|Home | About | Journals | Submit | Contact Us | Français|
The divorce-stress adjustment perspective defines divorce not as a single event, but as a process whose effects linger even after remarriage (Forste and Heaton 2004; Lorenz et al. 2006). Previous studies based on the divorce-stress adjustment perspective looked at divorce as a stressful process and analyzed how divorce can negatively affect health outcomes after the actual divorce had taken place. This perspective is a combination of various elements of stress frameworks that has been dominating the literatures on divorce (Amato 2000). However, the popularity of the stress framework resulted in paying less attention to studying divorce as an active or passive choice that some individuals make in their life course and the life event influences their social behaviors in later life, which could provide another possible explanation why divorce can negatively influence health even after remarriage.
Divorce is a life-altering decision that involves various thoughts before making the final legal decision. Many couples chose to try separation before making the decision as a way to think of the personal and social costs of divorcing their spouse as well as benefits and the necessity. Thus choosing to divorce as their final decision can be a sign of people taking a different path from those who did not or did not need to make the decision. If the path people take in choosing to divorce is less socially protected than the path of those who continued to be married take, not only lingering effects of divorce, but also differences in the life course patterns could also influence their health in later life.
Previous studies reported that less religious people are more likely to divorce since religious people are more likely to have stronger moral beliefs in marriage (Guerrero et al. 2007). Glen and Supancic (1984) found that frequent religious attendance has a strong negative correlation with divorce. Although most studies focused on looking at how religious attendance before or at the time of marriage reduces the chance of divorce, little attention was paid to examine how the divorce experience will reduce the chances of their religious attendance in later life controlling for their previous religious attendance. Religious involvement is discussed as having a protective effect including preventing and reducing unhealthy lifestyles (Idler and Kasl 1992; Rippentrop et al. 2005). Research by Hummer et al. (1999) indicates that people who are more deeply involved in religion are less likely to die than those who are not involved in religion. Thus, not simply divorce as a stressor continuously troubles health even after remarriage; those who experienced a divorce may be less likely to gain social protection through religious involvement in later life than those who continued to be married. Such differences in life course patterns of those who experienced a divorce and those who did not could also continuously influence health even after remarriage.
Despite extensive research of divorce and negative health outcomes, almost no attention has been paid to the study of the longitudinal relationship among previous divorce experience, later religious involvement, and possible linkage to healthy lifestyle. The major goal of this study is to examine married people regarding (1) whether their previous divorce experience decreases the likelihood of religious involvement, and (2) how religious involvement can be linked to health in later life.
The divorce-stress-adjustment perspective proposed a broader view of divorce as a stressor on one's life course rather than a simple causal view based on the comparison between divorced people with married ones at the time of the survey. The divorce-stress-adjustment perspective states that divorce as a process begins with the feelings of estrangement from their spouse, which leads to their decisions to separate, then followed by adjustment after divorce (Lorenz et al 2006). This perspective argues further that such lingering effects of divorce continue to disadvantage the well-being of previously divorced individuals even after they remarry (Amato 2000; Forste and Heaton 2004).
Previous research based on the divorce-stress adjustment perspective by Lorenz et al. (2006) found that the divorced women reported significantly higher levels of psychological distress immediately after divorce, and these women also reported higher levels of illness a decade later controlling for age, remarriage, education, income, and prior health status. Another study based on the divorce-stress adjustment perspective by Forste and Heaton (2004) indicates that ever-divorced or separated individuals are more likely to have been exposed to physical violence and threatened with a gun, and ever-divorced women are more likely to have been hospitalized (Forste and Heaton 2004). Forste and Heaton (2004) also emphasized the beneficial role remarriage plays in reducing negative effects of a previous divorce experience. They claimed that remarried individuals report higher levels of well-being compared with the still-divorced or separated (Forste and Heaton 2004). These studies based on the divorce-stress-adjustment perspective looked at divorce as a process to understand how divorce as a stressor produces negative effects on health even after the marriage has legally ended.
The divorce-stress-perspective includes elements of stress frameworks, especially sharing the major assumption that divorce is a stressful life transition that people must adjust to (Amato 2000). The largest number of studies of divorce begin with this assumption, thus many studies of divorce can be linked to stress perspectives and summarized in the divorce-stress-perspective (Amato 2000). However, the popularity of such perspectives kept out the view of divorce as a life event some individuals experience in their life course, which could influence their social behaviors in later life and health even after remarriage.
Since all couples enter marriage hoping for a mutually supportive and rewarding life-long relationship, initiation of uncoupling is often painful and they may spend considerable time to renegotiate the relationship by asking advice from others, separating for a period of time, or waiting for the problem to be solved with time (Amato 2000). In thinking of divorce, basically, there are those who continued to be married happily ever after, those who never divorced but may have gone through considerable thinking, and those who decided to divorce after considerable thinking. For the first two groups, traditional beliefs regarding the sanctity of marriage and the family value may serve as strong barrier to divorce, possibly even to thinking about divorce, relative to the last group (Booth and White 1980). In fact, previous studies show that more religious people are less likely to divorce (Guerrero et al. 2007). The State of Our Unions 2004 reported that religious affiliation reduced the percent in divorce by 14 percent, as well as 13 percent for higher education, and 30 percent for higher annual income (The National Marriage Project 2004). Early Years of Marriage project at the University of Michigan also reported that couples that attend church are less likely to divorce. These studies looked at how religious involvement before or at the time of marriage influences divorce. However, they have not examined how the divorce experience influences their religious attendance in later life controlling for their previous religious attendance.
Some people are religiously involved throughout their life course, but other people begin participating in religious organizations sometime in their life course, and divorce can be a significant life event that strongly affects their religious involvement in later life. Based on previous studies, controlling for previous religious involvement, whether or not the life event differentiates the life course patterns of religious involvement in later life cannot be unanswered.
Many studies have studied the significant and positive effect of religious involvement on the well-being of people (Musick 1996; Rippentrop et al. 2005; Brown 2002; Koenig 2002). Religious involvement is discussed to reduce “belief-based” behaviors including alcoholism, drug use, cigarette smoking, risky sexual activity, failure to wear seat belts, and drinking while driving (Koenig 2002). It is also argued that religiousness affects individual health through facilitating social support by and for religious group members, and providing coherent frameworks of meaning that provide comfort, coping, and understanding in the time of difficult transitions in life especially at old age (Idler 2004). Brown (2002) claims the relationship between religion and morbidity retain their significance even after controlling for age, sex, race, education, social support, level of physical functioning, severity of medical illness, and depressive symptoms. He goes further and claims that religion's ability to increase social support, reduce health risk behaviors, and enhance conformity with treatment cannot entirely explain the beneficial relationship between religion and health (Brown 2002).
Other studies found that religiously involved individuals are also more likely to have their medical problems diagnosed since being a part of a religious community increases social support and their likelihood of watching for future health problems by themselves as well as by people in the religious community (Koenig 2002). Koenig et al. (1999) showed 28% decrease in relative hazard of dying for frequent church attendees in comparison with infrequent attendees. Koenig et al. (2001) reviewed 138 studies examining the relationship between alcohol or drug use or abuse and religiousness, and found that 90% of these studies reported significantly lower substance use among those who were more religious. 24 out of 25 studies examined the relationship between religiousness and smoking, Koenig et al. (2001) reported that 24 found less smoking in those who were religious. However, most of these studies were conducted in young adults, not older adults (Koenig et al. 2001).
Finally, Idler (2002) demonstrated regional difference in mortality and religiousness. Nevada's death rate for males aged 40-49 was 54 percent higher than that of Utah's in 1974 (Idler 2002). More specifically deaths from cirrhosis of the liver were 111 percent higher and the deaths from cancer of the respiratory system were 296 percent higher in Nevada than Utah in 1974 (Idler 2002). The most recent data also showed the consistent result that Nevada has by far higher smoking rates in the U.S. (31.5 percent) compared with Utah, the lowest (13.9 percent) contributing to the lowest rate of lung cancer in Utah (Idler 2002).
These studies show the strong linkage between religious involvement and health in later life in the United States. Thus, it is possible that, controlling for previous religious involvement, those who continued to be married are more likely to attend religious services in later life than those who remarried, and not only lingering effects of divorce, but also differences in the life course patterns could influence their health in later life through health related lifestyle such as smoking.
Researchers have discussed the importance of demographic and socioeconomic factors influencing religions participation and smoking habit. These are age, sex, socioeconomic status, race/ethnicity, having a child, and physical condition.
Age is discussed to be an important factor influencing both smoking habit and religious attendance. Duke Medicine News and Communications (2006) reported that older people are more successful than younger ones in quitting smoking. As for the relation between age and religious attendance, various studies reported that religious attendance is higher for older adults than younger ones since people tend to experience crisis and challenges as they grow old (George et al. 2004). Sex is another important factor discussed to affect both religious attendance and smoking. Although the difference has been declining, men are more likely to smoke than women in the United States (Pampel 2001). As for religious involvement, previous studies found that women are more likely to be involved in religion than men (Barna 2002).
Income and education are two important socioeconomic variables influencing smoking as well as religious attendance. Researchers have found that people in lower socioeconomic groups are more involved in religion (Krause 2002). As for the effects of socioeconomic status on smoking habit, Paavola et al. (2004) showed a strong relationship between education and smoking. At the age of 28, 63% of those who had just a compulsory comprehensive education were smokers compared with 12% of those who had university level education (Paavola et al. 2004). Research by Laaksonen et al. (2005) also reported a clear inverse association with smoking and income.
There are many unanswered questions for understanding the relation between religion and health in relation to race differences (Chatters 2004). Religious institutions for African Americans have constituted distinctive social environments in the history and they have functioned as buffers of larger social forces (Chatters 2004). Religion for African Americans is connected with community and the public arena of life, significantly contributing to secular efforts and concerns including the establishment of education, social welfare, and political institutions. In other words, religion is deeply intertwined with human affairs for African Americans (Chatters 2004). African American churches can provide spiritual and emotional support to the people who go through adjustment in the time of stressful events like divorce (Orbuch and Brown 2005). Past studies comparing racial differences between health and religion showed inconsistent results in that some studies found religious involvement was associated with self-rated health for both Whites and African Americans. Others found significant effect for Whites, but not for African Americans (Chatters 2004). Further studies are needed to understand racial differences in the relationship between religious involvement and health in later life.
Having a child is discussed to be an important variable in smoking habits and religious attendance among married respondents. Becoming and raising children are good reasons to quit smoking, and research showed that married women with children were less likely to smoke than women without children (Jun and Acevedo-Garcia 2007). Having children also increases their opportunities for visiting religious services as part of children leaning religious socialization. In fact, a study found that people with children were more likely to attend religious services than people without children and singles (George et al 2004).
Finally, previous studies showed the positive effect of religious involvement on health in later life. Some studies warned of the possibility of a spurious relationship between religious involvement and health since disabled people are less likely to attend religious involvement (Levin and Vanderpool 1987). Thus, looking at the relationship between religious involvement and health related behaviors without controlling for their functional status is problematic.
George et al (2004) claims that little attention has been paid to understand possible links between life course patterns of religious participation and health. Better understanding of the longitudinal relationship among previous divorce experience, religious involvement in later life, and the well-being of currently married individuals can propose one possible mechanism to understand why previous divorce experience continues to differentiate the well-being among married people in later life. Going beyond the argument that the process of divorce functions as a long lasting stressor to their body and mind, the path people take in choosing a divorce may not be as protected as the one those who continued to be married take, which could also influence later health outcomes. In other words, adjustment of religious organizations to support for divorcing or divorced people may be as important as successfully going through the stressful life transitions to prevent or reduce their future health problems.
In this study, specifically two research questions are raised to be answered.
Research Question 1: Among married individuals, does previous divorce experience reduce the likelihood of regular and public religious involvement controlling for previous religious involvement?
Research Question 2: Among married individuals, does religious involvement influence their regular smoking habit?
Since little attention has been paid to studies that empirically look at the longitudinal relationship among married individuals, this research provides an initial step to understand how individual's previous divorce experience relates to their religious involvements, which could influence later health even after remarriage.
Data for the analysis are taken from the three waves of the Wisconsin Longitudinal Study (N=10317). All Spring 1957 high school graduates in Wisconsin filled out their first questionnaire in 1957. They were followed up to fill out their questionnaires in 1975 (when they were about 35 years old) and 1992 (about 53 to 54 years old). Various studies recognize the age from 35 to 55 as middle age compared with preceding age often labeled as young adulthood (Beck and Beck 1984). In other words, 1975 wave asks questions about respondents' behaviors at the beginning of middle age, and 1992 wave asks their behaviors at the end of middle age.
Of 10317 respondents, 67.7 percent of respondents answered that they are currently married. Of 6980 currently married respondents, 27.3 percent of respondents failed to provide answers to the questions or they passed away by the 1992 wave, resulted in a total sample size of 5076 married respondents to analyze the relationship between previous attendance and religious attendance in later life, and 5220 married respondents to examine the relationship between smoking and religious attendance.
Both 1975 and 1992 waves asks questions about respondents' frequency of religious attendance, which enables to look at the effects of divorce experience on their religious attendance in 1992, controlling for their religious attendance in 1975. In addition, Respondents were also asked about their regularity of smoking habit, which allows examining the relationship between religious attendance and smoking in 1992. Both waves also asked questions about respondents' marital status (number of marriages and current marital status).
The limitation with the Wisconsin Longitudinal Survey is that the survey includes so few non-whites (<2%), thus it is not suitable data to compare whites with non-whites and to study the relationship among SES, religion, race/ethnicity and health by race and ethnicity to understand the changing socio-demographic features of aging by race and ethnicity.
Frequency of religious attendance at the 1975 wave and 1992 wave are measured as follows. In 1975 wave, respondents' frequency of religious attendance was asked in the following format, “How often did you attend religious service last year?” Respondents choices were (1) one time per week, (2) two or three times per month, (3) one time per month, (4) a few times per year, (5) less than a few times per year, and (6) never. In 1992 wave, respondents were asked their number of religious attendance (0 to 365) as well as the unit for frequency of religious attendance (day, week, month, year) during the last year. These two variables were summarized into an ordinal variable, frequency of religious attendance during the last year. The value ranges from 0 (never or less than once a year) to 11 (approx. once a day or more, 321-730 times per year). To make religious attendance in 1975 and 1992 comparable, both variables were dichotomized indicating (0) no regular religious attendance (less than once a week) and (1) regular religious attendance (at least once a week). In 1975, 49.9 percent of people regularly participated in religious attendance (at least once a week), and the proportion was about the same (46.6 percent) in 1992. McNemar's test showed a significant statistical difference between 1975 and 1992 religious attendance among married people (Chi-Square 33.735, p<.001).
For a measure of the respondent's health related lifestyle, since smoking was discussed as a major example of unhealthy lifestyles in previous studies, a question asking about regular smoking habit at the 1992 wave was used. Respondents were asked, “Do you smoke regularly now?” 16.1 percent of married respondents at the time of 1992 survey smoked regularly. Those who smoked regularly were coded 1, and those who did not were coded 0.
Divorce experiences are measured by responses to multiple questions about their marital history (marital status and the number of marriages) as well as their 1975 and 1992 marital status. These questions contributed to create two variables, (1) whether a respondent experienced any divorces by 1975; (2) whether a respondent experienced any divorces between two waves. Those who experienced divorce are coded 1, if not they are coded 0. Among those who are married at the time of the 1992 survey, 10.8 percent experienced divorce by 1975, and 16.2 percent experienced divorce or additional divorce between two waves.
As control variables: two indicators of their biological information (age and sex); two indicators of respondents' socioeconomic status (income and education); information about respondent's total number of children; and respondents' number of physical symptoms are included. In this study, respondent's age in 1992 was measured as an interval variable (average 53.25 years). Sex was measured as a dichotomous variable (coded 1 for female, 0 for male), and 48.4 percent of respondents were men. Respondent's family income was measured at an interval level, and square transformation was used to reduce the positive skew (skew has changed from 2.172 to.752). Education was measured as summary years of education based on their most recent degree ranging from high school (12 years) to postdoctoral education (21 years).
The total number of children was measured at an interval level (average 3 children). Finally, to control for respondents' physical health status, the number of physical symptoms married respondents had during the past 6 months at the time of 1992 survey was included. The symptoms include 22 items: whether respondents experienced lack of energy, exhaustion, dizziness, numbness, ringing in ears, nauseated, shortness of breath, headaches, visual problems, upset stomach, vomited, constipation, diarrhea, urination problems, aching muscles, swollen joints, back pain, chest pain, excess sweating, respiratory problems, or skin problems. On average, respondents had 4 symptoms.
Logistic regression was used to examine two research questions: (1) whether previous divorce experience had a significant effect on married respondents' regular religious attendance at the 1992 wave controlling for their religious attendance at the 1975 wave; and (2) whether regular religious attendance at the 1992 wave had significant effect on respondents' regular smoking habit.
Table 1 reports an analysis of logistic regression looking at the effects of previous divorce on religious attendance among married individuals at the time of the 1992 survey. It shows, controlling for previous religious attendance, not the divorce experience by 1975 (approx. by age 35), but the divorce experience between 1975 and 1992 (approx. from age 35 to 53) significantly reduced the likelihood of attending religious services in 1992. Regardless of their religious involvement when they were younger adults as well as their physical health symptoms, divorce experience during the middle age significantly reduced the odds of married individuals attending religious services at the end of the middle age. Thus the first answer to the research question is that, among married individuals, previous divorce experience, especially divorce experience during the middle age, significantly reduced the odds of regular and public religious involvement in later life. In other words, without experiencing any divorces during middle age increases the chances of religious attendance at the end of middle age, the beginning of old age.
As expected, previous religious attendance had a strong and significant effect on current religious attendance at the time of the 1992 survey. The odds of those who attended the religious services regularly when they were younger adults to attend religious services regularly in later life were about 12.6 times the odds of those who did not attend the religious services regularly when they were younger adults. Although further inference is limited, such strong effects of religious attendance at the younger adult stage on their later religious attendance suggest religious socialization to be one of the strongest predictors of religious participation in later life.
Looking at control variables, education and the number of children respondents have slightly increased the likelihood of religious attendance, but age did not have a significant effect since the sample scheme of the Wisconsin Longitudinal Study is asking people at close ages (initially high school graduates). As previous studies suggested, women were about 20 percent more likely to attend religious services than men. Previous studies reported that people in lower socioeconomic groups are more involved in religion, however such relationship was not significant in this study especially because the WLS contains so few non-whites (<2%), which limited the detection of the complex effects of socioeconomic status and race on religious attendance. Finally, the number of physical symptoms did not have a significant effect of religious attendance in 1992. This is partly because many people do not start serious health problems at the age of 50, and on average respondents did have less than 5 physical symptoms out of 22 symptoms at the time of the survey.
Table 2 shows the result of logistic regression looking at the relationship between religious involvement and smoking among married respondents approx. at the age of 53 years old. Supporting previous studies, religious involvement greatly contributed to the healthy lifestyle. Regular religious attendance significantly reduced the odds of regular smoking. Examining married respondents at the age of around 53 years old, those who did not attend religious services regularly had almost 1.5 times the odds of regular smoking compared with those who regularly attended religious services (Table 2). Therefore, the answer to the second research question is that religious involvement had significant effect on regular smoking habit among married people. Previous studies showed that marriage is an important factor reducing risky health behaviors, and this result showed that regular religious attendance is a significant factor negatively linked to regular smoking habit among married people.
The bivariate correlation between regular smoking habit and previous divorce experience was less than .06, which suggests no direct relationship between these two variables. Thus, the results of this study do not suggest that remarried people are more likely to be unhealthy in later life compared with those continued to be married. The results of this study suggest who are more likely to be publicly involved with religion in later life, and how religious involvement benefits those people in later life. Since divorce experience during middle age was a significant factor determining regular religious attendance at the end of middle age, the results suggest one possible mechanism among many to explore why remarried individuals still get disadvantageous health outcomes in later life compared with continued to be married individuals. The reason may not be simply because they are still suffering from lingering effects of divorce. Further studies are necessary to explore the dynamic relationship in later life course.
Among married respondents, controlling for previous religious involvement, divorce experience during middle age reduced the chance of being publicly involved with religion at the end of middle age. Among the married, those who attend religious services regularly at the end of middle age were less likely to smoke regularly. These results suggest that not only lingering effects of divorce as a stressor negatively influences health outcomes even after remarriage, but also the protective life course of those who are religiously involved in later life course may differentiate health outcomes even after remarriage.
Divorce experience before entering middle age did not have a significant effect on religious attendance at the end of middle age, but divorce experience during middle age did. If the timing of divorce experience matters on later religious involvement, divorce experience during middle age may no longer be a significant factor on the religious involvement in old age. At old age, other factors, especially losing a spouse and friends, may trigger the likelihood of regular religious attendance. Future waves are necessary to further examine the longitudinal relationship between divorce experience, religious involvement, and health outcomes. In addition, other factors such as quality of relationship with a current spouse as well as support from families and friends should be included as substitutable protective factors on the well-being of remarried individuals in later life.
The longitudinal relationship between previous divorce experiences, religious attendance in later life, and well-being will be quite different by race and ethnicity. Other data sets which includes or focuses on African American respondents will make this study comparative, especially to understand whether the meaning and consequences of religious involvement differs between African Americans and Whites. Also, future study should further explore the longitudinal relationship comparatively looking at other marital groups including those who continued to be married or divorced, remarried, widowed, and never married individuals. Finally future studies should include respondents' religious affiliations to examine which religious affiliation is more influenced by respondents' divorce experience, and how the effect of religious involvement on the well-being in later life differs by religious affiliation.
Blomquiest (1985) conducted in-depth interviews of fifteen people who experienced or are experiencing divorce. Analyzing these interviews, she concluded that the divorce experience was a definite turning point in their spiritual lives, although the extent of their spiritual growth differs from one another depending on whether divorce experiences are perceived as traumatic (Blomquiest 1985). Some ever-divorced individuals may choose not to attend religious services in later life to be free from religious obligations; others seek for religious support but may not feel comfortable visiting or revising their religious organization because divorce became a barrier to attending religious services. Religious organizations adjusting to the social reality of increasing numbers of divorces can open doors for the latter as well as the former. Qualitative studies are essential to understand the context of divorce: including who initiated the divorce, the process of divorce, and the consequence of divorce on religiosity and health in later life.
Understanding spiritual or religious support, especially how it differs from other types of support, has not been intensively discussed partly because industrialized societies tend to suppress discussing the effects of spiritual or religious beliefs in public. However, this topic is important especially in later life when more people lose their spouses by death than by legal separation. To advance the research on divorce and remarriage, extending the study to understand the longitudinal dynamics of religion, health, and changes in marital status (divorce, remarriage, and widow/widowers) is essential.
The author wishes to acknowledge the Center for Demography of Health and Aging (P30 AG17266) and the Center for Demography and Ecology (R24 HD047873) for offering the NIA postdoctoral fellowship (T32 AG00129) to conduct this research. The author also would like to thank Jamison T. Arimoto, Esq. for useful insights.