadalla, 2008; Tach & Eads, 2015). Divorcees may have difficulty affording healthcare or maintaining their standard of living with more financial and emotional stress as a result (Waite & Gallagher, 2000). Additionally, spouses often exert social control over unhealthy behaviors, such as smoking or drinking alcohol (Ross et al., 1990; Umberson, 1992). Divorce removes this source of social control, leaving more room for individuals to engage in unhealthy behaviors (Das, 2013). Finally, spouses rely on each other for emotional and social support and companionship (Ross et al., 1990). The lack of this source of emotional and social support after divorce may also lead to a decline of individuals’ well-being (Williams & Dunne-Bryant, 2006).

In contrast, good physical and mental well-being have been linked to providing more informal help (e.g. Erlinghagen, 2010; Hank & Stuck, 2008; Li & Ferraro, 2005; Plagnol & Huppert, 2010). Although theory on this mechanism is sparse, it seems likely that poor well-being acts a barrier preventing individuals from engaging in informal helping. First, individuals with poor physical well-being may be unable to engage in physically straining helping activities, such as gardening or housework. Second, poor well-being may tap into an individual’s energy, leaving them with little capacity to notice that others need help. Third, individuals with low well-being likely will not be asked to help, because others do not want to burden them even more. Prior research indeed found that people with worse self-rated health (Erlinghagen, 2010; Plagnol & Huppert, 2010), health impairments (Hank & Stuck, 2008; Li & Ferraro, 2005) and mental health problems (Choi et al., 2007) provided less informal help.

Combining these two arguments, we expect individuals to reduce their informal helping after divorce (H1).

Household income as a moderator

The above-theorized reduction in well-being after divorce, and subsequent reduction in informal helping, may be less marked for people with a high household income. Although most people, including those with a high income, will likely experience some decline in income after a divorce (Gadalla, 2008; Tach & Eads, 2015), this reduction may have a less significant impact on the well-being of people with higher incomes. This is because people with higher incomes may still be able to afford decent health care and maintain a good standard of living after divorce. Moreover, their financial position after divorce might not be bad enough to cause too much stress. As a result, divorce is likely to be less detrimental to the well-being of people with a higher household income compared to those with a lower household income. Since decreased well-being has been linked to reduced informal helping (e.g. Erlinghagen, 2010; Hank & Stuck, 2008; Li & Ferraro, 2005; Plagnol & Huppert, 2010), we argue that the explanation of a differential impact of divorce for those with higher and lower incomes can be extended to informal helping. In other words, we expect that the higher a person’s household income, the less they reduce their informal helping after divorce (H2).

Positive consequences of divorce

Divorce also may have positive consequences for informal helping, according to research on social contacts conducted among both married and non-married individuals (e.g. Sarkisian & Gerstel, 2008, 2016). While the size of one’s social network may decline after divorce, this strand of literature argues that at the same time the need for close social relations increases (Kalmijn, 2012; Sarkisian & Gerstel, 2008, 2016). In this regard, two arguments come to the fore. First, marriage is often considered a ‘greedy’ institution (Kim & Dew, 2016; Sarkisian & Gerstel, 2016). That is, being married takes up time and energy that a person cannot devote to other social relations (Coser, 1974). When a person divorces, additional time and energy may become available that can be invested in alternative social relations. Second, married individuals have been found to rely mostly on their spouse for social contact and support (McPherson et al., 2006). After a divorce, the need for social contact and belonging likely has to be fulfilled through social relations with others (Baumeister & Leary, 1995). Therefore, people are expected to invest more in alternative social relations after a divorce.

Increased investment in alternative social relations also brings more opportunities for informal helping, as most people provide informal help only to those they know personally. More importantly, without the involvement of formal organizations, a person can only help when they are aware of others in need. People with more close social relations more easily pick up signals that others need help and may be more readily asked to provide help. Hence, those with more close social relations are likely to engage more in informal helping (Lee & Brudney, 2012; Wang et al., 2017).

As a result, we expect that individuals increase their informal helping after divorce (H3).

Adult children as moderator

The increased need for social relations after a divorce may be greater for those who during marriage were highly dependent on their spouse for social contact and support. According to Baumeister and Leary (1995), social need fulfillment has diminishing returns; for each new social relation, the additional positive impact is smaller. Conversely, losing a social relationship is more damaging for people with only a few close social contacts, compared to people with a lot of social relations. Hence, divorce is expected to be less damaging for people who during their marriage already have invested in close social relations next to their spouse. Consequently, people with feel less need to invest in new social relations after divorce. So, their opportunities for informal helping would increase less, compared to those who had fewer close social relations during their marriage (Lee & Brudney, 2012; Wang et al., 2017). Altogether, this implies that a smaller increase in informal helping after divorce is expected for those who were less dependent on their spouse for social contact during marriage.

One group of individuals that may be less dependent on their (previous) spouse are parents of adult children who live in a different household. Parents often rely on adult children for social contact and conversations (McPherson et al., 2006). While prior research has shown that the quality of the parent–child bond sometimes deteriorated after a parental divorce, these studies also demonstrated that the majority of parents kept frequent contact with at least one of their children, both before and after divorce (Daatland, 2007; Shapiro, 2003). This suggests that when parents of adult children divorce, their need for social contact will not be as large as it is for people without adult children. After a divorce, people with adult children may therefore invest less in social relations and thus increase their informal helping less.

As a result, we expect that parents of adult children who do not live in the same household increase their informal helping less after divorce, compared to divorcees without adult children (H4).

Full-time work as moderator

Another group, next to parents of adult children, that may be less dependent on spousal contact is people who work full time. Employed individuals generally report that approximately half of their daily contacts are with colleagues (McDonald & Mair, 2010). Full-time employed individuals interact with colleagues during the workweek, which makes it relatively easy for them to engage in close social relations with them. This is more difficult for individuals who work part-time, and impossible for those who are not employed. In other words, full-time employed individuals may be less dependent on their spouse for social contact because they can uphold close social relations with colleagues. Individuals who work full-time would therefore feel less need to invest more in social relations after divorce, compared to those who do not have a job or who work part time.

As a result, we expect that individuals who work full time increase their informal helping less after divorce than individuals who work part time or not at all (H5).

Data

To test our hypotheses, we used the first four waves of the Americans’ Changing Lives (ACL) survey study (House, 2018).¹ These data were collected in 1986, 1989, 1994 and 2002, and covered approximately 15 years of adult Americans’ life courses. Although a fifth wave was available, we were not able to use it in this study because of a change in measurement of our dependent variable after wave 4. Unlike earlier waves, the fifth wave did not give the example of childcare (for others) for informal helping. As a result, it is possible that people did not count childcare hours as informal helping hours in the fifth wave while they did so in earlier waves.

The ACL sampling procedure focused on a representative sample of the adult American population over 25 years of age (excluding residents of Alaska and Hawaii), but oversampled people older than 60 and Black people.² Respondents were selected through a multi-stage design. First, a selection of counties was made, followed by a selection of areas within these counties. Afterwards, housing units were selected, from which individual respondents were randomly picked and approached for a face-to-face interview. In 1986, 3,617 individuals (68%) participated. In 1989, all respondents were approached again for face-to-face interviews, and 2,867 respondents (83%) participated. In the third and fourth waves, face-to-face interviews were replaced by telephone interviews, with in-person interviews conducted only if the respondent could not be reached by phone. Moreover, in the third and fourth waves (in 1994 and 2002), respondents could indicate that they were unable to participate in the interview, for example, due to health impairments or disabilities. In that case, proxy respondents (mostly the partner or adult children) could answer the questions for them, although they were only asked about certain aspects of the original respondent’s life. These aspects did not include informal helping, so we excluded these respondents (259 observations) from our sample.

Each wave had 15% to 25% panel attrition. Reasons for attrition included death, refusal or having an undocumented address or phone number. A total of 1,593 individuals participated in all four waves. Because our models allowed for unbalanced panels, we included all respondents that participated in at least two waves. Since we were interested in the transition from marriage to divorce, we removed all observations in which individuals were single and never married (9.1%) or widowed (19.2%). We also removed all observations (2.6%) that indicated remarriage (i.e. being married after divorce or widowhood). Finally, we excluded all observations with one or more missing values (4.6%). Our final dataset contained 6,185 observations of 1,955 individuals.

Measurements

Informal helping was measured by asking respondents about help they had provided over the previous 12 months to friends, neighbors and relatives who did not live with them. The survey specified that it concerned help given without receiving payment in return. Specifically, respondents were asked whether they ‘provided transportation, shopped or ran errands’, ‘helped others with their housework or upkeep of their house, car or other things’, ‘did childcare without pay’ and ‘did any other things to help neighbors, friends or relatives who did not live with them’. Respondents who answered ‘yes’ to any of these items were then asked how many hours they spent doing these activities in total during the last 12 months. Response options were ‘less than 20 h’ (1), ‘20–39 h’ (2), ‘40–79 h’ (3), ‘80–159 h’ (4) and ‘160 h or more’ (5). Respondents who indicated that they had not given any help were assigned to the category ‘did not give any help’ (0). For 246 observations of 227 individuals (6.3%), the score on informal helping hours was invalid.

We measured marital status by differentiating between respondents who were married and those who were divorced or separated.³ As we were interested in changes in informal helping after divorce, the reference category was being married. There were two observations in which information on marital status was missing.

With respect to the moderating variables, we measured having adult children by asking respondents if they had any children living outside their own household and how old these children were.⁴ If children were older than age 18, the respondent was categorized as having adult children. Working full time was measured by asking respondents whether they had a paid job, and if so, how many hours they worked in their main job. Working more than 32 h a week was considered to be working full time. For 125 observations of 116 respondents (3.2%), we had no valid score on full-time employment. Household income was originally measured in ten categories ranging from ‘less than $5,000 per year’ to ‘more than $80,000 per year’. This included all household income contributions before taxes from employment, wealth, child support, social benefits and food stamps. The ACL dataset provided an imputed version of household income. We used this version to ensure that all respondents had a valid score on income. To correct for inflation between 1986 and 2002, respondents were assigned the midpoint of their category. These values were divided by the consumer price index (base year 2010) and multiplied by 100.⁵ Finally, the corrected household income was divided by 1,000 to facilitate interpretation.

Table 1 reports the descriptive statistics for the variables used in this study. This table also includes background information on the panel, such as average age and distribution of gender, educational attainment and ethnicity. Table 2 reports the prevalence of changes and transitions in the main variables between the four waves.

Table 1. Descriptive statistics per wave.

		Wave 1 (N = 1,919)		Wave 2 (N = 1,939)^a		Wave 3 (N = 1,439)		Wave 4 (N = 888)
	Range	Mean	SD	Mean	SD	Mean	SD	Mean	SD
Informal helping hours	0–5	2.13	1.60	2.24	1.59	2.02	1.48	2.04	1.50
Marital status
Married	0/1	0.80		0.77		0.77		0.78
Divorced or separated	0/1	0.20		0.23		0.23		0.22
Moderating variables
Having an adult child (vs. not having an adult child)	0/1	0.57		0.60		0.65		0.72
Working full time (vs. not working full time)	0/1	0.48		0.47		0.46		0.49
Income	3.03-179.05	56.50	42.72	56.35	42.10	59.80	41.92	68.31	41.42
Background variables
Male (vs. female)	0/1	0.41		0.41		0.43		0.45
Age^b	25–89	50.23	15.42	52.99	15.51	56.33	14.47	59.97	12.93
Educational attainment
No degree	0/1	0.26		0.25		0.21		0.15
High school degree	0/1	0.58		0.58		0.61		0.63
College degree	0/1	0.16		0.16		0.18		0.22
Ethnicity
White	0/1	0.65		0.65		0.68		0.73
Black	0/1	0.25		0.25		0.23		0.19
Hispanic	0/1	0.04		0.04		0.04		0.03
Native American	0/1	0.05		0.05		0.05		0.04
Other	0/1	0.01		0.01		0.01		0.02

Note: ^aMore individuals were included in the second wave than in the first wave. While striking, this was the result of removing all observations in which individuals had never been married. This marital status occurred most often in the first wave, causing those observations to be removed. These same respondents entered the sample if they had married before the next wave.

^b Although only age at Wave 1 is included in the hybrid panel models in Appendix A, here we present age at the time of the specific wave.

Source: ACL, 1986-2002.

Table 2. Prevalence of changes in informal helping and transitions to divorce by wave.

	Wave 1–2 (N = 1,939)	Wave 2–3 (N = 1,439)	Wave 3–4 (N = 888)	Total (N = 6,185)
Informal helping
Increase in helping	36.0% (699)	27.7% (398)	31.8% (282)	32.3% (1,379)
Decrease in helping	31.9% (619)	43.1% (620)	38.4% (341)	37.0% (1,580)
Marital status
Divorce	2.6% (65)	3.0% (56)	4.4% (50)	4.0% (171)

Source: ACL 1986–2002.

Method

To test our hypotheses, we first compared average scores on informal helping for people who experienced a divorce and those who stayed married. Potential differences between these scores were tested using one-way analysis of variance (ANOVA). Table 3 reports the results of this analysis for each wave separately.

Table 3. Informal helping (hours on scale from 0 to 5) after staying in marriage or transition to divorce per wave.

	Wave 2 (N = 1,939)	Wave 3 (N = 1,439)	Wave 4 (N = 888)	Total (N = 6,185)
Staying in marriage	2.24	2.01	2.09	2.13
Experienced divorce	2.22	2.43	1.84	2.18

Source: ACL 1986–2002.

Table 3 gives us a glimpse of the relationship between divorce and informal helping. However, these results may refer to between effects and within effects. In other words, they can reflect both differences between people who did and did not experience divorce and changes in informal helping after a divorce. To estimate the impact of changes in marital status (i.e. the divorce), which is the focus of our study, we estimated fixed effects models in R, using the plm package. These models are specifically suited to assess the impact of changes in marital status on informal helping, as they control for unobserved heterogeneity (Allison, 2009). Fixed effects models control for all time-invariant differences between persons and only report on effects of time-varying independent variables. Effects can thus be interpreted as changes in informal helping when an independent variable increases by 1. Accordingly, the effect of divorce is interpreted as the change in a person’s informal helping after divorce. Table 4 reports the results of the fixed effects analyses.

Table 4. Fixed effects models predicting informal helping (N observations  = 6,185, N individuals = 1,955).

	Model 1		Model 2		Model 3		Model 4
	B	SE	B	SE	B	SE	B	SE
Marital status (ref. = married)
Divorce	0.123***	0.122	0.230	0.157	0.258	0.140	−0.008	0.151
Household income			0.001	0.001
x divorce			−0.002	0.002
Adult children (ref. = no)					0.426***	0.075
x divorce					−0.333*	0.140
Full time employment							−0.168*	0.069
x divorce							0.206	0.131
Wave (ref. = wave 1)
Wave 2	0.107**	0.041	0.107**	0.041	0.097*	0.041	0.104*	0.041
Wave 3	−0.262***	0.046	−0.260***	0.046	−0.306***	0.047	−0.273***	0.047
Wave 4	−0.325***	0.056	−0.324***	0.056	−0.419***	0.059	−0.343***	0.057

* p < .05. **p < .01. *** p < .001.

Source: ACL 1986–2002.

Despite our focus on the impact of divorce on informal helping, we also estimated the between effects; that is, differences in informal helping between persons who did and who did not experience a divorce, by performing hybrid panel models. These models also report the effects of time-invariant background variables, such as gender, age, educational background and ethnicity, on informal helping. Appendix A presents the complete results of the hybrid models. The appendix also discusses how hybrid panel models work and how reported effects should be interpreted.

Results

Table 3 presents the differences in informal helping between individuals who experienced a divorce and those who remained married, per wave and in total. People who got divorced between the first and second wave reported an almost identical amount of informal helping hours as those who remained married. This was different for individuals who got divorced between the second and third wave, as they spent more hours on informal helping than people who remained married. Although this suggests an increase in informal helping after divorce, this difference proved not significant. In contrast, people who got divorced between Waves 3 and 4 reported slightly fewer hours of informal helping than those who remained married. Again, these differences were insignificant. Hence, we conclude from Table 3 that people who got divorced did not significantly differ in their informal helping from those who remained married.

Table 4 reports the results of the fixed effects models. In line with the results from Table 3, Model 1 in Table 4 indicates that informal helping did not change when individuals got divorced. In other words, there was no significant difference in informal helping between when respondents were married and when they were divorced. Hence, Hypothesis 1, which predicted a decrease, and Hypothesis 3, which predicted an increase, both were rejected.

Next we account for the situation that positive and negative consequences of marital transitions on informal helping may have canceled each other out, resulting in a non-significant effect of divorce in Model 1. To test whether positive or negative effects were present for specific groups, we tested interactions between divorce and the three moderators: household income, having adult children and being employed full time. These are presented, respectively, in Models 2, 3 and 4. As shown in Model 2 (Table 4), the effect of divorce on informal helping did not differ between income groups. This refutes Hypothesis 2, which suggests that the negative consequences of divorce for informal helping are smaller for people with higher incomes.

Model 3 concerns the interaction between divorce and having adult children outside the respondent’s household. Whereas people without adult children who experienced a divorce tended to increase their informal helping (by 0.26 on a scale from 0 to 5; not significant), people with adult children tended to reduce their informal helping (by 0.08; not significant). Although no estimate effect was significantly different from zero, the difference between the two groups in the effect of divorce (b = −0.333) was significant. Therefore, we conclude that divorce tends to increase informal helping those without adult children than for those with adult children. This is in line with Hypothesis 4.

Finally, Model 4 in Table 4 presents the results of the moderation-interaction between divorce and full-time employment. In short, the impact of divorce on informal helping did not differ depending whether respondents were in full-time employment. This contradicts Hypothesis 5, as divorce was expected to increase informal helping less among people who were in full-time employment, compared to those who worked less or were unemployed.

Robustness analyses

We performed several robustness checks to ensure that our measurement and modeling decisions did not substantially influence the results. Details and models of these robustness checks are available upon request.

In our first set of robustness analyses, we performed the analyses separately with ‘divorced’ and ‘separated’ respondents. Although our theoretical framework likely applies similarly to both groups, they may nonetheless have differed. The results of these analyses showed they did not, with one exception: the impact of separation on informal helping did not differ between those who had adult children and those who did not.

Second, we performed the analyses separately for the oversampled groups (Black people and people over age 60) to check the impact of this oversampling on our results. These analyses showed very little difference between the Black sample and the non-Black sample. The only exception was the interaction between divorce and having adult children, which was insignificant in the Black sample. However, this might have been due to a power problem, as only 25% of the sample was Black. If it was not a power problem, this difference could indicate an underestimation of the results presented in Table 4. With respect to the oversampling of people older than age 60, we also found little difference. Again, this means only that our conclusions were not affected by the oversampling of people older than age 60.

Third, we replaced dummy variables for wave with an interval variable indicating the time between waves. This was because the time between waves varied from 3 to 8 years, which may have affected our results. Including the time between waves in our model enabled us to assess the size of this potential impact. Our results, however, were not substantially different when the time between waves was included.

Discussion

This study examined the impact of divorce on informal helping using longitudinal data from four waves of the Americans’ Changing Lives (ACL) panel study. Prior research reported both positive and negative consequences of divorce for informal helping (Cramm & Nieboer, 2015; Helms & McKenzie, 2014; Li & Ferraro, 2005; Lim & Laurence, 2015; Plagnol & Huppert, 2010). We sought clarification by testing opposing hypotheses. Moreover, we examined whether one type of consequence was more prominent among people with particular characteristics in regard to household income, full-time employment and having adult children.

Our results indicate that divorce, on average, was unrelated to the amount of informal help a person provided. This suggests that concerns about rising divorce rates undermining community involvement, such as informal helping, may be largely unfounded (Hughes & Stone, 2006; Putnam, 2000). Even when we examined groups that were expected to be especially prone to the negative consequences of divorce, such as individuals with lower household incomes, we found no support for the expectation that divorce results in less informal helping. Rather, we established that people were rather stable in their informal helping (even) after experiencing a divorce. Note that this is possibly due to the fact that people with fewer economic resources might be more likely to have others in their social network with few economic resources. In such networks, the demand for informal help may be larger, creating more opportunities for help. Divorcees with lower incomes may thus capitalize more on their free time and desire for social contact than higher-income divorcees.

Our findings also indicate that people who were expected to be less dependent on their spouse for social contact and support (i.e. people with adult children) were less affected by divorce, as they did not invest more in close social relations after divorce. It must be noted that we expected a similar finding with regard to employment, namely, that those who worked full time and thus had more contact with colleagues would be less affected by divorce. Here, however, we found no difference in divorce’s impact. This finding may be due to a lack of close social relations between colleagues, meaning that despite contact with colleagues, full-time employed individuals may be as dependent on their partner for contact and support as other individuals. They would therefore have a similar need to invest in close social relations after divorce. Additionally, an alternative mechanism might explain this relationship between full-time work and informal helping. Because part-time or non-working people, in contrast to full-time working people, have had the time and energy to build close social relations during their marriage, especially the full-time workers may experience an increasing investment in close social relations after divorce. The loss of a partner relation freed up time and energy to spend on social contacts.

Our theoretical arguments were mostly based on social network size and the closeness of social interactions, though we did not have specific information about the size and closeness of respondents’ social networks. Based on previous findings, the size of the network was expected to decrease after divorce (Eckhard, 2021; Terhell et al., 2004; Wrzus et al., 2013), resulting in a decline in informal helping. In contrast, the closeness of (remaining or new) social relations was expected to increase after divorce (Kalmijn, 2012; Sarkisian & Gerstel, 2008, 2016), resulting in increased informal helping. One group of people who were expected to have less need of increased close social relations after divorce, that is, people with adult children, were indeed found to be less affected by divorce. This suggests that closeness may matter more than the size of the social network. We encourage other researchers to follow up on this proposition in future research, by measuring network size and the closeness of social interactions.

This study contributes to discussions about divorce and community life by showing the impact of divorce on informal helping. While informal helping is an important aspect of community life, and one that is perhaps more susceptible to the impact of divorce, it is not representative of all aspects of community life. Consequently, the results reported here cannot be generalized to all forms of community life. This is exemplified by prior studies, such as Lancee and Radl (2014) and Nesbit (2012), which found that divorce reduced people’s formal volunteering. Thus, in order to paint a full picture of the impact of divorce on community life, findings on various aspects of community life, such as formal volunteering, charitable giving and informal care, should be considered.

Although the present study improves on prior research by employing a longitudinal design, the study design has some limitations. First, our data was collected at four points in time that were rather far apart. For example, eight years passed between the collection of the third and fourth wave of the ACL data. As a result, this study may not capture all short-term effects of divorce on informal helping. Furthermore, the large gap between data points makes it difficult to disentangle the exact conditions under which divorce took place. For example, a person’s income may have been significantly reduced after divorce but restored to pre-divorce levels by the time of data collection (although De Vaus et al. (2017) found that divorced men and women in the US both remained at an economic disadvantage compared to their continuously married peers). Hence, a reduction in income may not always have been captured by the data. It would be advisable to replicate this study with yearly data on divorce and informal helping. Still, it must be noted that this study’s analyses likely underestimated the impact of divorce, as some people who got divorced will have already adjusted to their new situation by the time of data collection.

Second, the fourth round of ACL data was collected in 2002, making our most recent data point twenty years ago. Although divorce rates seem to have stabilized or even declined since then (NCHS, n.d.), other trends such as growing individualism and secularization have likely continued. Moreover, the last two decades have seen the rise of social media usage, making it easier to maintain social contacts between those who live farther apart – though it is probably still difficult to provide informal help to people who do not live nearby. Hence, close social relationships may bring increasingly less informal helping nowadays. We advise future research to replicate our study with more recent data. We also call on the Americans’ Changing Lives study to revise its measurement of informal helping in future waves, so that it is comparable to earlier waves.

This study suggests that the rising divorce rates in the 1980s and 1990s did not bring a decline in informal helping. People did not provide less informal help after divorce than before. Hence, concerns that the breakdown of traditional families because of rising divorce rates would result in the breakdown of community life seem unfounded.

Acknowledgements

The authors jointly developed the study idea and design. Ramaekers wrote the main manuscript and conducted the analyses. Verbakel, Kraaykamp and Van der Lippe substantially contributed to the manuscript.

Disclosure statement

No potential conflict of interest was reported by the author(s).

Additional information

Funding

This study is part of the research program Sustainable Cooperation – Roadmaps to Resilient Societies (SCOOP). The authors are grateful to the Netherlands Organization for Scientific Research (NWO) and the Ministry of Education, Culture and Science Netherlands (OCW) for generously funding this research in the context of its 2017 Gravitation program [grant number 024.003.025]; Nederlandse organisatie voor wetenschappelijk onderzoek (NWO).

Notes on contributors

Marlou J. M. Ramaekers

Marlou J. M. Ramaekers is a PhD candidate at the Department of Sociology at Radboud University Nijmegen. Her research focuses on informal person-to-person helping and the impact of family life and neighbors on the provision of help. Other interests of her include civic engagement and social identity.

Ellen Verbakel

Ellen Verbakel is full professor in Sociology at the Department of Sociology, Radboud University in Nijmegen, the Netherlands. She is a family sociologist with research interests in the interplay between family, work, and health, and has a special interest in informal care.

Gerbert Kraaykamp

Gerbert Kraaykamp is a Professor of Empirical Sociology in the Department of Sociology/ICS, Radboud University Nijmegen, the Netherlands. His research interests include educational inequality, health inequality and family socialisation. He has published widely on these subjects in international journals.

Tanja van der Lippe

Tanja van der Lippe is Professor of Sociology at the Department of Sociology/ICS of Utrecht University. Her research interests are in the area of work-family linkages in Dutch and other societies. She has published extensively on work and care of men and women, time use and time pressure in a comparative way, and the position of men and women on the labor market (including supervisory positions).

Notes

1 The ACL data were collected as a large-scale survey. This data collection method reflects a non-interventional study, and it was not necessary to gain approval by a research ethics committee.

2 Oversampling was performed by an oversampling of medium to high Black residential areas (based on the U.S. Census of 1980). Housing units and respondents were then sampled from such residential areas. As a result of this oversampling, Southern states are somewhat overrepresented in the data.

3 Cohabitors were not included in the ‘married’ category but were considered single (never married, divorced or widowed). Union dissolution after marriage might be a different experience than union dissolution after cohabitation, especially in the period under study. Robustness analyses in which cohabitation was equaled to marriage show similar results but with the exception that divorce has a significant and positive effect on informal helping. Robustness analyses in which cohabitors were removed however showed similar results.

4 As a robustness analysis, we included the number of adult children instead of having adult children (yes/no). This did not change our conclusions.

5 For example, respondents who earned $15,000 to $19,000 in 1989 were first assigned an income of $17,000. The price index for 1989 was 56.851. Hence, the income was corrected for inflation with the following formula: 17,000/56.851*100 = 29,903.