Age differences in Exposure and Reactivity to Interpersonal Tensions among Black and White Individuals across Adulthood

Abstract

The present study examined age differences in exposure and reactivity to interpersonal tensions among White and Black Americans. Participants from the National Study of Daily Experiences II (NSDE II, n= 1696 White and n = 239 Black; ages 34 to 84) reported their experiences of daily interpersonal tensions and well-being (positive and negative affect) over 8 days and provided salivary cortisol samples. A total of 40% of respondents reported having an argument and 62% reported avoiding an argument. Multilevel models estimated separately for Black and White respondents revealed that older people reported fewer interpersonal tensions (i.e., less exposure) than did younger people. However, age differences in reactivity to tensions (e.g., appraisals, coping strategies, implications of tensions for affect and cortisol) varied by race. Although older Black respondents reported tensions were less stressful than younger Black respondents, there were fewer age difference in reactivity to tensions overall among Black respondents compared to White respondents. Findings are consistent with the exposure reactivity model and gerontological theories of emotion regulation but show that the specific age differences vary by race which may indicate unique strengths and vulnerabilities among Whites and Blacks.

Gerontological research consistently shows that despite age related declines in health and cognition, there are age related improvements in social relationships (Carstensen, Fung, & Charles, 2003; Birditt, Jackey, & Antonucci, 2009). Older people report fewer interpersonal tensions, more avoidance, fewer arguments, and they are less reactive to interpersonal tensions than younger people (Birditt, Fingerman, & Almeida, 2005; Blanchard-Fields, Chen, & Norris; 1997). It is unclear, however, whether these age-related improvements vary by race. Black Americans may show different developmental patterns than White Americans due to their different life experiences. African Americans often report experiencing more stress than do White Americans (Mujahid, Diez Roux, Cooper, Shea, & Williams, 2011; Ross & Mirowsky, 2001; Williams & Mohammed, 2009). The accumulation of stress across the lifespan may lead to greater vulnerability among Black Americans and they may thus show fewer age-related improvements (Charles, 2010). In contrast, such experiences may lead to even greater age-related improvements among Black Americans due to their increased resilience (Neighbors, Hudson, & Bullard, 2009). Interestingly, the literature reveals racial disparities in physical health (Geronimus, Hicken, Keene, & Bound, 2006) but not mental health (Kessler, et al., 2005; Neighbors, Sellers, Zhang, & Jackson, 2011) which shows that increased vulnerability and resilience may exist simultaneously among Black Americans.

The present study examined age differences in daily experiences of arguments, avoidance of arguments, and their associations with daily self-reported well-being and diurnal cortisol among Black and White Americans. We include both self-reported and physiological measures to provide a comprehensive assessment of daily psychological and physical health. Cortisol provides an indication of the functioning of the HPA axis; chronic activation of which is linked to depression, heart disease, bone demineralization, loss of muscle mass, increased abdominal fat, and decreased hippocampal volume (Ader, 2001; Heim, Ehlert, & Hellhammer, 2000; McEwen, 1998; McEwen & Sapolsky, 1995; Repetti, Taylor, & Seeman, 2002; Sapolsky, 1996).

Theoretical Framework

The stressor exposure-reactivity model provides a framework for understanding coping with daily interpersonal tensions and their implications for well-being and cortisol (Almeida, 2005). We adapted the model for the purpose of this study and refer to it as the interpersonal exposure-reactivity model. Interpersonal tensions include problems and irritations in relationships. According to this model, there are variations in the number of problems people are exposed to as well as how they react to problems. Exposure and reactivity to daily stressors influence physiological systems and psychological well-being (Almeida, 2005; Almeida, McGonagle, & King, 2009; Bolger & Zuckerman, 1995). Exposure refers to the number of interpersonal tensions experienced and reactivity refers to appraisals of the tension, coping strategies used as well as the extent to which self-reported well-being or cortisol are altered by the experience of interpersonal tensions. Self reported and biological well-being represent separate but related dimensions of well-being.

Individuals vary in their emotional appraisals of situations and their behavioral reactions. Appraisals involve the meaning and severity attributed to the situation (Lazarus & Folkman, 1984). In the present study we define appraisals as the perceived stressfulness of the situation. Coping strategies are often defined along two dimensions in terms of whether they are active or passive (Folkman, Lazarus, Pimley, & Novacek, 1987; Lazarus, 1999). We considered avoidance of arguments as a passive response and engagement in arguments as an active response to potentially tense interpersonal interactions. Avoidance of arguments involves not confronting the stressor directly, such as accepting the situation as it is, reappraising the situation, and doing nothing (Birditt et al., 2005; Blanchard-Fields, Stein, & Watson, 2004). Engaging in arguments involves directly confronting the person regarding the problem.

Well-being comprises psychological and physical dimensions and we include self-reported assessments of mood (positive affect, negative affect) and a biological indicator of well-being (diurnal cortisol). Self-reported affect provides a good indicator of daily well-being because it fluctuates on a daily basis compared to more global measures (e.g., life satisfaction) which are more stable (Diener, Suh, Lucas, & Smith, 1999; Kahneman, Krueger, Schkade, Schwarz & Stone, 2006). It is important to examine affect on a daily basis because retrospective reports tend to overestimate the intensity of affect (Thomas & Diener, 1990).

Although self-reported well-being measures are associated with objective indicators of health such as mortality (Idler, Leventhal, McLaughlin, & Leventhal, 2004; Phillips, Der & Carroll, 2010), physiological measures of well-being such as cortisol may provide important information regarding the pathways by which daily interpersonal tensions influence overall health and well-being. Cortisol has a normal diurnal rhythm over the course of the day in which it begins to increase before waking, reaches a peak level at about 30 minutes after waking (cortisol awakening response[CAR]) and steadily declines thereafter until bedtime (daily decline; Fries, Dettenborn, & Kirshbaum, 2009; Pruessner et al., 1997). The CAR represents the anticipation of the coming day or a boost of energy to ready the person for the day (Adam, Hawkley, Kudielka, & Cacioppo, 2006; Almeida, Piazza, & Stawski, 2009; Fries et al., 2009). Having a blunted CAR is associated with burnout, fatigue, and post traumatic stress whereas a CAR that is too steep is associated with increased job stress and other life stress (Chida & Steptoe, 2009). Likewise, flatter daily declines are associated with increased stress and lower marital quality (Adam et al., 2006; Adam & Gunnar, 2001; Saxbe, Repetti, & Nishina, 2008) and higher mortality rates among women with breast cancer (Sephton, Sapolsky, Kraemer, & Spiegel, 2000). In addition, higher cortisol is associated with lower self reported well-being. (e.g., anger, stress, and anxiety; Adam, 2006; Adam et al., 2006; Adam, Klimes-Dougan, & Gunnar, 2007; Evans et al., 2007; Granger et al., 2006).

According to the exposure-reactivity model, exposure and reactivity to tensions vary by individual differences. Two particularly salient individual differences that are the focus of this study include age and race. Several gerontological theories suggest that interpersonal problems decline with age. Indeed, according to Socioemotional Selectivity Theory, as people age, they become less concerned with acquiring new information and more concerned with maintaining emotionally close relationships due to a decrease in future time perspective (how much time people perceive they have left to live; Carstensen & Charles, 1998). To achieve meaningful interactions, older people are more likely to regulate their emotional reactions (Gross et al., 1997; Carstensen, Isaacowitz, & Charles, 1999). Further, experience and knowledge of social partners may lead to greater acceptance of faults and improvements in relationships and greater resilience to problems in relationships (Hess, Osowksi, & Leclerc, 2005).

These age differences may also vary by race. According to the theory of cumulative disadvantage, because Black Americans experience greater stress across the lifespan, they may be more vulnerable to stress (George & Lynch, 2003; Thoits, 2010; Turner & Avison, 2003). The Strength and Vulnerability Integration model suggests that the experience of chronic stress can dampen or eliminate age related improvements in emotion regulation. (Charles, 2010). In contrast, the concept of resilience suggests that due to the greater stress experienced, Black Americans develop better coping strategies and are thus more resilient to stress (Neighbors et al., 2009; Schwartz & Meyer, 2010). According to the concept of resilience, Blacks may show greater age related improvements in emotion regulation than Whites. We discuss these ideas further below.

Age Differences in Exposure and Reactivity to Interpersonal Tensions

Older people report less exposure to interpersonal tensions than do younger people. For example, older people report that they have fewer problems in their relationships than do younger people (Akiyama, Antonucci, Takakashi, & Langfahl, 2003; Fingerman & Birditt, 2003). Older people also report fewer daily interpersonal tensions than do younger people (Birditt et al., 2005).

Older adults are also less reactive to interpersonal tensions when they do experience them. Older adults appraise interpersonal tensions as less stressful than do younger adults (Birditt et al., 2005). Older adults report more avoidant strategies and less destructive strategies (e.g., arguments) than do younger people in the spousal tie (Carstensen, Gottman, & Levenson, 1995), in the parent child tie (Fingerman, 1998), in response to interpersonal vignettes (Blanchard-Fields et al., 1997; Watson & Blanchard-Fields, 1998), and in retrospective reports of interpersonal problems across family and non-family relationships (Birditt & Fingerman, 2005; Blanchard-Fields et al., 2004). Using the first wave of data used in the present study (NSDE I), Birditt, Fingerman, and Almeida (2005) found that older adults were less likely to report arguments and more likely to report avoidance (i.e., to do nothing) in response to daily tensions than were younger adults.

Interpersonal tensions also appear to have a less detrimental effect on well-being among older adults. Charles and colleagues (2009) found that older adults report less negative affect on days in which they avoided arguments than younger adults (Charles, Piazza, Luong, & Almeida, 2009) but that there were no age differences in negative affect on days in which they reported arguments. The majority of these studies have predominately European American samples. There is little knowledge regarding whether these age differences exist across different ethnic/racial groups.

Age Differences in Interpersonal Tensions by Race

Age differences in tensions and reactivity to tensions may vary by race. Black Americans tend to be exposed to more stressors across the lifespan than White Americans (George & Lynch, 2003; Thoits, 2010; Turner & Avison, 2003). This may lead to two possible scenarios with regard to age differences in tensions.

First, older Black Americans may have increased vulnerability because of a lifetime of repeated exposure and reactivity to stressors (George & Lynch, 2003; Thoits, 2010; Turner & Avison, 2003). This increased vulnerability may lead to fewer age related improvements in emotion regulation. For example, according to research on health disparities, Black Americans remain at increased risk of morbidity and mortality (Williams & Jackson, 2005) and appear vulnerable to a disproportionate rate of stress-related diseases, such as cardiovascular disease (Woods-Giscombe & Lobel, 2008). Research also provides evidence of dysregulation in the functioning of the hypothalamic-pituitary-adrenocorticol (HPA) axis reflected in flatter diurnal cortisol rhythms among Black Americans compared to White Americans (Cohen et al., 2006). Together, these racial disparities may lead to fewer age related improvements in emotion regulation among Black Americans compared to White Americans, particularly in terms of physical and physiological reactivity to all types of tensions. Indeed, according to the strength and vulnerability integration model, age related improvements in emotion regulation are hampered or even eliminated when negative events are unavoidable, when stress is chronic, and when the HPA axis is dysregulated (Charles, 2010). In particular, the greater stress experienced among Black Americans may lead to either no age related improvements or age related decreases in emotion regulation (i.e., greater reactivity with age).

Second, older Black Americans may exhibit more resilience because they have been exposed to more stressors and over the years they have developed effective coping resources (Schwartz & Meyer, 2010). Throughout life, Black Americans are disproportionately exposed to economic stressors, racism, and discrimination (Mujahid et al., 2011; Ross & Mirowsky, 2001; Williams & Mohammed, 2009). Despite chronic stressor exposure, however, Black Americans report similar rates or lower rates of depression compared to Whites (Kessler et al., 2005; Williams et al., 2007). Research suggests that Black Americans engage in coping behaviors that mitigate the psychological consequences of stressors (Mezuket al., 2010; Taylor & Aspinwall, 1996). Therefore, older Black Americans may be particularly adept at coping with the interpersonal tensions they encounter in everyday life.

Further, there may be cultural differences in emotional expression that lead to variations in the age patterns by race. Previous research indicates that African Americans are socialized to value emotional expression, whereas White Americans are socialized to suppress anger and avoid conflict (Kochman, 1981; Mackey & O’Brien, 1998). Davidson (2002) found racial differences, where Black Americans responded to conflict with more confrontational behaviors and greater open expression of negative affect. These findings suggest that, even into later life, Black Americans may still be more likely to actively respond to interpersonal tensions by engaging in arguments, whereas White Americans may prefer to avoid conflicts.

Other Factors Associated with Interpersonal Tensions and Reactivity

This study also controls for factors that may lead to variations in the experience and implications of interpersonal tensions and cortisol including gender, self-rated physical health, socioeconomic status (education, financial status) and health behaviors. Women tend to be more reactive to interpersonal tensions than men (Almeida & Kessler, 1998). Socioeconomic status and self reported health are important predictors of stress and may influence daily well-being and cortisol (Almeida, 2005; Grzywacz, Almeida, Neupert, & Ettner, 2004; Steptoe et al., 2003). We also included several variables that are known to influence cortisol including smoking, wake time, medication use and whether the collection occurred on a weekend day (Almeida et al., 2009; Schlotz, Hellhammer, Schulz, & Stone, 2004; Thorn, Hucklebridge, Evans, & Clow, 2006)

Present Study

In the present study we examined daily accounts of interpersonal tensions and well-being to examine whether there were age differences in tensions among Black and White Americans. Although not a main focus of this study, we first examined whether there were race differences in exposure and reactivity to tensions followed by our focal interest which is an in depth examinations of age differences among Black and White respondents. We hypothesized age differences drawing from the literature reviewed above and further explored age differences by race. Research questions and hypotheses are as follows:

1. Are there race differences in exposure and reactivity to tensions? Because Black Americans experience greater stress and may have developed greater vulnerability to those stressors than White Americans, we predicted that Black Americans would report greater exposure and reactivity to tensions than White Americans.

2. Are there age differences in the number of interpersonal tensions among Black and White Americans? We predicted that older people would report fewer interpersonal tensions than younger people among both Black and White Americans (Birditt et al., 2005). We predicted that due to a lifetime of greater stress, there would be a smaller age difference among Blacks than Whites.

3. Are there age differences in stressor appraisals and coping strategies used (avoidance, arguments) among Black and White Americans? We predicted that older people would report less stress and greater avoidance and fewer arguments than younger people among both White and Black Americans (Birditt et al., 2005). We predicted that there would be fewer age differences among Blacks due to the greater stress over the life course.

4. Are there age differences in the implications of tensions (avoidance, arguments) for daily well-being (positive affect, negative affect, cortisol) among Black and White Americans? Because researchers have found variations in well-being depending on the coping strategy used, we examined variations in reactivity to arguments and avoidance of arguments. We predicted that older people would be less reactive to avoidance days (reporting lower negative affect, higher positive affect, and having a lower CAR, steeper daily decline, and lower cortisol levels) compared to non-tension days than younger people (Charles et al., 2009). We predicted that there would be fewer age differences in reactivity among Blacks due to the greater experience of stress among Blacks than Whites.

Method

Participants

Participants were from the second wave of the National Study of Daily Experiences (Almeida et al., 2009). The NSDE was conducted as part of the Midlife Development in the United States survey (MIDUS). The MIDUS is a national study of initially 7108 Americans in 1995 (aged 25 to 75; response rate of 70%; 87.3% White, 6.1% Black) and another wave of data were collected in 2004–2006 (n = 4963 aged 28 to 84; 90.1% White and 4.6% Black). A comparison of MIDUSI population to the Current Population Survey (CPS, 1995) revealed that African Americans were underrepresented in the MIDUS sample (see MIDUS, 1995). Thus, the MIDUS Milwaukee African American (n = 592, aged 34 to 85) study was conducted in 2005 to increase the sample of African Americans and to examine health disparities. The sample was selected from Milwaukee because the city is highly racially segregated and Blacks in Milwaukee report lower levels of education, lower income, poorer health, and higher unemployment than Blacks nationally (Levine, 2007; Massey & Denton 1993; Farley & Frey 1994). In addition, it was cost prohibitive to conduct oversampling in multiple cities around the U.S and the location facilitated inclusion of African Americans into other MIDUS II satellite studies, including biomarker and neuroscience assessments collected in Madison, WI. Using area probability sampling methods, participants were selected from areas with high concentrations of African Americans (based on the 2000 census). The sampling was stratified by age, gender and SES.

Participants in the MIDUS II and the Milwaukee MIDUS were asked to participate in the NSDE II and received $25 compensation. A total of 1755 participants from the MIDUS II (n = 1696 White; 59 Black) and 180 Black participants from the Milwaukee MIDUS participated in the NSDE II. Thus, 38% of the Whites and 26% of the Blacks from the national sample and 30% of the Milwaukee sample participated in the NSDEII. We removed the 87 participants who were a race other than White or Black American. Participants ranged in age from 34 to 84. See Table 1 for a description of the participants.

Table 1.

Sample Description

Variable	White (n = 1696)	Black (n = 239)
Age (M, SD)	56.65 (12.22)	54.10 (11.78)	t = 3.09, p < .01
Women (%)	56	68	χ2 = 11.87, p < .01
Education (M, SD)	7.41 (2.47)	6.19(2.55)	t = 7.11, p < .01
Financial status (M, SD)	6.62 (2.05)	5.05 (2.54)	t = 10.50, p < .01
Self rated health (M, SD)	3.64 (0.98)	3.09 (1.08)	t= 8.07, p < .01
Number of tensions each day	.24(.26)	.29(.33)	t=−2.57, p <.05
Appraised stress of arguments	1.97(.70)	2.12(.93)	t = −1.82, p = .07
Appraised stress avoidance	1.49(.80)	1.57(.92)	t = −1.10, p =.27
Proportion of days with avoidance	.13(.14)	.15(.20)	t = −2.28, p < .05
Proportion of days with arguments	.07(.12)	.07(.11)	t= 0.60, p = .55
Proportion of days with both avoidance and arguments	.02(.07)	.03(.12)	t=−2.75, p < .01
Positive affect	2.72(.70)	2.70(.83)	t= .58, p=.57
Negative affect	.20(.26)	.30(.39)	t= −5.53, p < .01
Waking cortisol	2.59(.69)	2.11 (.91)	t= 12.12, p < .01
30 minutes after wake cortisol	2.95(.66)	2.55(.86)	t= 10.65, p < .01
Lunch cortisol	1.75 (.67)	1.62(.70)	t= 3.41, p < .01
Bedtime cortisol	.61 (1.00)	1.09(.98)	t= −8.51, p < .01

Note. Education included 12 categories in which 1 = (no school), 6 = (1 to 2 years of college), and 12 = (Ph.D.). Financial situation rated from 0 (the worst possible financial situation) to 10 (the best possible financial situation). Self-rated health rated from 1 (poor) to 5 (excellent). All data were aggregated before calculating the descriptive statistics with the exception of cortisol. The cortisol scores are natural logged transformed and were calculated using the multilevel dataset after omitting the flagged scores.

In order to conduct multivariate analyses examining Black and White differences, we combined the Black national and Milwaukee samples to create a Black participant group (n = 239). The lower percentages of Black respondents who participated in the NSDEII may have been due to the recruitment procedures. A concentrated effort was made to recruit participants who had also participated in an extensive biomarker assessment conducted at a clinic which involved a 2-day visit including an overnight stay (for more information see: Love, Seeman, Weinstein, & Ryff, 2010). Because this time intensive protocol required respondents to take time off work and other family duties, it may have been more challenging for the Black respondents. We acknowledge that this recruitment strategy may limit the representativeness of the sample.

Procedure

The MIDUSII questionnaires involved phone interviews and leave behind questionnaires and the Milwaukee MIDUS was conducted via face-to face interviews and leave behind questionnaires. In the NSDE II, participants completed phone interviews every night for eight consecutive nights. White participants completed an average of 7.47 daily interviews and the Black participants completed an average of 6.74 interviews.

Participants were sent a Home Saliva Collection Kit a week before the study which included 16 salivette collection devices with small absorbent wads and an instruction sheet. Participants were asked to provide salivary samples four times a day: at waking, 30 minutes after waking, before lunch time, and bedtime for four of the diary days (days 2 through 5). After all tubes were ready to send, participants mailed the samples to the MIDUS biological core at the University of Wisconsin where they were stored in a −60C freezer for analysis.

Salivettes were thawed and centrifuged at 3000 rpm. The cortisol was measured with lumincence immunoassays (IBL, Hamburg, Germany); intra assay and inter assay coefficients were below 5 percent. Salivary cortisol can be affected by ph levels and the samples were tested and corrected if outside the normal range (ph 4 to 9). Participants were asked to provide saliva at least an hour after having a meal and to avoid dairy products at least 20 minutes before providing saliva. A total 88% of the White respondents and 73% of the Black respondents provided saliva.

Measures

Predictors

Race and age

Race was coded as 0 (White) or 1 (Black). Participants reported their birth date. Age was included as a continuous variable.

Outcomes

Engagement and avoidance of arguments

Participants were asked two questions each day regarding interpersonal tensions which included: Did you have an argument or disagreement with anyone since we spoke yesterday? And did anything happen that you could have argued about but you decided to let it pass in order to avoid a disagreement? These were coded as 0 (no) 1 (yes). We computed a sum of these two variables to create a number of tensions variable for each day (range 0 to 2). We also created a coping strategy variable and categorized each day into one of four categories: 1 (argument), 2 (avoidance), 3 (argue and avoid), or 4 (no interpersonal tension on that day). Having no interpersonal tensions on that day was the reference category.

Participants then rated how stressful the experience of arguments and the experience of avoiding arguments were from 1 (very stressful) to 4 (not at all stressful) which we recoded so that higher scores reflected greater stress. The two stressor appraisal variables were examined separately.

Self-reported affect

Participants completed 13 negative affect and 13 positive affect items derived from the Positive and Negative Affect Schedule (PANAS) and the Non Specific Psychological Distress Scale (Kessler et al., 2002; Watson, Clark, & Tellegen, 1988). Negative affect included items such as restless or fidgety, nervous, hopeless, ashamed, upset, angry, and frustrated. Positive items included emotions such as in good spirits, cheerful, extremely happy, calm and peaceful, active, and confident. Participants rated each item from 0 (none of the time) to 4 (all of the time). The negative and positive affect items were averaged to create two separate scales for each day. Alphas ranged from .83 to .85 across days for negative affect and .92 to .95 across days for positive affect.

Cortisol

Participants provided saliva at four time points: waking, 30 minutes after waking, before lunch and bedtime. Of the people who provided saliva samples 98.8 % provided samples on all four of the days. The cortisol scores were transformed with the natural log transformation. Days in which the cortisol data had errors were not included in the analysis. Errors included days in which 30 minute samples were provided at the wrong time (either less than 15 minutes or more than 60 minutes after waking), days in which participants were awake too long (more than 20 hours) or not awake long enough (less than 12 hours), days in which samples were above 120 nmol/l, days in which participants lunch scores were higher than their 30 minute scores by 10 nmol/l, days in which participants woke up before 4AM or after 12 noon, participants who did not follow instructions and provided saliva samples on non saliva sampling days, and days in which participants did not record times of saliva sample collection. A total of 1903 days received error flags which included 27% of the daily diary days that included cortisol (for reliability and validity of this protocol see Almeida et al., 2009).

Because omitting flagged data may have removed individuals who experienced more stress, we examined whether there were variations in cortisol between the flagged and non flagged individuals. We found that flagged individuals had lower waking and 30 minute cortisol and higher lunch and bedtime cortisol than non-flagged individuals. Thus, the data may be biased by excluding individuals with higher or lower cortisol however it is impossible to know whether the variations are due to errors in collection and/or greater stress.

Covariates

Covariates included gender, socioeconomic status (education, financial status), self-rated health, smoking, medication use, wake time, and whether the cortisol was collected on a weekend day. Gender was coded as 0 (man) or 1 (woman). Education included 12 categories in which 1 = (no school), 6 = (1 to 2 years of college), and 12 = (Ph.D.). Due to missing data regarding income, we used a financial status variable in which participants rated their financial situation from 0 (the worst possible financial situation) to 10 (the best possible financial situation). Self-rated health included how well the participant rated their overall health from 1 (poor) to 5 (excellent). Smoking included a combination of two variables: the number of cigarettes smoked during the eight day diary period and whether the participant reported being a regular smoker (0 = non smoker, 1 = smoker). We also included whether the participant was taking any of the following medications: steroid inhaler, steroid medications, medications including cortisone, birth control pills, other hormones, and anti depressants and anxiety medications (0 = no medication, 1 = at least one medication). Wake time included the time the first cortisol measurement was taken in military time. Weekend was coded as 0 (Monday thru Friday) or 1 (Saturday or Sunday).

Analysis Strategy

First, to describe the samples we examined whether all predictors, outcomes, and covariates varied by race (White vs. Black). We used t-tests to examine the continuous variables and chi-squares to examine the categorical variables. We then calculated correlations among the outcome variables (number of tensions, appraisals, coping strategies, affect, and cortisol) separately by race.

Two types of multilevel models were estimated using SAS PROC MIXED to examine race and age differences in exposure and reactivity to interpersonal tensions. Two level models were estimated to examine the variables that varied by day but not within day including number of tensions, stress appraisals coping strategy type, positive affect, and negative affect. Participants were the upper level and the days were the lower level. Models included a random intercept and an unstructured covariance matrix. Models examining race differences were conducted in two steps: 1) with race as the predictor, and 2) with race and the covariates including age, gender, education, self-rated health and financial status.. Analyses examining age differences were conducted separately for each racial group (White, Black) with age as a predictor and the covariates included gender, education, self-rated health and income.

Three level piecewise multilevel models were estimated to assess cortisol in which the lowest level referred to the cortisol measurement within day, the second level referred to the day, and the upper level referred to the participant (Almeida et al., 2009; Stawski et al., 2011). These piecewise models captured the within day patterns of cortisol with two predictors (aka pieces) that represented the cortisol awakening response (CAR) and the daily decline (DEC) centered on the 30 minute collection. Several models were estimated to determine which model had the best fit including random intercepts and pieces. The model with the best fit included a random intercept and two random slopes for CAR and DEC between participants and a random intercept and random daily decline slope within participant across days. To examine whether avoidance and arguments predicted variations in these scores we entered interactions between CAR and DEC and coping strategy type. To examine whether the associations between tensions and cortisol varied by age we entered three-way interactions among the CAR or DEC, age, and coping strategy type. Cortisol analyses controlled for between person variables including smoking, medicine use, gender, self-rated health, education and day level variables including wake time and whether the collection occurred on a weekend.

For each model estimated we calculated pseudo R²s in order to estimate the proportion of variance accounted for by the predictors. To do this we examined associations between the estimated predicted values and the actual values of the outcome variables using the method proposed by Singer and Willett (2003). It is important note, however that there is disagreement in the literature regarding the best methods for estimating R² in multilevel models. Thus, these statistics should be interpreted with caution.

Results

Description of the Data

The Black sample reported lower education levels, poorer financial status, lower self-rated health and they were younger than the White sample (Table 1). The Black sample also reported more interpersonal tensions, were more likely to report using avoidance and both avoidance and arguments, and reported greater negative affect than the White sample. Black individuals also had lower cortisol values at waking, 30 minutes after waking, and lunch time and higher cortisol at bedtime than did White individuals. These variations in cortisol may be due to the effects of chronic stress on the HPA axis. Overall, Black respondents reported greater disadvantage and demonstrated evidence of greater stress than did White respondents.

Similar correlations emerged among the variables for Black and White respondents (Table 2). Some of the higher correlations revealed that participants who appraised tensions as more stressful and who reported a greater number of tensions also reported greater negative and less positive affect. Positive and negative affect were negatively associated but the moderate correlation shows that these are distinct constructs. In addition, there were few correlations between cortisol and affect with most correlations revealing that greater stress appraisals, tensions, and greater negative affect were associated with lower cortisol levels. These seemingly contradictory findings are most likely due to examining individual cortisol scores rather than the diurnal rhythms.

Table 2.

Correlations among Number of Tensions, Appraisals, Coping Strategies, Self-reported Affect, and Cortisol

	Number of tensions	Appraised Stress of Argument	Appraised Stress of Avoidance	Argument	Avoid	Both	Negative Affect	Positive Affect	Cortisol- Wake	Cortisol- 30 mins after wake	Cortisol- Lunch	Cortisol- Bedtime
Number of tensions	1	−0.02	0.06	0.40**	0.60**	0.58**	0.33**	−0.19**	−0.06	−0.10	−0.01	−0.06
Appraised stress of Argument	0.04	1	0.67**	0.02	−	−0.02	0.43**	−0.37**	−0.12	0.02	−0.30	−0.18
Appraised stress of Avoidance	0.04	0.28**	1	−	−0.06	0.06	0.35**	−0.35**	−0.25	−0.39**	−0.17	−0.17
Argument	0.46**	−0.04	−	1	− 0.11**	−0.05	0.19**	−0.12**	−0.06	−0.05	−0.04	0.01
Avoidance	0.63**	−	−0.04	−0.10**	1	− 0.07**	0.12**	−0.08**	−0.03	−0.01	0.11	−0.03
Both	0.52**	0.04	0.04	−0.04**	− 0.05**	1	0.23**	−0.12**	−0.02	−0.12*	−0.10	−0.07
Negative Affect	0.32**	0.27**	0.37**	0.20**	0.16**	0.17**	1	−0.51**	−0.09	−0.18**	−0.14*	−0.11
Positive Affect	−0.17**	−0.26**	−0.32**	−0.10**	− 0.08**	− 0.09**	−0.49**	1	0.07	0.14*	0.04	0.04
Cortisol-Wake	−0.04*	−0.03	−0.01	−0.03*	−0.03	0.01	−0.02	0.01	1	0.54**	0.35**	0.22**
Cortisol- 30 mins after wake	0.01	0.01	−0.01	−0.02	0.01	0.02	−0.01	0.02	0.52**	1	0.41**	0.25**
Cortisol- Lunch	−0.01	−0.05	−0.10*	−0.02	−0.01	0.02	−0.01	0.03	0.29**	0.38**	1	0.38**
Cortisol- Bedtime	−0.04*	0.01	0.04	−0.04*	−0.01	−0.02	0.03	0.01	0.16**	0.19**	0.31**	1

Note. Correlations for Whites are below the diagonal and correlations for Blacks are above the diagonal. Cortisol values are the natural logged transformed and omit the flagged scores.

Research Question 1: Race Differences in Exposure and Reactivity to Tensions

First, we conducted a series of multilevel models examining race differences in the number of interpersonal tensions, appraisals, coping strategies, and well-being. In the interest of space, these findings are not tabled. The models without covariates revealed that Black respondents were more likely to report using both coping strategy types (avoidance of arguments and arguments) in the same day, they reported greater negative affect, and they had lower cortisol scores and a flatter decline over the course of the day than White respondents. After controlling for socioeconomic status, health, gender, and age, there were no significant differences between Black and White participants with the exception of cortisol. Black respondents had lower cortisol and flatter daily declines in cortisol than White respondents.

We also conducted analyses to examine whether there were race differences in the associations between tensions and well-being which revealed racial similarities in associations between tensions and well-being tensions with a few exceptions. The models without covariates and with covariates revealed that Black participants reported greater negative affect, higher cortisol, and a flatter CAR on days when they reported both types of coping strategies (avoidance, arguments) than did White participants.

Research Question 2: Age Differences in the Number of Interpersonal Tensions by Race

Multilevel models were estimated to examine the association between age and the number of interpersonal tensions reported each day separately for each racial group (Table 3). As predicted, older people reported fewer interpersonal tensions than younger people among both Black and White respondents. Unlike we hypothesized, the age differences appeared to be similar if not greater among Black respondents than among White respondents.

Table 3.

Multilevel Models Examining Number of Tensions by Age Separately for Whites and Blacks

	Number of Tensions each Day
Predictor	White (n = 1615) B (SE)	Black (n = 229) B (SE)
Intercept	0.432 (0.045)***	0.521 (0.125)***
Age	−0.004 (0.000)***	−0.005(0.001)**
Covariates
Gender	0.030 (0.012)**	0.033 (0.037)
Education	0.012 (0.002)***	0.015 (0.007)*
Financial situation	−0.012 (0.003)***	−0.028 (0.007)***
Self-rated health	−0.006 (0.006)	0.001 (0.017)
Pseudo R2	.02	.04

^*p < 0.05,

^**p < 0.01,

^***p < 0.001.

Research Question 3: Age Differences in Stressor Appraisals and Interpersonal Coping by Race

Multilevel models were conducted to examine whether appraised stress of arguments and avoidance varied by age among White and Black participants (Table 4). As we hypothesized, older people reported that avoidance was less stressful than younger participants among both White and Black respondents. Arguments were also rated as less stressful with age among Black respondents but not among White respondents. This finding was not consistent with our prediction that Black respondents would show fewer age differences than White respondents.

Table 4.

Appraised Stress of Arguments and Avoidance of Arguments as a Function of Age among Whites and Blacks

	Appraised stress of argument		Appraised stress of avoidance
Predictor	White (n = 1615) B (SE)	Black (n = 229) B (SE)	White (n = 1615) B (SE)	Black (n = 229) B (SE)
Intercept	2.071 (0.201)**	2.598 (0.626)**	1.957 (0.177)**	2.690 (0.468)**
Age	−0.004 (0.002)	−0.021 (0.008)*	−0.009 (0.002)**	−0.017 (0.006)**
Covariates
Gender	0.252 (0.053)**	0.606 (0.209)**	0.384 (0.046)**	0.331 (0.149)*
Education	−0.008 (0.011)	−0.000 (0.039)	0.015 (0.010)	0.020 (0.029)
Financial situation	−0.009 (0.014)	−0.042 (0.036)	−0.035 (0.012)**	−0.060 (0.029)*
Self-rated health	−0.039 (0.028)	−0.097 (0.090)	−0.119 (0.026)***	−0.195 (0.067)**
Pseudo R2	.04	.17	.09	.12

^*p < 0.05, ** p < 0.01

Multilevel models were estimated to examine the association between age and each coping strategy type separately for Blacks and Whites (Table 5). Among White participants, as we hypothesized, older people were less likely to report arguments and more likely to report avoidance than younger people. There were no associations among age and any of the three coping strategy types (i.e., arguments, avoidance, and both arguments and avoidance) among the Black participants. Thus, this finding was consistent with our hypothesis that there would be fewer age differences among Black respondents.

Table 5.

Multilevel Models Examining Coping Strategies by Age in each Racial Group among Individuals who Experienced Interpersonal Tensions

	Argument		Avoidance		Both argument and avoidance
Predictor	White (n = 1174) B (SE)	Black (n = 163) B (SE)	White (n = 1174) B (SE)	Black (n = 163) B (SE)	White (n = 1174) B (SE)	Black (n = 163) B (SE)
Intercept	0.082 (0.346)	−0.598 (0.931)	−0.545 (0.333)	−0.230 (0.893)	−2.523 (0.700)**	−2.269 (1.466)
Age	−0.012 (0.004)**	−0.002 (0.012)	0.017 (0.004)**	0.010 (0.012)	−0.024 (0.008)**	−0.019 (0.020)
Covariates
Gender	−0.101 (0.091)	0.415 (0.306)	0.005 (0.088)	−0.255 (0.289)	0.291 (0.185)	−0.197 (0.472)
Education	0.022 (0.019)	−0.134 (0.059)*	−0.040 (0.018)*	0.052 (0.056)	0.073 (0.038)	0.181 (0.092)
Financial situation	0.004 (0.023)	−0.035 (0.054)	0.001 (0.022)	0.111 (0.053)*	−0.022 (0.008)**	−0.210 (0.090)*
Self-rated health	−0.052 (0.050)	−0.017 (0.131)	0.076 (0.048)	−0.097 (0.126)	−0.111 (0.098)	0.216 (0.206)
Pseudo R2	.01	.03	.01	.04	.01	.05

^*p < 0.05,

^**p < 0.01

Research Question 4: Age Differences in the Implications of Tensions for Daily Well-being by Race

Next, we conducted a series of multilevel models to examine whether there were age differences in the implications of tensions for well-being among White and Black respondents (Table 6). We conducted analyses examining whether negative affect, positive affect, and cortisol varied by age, coping strategy type (engagement in arguments, avoidance of arguments), and age X coping strategy type separately for each racial group. Overall, there were fewer age differences among Black respondents than White respondents.

Table 6.

Multilevel Models Predicting Negative and Positive Affect as a Function of Tension and Age in each Racial Group

	Negative Affect		Positive Affect
Predictor	White (n = 1615) B (SE)	Black (n = 229) B (SE)	White (n = 1615) B (SE)	Black (n = 229) B (SE)
Intercept	0.483 (0.043)**	0.889 (0.161)***	1.378 (0.125)***	1.132 (0.358)**
Argument	0.283 (0.037)**	0.456 (0.140)**	−0.070 (0.068)	−0.628 (0.247)*
Avoidance	0.208 (0.030)**	0.243 (0.107)*	−0.099 (0.056)	−0.373 (0.189)
Both argument and avoidance	0.607 (0.072)**	1.151 (0.244)**	0.022 (0.133)	−0.362 (0.432)
No interpersonal tension
Age	−0.002 (0.000)**	−0.004 (0.002)*	0.009 (0.001)***	0.015 (0.004)**
Age X Argument	−0.001 (0.001)†	−0.004 (0.003)	−0.002 (0.001)†	0.007 (0.005)
Age X Avoidance	−0.002 (0.001)**	−0.002 (0.002)	0.000 (0.001)	0.005 (0.004)
Age X Both	−0.006 (0.001)**	−0.014 (0.005)**	− 0.004 (0.003)†	0.002 (0.009)
Age X none
Covariates
Gender	0.029 (0.011)**	−0.007 (0.046)	0.008 (0.032)	0.050 (0.103)
Education	0.005 (0.002)*	−0.008 (0.009)	−0.036 (0.007)***	−0.025 (0.020)
Financial situation	−0.020 (0.003)**	−0.032 (0.009)**	0.075 (0.008)***	0.092 (0.020)**
Self-rated health	−0.020 (0.003)**	−0.070 (0.021)**	0.172 (0.017)***	0.144 (0.047)**
Pseudo R2	.16	.21	.15	.19

^†p < 0.10,

^*p < 0.05,

^**p < 0.01

Negative affect

There was a significant interaction between age and having both tensions among both Blacks and Whites when predicting negative affect which indicates that older people reported lower negative affect in response to having both types of tensions on the same day compared to younger people. In addition, among the White sample, there were significant interactions between age and arguments and age and avoidance which indicate that older people reported lower negative affect on days in which they reported engaging in arguments and days in which they avoided arguments than did younger people.

Positive affect

When predicting positive affect there were interactions that approached significance among White and not Black respondents. Consistent with our hypothesis, older White individuals reported higher positive affect on days in which they reported arguments and days in which they reported both tensions compared days in which they had no tensions than did younger participants who reported lower positive affect

Cortisol

The analyses of cortisol revealed age differences in reactivity among Whites and not Blacks (Table 7). Older White people reported a greater cortisol awakening response than younger people on days in which they avoided tensions compared to days in which they had no tension. Thus, it would appear that older people are more physiologically reactive on days in which they avoid tensions compared to days in which they have no tension. The increased CAR may represent a boost in energy to prepare for the problems ahead and/or it may represent increased stress.

Table 7.

Multilevel Models Predicting Cortisol Patterns by Tensions and Age in each Racial Group

	White (n = 1287)			Black (n = 109)
Predictor	Intercept B (SE)	CAR B (SE)	DEC B (SE)	Intercept B (SE)	CAR B (SE)	DEC B (SE)
Intercept	2.809 (0.113)**	0.281 (0.119)*	−0.298 (0.008)**	2.323 (0.411)**	0.097 (0.506)	−0.194 (0.032)**
Argument	−0.020 (0.147)	−0.438 (0.351)	−0.005 (0.019)	−0.884 (0.622)	−0.524 (1.359)	0.121 (0.074)
Avoidance	−0.155 (0.113)	−0.473 (0.279)†	0.013 (0.015)	0.169 (0.444)	−1.015 (1.087)	−0.085 (0.051)†
Both Arg. And Avoid	0.244 (0.271)	0.949 (0.679)	−0.051 (0.036)	−1.222 (0.984)	−2.405 (1.984)	0.014 (0.110)
No interpersonal tension
Age	0.006 (0.001)**	−0.004 (0.002)*	0.001 (0.000)**	0.005 (0.005)	0.011 (0.009)	−0.000 (0.001)
Age X Arg.	0.001 (0.003)	−0.005 (0.006)	0.000 (0.000)	0.015 (0.011)	0.008 (0.024)	−0.002 (0.001)
Age X Avoid	0.003 (0.002)	0.010 (0.005)*	−0.000 (0.000)	−0.002 (0.008)	0.025 (0.021)	0.001 (0.001)
Age X Both	−0.002 (0.005)	−0.012 (0.013)	0.001 (0.001)	0.013 (0.016)	0.025 (0.033)	−0.000 (0.002)
Age X None
Covariates
Waking Time	−0.052 (0.007)***			−0.031 (0.023)
Smoker	0.065 (0.036)			0.149 (0.128)
Medicine User	−0.050 (0.026)*			0.136 (0.117)
Weekend	−0.060 (0.013)**			0.058 (0.046)
Gender	−0.071 (0.025)**			−0.068 (0.105)
Self-rated health	0.056 (0.014)**			0.023 (0.048)
Education	0.001 (0.005)			0.007 (0.020)
Financial situation	−0.003 (0.007)			0.010 (0.023)
Pseudo R2	.87			.77

^†p < 0.10,

^*p < 0.05,

^**p < 0.01

Discussion

This study examined age differences in exposure and reactivity to daily interpersonal tensions among Black and White individuals. The purpose of the study was to examine whether there were age differences in the number of interpersonal tensions, appraised stress of those tensions, coping strategies used, and links between tensions and well-being among Black and White individuals. This study revealed that age and race are important components of the interpersonal exposure reactivity model. Among both Black and White Americans older people reported fewer tensions and less reactivity to those tensions than younger people. There were variations, however, in the specific age differences by race that may reflect differences in life experiences and cultural norms for emotional expression.

Race Differences in Exposure and Reactivity to Tensions

Black respondents reported greater disadvantage than did White respondents, reporting lower socioeconomic status and poorer health. Description of the data also revealed that Black respondents reported greater number of tensions each day, were more likely to use avoidance and both types of coping strategies on the same day, reported greater negative affect, and had either lower or higher cortisol levels did White respondents. The multilevel models revealed however that some of the race differences were eliminated especially after controlling for indicators of disadvantage. In particular, there were no longer race differences in coping strategies and negative affect. The stress and anxiety associated with economic disadvantage and poor health may create a context ripe for interpersonal tensions and may cause feelings of greater negative affect (Gryzywazc et al., 2004; Almeida, Neupert, Banks, & Serido, 2005).

There were race differences in cortisol and race differences in the links between tensions and well-being that remained after including covariates. In particular, Black participants reported lower cortisol and a flatter daily decline in cortisol over the course of the day. This finding is consistent with previous work indicating that Black participants have higher cortisol levels in the evenings even after controlling for sociodemographic characteristics (Cohen et al., 2006). Bedtime cortisol levels may be the most sensitive to chronic stress due to reduced ability to unwind after experiencing high levels of stress. Researchers have postulated race differences in cortisol may have a genetic or heritable component (Cohen et al., 2006). However, there are most likely many other factors accounting for race differences in cortisol that were not considered in the present study such as early life events, chronic environmental and economic stress, and health behaviors, to name a few.

Further, Black respondents were more reactive (greater negative affect, higher cortisol) to days in which they reported both types of coping strategies (arguments, avoidance of arguments) than White respondents and these effects remained after controlling for the indicators of disadvantage. This finding indicated that as hypothesized, the Black individuals may be more vulnerable to stress than Whites perhaps because of the experience of chronic stress.

Age Differences in Exposure to Tensions by Race

Older people reported fewer tensions among both Black and White Americans. This is consistent with gerontological theory and previous research using predominately White samples (Birditt et al., 2005; Carstensen et al., 1999). Older people may report fewer tensions because of age related improvements in emotion regulation. For example, older people experience less anger in response to problems with social partners than do younger people (Birditt & Fingerman, 2003). Research also shows that older people pay less attention to and are less likely to remember negative information than are younger people (Carstensen, 2006).

However, these age differences may also be due to changes in social roles that lead to less exposure to tensions among older people. For example, older people are less likely to be employed full time which may reduce the number of interactions that occur on a daily basis. It is also possible that age differences reflect cohort differences rather than developmental changes. The older generations may have been exposed to more stress in general or feel it is less appropriate to describe irritations in their relationships than younger cohorts due to the historical experiences of their cohort such as the Great Depression.

Age Differences in Appraisals and Coping Strategies by Race

Appraisals varied by age among White and Black participants. Older people reported that avoidance was less stressful than younger participants among both White and Black respondents. Arguments were also rated as less stressful with age among Black respondents but not among White respondents. This finding may be due to greater resilience among Black Americans as they age. Evidence of this resilience is found in studies of chronic stress, where Black caregivers appraise the stressors associated with family caregiving more favorably compared to White caregivers (Gallagher-Thompson, 2006; Pinquart & Sorensen, 2005). Our findings suggest this resilience also characterizes Black Americans’ appraisal of their daily experiences. Throughout life, Black Americans are exposed to more stress across the lifespan and thus may view tensions as less stressful as they age, especially in comparison to the other stressors older Black Americans face, such as financial concerns or chronic health conditions (Mujahid et al., 2011; Ross & Mirowsky, 2011; Williams & Mohammed, 2009).

Gerontological research consistently notes that when older people do experience tensions, they are more likely to avoid tensions than to engage in arguments (Birditt et al., 2005; Blanchard-Fields et al., 1997). This study indicates that these age patterns exist among White Americans and not among Black Americans. The variations in age differences by race do not appear to be due to race differences in the coping strategies reported. We found no differences in the frequency with which White and Black Americans reported avoidance or engagement in arguments. There may be differences between racial groups in the ways in which their relationships and emotion regulation improve with age. In addition, the measures we employed may not have captured the types of strategies that older Black Americans use.

However, it is also possible that this finding reflects racial differences in stress exposure and vulnerability (George & Lynch, 2003; Thoits, 2010; Turner & Avison, 2003). Although older adults experience fewer problems and rate experiences as less stressful among the Black participants, they may be less able to change their behavioral reactions to those tensions. Indeed, the Strength and Vulnerability Integration model suggests that age related improvements in emotion regulation are dampened when individuals have experienced chronic stress and stress is unavoidable (Charles, 2010). There may also be cultural differences in emotion regulation. For example, emotional expression is more valued among Black Americans than White Americans which may lead to fewer age related changes in coping strategies among Black Americans (Kochman, 1981; Mackey & O’Brien, 1998).

Age Differences in Reactivity by Race

Consistent with the exposure reactivity model and gerontological theories, we found that older people tend to be less reactive to tensions than younger people (Birditt et al., 2005; Charles et al., 2009). There were variations in the specific age differences by race. Overall we found fewer age differences in reactivity among Blacks than Whites.

When examining self reported affect, among the White American sample, older people reported lower negative affect on days in which they experienced avoidance of arguments, engagement in arguments, or both arguments and avoidance of arguments compared to younger people. In addition, older people reported higher positive affect on days in which they reported arguments or both arguments and avoidance of arguments than did younger people. This is somewhat consistent with work on the first wave of the NSDE. Charles and colleagues (2009) found that older people were less reactive to avoidance of arguments than younger adults. However they did not find age differences in the reactivity to arguments. It is possible that we found age differences in response to both tensions because we had a larger age range. Indeed our previous work shows that individuals over 80 have distinct advantages with regard to interpersonal tensions, reporting less anger and more avoidance (Birditt & Fingerman, 2003; 2005).

Among the Black sample, older people reported lower negative affect on days in which they reported both arguing and avoidance of arguments compared to younger people. There were no age differences in reactivity when examining positive affect. Although older Black Americans may appraise arguments as less stressful relative to White Americans, our findings suggest these tensions are still stressful enough to elicit an emotional response from older Black Americans. The few age differences in reactivity observed among Black Americans suggests that then may not benefit from age-related improvements in emotion regulation. Perhaps a lifetime of repeated stressor exposure and reactivity may deplete the resources older Black Americans have to cope with interpersonal tensions (Charles, 2010).

The cortisol analyses also revealed variations in age in cortisol reactivity among the White sample and not the Black sample. Among the White sample, older people reported a greater cortisol awakening response than younger people on days in which they avoided tensions compared to days in which they had no tension. The increased CAR can represent a positive boost for the course of the day or it may indicate increased stress on days in which conflict will be avoided (Chida & Steptoe, 2009; Fries et al., 2009). If it is an indication of increased stress, this finding may indicate a disconnect between what older adults are reporting in the self- reported affect measure (lower negative affect) and what they are experiencing physiologically. It is also important to note here that the reports of conflict avoidance most likely occur after the CAR in the morning. Thus, it may be that the increased CAR leads to more conflict avoidance. However, it may also mean that there is an ongoing problem in an individual’s relationship that they are preparing to deal with.

Limitations and Future Research Directions

There are several directions to pursue in future research. Because the data are cross sectional it is unclear whether the effects are due to cohort differences and/or aging. In addition, there is most likely a bidirectional relationship between interpersonal tensions, affect and cortisol. For example the CAR may predict variations in the experience of interpersonal tensions. Further research should be conducted to examine these associations over time. Although there were more Black Americans in the MIDUS II than in MIDUS I, future studies of daily interpersonal tensions should include a larger and more representative national sample of Black Americans. It is unclear how applicable these results are to Black Americans in general. Future research should also consider more complex measurements of coping strategies. The coping strategies were limited to either the avoidance of arguments or engagement in arguments. In addition, it is unclear what participants mean by the avoidance of arguments or engagement in arguments. For example, avoidance may include a variety of behaviors (e.g., keeping quiet, cognitive reappraisal, drinking alcohol) just as arguments may include a variety of behaviors (e.g., heated discussion, screaming, insults). A more nuanced measure of coping strategies may provide information on age differences in coping among Black Americans. In addition, we know little about race differences in the ideals with regard to coping strategies. Future studies should examine what participants did as well as what they believe they should have done in response to interpersonal tensions which would allow for an examination of cultural and age differences in beliefs about appropriate emotional expression. This type of study would provide another necessary layer of information for understanding variations in the experience of interpersonal tensions.

Overall, this study shows that age and race are important components of the interpersonal exposure reactivity model. The research to date has often focused on White Americans. This study shows that while there may be age related improvements among both Black and White Americans, the specific nature of those improvements may vary by race. We hope that this study leads to more research in this area with regard to the examination of age differences among individuals from different cultural and ethnic groups.