The Influence of Divorce and Parents’ Communication Skills on Adolescents’ and Young Adults’ Stress Reactivity and Recovery

Abstract

This study examined whether the impact of parents’ marital status (divorced/married) on children’s hypothalamic-pituitary-adrenocortical (HPA) axis (measured through salivary cortisol) and sympathic nervous system (SNS; measured through salivary alpha-amylase or sAA) response and recovery patterns was moderated by parents’ communication skills and the age of the child. One hundred eighteen parent–child (ages 15-22) dyads talked about something stressful related to the parents’ relationship. Children who thought their parents were more communicatively incompetent had higher pre-interaction sAA levels. In addition, children whose parents were communicatively skilled (i.e., socially supportive, communicatively competent, children felt less caught between them) were able to down regulate quickly after the discussion, regardless of the marital status of the parent and the age of the child. The results for sAA, and somewhat for cortisol, revealed that parents’ marital status and the age of the child were important in determining differences in children’s physiological response and recovery patterns only when parents were less communicatively skilled.

Keywords

HPA SNS salivary cortisol salivary alpha-amylase (sAA)interparental conflict social support communication competence parent-child relationships adolescents young adults

There is no doubt that divorce is a stressful experience and that it can have negative consequences for children (see Amato, 2001). Research, however, suggests that children vary considerably in their responses to divorce, with some experiencing few effects, others suffering from long-term personal and relational health consequences, and still others benefiting from being removed from a tumultuous family environment (Hetherington, 1999). The most important factor influencing children’s postdivorce adjustment is parents’ communication. In fact, interparental conflict is often more important than divorce in predicting children’s physical and mental health (Amato & Sobolewski, 2001; Booth & Amato, 2001; Jekielek, 1998).

While there is a wealth of research on the impact of divorce on children, only recently have researchers begun to investigate the impact of divorce on children physiologically. Incorporating biosocial markers into the study of divorce is important because how children respond physiologically to their family environment can affect their personal and relational health (Granger et al., 1998). Children often avoid talking about their parents’ discordant relationship and the way their parents communicate to maintain harmony in the family (Afifi, Afifi, Morse & Hamrick, 2008), but their bodies might be churning with anxiety and stress. Children’s self-reports of how they are adapting to a divorce may be positively biased because they want others and/or themselves to believe they are coping effectively. Physiological measures, when combined with self-reports, can provide a more accurate and holistic assessment of children’s responses to divorce and parents’ communication patterns (Gordis, Granger, Susman, & Trickett, 2008). Even though some of these physiological indicators are relatively new (e.g., salivary alpha amylase; oxytocin receptor genes), the more researchers examine them and combine them with self-reports and other physiological measures that have a long history of reliable results (e.g., cortisol), the better they will be able to advance the understanding of the human stress response.

A significant amount of research has been focused on how parents’ communication skills contribute to the dysregulation (e.g., abnormally high or low basal cortisol levels, inability to recover from an acute stress task, irregular diurnal rhythms) of children’s hypothalamic-pituitary-adrenocortical (HPA) axis through the activation of the hormone cortisol. Unfortunately, this dysregulation can negatively affect children’s health (Davies, Sturge-Apple, Cicchetti, & Cummings, 2007; Granger et al., 1998). A handful of studies have also shown that children whose parents are divorced are at greater risk for dysregulated HPA activity compared with children whose parents are married (e.g., Bloch, Peleg, Koren, Aner, & Klein, 2007; Meinlschmidt & Heim, 2005). These studies on divorce, however, tend to compare the basal cortisol levels of children from divorced and nondivorced families without examining thoroughly how parental communication patterns, as well as other contextual factors such as the age of the child, affect children’s physiological responses.

The present study uses a naturally occurring, stress-inducing discussion between parents and children from divorced and nondivorced families to provide a more complete understanding of how parents’ marital status (divorce/married) interacts with their communication skills and the age of the child (i.e., adolescents vs. young adults) to influence children’s physiological responses. We argue that when children believe their parents are communicatively skilled (i.e., more socially supportive, communicatively competent, feel less caught between them), they should respond to a stressful discussion with their parent with a roughly quadratic physiological response and recovery (i.e., a slight physiological reaction to the stress task and a steady trajectory downward to baseline) regardless of whether the parents are divorced. High levels of parental support and responsiveness have been shown to buffer against much of the impact of childhood adversity on children’s HPA systems (Evans, Kim, Ting, Tesher, & Shannis, 2007). From an attachment theory perspective (Bowlby, 1982), children whose parents are communicatively skilled should recover quickly from a stressful discussion with their parent because the parents’ communication provides a sense of security.

When parents lack these essential communication skills, divorce is likely to have a much greater impact on children’s physiological responses, which is probably also magnified by the age of the child. When parents are less skilled communicatively, the “foundation” of the child’s secure base might already be uncertain and other circumstances of the child’s environment may then accelerate the child’s physiological stress responses. According to the emotional security hypothesis (Davies & Cummings, 1994), when children are immersed in an environment of negative parent affect and interparental conflict, it threatens children’s emotional security, resulting in over- or under-active physiological responses (Davies et al., 2007). Children of divorced and married parents most likely experience an inability to recover from a stressful interaction when their parents lack certain communication skills because their emotional security is threatened. Children whose parents are still married who lack communication skills, however, may actually have more difficulty recovering from a stressful interaction than children of divorced parents because they cannot escape their parents’ communication skill deficiencies (Amato & Afifi, 2006). This situation may be especially difficult for adolescents, because they live with their parents and are surrounded by this communicative environment, while simultaneously still developing their attachments with their parents and their own identities as adults (see Buchanan, Maccoby, & Dornbusch, 1991, 1996). Nonetheless, young adults who live out of the home (e.g., college) may still experience their parents’ communication patterns as stressful, particularly if their parents disclose more adult issues to them (Afifi & Schrodt, 2003). Such intriguing findings regarding adolescents (defined here as ages 12-17) and young adults (18-22) highlight the need to further distinguish between these two populations in this body of work.

Given that the research on divorce and children’s physiological responses has only examined the HPA system and cortisol, it would also be useful to examine whether similar patterns emerge in the sympathic nervous system (SNS). As a result, we use a multi-systems approach to explore the influence of parents’ communication skills, marital status, and the age of the child on children’s HPA axis, as measured through salivary cortisol, and SNS, as measured through salivary alpha-amylase (sAA). The HPA and SNS systems are responsive to communication patterns, but in slightly different ways (Gordis et al., 2008; Pendry & Adam, 2007). Cortisol has long been a valid and reliable marker of the HPA system (Byrd-Craven, Auer, Granger, & Massey, 2012). When faced with an acute stressor or threat, the HPA axis releases the hormone cortisol to combat the stressor (McEwen, 1998). A healthy cortisol response to an acute stress task is characterized by a rise in cortisol approximately 15 minutes after experiencing a stressor followed by quick and steady down regulation once the threat of the stressor has passed (approximately 30 minutes; Gordis et al., 2006; McEwen, 2001; Laurent & Powers, 2006). Too much or too little cortisol in response to an acute stressor, an irregular pattern of ups and downs, and/or the inability to recover from the stress task can indicate that the body is experiencing difficulty coping with the stressor.

Recently, investigators have incorporated sAA as a surrogate marker of the SNS system among adolescents (Granger, Kivlighan, El-Sheikh, Gordis, & Stroud, 2008). sAA is an enzyme produced by the salivary glands that aids in digestion and is an indicator of SNS activity (see Nater & Rohleder, 2009). Whereas cortisol has a delayed response to stress, sAA rises and recovers more quickly in response to changing environmental demands (Gordis et al. 2006). According to Engert et al. (2011), sAA reaches its peak approximately 13.5 minutes earlier than cortisol. SNS reactivity reflects the body’s “flight or fight” response and is believed to be most reactive for threats that are perceived as controllable or something one can fight against (Henry, 1992). In contrast, cortisol is often considered most reactive to distressing and defeating situations (see Laurent, Ablow, Measelle, 2012). Research shows that sAA is an indicator of physical and psychological stress like cortisol (Chatterton, Vogelsong, Lu, Ellman, & Hudgens, 1996; Rohleder, Nater, Wolf, Ehlert, & Kirshbaum, 2004), but is also positively associated with anxiety (Allwood, Handwerger, Kivlighan, Granger, & Sroud, 2011) and arousal necessary for challenge-related coping (see Adam, Hoyt, & Granger, 2011). In general, dysregulation of the SNS and HPA systems have been predictive of internalizing and externalizing problems in children (El-Sheikh, Erath, Buckhalt, Granger, & Mize, 2008).

Given that stress affects several parts of the body simultaneously, understanding the psychobiology of the stress response necessitates a multiple system approach (Bauer, Quas, & Boyce 2002; Granger et al., 2007). Using measures of the HPA axis and SNS helps capture the synchrony, or lack thereof, of different biological systems in response to a stressor, providing a more comprehensive picture of how the body responds to stress (Bauer et al., 2002; Gordis et al., 2008). The SNS and HPA systems directly influence each other through feedback loops among the different glands in the body. Consequently, examining how physiological responses in both systems are connected provides a more holistic assessment of the stress response. Research suggests that sAA and cortisol respond similarly to the same acute stressor task, but might not be correlated when one takes an average correlation across time points, because they are released at different time points (Engert et al., 2011; Rohleder et al., 2004). Salivary cortisol and sAA are typically positively correlated when the time lag is considered (Engert et al., 2011). As a result, we argue that sAA and cortisol should follow similar response and recovery patterns to an acute stressor, with cortisol perhaps having a more delayed effect. Because we are using an acute, stress-inducing conversation between the parent and child (for up to 20 minutes) as the stimulus, a typical cortisol response should involve a reaction immediately after the conversation (given the longer length of the conversations) or 15 minutes later, but then a steady downward trajectory and a recovery by 40 minutes after the conversation. In contrast, a typical physiological response for sAA should be a reaction throughout and immediately after the conversation (given that a conversation about a sensitive topic is probably still stressful during the conversation), with a steady downward trajectory thereafter.

Divorce and Physiological Stress Responses

While there are many short-term and long-term effects of divorce (see Amato, 2000), meta-analyses reveal that the differences between children of divorced and nondivorced families are relatively small, primarily due to individual, interpersonal, and contextual differences in response to divorce (Amato, 2001; Amato & Keith, 1991). The most important factor that determines how well children function after divorce is parents’ communication—with each other and their children. For example, research shows that many of the long-term effects of divorce dissipate if conflict between former spouses is kept at a minimum (see Amato, 2000, 2001; Hetherington, 1999). As Amato and his colleagues (2001; Amato & Afifi, 2006; Amato & Sobolewski, 2001; Booth & Amato, 2001) note, children whose parents have a conflicted marriage and who remain married often have poorer psychological well-being than children whose parents have a conflicted marriage and divorce. It is not only parents’ conflict, however, that affects children. Parents’ general communication competence, or the ability to use communication in an appropriate and effective manner for a given situation (Spitzberg, Canary, & Cupach, 1994), and social support also help shield children from the negative effects of divorce (see Afifi & McManus, 2010; Gately & Schwebel, 1991).

When considering divorce, the age of the child has also been shown to be a strong predictor of children’s adaptation. Adolescents seem to be most at risk for divorce-related externalizing and internalizing problems. Not only are they coping with their parents’ divorce, but they are attempting to establish their identity as emerging adults and are vying for autonomy from their parents (Hetherington, 1999). Adolescents are also living with their parents while the divorce process unfolds, increasing the likelihood that they are exposed to their parents’ ongoing conflict compared with young adults who are living on their own (Buchanan et al., 1991). Even though parents may disclose more inappropriate information to their young adult children, research shows that adolescents are more likely than young adults to feel caught between their parents’ conflict (Afifi & Schrodt, 2003).

The research that is available on divorce and children’s physiological stress response systems suggests that divorce affects children’s stress responses (e.g., Kraft & Luecken, 2009; Luecken, Kraft, Appelhans, & Enders, 2009). There are numerous theoretical reasons why parents’ marital status might interact with their communication skills and the age of the child to affect children physiologically. One primary explanation could be the emotional security hypothesis (Davies & Cummings, 1994, 1998), which is rooted in attachment theory (Ainsworth, Blehar, Waters & Wall, 1978; Bowlby, 1982). According to attachment theory, how people approach relationships is influenced largely by the emotional availability and responsiveness of the caregivers who raised them (Jang, Smith, & Levine, 2002). Children’s attachments with their caregivers not only affect children’s ability to form loving and trusting relationships later in life, but their ability to manage stress (Guerrero, Faranelli, & McEwan, 2009; McCarthy, Lambert, & Moller, 2006). The primary caregiver’s continued responsiveness to a child’s needs tend to foster a secure attachment in the child. The child learns to rely on the caregiver as a safe haven or someone who can protect them from distress and threatening situations (Bowlby, 1982; Hazan & Shaver, 1994b). The way parents communicate with their children, even as young adults, provides a secure base from which children can easily recover from stressful situations (Bowlby, 1982).

Davies and Cummings (1998) argue in their emotional security hypothesis that children’s primary goal is to preserve their emotional security when it is threatened by parents’ negative communication patterns. Emotional security consists of children’s emotional attachment/security with their parents and their internal representations of marriage and family life. When children feel as if their emotional security is threatened, they are likely to react physiologically (with abnormally high or low or erratic HPA and SNS responses, and/or inability to recover from a stress task; Davies et al., 2007), psychologically, and behaviorally (e.g., fight or flight responses in the form of avoidance, fear or aggression; see also Afifi et al., 2008; Afifi, Schrodt, & McManus, 2009). When children feel caught between their parents’ disputes, and/or parents are unsupportive and lack general communication competence, it most likely jeopardizes children’s emotional security, which prompts children to respond physiologically. The emotional security hypothesis has more to do with the parents’ communication skills than the parents’ marital status. Children whose parents stay married and lack communication skills may, in fact, feel even more emotionally threatened than children of divorced parents who had a conflicted marriage (see Amato & Afifi, 2006). Both groups of children most likely over react to stressful conversations with a parent, but the children of married parents might have more difficulty recovering because they are continually exposed to their parents’ ineffective communication patterns through the parents’ marriage (Amato & Afifi, 2006).

Research has confirmed that interparental conflict is associated with over- or under-activation of the HPA and SNS systems. Most of the studies that have found evidence for under activation of the HPA and SNS systems (or lower levels of cortisol and sAA) have been tested with samples of chronic and intense conflict, maltreated youth, clinically referred populations, or individuals with extreme living conditions (e.g., Davies et al., 2007; Gordis et al., 2006, 2008; Granger et al., 1998; Luecken, Kraft, & Hagan, 2009). McEwen (2001) argues that the body has a natural allostasis or way of adapting to stress that can get disrupted when exposed to chronic stress. When individuals experience chronic stress, the stress slowly “wears and tears” on the body or produces allostatic load (McEwen, 1998). Over time, the body can become habituated to the stress and no longer adapt to it (e.g, the body “shutting down” and not reacting to parents’ chronic conflict anymore). In contrast, evidence suggests that acute and moderate to low levels of interparental conflict, as is the case with the means for children’s feelings of being caught between their parents in present sample from the general population, is predictive of heightened pre-laboratory stress task and basal cortisol levels (e.g., Laurent & Powers, 2006; Luecken & Appelhans, 2006). Samples from the broader population, in which there is considerable variance in parents’ communication skills, may result in most children “over reacting” to their parents’ communication during an acute stress task even though some of the sample might have grown up with more intense, and potentially chronic, conflict.

On the other hand, social support and similar constructs like affection (see Floyd et al., 2007; Floyd & Riforgiate, 2009; Priem & Solomon, 2013) have consistently been shown to be stress buffers. Research has found that parental social support, warmth, involvement, and responsiveness allow children to recover more quickly from stressful situations (e.g., Blair, Granger, Willoughby, & Kivlighan, 2006). For instance, Pendry and Adam (2007) found that poor marital functioning (e.g., low marital satisfaction, interparental conflict) was associated with higher average daily and wakeup levels of cortisol in adolescent and kindergarten-age children. On the other hand, higher marital functioning was associated with a lowering of bedtime cortisol levels and steeper diurnal slopes for kindergarten age children compared with those of adolescents. They also found that mothers’ warmth and involvement was associated with quicker recovery or a steeper slope of diurnal cortisol for children of both ages.

The limited studies on divorce and the HPA axis have found that young adults who reported severe separation or loss of a parent as a child as a result of divorce experienced lower cortisol awakening responses (or the rise in cortisol that occurs 20-30 minutes after waking) (Bloch et al., 2007; Meinlschmidt & Heim, 2005) and lower pre-task cortisol levels (Tyrka et al., 2008). In addition, the work by Luecken and colleagues (e.g., Kraft & Luecken, 2009; Luecken, Kraft & Hagan, 2009) examined the larger family interaction patterns of divorced and nondivorced families. Kraft and Luecken (2009) evaluated cortisol in 94 young adults from divorced and intact families before and three times after a challenging speak task. They discovered that young adults from divorced families had significantly lower cortisol levels across the time periods compared with young adults from intact families, even after controlling for family conflict, depression, and anxiety. They also found that divorce was associated with attenuated cortisol levels, but partially due to lower income in divorced families. Therefore, it could be that divorced families’ lower cortisol levels was due to allostatic load produced by chronic financial stress rather than divorce per se. In another study, Luecken, Rodriguez, Bradley and Applehans (2005) assessed systolic blood pressure, diastolic blood pressure, and heart rate responses to a stressful speech task with 150 young adults who experienced the death of a parent, divorce, or were from intact marriages. They found that young adults with positive family relationships, regardless of family type, which were characterized by affection, minimal conflict, cohesion, had a stronger recovery in systolic blood pressure after the speech task. Consequently, children’s responses to a stressful interaction with their parent may more closely revolve around the parents’ communication than their marital status.

In many of the studies mentioned above, however, the young adults seemed to have experienced rather severe divorce circumstances. These samples were also not assessing responses to natural conversations with a parent. A more general population of adolescents and young adults who are reacting to a single stressful interaction with their parent who they perceive as less communicatively skilled may result in elevated, rather than attenuated, baseline cortisol and sAA levels. When children are asked to talk with their parent about something stressful about their parents’ relationship, it is likely stressful for many children because they are reacting to an often negatively valenced conversation that might have occurred in the past (or a similar type of conversation has occurred), but most of them have probably not been exposed to chronic levels of conflict, violence, trauma or loss that would make them completely unresponsive. The parents’ communication skills likely interact with the parents’ marital status to predict children’s pre-interaction or basal cortisol and sAA levels.

Children whose parents are divorced and children whose parents are married and are less communicatively skilled should also both respond physiologically after the discussion task in sAA and cortisol, but children whose parents are still married and lack communication skills may have more difficulty recovering from such a discussion because they cannot escape their parents’ negative communication skills. Therefore, it is predicted that marital status will interact with the parents’ communication skills (i.e., communication competence, socially social support, and children’s feelings of being caught between their parents’ conflict) to affect children’s physiological responses. These physiological responses should also depend on the age of the child, with adolescents potentially having the greatest difficulty recovering from the stressful discussion because they still live with their parents. Therefore, we hypothesize that:

Hypothesis 1: The impact of parents’ marital status on children’s sAA and cortisol levels prior to a stressful discussion task (or basal levels) should be moderated by parents’ communication skills (i.e., feeling caught, social support, communication competence) and the age of the child (adolescent/young adult).

Hypothesis 2: The impact of parents’ marital status on children’s sAA and cortisol reaction and recovery patterns should be moderated by parents’ communication skills and the age of the child.

Method

Participants

The sample was comprised of 118 parent-child dyads (one parent and child per family; children ages 15-22, M = 16.5). Approximately half of the parents and adolescents/young adults were from intact families (n = 57) and half were from separated/divorced families (n = 61). The sample consisted primarily of mothers (n = 91 or 71%; fathers = 27 or 23%), but had a relatively equal number of female (n = 64 or 55%) and male (n = 53 or 45%) adolescents. The majority of the adolescents (n = 73 or 62%) and parents (n = 82 or 70%) were White or Latino (adolescents = 27 or 23%; parents = 24 or 20%). All but one of the parents had some college or an Associate’s Degree (n = 42 or 36%) or a Bachelor’s Degree (n = 36 or 31%). The household income ranged from $10,000 to over $100,000, but the median household income was $65,000.

For most of the divorced families, the divorce occurred quite recently, with the average time since physical separation being 3.3 years (range = 1 month to 8 years). Six of the parents were physically separated and had filed for divorce, even though the divorce was not official. These families were grouped under the divorced category. The average time since the divorce was 2.9 years. The majority of the parents (n = 45 or 75%) had primary physical custody of their children, but some had joint physical custody (n = 16 or 26%). Nine of the divorced parents were remarried.

Procedures

The families were recruited through newspaper advertisements and flyers placed in local nonprofit organizations, businesses, and a major university. For the separated/divorced families, the parents had to have primary or joint custody and separated or filed for divorce within the past 6 years. Their participation entailed a one-time visit to a research laboratory.

The study involved a pre-interaction survey, a stressful discussion between the parent and child, and a post-interaction survey. After completing their consent forms, the parent and child sat on a couch in a living room setting in the laboratory and engaged in small talk for 7 minutes. Following the small talk, the parent-child dyads engaged in a challenge task where they were asked to talk to one another about something stressful and conflict inducing related to the parents’ relationship. This task was meant to elicit stress and anxiety through a naturally occurring discussion between the parent and child rather than through artificial stress tests (e.g., public speaking tasks, difficult math problems). In a previous examination of this data, Afifi, Granger, Denes, Joseph and Aldeis (2011) found that 52% of the children were considered “overreactors” to the challenge task (or had sAA levels that rose by 10% after the discussion task), 16% had no reactivity across the task, and 22% were in the down regulator group (increased in their scores, but not to 10% and then declined). For cortisol, there was no middle group, but 20% over reacted to the task and 66% down regulated after the task. Thus, the task seemed to be stressful and anxiety producing, but with considerable variance in the responses. The parents and adolescents discussed this topic in private, but were informed that their discussion was being videotaped. They were told to discussthe topic for as long as they felt comfortable for up to 20 minutes, and to signal the researcher when they were finished and the researcher would reenter the room. The discussions lasted approximately 16 minutes (range = 10-20 minutes). After the saliva samples were collected, the participants completed a survey about the interaction that just occurred. All of the families were debriefed.

Physiological Procedures and Measures

Participants were not included in the study if they had bi-polar disorder, an endocrine disorder (e.g., diabetes, a thyroid disorder), or were taking endocrine-related medications that could have altered their cortisol or sAA levels (e.g., strong anti-inflammatory medications, hormone replacement therapies; Granger, Hibel, Fortunato, & Kapelewski, 2009). Given that many people who are experiencing divorce take antidepressants, participants were included if they were on other antidepression and anxiety medications, enhancing the external validity of the study. Participants were told not to drink alcohol or use illegal drugs, or visit the dentist within 48 hours of participating in the study (e.g., left over blood from dentistry could get into the saliva). They were also asked not to exercise rigorously, consume alcoholic or caffeinated beverages, or smoke that day. Finally, they were asked not to brush their teeth up to 2 hours or consume a meal up to 1 hour prior to coming to the lab (Gibson et al., 1999).

On arriving at the laboratory, the participants were asked to rinse their mouth with a small amount of water and drink the remaining 6 ounces before completing their pre-interaction survey to remove bacteria that could affect the hormone results. Four saliva samples were collected via passive drool (see Granger et al., 2007) to measure the participants’ HPA and SNS reactivity and recovery to the discussion task. Saliva was collected prior to the parent-child discussion task, immediately on completion of the discussion task (when sAA typically peaks), 20 minutes after the completion of the task (when cortisol typically peaks), and 40 minutes after the completion of the task (for recovery; see also Allwood et al., 2011; Gordis et al., 2006). The participants were asked to imagine that they were chewing their favorite food and passively drool into a straw that was inserted into a 2 mL cryogenic vial. Once the required amount of saliva was collected, the vials were sealed and immediately frozen (−20° C). The samples were then shipped overnight to Salimetrics, where they were frozen at −80° C until analyzed. Following Granger et al. (2007), all of the samples were assayed for salivary cortisol, by enzyme immunoassay, and sAA, by kinetic reaction assay, without modification to the manufacturers (Salimetrics, State College, PA) recommended protocol. Intra- and inter-assay coefficients of variation were, on average, less than 10% and 15%, respectively. sAA is reported in units of activity per milliliter (U/mL) and cortisol in micrograms per deciliter (ug/dL). To restrict variation in the salivary analytes due to diurnal patterns, all of the samples were collected in the late afternoon (3 p.m. on) or early evening (before 7 p.m.).

Self-Report Measures

Because only the child’s perception of the parent’s communication skills were of interest and the parents’ perception of their own communication skills did not affect the child’s physiological responses, only the child’s self-reports are used in this study. For Hypothesis 1, the child’s perceptions of the parent’s communication skills prior to the interaction were used (given that we were determining how the parent’s communication skills in general are related to the child’s baseline physiological scores). For Hypothesis 2, the child’s perceptions of the parent’s communication during the interaction were used (given that what happened in the interaction is likely to have the greatest impact on physiological responses after the interaction).

Children’s feelings of being caught

Buchanan et al.’s (1991) Feeling Caught scale was used to measure adolescents’ and young adults’ feelings of being caught between their parents’ disputes in the pre- and post-interaction surveys. The original scale consists of 11 items that ask children about the degree to which they feel put in the middle of their parents using a 5-point Likert-type scale that ranges from (1) never to (5) very often. The scale was reduced to five items (to reduce fatigue and because some of the items from the existing scale were not applicable to a single discussion task) and asked the children about their feelings of being caught in general (pre-interaction; for example, “How often do you feel like if you are loyal to one parent, you are being disloyal to the other parent?” “How often do you feel caught between your mother and father?” “How often do you feel torn between your mother and father?”) and in the specific interaction that just occurred (post-interaction; for example, “How often did you feel like if you were loyal to one parent, you were being disloyal to the other parent?” “How often did you feel caught between your mother and father?” “How often did you feel torn between your mother and father?”). Larger numbers represent greater feelings of being caught. The alpha coefficient was .87 for the pre-interaction and .79 for the post-interaction.

Communication incompetence

The children’s perception of their parent’s communication competence in general before the discussion task (pre-interaction) and during the discussion task (post-interaction) was assessed with Guerrero’s (1994) Communication Competence scale. The scale consisted of six items (e.g., “My parent has [or had] a hard time communicating his/her feelings clearly,” “My parent is [or was] a good listener,” “My parent is [or was] a good communicator”) on a Likert-type scale that ranged from (1) Strongly disagree to (5) Strongly agree with larger numbers indicating greater competence. The Cronbach’s alphas were .87 for the pre-interaction survey and .86 for the post-interaction survey.

Social support

The children’s perception of their parent’s social support as a whole before the discussion task (pre-interaction) and during the discussion task itself (post-interaction) was measured with a revised version of the Recipient Support Perception Scale (Trees, 2000). Eleven of the original 18 items that focused on empathy, perspective taking, listening, receptivity, and confirmation were used (e.g., “My parent made [or makes] me feel loved,” “My parent made [or makes] me feel cared for,” “My parent expressed [or expresses] confidence in my ability to handle to problems we discussed [or discuss]).” The scale was reduced to prevent participant fatigue. The Likert-type scale ranged from 1 to 7, with 1 being “strongly disagree” and 7 being “strongly agree.” Larger numbers indicate greater perceptions of social support. The Cronbach’s alphas were .90 for the pre-interaction and .88 for the post-interaction.

Self-reported stress and anxiety

Children’s self-reported stress and anxiety during the interaction were measured with five items in the post-interaction survey. The two stress items asked the participants how stressed and tense they felt during the interaction. The three anxiety items asked them how anxious, uneasy, and uncomfortable they felt during the discussion task. The items were positioned on a five-point semantic differential scale that ranged from 1 to 7, with larger scores indicating greater stress. The Cronbach’s alpha for stress and anxiety was .81.

Realism

A five item measure of interaction realism was included in the post-interaction survey. The children were asked about the typicality of the conversation that just transpired using a semantic differential scale that ranged from 1 to 7, with higher numbers representing more realistic, natural, and typical conversations. The children found the interaction to be quite naturalistic (M = 5.30, SD = 1.17). The alpha coefficient was .71.

Results

Preliminary Analyses

The distributions of cortisol and sAA scores were analyzed. Of the original 118 dyads, 9 of the children had outliers greater than 3 standard deviations from their respective mean. These were kept in the analyses by recoding to the next highest value in their distribution. One of the children’s cortisol and sAA scores was also missing and the 40-minute cortisol and sAA collections for one of the dyads were not collected due to a technical difficulty. Thus, the final N for most of the analyses was 116 parents and 116 children. A natural log transformation was used to address substantial skew in the cortisol scores (average skewness = 8.66, SE = .22; transformed skew = 0.54, SE = .15). A square root transformation was used to address a moderate skew in the sAA scores (average skewness = 1.46, SE = .22; transformed skew = 0.46, SE = .15).

The means, standard deviations, and correlations among the variables are located in Table 1. Of particular interest is the robust pattern of negative correlations among the communication variables and sAA and the lack of significant associations involving cortisol. To further explore these correlations, separate correlations for sAA and cortisol and the variables of interest were calculated for children from divorced and nondivorced families (see Table 2). Once again, there was a more consistent pattern of significant associations for sAA than cortisol. There was one significant correlation for feeling caught and cortisol (immediate and 20 minutes post) after the discussion task for children whose parents were married. In addition, unlike what one might expect, there were no significant correlations for feeling caught before or after the conversation for divorced children. However, almost all of the correlations for feeling caught and sAA were significant for children whose parents were married. For children whose parents were married, the more they felt caught between their parents, the higher their sAA.

Table 1.

Descriptive Statistics and Correlations for the Primary Variables in the Analyses.

Variable	M	SD	1	2	3	4	5	6	7	8	9	10	11	12	13	14
1. Communication competence T1	3.70	0.83	—
2. Social support T1	5.57	1.01	.76**	—
3. Feeling caught T1	1.81	0.65	−.45**	−.40**	—
4. Communication. competence T2	4.16	0.71	.72**	.66**	−.39**	—
5. Social support T2	5.46	1.00	.62**	.67**	−.37**	.74**	—
6. Feeling caught T2	1.72	0.70	−.45**	−.34**	.61**	−.45**	−.44**	—
7. sAA pre-task	8.68	3.59	−.22*	−.15	.16	−.20*	.02	.07	—
8. sAA imm. post	8.51	3.74	−.27**	−.16	.17	−.20*	−.05	.12	.87**	—
9. sAA 20-min post	7.93	3.48	−.26*	−.22*	.20*	−.26**	−.07	.15	.84**	.88**	—
10. sAA 40-min post	7.71	3.27	−.27*	−.23*	.15	−.27**	−.11	.17	.85**	.83**	.88**	—
11. Cortisol pre-task	−0.92	0.25	.02	−.01	−.05	−.01	.08	−.07	.06	.01	.05	.06	—
12. Cortisol imm. post	−1.02	0.24	003	−.04	−.08	−.01	.06	−.14	−.05	−.06	−.03	−.02	.85**	—
13. Cortisol 20-min post	−1.07	0.27	−.01	−.09	−.02	−.06	−.01	−.08	−.01	−.08	.04	.03	.80**	.85**	—
14. Cortisol 40-min post	−1.11	0.26	−.02	−.11	−.05	−.05	−.01	−.14	−.10	−.18*	−.09	−.04	.72**	.82**	.87***	—

Note. sAA = salivary alpha-amylase; imm. = immediate.

p < .05. **p < .01. ***p < .001.

Table 2.

Breakdown of Correlations for sAA and Cortisol with the Primary Variables for Children of Divorced and Nondivorced Families.

Variable	Communication competence T1	Social support T1	Feeling caught T1	Communication competence T2	Social support T2	Feeling caught T2
Divorced families
1. sAA pre-task	−.22*	−.22*	.05	−.24*	.01	−.08
2. sAA imm. post	−.28**	−.19	.12	−.17	−.01	−.01
3. sAA 20 min post	−.23*	−.24*	.10	−.30**	−.10	.03
4. sAA 40 min post	−.21	−.19	.03	−.26**	−.05	.03
5. Cortisol pre-task	−.08	−.10	.04	−.10	.05	.04
6. Cortisol imm. post	−.07	−.19	.09	−.08	−.03	.01
7. Cortisol 20 min post	−.09	−.17	.15	−.17	−.13	.06
8. Cortisol 40 min post	−.05	−.14	.08	−.11	−.07	−.07
Nondivorced families
1. sAA pre-task	−.21	−.06	.26**	−.16	.06	.25*
2. sAA imm. post	−.25*	−.13	.22*	−.22*	−.09	.26**
3. sAA 20 min post	−.28**	−.18	.30**	.29**	−.02	.29**
4. sAA 40 min post	−.34***	−.27**	.27**	−.29**	−.17	.35***
5. Cortisol pre-task	.14	.13	−.15	.09	.11	−.21
6. Cortisol imm. post	.12	.06	−.19	.06	.11	−.28**
7. Cortisol 20 min post	.06	−.04	−.17	.05	.11	−.24*
8. Cortisol 40 min post	−.01	−.13	−.12	−.01	.02	−.19

Note. sAA = salivary alpha-amylase; imm. = immediate.

p < .10. **p < .05. ***p < .01.

Preliminary ANOVAs were also computed to compare the self-reported measures of children from divorced and nondivorced families (see Table 3). Children of divorced parents were significantly more stressed and anxious than children whose parents were married. Overall, children of divorced parents reported being anxious, but only moderately stressed, during the conversation. Nevertheless, there was quite a bit of variance in the children’s self reports of stress and anxiety. Interestingly, children of divorced parents did not report feeling caught between their parents during the conversation more than children whose parents were married, but they did report significantly greater feelings of being caught in general before the conversation. The children, as a whole, did not report feeling very caught between their parents. Children from divorced and intact families did not significantly differ in their perceptions of their parent’s social support or communication competence before or after the interaction. As a result, Table 3 is limited to broader comparisons between children of divorced and intact families. The average cortisol and sAA scores for adolescents/young adults were not significantly correlated. Furthermore, the average cortisol and sAA scores for adolescents were not significantly correlated with their self-reported stress and anxiety, although the correlation between sAA and self-reported anxiety was approaching significance (p < .10)

Table 3.

Differences in Means for Children from Divorced and Nondivorced Families.

	Divorced		Nondivorced
Variables	M	SE	M	SE	F
Feeling caught T1	1.63	.09	1.99	.08	8.98**
Feeling caught T2	1.79	.10	1.61	.09	1.77
Communication competence T1	3.72	.11	3.76	.12	0.09
Communication competence T2	4.19	.09	4.20	.10	0.004
Social support T1	5.53	.14	5.68	.14	0.59
Social support T2	5.38	.14	5.58	.14	1.03
Self-reported stress T2	3.13	.23	2.30	.23	7.23**
Self-reported anxiety T2	4.83	.20	2.67	.20	57.84***

p < .05. **p < .01. ***p < .001

Primary Results

Pre-task sAA and cortisol levels

Hierarchical regression was used to assess whether parents’ marital status (divorce/married) interacted with children’s perceptions of their parents’ communication skills in general before the interaction task and the age of the child to predict the children’s pre-interaction sAA and cortisol scores (H1). The main effects for the specific communication pattern of interest, the age of the child, and marital status were entered into the regression model first. The two-way interactions among the variables were entered in the second step, followed by the three-way interaction for the communication variable, age, and divorce in the third step.

There were no significant three-way interactions for any of the models (providing a lack of support for H1; see Table 4). The only significant interaction was between divorce and the age of the child for all three sAA models. Follow-up analyses revealed that young adults from nondivorced families (M = 6.79; SD = 0.81) had significantly lower sAA pre-task scores compared with adolescents from nondivorced families (M = 9.18; SD = 0.57; t = 2.60, p < .01) and adolescents (M = 8.96; SD = 0.57; t = 2.88, p < .001) and young adults (M = 8.94; SD = 0.75; t = 2.42, p < .05) from divorced families.

Table 4.

Communication Skills, Age of the Child, and Marital Status of Parents Predicting Children’s Salivary Alpha-Amylase and Cortisol Pre-Interaction Scores.

	sAA				Cortisol
Models	R ²	R ^2change	β	T	R ²	R ^2change	β	t
Feeling caught models
Step 1	.05	.05			.08	.08
Age			−.16	−1.73			.28	3.12**
Divorce			.05	0.49			−.03	−0.35
Feeling caught			.27	1.80*			−.06	−0.56
Step 2	.08	.03			.11	.02
Age x Divorce			.32	2.04**			.10	0.63
Age × Feeling caught			−.17	−0.56			.21	1.39
Divorce × Feeling caught			−.29	−0.87			.20	0.63
Step 3	.09	.007			.11	.001
Age × Divorce × Feeling caught			.52	0.94			−.16	0.30
Competence models
Step 1	.07	.07			.08	.08
Age			−.11	1.18			.28	3.02**
Divorce			.09	0.93			−.05	−0.55
Communication competence			−.20	−2.15**			−.02	−0.23
Step 2	.11	.04			.12	.05
Age × Divorce			.33	2.21**			.25	1.69***
Age × Competence			.11	0.26			−.31	−0.71
Divorce × Competence			−.16	−0.26			−.70	−1.66
Step 3	.11	.002			.13	.003
Age × Divorce × Competence			−.36	−0.50			−.45	−0.63
Social support models
Step 1	.04	.04			.08	.08
Age			−.11	−1.11			.28	3.02**
Divorce			.05	0.55			−.06	−0.59
Social support			−.14	−1.46			−.03	−0.28
Step 2	.10	.06			.10	.02
Age × Divorce			.35	2.31**			.23	1.50
Age × Support			−.02	−0.04			−.17	0.31
Divorce × Support			.38	0.86			.08	0.16
Step 3	.10	.004			.10	.002
Age × Divorce × Support			−.58	−0.70			.36	0.44

p < .10. **p < .05. ***p < .01.

Given that the results also suggested that there might be a main effect for communication competence (see Table 4), we examined the bivariate correlation between communication competence and the pre-task sAA (which is equivalent to re-running the regression model with only communication competence). As Table 1 suggests, there was a significant, negative correlation between communication competence and pre-task sAA. Children whose parents were more communicatively competent in general had lower pre-interaction sAA scores. Stated differently, children whose parents were more communicatively incompetent had higher pre-interaction sAA scores. Even though there were no significant interactions with divorce, children of divorced parents had sAA scores that were at least 10% higher than children of married parents. A 10% increase in sAA is considered by many researchers to be a meaningful difference (see Afifi, Granger, Denes, Joseph, & Aldeis, 2011). There was also a significant main effect for the age of the child and cortisol (β = .21, t = 2.21, p < .05). Follow-up analyses revealed that adolescents (M = −0.97; SD = 0.24) had lower pre-interaction cortisol levels than young adults (M = −0.82; SD = 0.25; F = 9.65, p < .01).

Parents’ communication skills and children’s reactivity and recovery to the discussion task

Multi-level modeling (MLM) or growth curve analysis (using the SPSS Mixed Procedure) was used to examine response and recovery patterns in sAA and cortisol within and between the adolescents/young adults as a function of the parents’ communication skills and marital status (H2). The children’s sAA and cortisol scores over time (i.e., pre-interaction, immediate post-interaction, 20 minutes post-interaction, and 40 minutes post-interaction) were the Level 1 factors (coded as 0, 1, 2, 3). The age of the child (0 = adolescent, 1 = young adult), whether the parents were divorced (0 = married, 1 = divorced), and the child’s perception of the parent’s communication skills (support, communication competence, feeling caught) during the discussion task were Level 2 factors. The age of the child was dichotomized rather than used as a continuous variable to best match the literature that divides the two groups—research that shows that adolescents tend to be quite different from young adults due to adolescents still living with their parents, being directly parented and monitored, and experiencing more dramatic fluctuations in hormones. Separate analyses were conducted for each communication process for cortisol and sAA. The time of data collection, length of the interaction, and use of medications did not significantly influence the results and therefore were not included as control variables. The variables were grand mean centered.

Before conducting the primary analyses, two unconditional models with no predictors (or intercepts only/null models) were tested to assess whether there was significant variability in the random intercepts for sAA and cortisol. There was significant variability in the sAA scores, z = 7.05, p < .001, and the cortisol scores, z = 7.03, p < .001, over time. In the primary models, models for the main effects for time, divorce, the age of the child, and the particular communication variable of interest were assessed first, followed by tests for the interactions. The final models include the main effects and the interactions together in the same model. The final models presented in this manuscript are the best fitting models for each communication variable. If the higher order interactions (i.e., three or four way interactions) were not significant, they were removed from the model to preserve power for the lower level interactions. For all of the models, there were significant linear and quadratic effects. However, because the graphing of the results suggested quadratic trends and these trends were what was hypothesized, only the quadratic results are reported. The final quadratic models for sAA and cortisol are provided in Tables 5 and 6. Overall, the random effects for all of the communication variables showed that growth varied significantly across children for sAA and cortisol. When determining which processes explain the trajectories among children, several significant interactions emerged. The models for sAA will be discussed first.

Table 5.

Final Quadratic Models of Fixed and Random Effects for Communication Skills and Salivary Alpha-Amylase.

	B	SE	t or Wald Z	Significance
Model 1 for Communication competence
Fixed effects
Intercept	9.15	0.57	16.14	.001
Age	−2.4	0.98	−2.46	.01
Divorce	−0.5	0.81	0.62	.53
Time	−0.08	0.04	−2.16	.03
Communication Competence T2	−0.95	0.76	−1.26	.22
Age × Time	0.02	0.27	0.23	.82
Divorce × Age	3.02	1.39	2.17	.03
Comp T2 × Age	0.44	1.21	0.36	.72
Divorce × Time	−0.02	0.06	−0.41	.68
Comp T2 × Time	−0.09	0.05	−1.77	.08
Divorce × Comp T2	0.19	1.09	0.18	.86
Divorce × Age × Time	−0.04	0.1	−0.46	.65
Comp T2 × Age × Time	0.14	0.07	1.98	.05
Divorce × Comp T2 × Age	−2.57	1.8	−1.43	.16
Divorce × Comp T2 × Time	0.06	0.07	0.86	.39
Random effects
Intercept	11.16	1.62	6.87	.001
Slope	0.05	0.08	6.16	.001
Model 2 for Social Support
Fixed effects
Intercept	9.1	0.56	16.43	.001
Age	−2.35	0.95	−2.47	.01
Divorce	−0.42	0.79	−0.53	.6
Time	−0.08	0.04	−2.09	.03
Support T2	0.25	0.53	0.47	.6
Age × Time	0.007	0.06	0.11	.91
Divorce × Age	2.29	1.32	1.73	.08
Social Support T2 × Age	−0.65	0.91	−0.71	.48
Divorce × Time	−0.02	0.05	−0.39	.7
Social Support T2 × Time	−0.1	0.04	−2.72	.08
Divorce × Social Support T2	−0.1	0.78	−0.12	.9
Divorce × Age × Time	−0.02	0.09	−0.21	.84
Social Support T2 × Age × Time	0.12	0.06	1.9	.06
Divorce × Social Support T2 × Age	0.002	1.36	0.001	.1
Divorce × Social Support T2 × Time	0.11	0.05	2.12	.03
Divorce × Social Support T2 × Age × Time	−0.21	0.09	−2.33	.02
Random effects
Intercept	10.46	1.53	6.85	.001
Slope	0.03	0.01	2.76	.006
Model 3 for feeling caught
Fixed effects
Intercept	9.24	0.56	16.57	.001
Age	−2.41	0.92	−2.62	.01
Divorce	−0.48	0.78	−0.61	.54
Time	−0.08	0.03	−2.22	.001
Feeling Caught T2	1.79	0.73	2.44	.01
Age × Time	0.01	0.04	0.21	.83
Caught T2 × Age	−1.36	0.88	−1.55	.12
Divorce × Age	2.08	1.18	1.76	.08
Feeling Caught T2 × Time	0.03	0.03	0.91	.37
Divorce × Time	−0.03	0.04	−0.73	.36
Divorce × Caught T2	−1.56	0.82	−1.9	.06
Random effects
Intercept	11.23	1.62	6.94	.001
Slope	0.05	0.01	3.87	.001

Table 6.

Final Quadratic Models of Fixed and Random Effects for Communication Skills and Cortisol.

	B	SE	t or Wald Z	Significance
Model 1 for Communication competence
Fixed effects
Intercept	−0.92	.05	19.03	.001
Age	0.12	.08	1.50	.13
Divorce	−0.08	.07	−1.18	.24
Time	−0.02	.002	−8.53	.001
Communication competence T2	−0.06	.06	0.97	.34
Age × Time	0.006	.003	1.67	.09
Divorce × Age	−0.04	.11	−0.35	.72
Comp T2 × Age	0.05	.08	0.58	.56
Divorce × Time	0.005	.003	1.58	.12
Comp T2 × Time	−0.001	.002	−0.56	.57
Divorce × Comp T2	−0.09	.08	−1.16	.25
Random effects
Intercept	0.08	.01	6.83	.001
Slope	0.002	.0004	5.31	.001
Model 2 for Social support
Fixed effects
Intercept	−0.92	.05	−19.37	.001
Age	−2.35	.95	−2.47	.01
Divorce	0.13	.08	1.65	.10
Time	−0.02	.002	−8.45	.001
Support T2	0.05	.04	1.29	.20
Age × Time	0.005	.003	1.62	.11
Divorce × Age	−0.02	.11	0.19	.85
Social Support T2 × Age	−0.02	.06	−0.45	.66
Divorce × Time	0.004	.003	1.54	.13
Social Support T2 × Time	−0.001	.001	−0.62	.54
Divorce × Social support T2	−0.14	.05	−2.53	.01
Random effects
Intercept	0.10	.01	6.92	.001
Slope	0.002	.0005	5.39	.001
Model 3 for feeling caught
Fixed effects
Intercept	−0.92	.05	−18.81	.001
Age	0.15	.09	−1.75	.08
Divorce	−0.07	.07	−0.98	.33
Time	−0.02	.003	−7.85	.001
Feeling caught T2	−0.06	.08	−0.81	.42
Age × Time	0.004	.005	0.81	.42
Caught T2 × Age	0.14	.13	1.06	.29
Divorce × Age	−0.07	.12	−0.60	.55
Feeling Caught T2 × Time	0.003	.004	0.74	.46
Divorce × Time	0.003	.004	0.77	.44
Divorce × Caught T2	0.05	.10	0.53	.60
Divorce × Age × Time	0.003	.007	0.51	.61
Caught T2 × Age × Time	−0.01	.005	−2.11	.03
Divorce × Caught T2 × Age	−0.11	.17	−0.66	.51
Divorce × Caught T2 × Time	0.007	.005	1.50	.14
Random effects
Intercept	0.08	.01	6.81	.001
Slope	0.0002	.0004	5.12	.001

Note. Results for sAA response and recovery patterns. sAA = salivary alpha-amylase.

Results for sAA response and recovery patterns

When analyzing the model for communication competence and sAA, there was only a significant three-way interaction for communication competence, the age of the child, and time (at p = .05; see Table 5, providing partial support for H2). Adolescents whose parents were less communicatively competent during the discussion had sAA scores that tended to increase slightly after the talk, decline at 20 minutes, but then increase again at 40 minutes (see Figure 1). Young adults whose parents were less competent had sAA scores that decreased right after the discussion, demonstrated a delayed increase at 20 minutes, and a decrease again at 40 minutes (see Figure 1). In contrast, adolescents whose parents were more communicatively skilled experienced a very slight increase in sAA immediately after the task, but then naturally declined in their sAA over time. Young adults whose parents were more communicatively skilled did not even experience an increase in sAA, but only a steady decrease in sAA from the first time point. Given that sAA tends to increase immediately after a stressor, a delayed reaction at 20 and 40 minutes after the conversation for young adults and adolescents whose parents lacked skills suggests more difficulty recovering from the task when parents were less skilled. Adolescents whose parents were less communicatively competent had the most difficult time recovering from the conversation.

Figure 1.

Alpha amylase for adolescents and young adults, whose parents were lower and higher in communication competence.

When examining the results for social support and sAA patterns, there was a significant interaction for divorce, social support, age, and time (see Table 5). As it was predicted (H2), when parents were socially supportive, adolescents from divorced and nondivorced families showed a slight immediate reaction in sAA, followed by a steady downward trend in their sAA (see Figure 2). Young adults from divorced and nondivorced families showed no immediate sAA response to the conversation, but a downward trend from the pre-interaction sAA. On the other hand, children from divorced and nondivorced families whose parents were less socially supportive during the discussion task experienced more reactance than children whose parents were more socially supportive (see Figure 3). This reaction depended on whether their parents were divorced or married and whether the child was an adolescent or young adult. Young adults from divorced families whose parents were less socially supportive had a delayed peak reaction at 20 minutes after the interaction, but their slope declined somewhat by 40 minutes, whereas adolescents from divorced families did not have much of a reaction across the four time points. In contrast, young adults whose parents were married but less socially supportive showed their peak reaction immediately after the conversation followed by a rapid trajectory downward by 20 minutes and even further downward by 40 minutes, which would be indicative of a natural or expected sAA response. Adolescents whose parents were still married and less supportive, however, experienced an immediate post-interaction sAA response, followed by a decrease at 20 minutes, but another increase in sAA at 40 minutes—demonstrating a reaction and potential inability to recover from the discussion task.

Figure 2.

Alpha amylase for adolescents and young adults from divorced and nondivorced families, whose parents were high in social support.

Figure 3.

Alpha amylase for adolescents and young adults from divorced and nondivorced families, whose parents were lower in social support.

Finally, the results for the feeling caught and sAA model revealed an interaction between divorce and feeling caught that was approaching significance (see Table 5). Children (regardless of age or their parents’ marital status) who felt less caught between their parents did not even demonstrate an sAA response to the discussion task, but showed a stead trajectory downward from the pre-interaction sAA score (suggesting modest support for H2; see Figure 4). On the other hand, children who were more likely to feel caught between their married parents during the interaction were more likely to experience an immediate sAA response, followed by a decrease in sAA at 20 minutes, and then an increase at 40 minutes, suggesting an inability to recover from the conversation (see Figure 4). Children of divorced parents who were more likely to feel caught during the interaction down regulated right after the talk and experienced a delayed sAA response at 20 minutes, and then recovered by 40 minutes.

Figure 4.

Alpha amylase for children from divorced and nondivorced families with higher and lower feeling of being caught between their parents.

Results for cortisol response and recovery patterns

There was only a significant interaction between divorce and support for cortisol (see Table 6). Children from divorced and nondivorced families whose parents were higher in social support experienced no cortisol response and only declined in cortisol over time from their pre-interaction scores, with children from nondivorced families experiencing a slightly more trajectory downward (see Figure 5). When children from divorced and nondivorced families perceived their parent as less supportive, they both experienced an increase in cortisol at 20 minutes (see Figure 5). This increase, however, did not exceed their pre-task cortisol levels. This response is also considered a normal reaction to a stress task given that they still recovered to baseline by 40 minutes. This suggests that there was some change in their stress levels as a result of the conversation, but not enough to exceed their stress levels going into the conversation. There was also a significant three-way interaction for feeling caught, age, and time for cortisol. When analyzing the graphs, the young adults and adolescents all experienced a steady decrease in cortisol, with only the adolescents who felt caught experiencing a slight delayed increase in cortisol at 40 minutes, suggesting a slight lack of ability to recover from the discussion task. This delayed response, however, also did not exceed their pre-interaction cortisol level.

Figure 5.

Cortisol for children from divorced and nondivorced families whose parents were lower and higher in social support.

Discussion

The purpose of this study was to examine the extent to which the impact of divorce on children’s HPA and SNS reactivity and recovery patterns is moderated by parents’ communication skills and the age of the child. The results suggest that when children perceived their parents as more communicatively skilled (i.e., more communicatively competent, socially supportive, children felt less caught between them) when talking about something stressful related to the parents’ relationship, the children were able to recover physiologically from the interaction. Some of the children of parents who were highly skilled did not even react physiologically to the discussion task (but still down regulated), and others demonstrated the expected curvilinear trend. This downward trend was consistent across the communication skills and did not depend on the marital status of the parent or the age of the child. In contrast, when parents were perceived as less skilled communicatively, the parents’ marital status and the age of the child played an important role in determining children’s physiological responses. When children thought their parents were less communicatively skilled, they had more reactance and had a more difficult time recovering from the discussion task for sAA and somewhat for cortisol. The results also showed in terms of children’s pre-task sAA levels that when parents (regardless of marital status) were more communicatively incompetent, regardless of the parents’ marital status, the children had higher sAA before the interaction. These findings are explained in greater detail in the following.

Divorce, Parents’ Communication Skills, and Pre-Task Levels of Cortisol and sAA

The few previous studies that have been conducted on divorce and the HPA axis have found that young adult children of divorce have lower cortisol awakening responses and basal cortisol levels than young adults whose parents are married (Bloch et al., 2007; Meinlschmidt & Heim, 2005; Tyrka et al., 2008). Why, then, did we find no significant pre-task differences in cortisol or sAA as a function of divorce? One possible explanation is that other studies often do not examine the interaction between marital status and parents’ communication skills (for exceptions see work of Luecken and colleagues). The results from the present study suggest that this may be fundamental to explaining the impact of divorce on children’s physiological responses. In terms of pre-task sAA levels, the results from our study showed that when children thought their parent was more communicatively incompetent, regardless of the parents’ marital status, they had higher pre-task sAA levels. A second explanation could be that prior studies on divorce and children’s physiology examined traumatic divorce situations (e.g., no contact with the noncustodial parent) that might contribute to more chronic stress than typical divorce circumstances where there is greater variance in children’s relationships with their parents. Granted, we did not collect information on the parents’ marital history. However, the pre-task measures indicated that the children, on average, reported low levels of feeling caught between their parents and they thought their parents were quite skilled.

The findings of the present study also suggest that young adults from nondivorced families have lower sAA scores than adolescents from nondivorced families and adolescents and young adults from divorced families. It could be that young adults whose parents are married are no longer living with their parents and are probably less likely, on average, to witness interparental conflict compared with adolescents who are living with their parents and young adults from divorced families who probably still experience conflict at higher rates. Our findings also provide partial support for the self-report literature on divorce (see Amato, 2000), in that parents’ general communication competence, and somewhat for parents’ conflict, may be more important in predicting children’s pre-task sAA levels than divorce. Nevertheless, even though the pre-task levels of sAA were not significantly different between children of divorced and nondivorced families, it should be noted that children of divorce had pre-task sAA scores that were more than 10% higher than children whose parents were married.

One plausible explanation for the more pronounced findings for sAA compared with cortisol is that they were measuring different physiological responses. The adolescents and young adults in the present sample, especially if their parents were divorced, reported being more anxious than stressed. The correlation between sAA and self-reported anxiety (but not stress) was approaching significance. Cortisol was not significantly associated with self-reported anxiety or stress. Therefore, it is possible that the results for sAA may be more of a reflection of anxiety and arousal than stress. Research has found that sAA is an indicator of stress (e.g., Chatterton et al., 1996; Rohleder et al., 2004), but as Nater and Rohleder (2009) suggest in their review of the research on sAA, it is also an indicator of anxiety and arousal. In fact, Van Veen et al. (2008) found that patients with anxiety disorders had higher basal levels of sAA, but not cortisol, compared with a control group. Given that sAA is released by the SNS, another interpretation of the sAA results could be that a child’s fluctuations in sAA meant that the child experienced more of a “fight” response to a situation that felt controllable. At a general level, similar trends emerged for parents’ communication skills and cortisol—just not with the same nuanced interaction effects. Perhaps sAA is a more sensitive or responsive biosocial marker than cortisol. Using a multi-systems approach may shed light on the different types of physiological responses children experience. While the present study did not measure chronic anxiety, the results pose interesting questions about how anxiety affects children of divorce long-term. Even though this was only one interaction, what might it affect children who have these types of interactions on a regular basis? If children of divorce have chronically elevated (or attenuated) anxiety, how does this contribute to their personal and relational health?

Divorce, Parents’ Communication Skills, and Changes in Cortisol and sAA

Similar to previous research (e.g., Floyd et al., 2007; Floyd & Riforgiate, 2009; Pendry & Adam, 2007), our findings suggest that communication skills that are supportive help buffer stress and anxiety. Across all three communication skills, when parents were highly skilled the children showed an ability to recover quickly in their sAA from the discussion task, regardless of the age of the child or the marital status of the parents. When children thought their parents were highly skilled, they also tended to not “react” as much to the discussion task or did not perceive the task to be very stressful.

The patterns of reactivity and recovery when the parents lacked communication skills were quite different from the downward trends that resulted when the parents were skilled. When parents were less skilled communicatively, the parents’ marital status and the age of the child became important in determining children’s physiological responses. The interaction among marital status, parents’ communication skills, and the age of the child was most pronounced for social support. This three-way interaction showed how differently children from divorced and nondivorced families can respond physiologically when their parents are less supportive. For example, young adults whose parents were married and who were less socially supportive had a typical sAA response to the stress task (i.e., an immediate reaction after the conversation followed by a quick recovery to baseline), whereas adolescents from similar families had difficulty recovering from the conversation. If the children were from a divorced family and the parent was less socially supportive, young adults showed a delayed response in sAA at 20 minutes and recovery, but adolescents with the same circumstances remained stable across the four time points. The results for social support suggest that adolescents and young adults from divorced and married families respond differently when their parents are less socially supportive, with adolescents whose parents are still married potentially having the most difficulty recovering from a stressful interaction.

The results from the present study suggest that when parents are less socially supportive or less communicatively competent, adolescents may have greater difficulty recovering from a stressful task than young adults, at least in terms of sAA. A similar trend was that children from divorced families may experience a more delayed sAA reaction to the discussion task, whereas children whose parents are still married may have a more difficult time recovering from it. In essence, the parents’ marital status and the age of the child are important in determining children’s irregular sAA responses and potential inability to recover from stressful discussions with their parents, but only when parents are less skilled communicatively.

Adolescents and young adults from married parents who felt caught between their parents also had a slightly more difficult time fully recovering in sAA from the discussion task (p < .10), suggesting that these children may have the most difficult time adapting to their parents’ communication patterns. The finding for children’s feelings of being caught, combined with the fact that adolescents from married parents who also had an inability to recover from the discussion task, suggests that some children whose parents are still married who are less skilled may have difficulty adapting to their parents communication patterns. These results provide initial support for previous self-report research that children whose parents remain in a turbulent relationship may have the poorest well-being, because they are unable to escape their parents’ conflict (Amato & Sobolewski, 2001; Booth & Amato, 2001; Jekielek, 1998). Children of divorced parents who have a conflicted relationship probably also have become accustomed to talking about their parents’ relationship because of the divorce, helping them eventually recover from the discussion task even though they tended to have a delayed sAA response.

What was particularly interesting about the findings was that children whose parents were married had an immediate sAA reaction when their parents had less communication skills, while children of divorce, particularly young adults, had a delayed peak sAA reaction at 20 minutes. These findings are noteworthy, because sAA typically peaks immediately following a stressor (Gordis et al., 2006). While purely speculative, one explanation for the delayed reaction for children of divorce may be due to rumination. Children whose parents are divorced may have been mulling over their conversation with their parent after it occurred. Children of divorce often avoid talking about their parents’ relationship in an effort to prevent conflict and protect themselves and their relationships (e.g., Afifi, 2003; Afifi et al., 2008). This avoidance may initially serve an important protective function. However, in their attempt to avoid the conversation and block it out of their mind, the opposite effect might occur. As the research on disclosure suggests, sometimes when people attempt to suppress a thought, it actually encourages more intrusive thoughts, rumination, and thought suppression (Wegner, Schneider, Carter, & White, 1987). Because young adults are often the recipients of their parents’ negative disclosures about each other more often than adolescents (Afifi & Schrodt, 2003), it could be that they are avoiding making the conversation worse, but then begin to ruminate about their parents’ disclosures as time passes. Future research ought to investigate the role of rumination, parents’ communication patterns, and children’s physiological responses.

The practical and theoretical implications of this study must be couched within its limitations. The present investigation was only one conversation with a parent. To truly assess the impact of divorce and parental communication patterns on children, longitudinal research over a greater duration is required. A larger sample is also necessary to increase the power to detect significant differences. We chose to report some results that were “approaching significance” because the results were most likely conservative estimates due to a lack of power. Scholars have long argued that less emphasis should be placed on the arbitrary nature of significance testing with more emphasis placed on effect sizes, the practical importance of the findings, and replication (see Glaser, 1999; Krantz, 1999; Schmidt, 1996). Even though some of the effects in the present study are rather small, the trends are novel, in the hypothesized direction, and practically and theoretically important in terms of how researchers, practitioners, and the lay public perceive the impact of divorce and parents’ communication skills on children.

A final limitation of the present study is the variance in the length of the interactions, which ranged from 10 to 20 minutes. Our study used methods similar to previous studies that have used naturalistic discussion tasks (Laurent & Powers, 2006). Similar to these previous studies, we attempted to enhance the external validity of the study by using a naturalistic conversation rather than an artificial stress task (e.g., Trier tasks). We chose to have the parents and children talk for up to 20 minutes to elevate the cortisol levels. Luckily, the length of the conversation was not a significant control variable and was subsequently removed from the models to preserve power. However, limiting the length of the conversations to 10 or 15 minutes instead of 20 minutes might have enhanced the interpretation of the cortisol results. It is possible, for example, that some participants reached their peak cortisol reaction during the conversation if the conversation was 20 minutes. Nevertheless, the participants should have still been responding to the discussion after it ended and up to 20 minutes, and they all should have recovered from the task by 40 minutes for a normal cortisol trajectory.

As a whole, this study demonstrates that divorce makes a difference in some ways and not in others. Divorce and other contextual differences, such as the age of the child, may become more important when parents are less communicatively skilled. When parents were communicatively skilled, it had a positive, soothing effect for children in general, regardless of divorce or the age of the child. The results also highlight the importance of incorporating a multi-systems approach to the study of psycho-biology. In particular, this study illustrates the influential role that sAA may play in physiological research in the social sciences. As Booth, Johnson and Granger (2005) contend, cutting edge research in the social sciences now involves the use of biological markers as alternatives or additions to traditional models of measuring individual responses and changes in behavior. The integration of sAA into this equation provides a host of new research possibilities.

Footnotes

Declaration of Conflicting Interests

The authors declared the following potential conflicts of interest with respect to the research, authorship, and/or publication of this article: In the interest of full disclosure, we note that DAG is the founder and Chief Scientific and Strategy Advisor at Salimetrics LLC and SalivaBio LLC. These relationships are managed by the policies of the committee on conflict of interest at Johns Hopkins School of Medicine and the Office of Research Integrity and Adherence at Arizona State University

Funding

The authors disclosed receipt of the following financial support for the research, authorship, and/or publication of this article: This research was supported by internal funding from the Institute for Social, Behavioral, and Economic Research, the Faculty Senate, and the Chicano Studies Department at the University of California, Santa Barbara

Author Biographies

Tamara D. Afifi is a Professor in the Department of Communication Studies at the University of Iowa. Her research focuses on communication patterns that foster risk and resiliency in families and other interpersonal relationships, with particular emphasis on: (1) information regulation (privacy, secrets, disclosure, avoidance), and (2) communication processes related to uncertainty, stress, resiliency and communal coping. Her recent research examines the impact that various factors like divorce, economic uncertainty and parents’ communication patterns (e.g., conflict, reoccurring stressful disclosures, social support) have on adolescents’ physiological stress responses.

Douglas A. Granger is a psychoneuroendocrinology researcher who is well known for his development of methods related to saliva collection and analysis and the theoretical and statistical integration of salivary measures into developmental research. At Arizona State University, he is a Foundation Professor in the Department of Psychology. He also holds an Adjunct Faculty position at the Johns Hopkins University School of Nursing, School of Medicine, and Bloomberg School of Public Health. His studies have been instrumental in the conceptualization and analysis of biosocial relationships involving child well-being, parent-child and family relationships, as well as how these biosocial links moderate and mediate the effects of early adversity and stress on children’s adjustment. Dr. Granger is a leading expert engaged in work focused on the discovery, measurement, and application of analytes (hormones, antibodies, chemicals, DNA) in saliva.

Andrea Joseph is a Senior Analyst at Applied Marketing Science in Boston, Massachusetts. She received her PhD from the Department of Communication at the University of California Santa Barbara. She specializes in research methods, statistics and examines social support in interpersonal relationships.

Amanda Denes is an Assistant Professor in the Department of Communication at the University of Connecticut. Her research focuses on communication in various types of interpersonal relationships such as romantic relationships, parent-child relationships, and friendships. Much of her work examines the association between communication in interpersonal relationships and people’s physiological, psychological, and relational health. In particular, she is interested in why individuals disclose information about themselves to others, how they disclose that information, and the effects of such disclosures on individuals and their relationships.

Desiree Aldeis received her PhD from the Department of Communication at the University of California Santa Barbara where she is currently a lecturer. Her research interests reside in the darkside of interpersonal and family communication, with particular emphasis on secrecy in relationships and the conditions in which people reveal secrets in close relationships and their effects.

References

Adam

E. K.

Hoyt

L. T.

Granger

D. A.

(2011). Diurnal alpha amylase patterns in adolescents: Associations with puberty and momentary mood states. Biological Psychology, 88, 170-173. doi:10.1016/j.biopsycho.2011.07.007

Afifi

T. D.

(2003). “Feeling caught” in stepfamilies: Managing boundary turbulence through appropriate privacy coordination rules. Journal of Social and Personal Relationships, 20, 729-756. doi:10.1177/0265407503206002

Afifi

T. D.

Afifi

W. A.

Morse

Hamrick

(2008). Adolescents’ avoidance tendencies and physiological reactions to discussions about their parents’ relationship: Implications for post-divorce and non-divorced families. Communication Monographs, 75, 290-317. Retrieved from https://dx-doi-org.web.bisu.edu.cn/10.1080/03637750802342308

Afifi

T. D.

Granger

Denes

Joseph

Aldeis

(2011). Parents’ communication skills and adolescents’ salivary α-amylase and cortisol response patterns. Communication Monographs, 78, 273-295. Retrieved from https://dx-doi-org.web.bisu.edu.cn/10.1080/03637751.2011.589460

Afifi

T. D.

McManus

(2010). Divorce disclosures and adolescents’ physical and mental health and parental relationship quality. Journal of Divorce & Remarriage, 51, 83-107. Retrieved from https://dx-doi-org.web.bisu.edu.cn/10.1080/10502550903455141

Afifi

T. D.

Schrodt

(2003). “Feeling caught” as a mediator of adolescents’ and young adults’ avoidance and satisfaction with their parents in divorced and non-divorced households. Communication Monographs, 70, 142-173. doi:10.1080/0363775032000133791

Afifi

T. D.

Schrodt

McManus

(2009). The divorce disclosure model (DDM): Why parents disclose negative information about the divorce to their children and its effects. In Afifi

T. D.

Afifi

W. A.

(Eds.), Uncertainty, information management, and disclosure decisions: Theories and Applications (pp. 403-425). Taylor & Francis.

Ainsworth

M. D. S.

Blehar

M. C.

Waters

Wall

(1978). Patterns of attachment: A psychological study of the Strange Situation. Hillsdale, NJ: Lawrence Erlbaum.

Allwood

M. A.

Handwerger

Kivlighan

K. T.

Granger

D. A.

Stroud

L. R.

(2011). Direct and moderating links of salivary alpha-amylase and cortisol stress reactivity to youth behavioral and emotional adjustment. Biological Psychology, 88, 57-64. doi:10.1016/j.biopsycho.2011.06.008

10.

Amato

P. R.

(2000). The consequences of divorce for adults and children. Journal of Marriage and Family, 62, 1269-1287. doi:10.1111/j.1741-3737.2000.01269.x

11.

Amato

P. R.

(2001). Children of divorce in the 1990s: An update of the Amato and Keith (1991) meta analysis. Journal of Family Psychology, 15, 355-370. doi:10.1037//0893-3200.15.3.355

12.

Amato

P. R.

Afifi

T. D.

(2006). Feeling caught between parents: Adult children’s relations with parents and subjective well-being. Journal of Marriage and Family, 68, 222-235. doi:10.1111/j.1741-3737.2006.00243.x

13.

Amato

P. R.

Keith

(1991). Separation from a parent during childhood and adult socioeconomic attainment. Social Forces, 70, 187-206. doi:10.2307/2580068

14.

Amato

P. R.

Sobolewski

J. M.

(2001). The effects of divorce and marital discord on adult children’s psychological well-being. American Sociological Review, 66, 900-921. doi:10.1111/j.1741-3737.2001.01038.x

15.

Bauer

A. M.

Quas

J. A.

Boyce

W. T.

(2002). Associations between physiological reactivity and children’s behavior: Advantages of a multi-system approach. Journal of Developmental & Behavioral Pediatrics, 23, 102-113. doi:10.1097/00004703-200204000-00007

16.

Blair

Granger

D. A.

Willoughby

Kivlighan

K. T.

, & The Family Life Project Investigators. (2006). Maternal sensitivity is related to hypothalamic-pituitary-adrenal stress reactivity and regulation in response to emotion challenge in 6-month-old infants. Annals of the New York Academy of Science, 1094, 263-267. doi:10.1196/annals.1376.031

17.

Bloch

Peleg

Koren

Aner

Klein

(2007). Long-term effects of early parental loss due to divorce on the HPA axis. Hormones and Behavior, 51, 516-523. doi:10.1016/j.yhbeh.2007.01.009

18.

Booth

Amato

P. R.

(2001). Parental predivorce relations and offspring postdivorce well-being. Journal of Marriage and Family, 63, 197–212. doi:10.1111/j.1741-3737.2001.00197.x.

19.

Booth

Johnson

D. R.

Granger

D. A.

(2005). Testosterone, marital quality, and role overload. Journal of Marriage and Family, 67, 483-499. doi:10.1111/j.0022-2445.2005.00130.x

20.

Bowlby

(1982). Attachment theory and its therapeutic implications. Adolescent Psychiatry, 6, 5-33.

21.

Buchanan

C. M.

Maccoby

E. E.

Dornbusch

S. M.

(1991). Caught between parents: Adolescents’ experience in divorced homes. Child Development, 62, 1008-1029. doi:10.2307/1131149

22.

Buchanan

C. M.

Maccoby

E. E.

Dornbusch

S. M.

(1996). Adolescents after divorce. Cambridge, MA: Harvard University Press.

23.

Byrd-Craven

Auer

B. J.

Granger

D. A.

Massey

A. R.

(2012). The father–daughter dance: The relationship between father–daughter relationship quality and daughters’ stress response. Journal of Family Psychology, 26, 87-94. doi:10.1037/a0026588

24.

Chatterton

R. T.

Vogelsong

K. M.

Ellman

A. B.

Hudgens

G. A.

(1996). Salivary α-amylase as a measure of endogenous adrenergic activity. Clinical Physiology, 16, 433-448. Retrieved from https://dx-doi-org.web.bisu.edu.cn/10.1111/j.1475-097X.1996.tb00731.x

25.

Davies

P. T.

Cummings

E. M.

(1994). Marital conflict and child adjustment: An emotional security hypothesis. Psychological Bulletin, 116, 387-411. doi:10.1037//0033-2909.116.3.387

26.

Davies

P. T.

Cummings

E. M.

(1998). Exploring children’s emotional security as a mediator of the link between marital relations and child adjustment. Child Development, 69, 124-139. doi:10.2307/1132075

27.

Davies

Sturge-Apple

Cicchetti

Cummings

E. M.

(2007). The role of child androcorticol functioning in pathways between interparental conflict and child maladjustment. Developmental Psychology, 43, 918-930.

28.

El-Sheikh

Erath

S. A.

Buckhalt

J. A.

Granger

D. A.

Mize

(2008). Cortisol and children’s adjustment: The moderating role of sympathic nervous system activity. Journal of Abnormal Child Psychology, 36, 601-611.

29.

Engert

Vogel

Efanov

S. I.

Duchesne

Corbo

Ali

Pruessner

J. C.

(2011). Investigation into the cross-correlation of salivary cortisol and alpha-amylase responses to psychological stress. Psychoneuroendocrinology, 36, 1294-1302. doi:10.1016/j.psyneuen.2011.02.018

30.

Evans

G. W.

Kim

Ting

A. H.

Tesher

H. B.

Shannis

(2007). Cumulative risk, maternal responsiveness, and allostatic load among young adolescents. Developmental Psychology, 43, 341-351. Retrieved from https://dx-doi-org.web.bisu.edu.cn/10.1037/0012-1649.43.2.341

31.

Floyd

Mikkelson

A. C.

Tafoya

M. A.

Farinelli

La Valley

A. G.

Judd

. . . Wilson

(2007). Human affection exchange: XIII. Affectionate communication accelerates neuroendocrine stress recovery. Health Communication, 22, 123-132. doi:10.1080/10410230701454015

32.

Floyd

Riforgiate

(2009). Affectionate communication received from spouses predicts stress hormone levels in healthy adults. Communication Monographs, 75, 351-368. doi:10.1080/03637750802512371

33.

Gately

D. W.

Schwebel

A. I.

(1991). The challenge model of children’s adjustment to divorce: Explaining favorable post-divorce outcomes in children. Journal of Family Psychology, 5, 60-81. doi:2048/10.1037/0893-3200.5.1.60

34.

Gibson

E. L.

Checkley

Papadopoulos

Poon

Daley

Wardle

(1999). Increased salivary cortisol reactivity induced by a protein-rich midday meal. Psychosomatic Medicine, 61, 214-224.

35.

Glaser

D. N.

(1999). The controversy of significance testing: Misconceptions and alternatives. American Journal of Critical Care, 8, 291-296.

36.

Gordis

E. G.

Granger

Susman

E. J.

Trickett

P. K.

(2006). Asymmetry between salivary cortisol and alpha-amylase reactivity to stress: Relation to aggressive behavior in adolescents. Psychoneuroendocrinology, 31, 976-987. Retrieved from https://dx-doi-org.web.bisu.edu.cn/10.1016/j.psyneuen.2006.05.010

37.

Gordis

E. G.

Granger

Susman

E. J.

Trickett

P. K.

(2008). Salivary alpha amylase-cortisol asymmetry in maltreated youth. Hormones and Behavior, 53, 96-103. doi:10.1016/j.yhbeh.2007.09.002

38.

Granger

D. A.

Blair

Willoughby

Kivlighan

K. T.

Hibel

L. C.

Fortunato

C. K.

Wiegand

L. E.

(2007). Individual differences in salivary cortisol and alpha-amylase in mothers and their infants: Relation to tobacco smoke exposure. Developmental Psychobiology, 10, 692-701. doi:10.1002/dev.20247

39.

Granger

D. A.

Hibel

L. C.

Fortunato

C. K.

Kapelewski

C. H.

(2009). Medication effects on salivary cortisol: Tactics and strategy to minimize impact on behavioral and developmental science. Psychoneuroendocrinology, Retrieved from https://dx-doi-org.web.bisu.edu.cn/10.1016/j.psyneuen.2009.06.017

40.

Granger

D. A.

Kivlighan

K. T.

El-Sheikh

Gordis

E. B.

Stroud

L. R.

(2008). Assessment of salivary α-amylase in biobehavioral research. In Luecken

L. J.

Gallo

L. C.

(Eds.), Handbook of physiological research methods in health psychology (pp. 95-114). Thousand Oaks, CA: Sage.

41.

Granger

D. A.

Serbin

L. A.

Schwartzman

Lehoux

P. E.

Cooperman

Ikeda

(1998). Children’s salivary cortisol, internalising behaviour problems, and family environment: Results from the Concordia Longitudinal Risk Project. International Journal of Behavioral Development, 22, 707-728. Retrieved from https://dx-doi-org.web.bisu.edu.cn/10.1080/016502598384135

42.

Guerrero

L. K.

(1994). “I’m so mad I could scream”: The effects of anger expression on relational satisfaction and communication competence. The Southern Communication Journal, 59, 125-141. doi:10.1080/10417949409372931

43.

Guerrero

L. K.

Farinelli

McEwan

(2009). Attachment and relational satisfaction: The mediating effect of emotional communication. Communication Monographs, 76, 487-514.

44.

Hazan

Shaver

P. R.

(1994b). Deeper into attachment theory: Reply to commentaries. Psychological Inquiry, 5, 68-79.

45.

Henry

J. P.

(1992). Biological basis of the stress response. Integrative Physiological & Behavioral Science, 27, 66-83. doi:10.1007/BF02691093

46.

Hetherington

E. M.

(1999). Family functioning and the adjustment of adolescent siblings in diverse types of families. In Hetherington

E. M.

Henderson

S. H.

Reiss

(Eds.), Adolescent siblings in stepfamilies: Family functioning and adolescent adjustment (pp. 1-25). Ann Arbor, MI: Monographs of the Society for Research in Child Development.

47.

Jang

S. A.

Smith

Levine

(2002). To stay or to leave? The role of attachment styles in communication patterns and potential termination of romantic relationships following discovery of deception. Communication Monographs, 69, 236-252.

48.

Jekielek

(1998). Parental conflict, marital disruption and children’s emotional well-being. Social Forces, 76, 905-936. Retrieved from https://dx-doi-org.web.bisu.edu.cn/10.2307/3005698

49.

Kraft

A. J.

Luecken

L. J.

(2009). Childhood parental divorce and cortisol in young adulthood: Evidence for mediation by family income. Psychoneuroendocrinology, 34, 1363-1369. doi:10.2307/3005698

50.

Krantz

D. H.

(1999). The null hypothesis testing controversy in psychology. Journal of American Statistical Association, 94, 1372-1381. Retrieved from https://dx-doi-org.web.bisu.edu.cn/10.2307/2669949

51.

Laurent

H. K.

Ablow

J. C.

Measelle

(2012). Taking stress response out of the box: Stability, discontinuity, and temperament effects on HPA and SNS across social stressors in mother-infant dyads. Developmental Psychology, 48, 35-45. doi:10.1037/a0025518

52.

Laurent

H. K.

Powers

(2006). Social-cognitive predictors of hypothalamic-pituitary-adrenal reactivity to interpersonal conflict in emerging adult couples. Journal of Social and Personal Relationships, 23, 703-720. Retrieved from https://dx-doi-org.web.bisu.edu.cn/10.1177/0265407506065991

53.

Luecken

Appelhans

(2006). Early parental loss and salivary cortisol and young adulthood: The moderating role of family environment. Development and Psychopathology, 18, 295-308. doi:10.1017/S0954579406060160

54.

Luecken

L. J.

Kraft

Appelhans

B. M.

Enders

(2009). Emotional and cardiovascular sensitization to daily stress following childhood parental loss. Developmental Psychology, 45, 296-302. doi:10.1037/a0013888

55.

Luecken

L. J.

Kraft

Hagan

M. J.

(2009). Negative relationships in the family-of-origin predict attenuated cortisol in emerging adults. Hormones and Behavior, 55, 412-417. doi:10.1016/j.yhbeh.2008.12.007

56.

Luecken

L. J.

Rodriguez

A. P.

Bradley

Applehans

M. A.

(2005). Cardiovascular stress responses in young adulthood associated with family-of-origin relationship experiences. Psychosomatic Medicine, 67, 514-521. doi:10.1097/01.psy.0000160466.10397.18

57.

McCarthy

C. J.

Lambert

R. G.

Moller

(2006). Preventive resources and emotion regulation expectancies as mediators between attachment and college students’ stress outcomes. International Journal of Stress Management, 13(1), 1-22.

58.

McEwen

B. S.

(1998). Protective and damaging effects of stress mediators. New England Journal of Medicine, 338, 171-179. Retrieved from https://dx-doi-org.web.bisu.edu.cn/10.1016/S0079-6123(08)62128-7

59.

McEwen

B. S.

(2001). Plasticity of the hippocampus: Adaptation to chronic stress and allostatic load. Annals of the New York Academy of Science, 933, 265-277. doi:10.1111/j.1749-6632.2001.tb05830.x

60.

Meinlschmidt

Heim

(2005). Decreased cortisol awakening response after early loss experience. Psychoneuroendocrinology, 30, 568-576. doi:10.1016/j.psyneuen.2005.01.006

61.

Nater

Rohleder

(2009). Salivary alpha-amylase as a non-invasive biomarker for the sympathetic nervous system: Current state of the research. Psychoneuroendocrinology, 34, 486-496. doi:10.1016/j.psyneuen.2009.01.014

62.

Pendry

Adam

(2007). Associations between parents’ marital functioning, maternal parenting quality, maternal emotion, and child cortisol levels. International Journal of Behavioral Development, 31, 218-231. Retrieved from https://dx-doi-org.web.bisu.edu.cn/10.1177/0165025407074634

63.

Priem

J. S.

Solomon

D. H.

(2013). The illocutionary force of supportive communication: The effects of message features on physiological stress recovery. Manuscript submitted for publication.

64.

Rohleder

Nater

U. M.

Wolf

U. E.

Ehlert

Kirshbaum

(2004). Psycho-social stress-induced activation of salivary alpha-amylase. Annual New York Academy of Sciences, 1032, 258-263. doi:10.1196/annals.1314.033

65.

Schmidt

F. L.

(1996). Statistical significance testing and cumulative knowledge in psychology: Implications for training of researchers. Psychological Methods, 2, 115-129. Retrieved from https://dx-doi-org.web.bisu.edu.cn/10.1037//1082-989X.1.2.115

66.

Spitzberg

B. H.

Canary

D. J.

Cupach

W. R.

(1994). A competence-based approach to the study of interpersonal conflict. In Cahn

(Ed.), Conflict in personal relationships (pp. 183-202). Hillsdale, NJ: Lawrence Erlbaum.

67.

Trees

A. R.

(2000). Nonverbal communication and the support process: Interactional sensitivity in interactions between mothers and young adult children. Communication Monographs, 67, 239-261. doi:10.1080/03637750009376509

68.

Tyrka

Wier

Price

Ross

Anderson

Wilkinson

Carpenter

(2008). Childhood parental loss and adult hypothalamic-pituitary-adrenal function. Biological Psychiatry, 63, 1147-1154. doi:10.1016/j.biopsych.2008.01.011

69.

Van Veen

J. F.

van Vliet

I. M.

DeRijk

R. H.

van Pelt

Mertens

Zitman

F. G.

(2008). Elevated alpha-amylase but not cortisol in generalized social anxiety disorder. Psychoneuroendocrinology, 33, 1313-1321. Retrieved from https://dx-doi-org.web.bisu.edu.cn/10.1016/j.psyneuen.2008.07.004

70.

Wegner

D. M.

Schneider

D. J.

Carter

S. R.

White

T. L.

(1987). Paradoxical effects of thought suppression. Journal of Personality and Social Psychology, 53, 5-13. doi:10.1037//0022-3514.53.1.5