Dyadic perceptions of depression and anxiety in romantic relationships

Abstract

This study investigated dyadic perceptions of the causes of mental health symptoms within dating couples. We tested the extent to which a participant’s beliefs about their romantic partner’s attributions about mental health were accurate and/or biased and the extent to which causal attributions about mental health predicted relationship quality. Ninety-four dyads (N = 188) reported on biological and psychological causes of mental health symptoms and relationship conflict and support. Consistent with hypotheses, the actor–partner interdependence model revealed more bias than accuracy in perceptions of the causes of mental health symptoms. Results also showed more effects of causal attributions on relationship quality for the romantic partner than for the participant with symptoms and more effects of controllable psychological causes on relationship quality relative to uncontrollable psychological or any biological causes. While relationship quality was generally positive, endorsement of most causes of mental health symptoms predicted higher conflict and lower support. These accuracy and bias effects as well as the effects on the romantic partner affirm the importance of studying both members of the romantic relationship because they would have been missed completely if only studying the participant with mental health symptoms. In terms of clinical practice, working with couples whose attributions align might look differently than working with couples who are not so aligned.

Keywords

Actor–partner interdependence model anxiety causal attributions depression relationship quality

Depression has been well-studied with respect to its prevalence, symptomology, and personal impact (Anxiety and Depression Association of America, 2014). Even more prevalent, anxiety disorders are currently the most common mental health issue in the U.S., affecting roughly 40 million American adults (Anxiety and Depression Association of America, 2014). Despite thorough investigation of the individual experience of depression and anxiety, it is equally important to understand the interpersonal nature of these disorders.

Negative interpersonal effects of depression and anxiety have been well-documented in marriages (e.g., Benazon, 2000; Kessler, Walters, & Forthofer, 1998; McLeod, 1994; Whisman, 2001; Zaider, Heimberg, & Iida, 2010) and to a lesser degree in dating relationships. In one sample of dating couples, Segrin, Powell, Givertz, and Brackin (2003) found that depressive symptoms were associated with lower relationship quality, which was then associated with loneliness. Katz, Beach, and Joiner (1999) found that people in dating relationships with individuals with depression were at a higher risk of developing depression themselves, regardless of relationship quality. Depressive symptoms were especially contagious when the person with depression burdened his or her partner with frequent prompts for reassurance, again supporting prior models of depression (Katz, Beach, & Joiner, 1999; Stewart & Harkness, 2015). Similarly, Porter and Chambless (2014) found that social anxiety was inversely related to perceptions of intimacy for both male and female undergraduates. When social anxiety was high, relationship intimacy in general was seen as riskier and the relationship itself was seen as less intimate. Women high in social anxiety (relative to those with lower social anxiety) reported giving, getting, and desiring less support, being less satisfied in the relationship, and disclosing less to their partners (Porter & Chambless, 2014). Many studies control for overall relationship quality before modeling the influences of depression and/or anxiety; thus, being in a low-quality relationship is likely not the sole explanation for the relational problems that many people with depression and/or anxiety experience. Consistent with systemic couples’ therapy (Leff et al., 2000) where “close relationships are regarded as both influencing, and being influenced by, the patient and his/her symptoms” (p. 96), attributions of depression or anxiety could be made by one or both members of the couple thereby influencing the experience of the relationship for one of both of them.

Interpersonal models of depression and anxiety have found broad support to help explain the negative interpersonal effects of depression and anxiety. Joiner, Alfano, and Metalsky’s (1993) integrative theory of depression suggested that people with depression may unknowingly use a fluctuating approach toward interacting with others, oscillating between pursuits of positive, self-enhancing feedback and negative, self-consistent information (see also Swann, Griffin, Predmore, & Gaines, 1987). People with depression seek reassurance as to whether others truly care about them, but this self-enhancing information may only satisfy their discomfort temporarily (Coyne, 1976). Once the person processes the positive feedback, he or she realizes its incompatibility with a negative self-concept. Doubtful, the person with depression then searches for negative, self-consistent information, and the pattern continues at his or her detriment. Joiner et al. (1993) found that these information-seeking behaviors, combined with the presence of depression, predicted a higher risk of being negatively evaluated by others. This interpersonal style may be confusing or aggravating for social network members of a person with depression, which might explain why these factors are associated with rejection (Joiner, Alfano, & Metalsky, 1993; Stewart & Harkness, 2015). With regard to anxiety, Hofmann (2014) claims that anxious individuals use people in their lives to help manage their irrational worries. Specifically, these “safety people” (p. 489) are those who create feelings of safety and reassurance to mitigate the emotional discomfort that a person with an anxiety disorder experiences. This allows the anxious person to avoid dealing with their anxiety directly, thus helping to maintain the symptoms of the disorder (Hofmann, 2014). Aside from seeking reassurance from others for unfounded worries, people with anxiety tend to engage in behaviors which negatively impact others around them.

Depression and anxiety influence interpersonal dynamics across a number of close relationships, including friendships, roommate relationships, and romantic relationships. Starr and Davila (2008) examined excessive reassurance seeking, interpersonal rejection, and depression across 38 studies in a meta-analysis of relational difficulty for those with depression. The authors suggested that not only do people with depression experience the intrapersonal difficulties associated with the disorder, but they also impact others in a reciprocal and meaningful way (Starr & Davila, 2008). Of note, the researchers found a stronger reciprocal effect for romantic relationships compared to nonromantic relationships (e.g., roommates). The romantic relationship (which includes marriage and dating relationships) tends to be the most significant interpersonal connection for adults relative to other nonromantic relationships (Doyle & Molix, 2014). Dyadic literature examining mental illness remains wanting, especially that which addresses romantic couples who must inevitably navigate symptoms of mental illness collectively. By dyadic literature, we are specifically referring to research which includes data from both members of the couple or dyad.

While the simple presence of mental health symptoms can impact relationships, it is likely that the beliefs about the onset of those symptoms are also consequential in romantic relationships. Weiner’s (1985) attributional theory posits that people develop beliefs about the cause of an individual’s behavior based upon multiple dimensions. Locus of causality refers to the internal or external nature of the behavior, controllability refers to whether or not the person is responsible for behavior, and stability refers to how much the cause of the behavior can change over time. Weiner, Perry, and Magnusson (1988) applied the original attribution model to stigmas of physical and mental health. They found that stigmas which were seen as controllable (e.g., AIDS) elicited feelings of blame and anger, but conditions viewed as uncontrollable (e.g., Alzheimer’s disease) prompted feelings of pity and liking. Cronan, Key, and Vaughn (2016) expanded those stigmas to include depression and anxiety. While most stigmas are dichotomized in terms of physical or mental health (i.e., biological or psychological), Cronan and colleagues found that depression and anxiety clustered in the midrange for both controllability and stability with other physical stigmas like heart disease and stroke. Furthermore, when additional information was presented suggesting the target was more responsible for their condition (i.e., higher controllability), feelings of blame and anger were elicited. This suggests that the perception of the onset of mental health symptoms could be influential in romantic relationships in that it could predict relationship satisfaction.

Blais and Renshaw (2014) studied perceptions of depression within people who were in dating relationships by asking the participant with depressive symptoms what they thought their romantic partner thought was the cause of their symptoms. Beliefs that their partner thought their depression had a controllable, psychological source (e.g., event interpretation) were associated with higher perceived relationship conflict and lower perceived relationship support while beliefs that depression had a controllable, biological source (e.g., diet) were associated with higher perceived conflict. One noteworthy limitation was that Blais and Renshaw’s (2014) study was not dyadic. They did not collect data from the corresponding romantic partner so there was no way of confirming whether or not the romantic partner of the individual with depressive symptoms actually held those beliefs about the source of their partner’s symptoms. Previous research on perception within couples has shown less accuracy and more bias (or projection; Agnew, Loving, & Drigotas, 2001; Kenny & Acitelli, 2001; MacDonald & Ross, 1999) but these studies have not specifically addressed attributions of mental health symptoms. Kenny and Acitelli (2001) found evidence for both bias and accuracy across a number of perceptions; however, the bias effects were stronger especially for variables related to the relationship (as opposed to variables outside of the relationship like job satisfaction). An additional limitation to the work of Blais and Renshaw (2014) was that, because the study was not dyadic, it is impossible to measure the effects of attributions made by the participant with the depressive symptoms or their romantic partner on either the participant with depressive symptoms or the romantic partner. Previous research has shown that negative attributions (e.g., blame, responsibility) have negative effects on relationship outcomes (Furman, Luo, & Pond, 2017). It is prudent to study the dyadic effects of mental health attributions within romantic relationships because assuming these associations exist, there are areas of potential clinical intervention for working with people with mental health symptoms.

Therefore, the present study examined perceptions of mental health symptoms in dating couples, investigating attributions made by both members of the dyad. Participants with symptoms of depression and/or anxiety reported what they thought was the cause of their symptoms and what they thought their romantic partners thought was the cause of their symptoms. To explore the accuracy and bias (or projection) of these ratings, we also asked their romantic partners what they thought was the cause of the participants’ symptoms. We used all three pieces of information to predict relationship conflict and support using the actor–partner interdependence model (APIM; Kenny, Kashy, & Cook, 2006). Because the two members of a dating couple are not simply two independent individuals but instead are part of a dyad/couple, the APIM takes into account the nonindependence of their data (Kenny et al., 2006). Since no study (to date) has explored the accuracy or bias of causal attributions of depressive and/or anxiety symptoms within romantic relationships, it was hypothesized that more bias would be found in these ratings relative to accuracy (Agnew et al., 2001; Kenny & Acitelli, 2001; MacDonald & Ross, 1999). Consistent with previous research (Blais & Renshaw, 2014), it was also hypothesized that beliefs that depressive and/or anxiety symptoms had controllable (biological and psychological) causes would be related to higher conflict and lower support for both members of the dyad (in terms of actors effects; a description of actor and partner effects is found in the “Analytic strategy” section).

Method

Participants

The sample consisted of 94 dating couples (total N = 188). Undergraduates who met prescreen criteria were recruited from the psychology and communication participant pools at a large southwestern public university. In order to be eligible for participation, students had to (1) be at least 18 years old, (2) report at least mild depressive or anxiety symptoms (or both), and (3) be in a current romantic relationship where their romantic partner was also willing to participate in the study. Sixteen couples were in long-distance relationships (i.e., ones in which romantic partners were willing to participate but could not physically come to the lab with the participant). These long-distance couples did not differ from in-person couples on any variables of interest or demographics, so they were combined for all analyses.

The participants were primarily female (85%) and had an average age of 19.70 years (SD = 2.26) while the romantic partners were primarily male (83%) and had an average age of 20.09 years (SD = 2.45). Participants reported being in their current romantic relationships for an average of 81 weeks or a little over a year and a half, and these reports were consistent across both members of the couple (r = .96, p < .001). Participants identified as White (40%), Asian American/Pacific Islander (29%), Latino American (19%), and African American (6%), while romantic partners identified as White (47%), Latino/Hispanic (19%), Asian American/Pacific Islander (16%), and biracial/multiracial (8%). More than half of the dyads (58 or 62%) were of the same ethnic group while 36 (38%) were from different ethnic groups. Ethnic composition of the dyad (matched vs. mismatched) did not predict any variables of interest and was therefore not included in any model. The majority of the participants were in heterosexual relationships (95%). One gay couple and four lesbian couples participated in the current study. These same-sex couples did not differ from heterosexual couples on variables of interest or other demographics, so they were combined for all analyses.

Materials

The 21-item Depression Anxiety Stress Scale (DASS; Lovibond & Lovibond, 1995) was used as a prescreen measure to identify participants who reported at least mild depressive or anxiety symptoms (or both). The DASS has three subscales (each containing 7 items) to measure states of depression, anxiety, and stress in the last week, yet only the depression and anxiety subscales were considered in the current study. Responses were indicated on a 4-point scale (0 = did not apply to me at all, 3 = applied to me very much). The 7 items of each subscale were summed to provide a total score with higher scores representing more symptoms of depression or anxiety. A sample item of the depression subscale included: “I couldn’t seem to get any enjoyment out of the things I did.” A sample item of the anxiety subscale included: “I had a feeling of shakiness (e.g., legs going to give way).” To be eligible for the study, participants must have scored between 5 points and 21 points on the depression subscale (mild to severe depressive symptoms) or between 4 points and 21 points on the anxiety subscale (mild to severe anxiety symptoms) (Blais & Renshaw, 2014; Lovibond & Lovibond, 1995). They were considered in the comorbid symptoms group if they met both cutoff scores. In the present sample, the depression subscale was reliable (α = .81) while the anxiety subscale had slightly lower reliability (α = .67). Respondents to the depression subscale reported an average score of 8.15 (SD = 3.63). Respondents to the anxiety subscale reported an average score of 7.58 (SD = 3.20). Since the DASS was part of the prescreen materials, it should be noted that we did not assess depression or anxiety symptoms in the romantic partners.

The 13-item Biological and Psychological Attribution Scale for Depression (BPASD; Blais & Renshaw, 2012) asked respondents to indicate the extent to which they believed certain factors caused their depressive (and/or anxiety) symptoms on a 5-point scale (1 = not at all, 5 = a great deal). Six items assessed biological causes of a person’s depressive (or anxiety) symptoms, including three which were considered controllable (i.e., diet, exercise, alcohol/drugs) and three which were considered uncontrollable (i.e., chemical imbalance, hormone changes, genetic predisposition). Seven items assessed psychological causes of a person’s depressive (or anxiety) symptoms, including four which were considered controllable (e.g., personal inadequacies, personality, behavior, event interpretation) and three which were considered uncontrollable (i.e., family environment, upbringing, situational factors). Similar to Blais and Renshaw (2012), participants first responded to his or her self-view (how they viewed the cause of their own depressive and/or anxiety symptoms), and then their beliefs about their romantic partner’s view of their symptoms (how they thought their romantic partners viewed the cause of their depressive and/or anxiety symptoms). The romantic partners responded to how they viewed the cause of their partner’s depressive and/or anxiety symptoms.

The 25-item Quality of Relationships Inventory (Pierce, Sarason, & Sarason, 1991) was used to assess individual perceptions of relationship quality as it pertained to a specific relationship (in this case, the romantic dating relationship). The scale was divided into subscales based on perceptions of conflict, social support, and depth with all responses rated on a 4-point scale (1 = not at all, 4 = very much). The conflict subscale contained 12 items such as: “How angry does this person make you feel?” The perceived social support subscale contained 7 items such as: “To what extent could you turn to this person for advice about problems?” The depth subscale was not used in the present study. Subscale items were averaged with higher scores representing more conflict and more social support. In the present sample, the conflict subscale was reliable (participants: α = .86, romantic partners: α = .87) while the social support subscale had slightly lower reliability (participants: α = .72, romantic partners: α = .68).

Procedure

The survey was created using Qualtrics software (Provo, Utah, USA; Qualtrics, 2018) tailored specifically for each type of mental health symptom: depression, anxiety, or both (comorbid). Upon arrival to the lab, participants and their romantic partners went through the informed consent process with a researcher and completed the survey on a computer. Participants were in separate cubicles to ensure confidentiality of responses. Long-distance romantic partners were recruited by the participants (no initial contact from researchers) and participated via a private survey link (sent to them via e-mail from the researcher) while the participants came to the lab alone. Once both dyad members completed the survey, they were thanked and compensated for their participation. Participants were given partial course credit for participation (1½ hr of credit in the SONA participant pool system), while their romantic partners were paid US$10 in cash for their participation. Long-distance romantic partners participated so that the participants could earn the course credit and none of them opted for a long-distance compensation (i.e., none of them asked us to mail their compensation payment although it was offered).

Results

Descriptive statistics

There were no differences found between participants with depressive symptoms (n = 19), participants with anxiety symptoms (n = 44), and participants with both depressive and anxiety symptoms (n = 31) on ratings of relationship support or conflict, beliefs about the causes of mental health symptoms, or any demographic variables (ps range from .088 to .981; exact results available from the corresponding author by request). Therefore, the dyadic models described below were run on all 94 couples.

See Table 1 for means, standard deviations, and measures of nonindependence for all variables of interest for the participants and their romantic partners. Participants and romantic partners did not differ on relationship conflict or support, and overall, they reported more support than conflict in their relationships. Participants and romantic partners only differed on one group of causal attributions—that is, controllable psychological causes. Overall, both participants and romantic partners endorsed similar patterns of causal attributions for the participants’ symptoms. Specifically, they endorsed uncontrollable psychological causes the most, followed by controllable psychological causes, followed by uncontrollable biological causes, and finally controllable biological causes. Table 2 presents the correlations among all variables within the participants (below the diagonal) and within the romantic partners (above the diagonal), while Table 3 presents the correlations between the participant and the romantic partner.

Table 1.

Descriptive statistics (including measures of nonindependence) comparing relationship quality and causal attributions between participants and romantic partners.

	Participant, M (SD)	Romantic partner, M (SD)	t	r
QRI conflict	1.92 (.52)	1.99 (.52)	−1.33	.47***
QRI support	3.56 (.40)	3.63 (.35)	−1.34	.20*
Controllable biological	2.07 (.87)	1.97 (.74)	1.03	.31**
Uncontrollable biological	2.39 (.86)	2.30 (.90)	0.81	.29**
Controllable psychological	2.96 (.85)	2.61 (.88)	2.75**	−.03
Uncontrollable psychological	3.04 (.90)	3.06 (.93)	−0.16	.26*

Note. M = mean; SD = standard deviation; t = paired-samples t-test; r = Pearson correlation as a measure of nonidependence between both members of the dyad; QRI = Quality of Relationships Index. QRI scores range from 1 to 4. All other attribution scores range from 1 to 5.

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

Table 2.

Within-partner correlations among measures of communication privacy management and relationships quality for women (below diagonal) and for men (above diagonal).

	1	2	3	4	5	6
1. CPM boundary permeability		.26*	−.13	−.13	.41**	.43**
2. CPM ownership/boundary control	.48**		.04	−.29**	.35**	.20
3. CPM boundary linkages	.11	−.01		.11	.09	−.02
4. QRI conflict	−.31**	−.26*	.38**		−.45**	−.09
5. QRI support	.44**	.33**	.02	−.35**		.53**
6. QRI depth	.48**	.30**	.17	−.04	.52**

Note. CPM = Communication Privacy Management-Romantic Relationships scale; QRI = Quality of Relationships Index.

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

Table 3.

Between-partner correlations among measures of relationships quality and causal attributions of mental health symptoms.

	Romantic partner
Participant	1	2	3	4	5	6
1. QRI conflict	.47**	−.30**	−.06	−.05	−.07	.03
2. QRI support	−.18	.20	.039	−.01	.14	−.02
3. Controllable biological	.13	−.03	.31**	.15	−.01	.04
4. Uncontrollable biological	.02	−.19	.11	.29**	.12	.05
5. Controllable psychological	−.01	−.18	.00	.02	−.03	.08
6. Uncontrollable psychological	.10	−.13	−.05	.12	−.15	.26*

Note. QRI = Quality of Relationships Index.

*p < .05; **p < .01.

Analytic strategy

Dyadic effects were explored using the APIM in Mplus (version 8; Muthén & Muthén, 2017). Generally speaking, an actor effect occurs when one person’s predictor variable predicts that same person’s outcome variable. A partner effect occurs when a partner’s predictor variable predicts the participant’s outcome variable (Kenny et al., 2006). Specific to the present study, if a woman has symptoms of anxiety, an actor effect exists if her causal attributions about the source of her anxiety predict her perceptions of relationship conflict. A partner effect exists if her romantic partner’s causal attributions about the cause of her anxiety predict her perceptions of relationship conflict.

In the current study, we were interested in whether perceptions of causal attributions of mental health symptoms (from either member of the dyad) predicted relationship quality, so we included (1) participant’s perceptions of what they thought was the cause of their symptoms, (2) romantic partner’s perceptions of what they thought was the cause of participant’s symptoms, and (3) participant’s perceptions of what they thought their romantic partners thought was the cause of their symptoms (see Figure 1). This allowed not only for traditional actor and partner effects but we could also explore how participant’s perceptions of what they thought their romantic partners thought was the cause of their symptoms also predicted relationship quality.

Figure 1.

Actor–partner interdependence model of symptom causes and relationship quality.

Furthermore, we were also interested in the accuracy and bias of those ratings (West & Kenny, 2011). A full accuracy-bias model was not possible to run since we did not ask romantic partners to guess what they thought the participants viewed as the cause of their own symptoms. However, we could explore the accuracy of romantic partner’s ratings. That is, the degree to which the participant’s ratings of how they thought their romantic partners viewed the cause of their symptoms actually mapped onto how the romantic partners actually rated the cause of the participants symptoms. Similarly, we could explore the bias of the participant’s ratings. That is, the degree to which the participant’s ratings of how they viewed the cause of their own symptoms actually mapped onto their perceptions of how they thought their romantic partners viewed the cause of their symptoms. Figure 1 shows an example of one model. Each model had a total of eight paths: two actor effects, two partner effects, one accuracy effect, one bias effect, and two effects of how the participant thought their romantic partners viewed the cause of their symptoms on relationship quality. Four models were run with relationship conflict as the outcome variable while four models were run with relationship support as the outcome variable (eight models total).

Actor–partner interdependence models

Accuracy and bias

Unsurprisingly, bias was found for all four types of causes (Bs ranged from .60 to .77, all ps < .0001; West & Kenny, 2011). For example, if the participant thought their mental health symptoms were due to controllable psychological causes, they also expected their romantic partner to endorse controllable psychological explanations as the cause of their mental health symptoms. Accuracy was found for controllable biological causes (B = .27, p = .004). If the romantic partner thought the participant’s mental health symptoms were due to controllable biological causes, the participant also thought that their romantic partner thought their mental health symptoms were due to controllable causes. Another marginal accuracy effect is worth noting for controllable psychological causes (B = .13, p = .068). If the romantic partner thought the participant’s mental health symptoms were due to controllable psychological causes, the participant also thought that their romantic partner thought their mental health symptoms were due to controllable psychological causes.

Conflict

Results of the APIM analyses examining perceptions of the causes of mental health symptoms on conflict are displayed in the top half of Table 4 (Models 1 to 4). Actor effects were found for all four types of causes for the romantic partner. Specifically, the more the romantic partner endorsed any biological (controllable: p = .002; uncontrollable: p < .001) or psychological (controllable: p < .001; or uncontrollable: p < .001) causes of mental health symptoms in the participant, the more conflict the romantic partner reported in the relationship. There were no partner effects for either the participant or the romantic partner. Finally, there were three significant effects of the participant’s perceptions of what they thought their romantic partner thought was the cause of their symptoms. Specifically, the more the participant perceived that their romantic partner endorsed controllable psychological causes of their mental health symptoms, the more conflict the participant and romantic partner reported in the relationship (ps = .026 and .017, respectively). Similarly, the more the participant perceived that their romantic partner endorsed uncontrollable psychological causes of their mental health symptoms, the more conflict the participant reported in the relationship (p = .013).

Table 4.

Actor effects, partner effects, and effects of participant’s perceptions of romantic partners’ ratings in the relationship between symptom cause and relationship quality for participants and their romantic partners.

	Participant						Romantic partner
	Actor effect		Partner effect		P’s perceptions of RP effect		Actor effect		Partner effect		P’s perceptions of RP effect
	B	SE	B	SE	B	SE	B	SE	B	SE	B	SE
Conflict
1. Controllable biological	.07	.097	−.08	.086	.08	.094	.25**	.081	−.03	.092	.04	.089
2. Uncontrollable biological	.05	.088	.04	.068	.01	.086	.26***	.062	−.11	.078	.04	.078
3. Controllable psychological	−.02	.098	−.04	.064	.22*	.097	.27***	.057	−.15	.086	.21*	.087
4. Uncontrollable psychological	−.07	.094	−.03	.063	.21*	.084	.23***	.059	−.02	.089	.04	.080
Support
5. Controllable biological	−.15*	.074	.01	.065	.13	.071	−.08	.056	.01	.064	.01	.062
6. Uncontrollable biological	−.07	.066	−.01	.052	.06	.065	−.12**	.043	−.03	.054	.00	.054
7. Controllable psychological	.02	.077	.06	.051	.00	.077	−.13***	.039	.06	.058	−.17**	.060
8. Uncontrollable psychological	.01	.075	−.01	.050	−.03	.067	−.03	.043	.00	.064	−.02	.057

Note. P’s perceptions of RP = participant’s guess of what their romantic partner thought was the source of their mental health symptoms. Covariates (age_P, gender_P, age_RP, gender_RP) are included in all models.

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

Support

Results of the APIM analyses examining perceptions of the causes of mental health symptoms on support are displayed in the bottom half of Table 4 (Models 5 to 8). One actor effect was found for controllable biological causes for the participant. Specifically, the less the participant endorsed controllable biological causes (e.g., chemical imbalance) of mental health symptoms, the more support the participant reported in the relationship (p = .046). Actor effects were also found for two of the four types of causes for the romantic partner. Specifically, the less the romantic partner endorsed an uncontrollable biological cause or controllable psychological causes of mental health symptoms in the participant, the less support the romantic partner reported in the relationship (ps = .005 and .001, respectively). There were no partner effects for either the participant or the romantic partner. Finally, there was one significant effect of the participant’s perceptions of what they thought their romantic partner thought was the cause of their symptoms on the romantic partner. Specifically, the less the participant perceived that their romantic partner endorsed controllable psychological causes of their mental health symptoms, the more support the romantic partner reported in the relationship (p = .005).

Discussion

The purpose of the present study was to examine the perceptions of the causes of mental health symptoms and their effects on relationship quality within dating couples. More specifically, we examined the perception of the causes of depressive and anxiety symptoms from the participant with said symptoms: what they thought was the cause of their symptoms and what they thought their romantic partner thought was the cause of their symptoms. We explored the level of accuracy and bias of the latter by obtaining the romantic partner’s perception of the cause of the participant’s symptoms. Finally, we observed how each of these three ratings related to perceptions of relationship conflict and social support by both members of the dyad using the APIM. Overall, we found some noteworthy trends: (1) there was more bias (projection) than accuracy for the participant with mental health symptoms in attributions perceptions of mental health symptoms, (2) there were more actor effects than any other effects (i.e., partner effects, effects of the participant’s perceptions of their romantic partner’s ratings), (3) there were more effects of attributions on relationship quality for the romantic partner than the participant, and (4) results were more consistently found for controllable psychological causes on relationship quality than any others (i.e., uncontrollable psychological, controllable, and uncontrollable biological).

One unique contribution of the current study was to explore accuracy and bias in romantic couples’ perceptions of the sources of depressive and/or anxiety symptoms. As expected, the findings showed a high degree of perceptual bias or projection such that participants assumed similarity between their romantic partner’s responses and their own for all causes. These bias effects suggest that depression and anxiety share characteristics of individual and social dysfunction, providing support for suggestions that depression and anxiety are similar manifestations of underlying neuroticism (Hettema, Neale, Myers, Prescott, & Kendler, 2006; Zinbarg et al., 2016). Accuracy was found for controllable biological causes suggesting that most people (in general) know the role that diet, exercise, and substances play on mental health symptoms (i.e., actual accuracy; West & Kenny, 2011). That accuracy trends were also found for controllable psychological causes imply that people also recognize (via cues) the role that things such as personal inadequacies, personality, behavior, and event interpretation play on mental health symptoms (i.e., indirect accuracy; West & Kenny, 2011). These findings also suggest avenues for clinical interventions. That is, working with couples where one person views the cause of their mental health symptoms in a similar way as their romantic partner might look differently than working with couples who are not so aligned.

A discussion of actor effects reflects much of what has been traditionally explored in the literature, but the interesting nuance here is that the majority of the actor effects were found for the romantic partner. By and large, one’s greater belief in biological and psychological causes of mental health symptoms was related to one’s own lower ratings of relationship support and higher ratings of relationship conflict. This is consistent with prior work on mental illness in that the mere presence of depressive and/or anxiety symptoms has often been linked with perceptions of low relationship quality from the perspectives of one or both partners (Benazon, 2000; Blais & Renshaw, 2014; McLeod, 1994; Porter & Chambless, 2014; Whisman, 2001; Zaider et al., 2010).

It is worth mentioning that the participant’s perceptions of their romantic partner’s view of their symptoms were predictive of relationship quality in a few instances. These perceptions were predictive of relationship conflict and support when they were controllable psychological in nature. If a person with mental health symptoms thinks their romantic partner thinks it is psychological (i.e., “in their head”) and controllable, they might also expect their partner to think that they can simply change which would lead to conflict within the relationship because they might not think this change is possible. Because public stigma tends to portray sufferers of mental health symptoms as responsible for their problems (Corrigan & Watson, 2002; Weiner, Perry, & Magnusson, 1988), participants might logically anticipate judgments of responsibility from their romantic partners as to the causes of their symptoms, despite their own self-endorsements of uncontrollability.

While it is worth mentioning that absolute levels of relationship conflict and support were equivalent across the dyad, there were more effects of causal attributions on relationship quality for the romantic partner compared to the participant. In general, the more the romantic partner endorsed any class of attribution (biological or psychological, controllable or uncontrollable), the worse the relationship quality. Given the sternness of mental illness stigma, it is possible that romantic partners viewed many explanatory offerings as self-indulgent, or that symptoms were seen as minute barriers to simply cheering up in the case of depression (Martinez, Piff, Mendoza-Denton, & Hinshaw, 2011) or calming down in the case of anxiety. It may be that partners have difficulty understanding ongoing and fluctuating mental health symptoms if they lack personal familiarity not allowing attributional empathy toward their partners. If they have never experienced depressive or anxiety symptoms, efforts to understand symptoms (i.e., considering a variety of sources) may be bewildering and end in discontent which is consistent with the findings that depressive and anxiety symptoms appeared to take a toll on romantic partner’s relationship quality in the present study. These findings are consistent with previous dyadic research assessing the romantic partner’s experience of being in a relationship with someone with mental health symptoms (Katz et al., 1999; Porter & Chambless, 2014; Stewart & Harkness, 2015) and also affirm the importance of studying mental illness dyadically. Future research should assess symptomology in both members of the couple such that one’s familiarity with mental health symptoms outside of the relationship could be compared with that of their personal experience within the relationship.

Consistent with our hypothesis, there were more effects of controllable causes than uncontrollable causes (Blais & Renshaw, 2014). Controllable psychological causes were the most relevant for relationship quality. This suggests that psychological origins of mental health symptoms were viewed more consequentially to relationships than biological origins. This is intuitive considering that the present study was advertised as involving “romantic relationships and mental health” and that psychological illness has historically been seen as outside of the biological realm (Weiner et al., 1988). People often lack “psychological flexibility” (p. 1255) in understanding mental illness, especially if they experience personal psychological displeasure, so it may be that romantic partners were internally biased away from fully considering the less-stigmatized biological and/or uncontrollable sources of symptoms (Masuda, Price, Anderson, Schmertz, & Calamaras, 2009).

The present study was not without limitations. First, we used a college student sample of dating couples and these might be inherently different from noncollege samples at other relationship stages (e.g., engaged, married). Second, these data were cross-sectional. Future research should explore longitudinal data to make conclusions about the causal relationship between attributions and relationship quality. Third, symptom levels of anxiety and depression examined in the current study may not generalize to more severely symptomatic individuals or clinical populations. Based on the cutoffs for each subscale of the DASS, the current sample reported moderate depressive symptoms and severe anxiety symptoms. Fourth, we did not assess depression or anxiety in the romantic partner. One study found that students with mental health symptoms (e.g., psychological distress, depression) were less able to recognize depression compared to those with no symptoms (Kim, Saw, & Zane, 2015). It is safe to assume that if the romantic partner was also experiencing mental health symptoms, their attributions (and possibly relationship quality) may be different than if they were not so future research should assess for depression and anxiety symptoms (and related attributions) in both members of the dyad. This would also have the additional benefit of testing the full accuracy-bias model. Fifth, future researchers might benefit by using a different scale to investigate beliefs about the causes of symptoms. Specifically, some of the causes (e.g., personal inadequacies, personality) may have been confusing to participants. It could be argued that “personal inadequacies” is an ambiguous notion that overlaps with personality or behavior depending on the respondents’ subjective perception. Similarly, “family environment” and “upbringing” may have been viewed as synonymous while “situational factors” is again broad and somewhat ambiguous. With regard to scoring, “personal inadequacies” and “personality” are considered controllable causes in the BPASD but some might view these as stable and therefore uncontrollable. The psychometric properties of this instrument should be explored more. Sixth, some of the scales used in the current demonstrated lower α levels. This could be that those subscales had fewer items (7) as opposed to 12 and/or it could be because not all of the items truly “fit” the factor structure. In the interest of using the scales as previously published, these lower αs could be weakening some of the effects found in the current study and future research should explore the psychometric properties of these scales more in depth. Finally, it would benefit future researchers to collect data regarding the perceived locus of control of the symptomatic participant as well as emotional ratings of pity, anger, blame, and so on that are classically studied in much of attribution research (see Cronan, Key, & Vaughn, 2016; Weiner et al., 1988). This would allow for comparison of constructs between prior work and more contemporary endeavors within relationship research that are important to gaining knowledge about mental illness in an interpersonal context.

Understanding perceptions of depressive and anxiety symptoms from both partners within romantic relationships is important not only because of the widely documented impact of perceptions of mental health on the person living with the disorder but especially because of the impact on their romantic partner as shown in this study. Based on the aforementioned findings, this study has meaningful implications for future work on the stigma of mental illness within romantic couples. These findings add to the body of research on stigma toward mental health issues and could possibly help consumers of psychological research to better understand the experience of depression and anxiety when the person is part of a romantic relationship. The current study demonstrated that causal attributions of symptoms of depression and anxiety function somewhat differently within interpersonal partnerships, and this is important not only for the treatment of illness in clinician–patient relationships but also for helping sufferers of mental illness to be understood and supported within romantic relationships which are often the most substantive in their lives. That their perceptions of the romantic partner’s perceptions were relationship-relevant (regardless of the accuracy of those ratings) highlight interesting avenues for research and practice. Similarly, the often contagious and bidirectional nature of depressive and/or anxiety symptoms posits that understanding the romantic partners’ experience of living with another person’s symptoms is also worthwhile. This is important in order to prevent future illness and to help partners understand how to behave in ways that do not stigmatize sufferers or contribute to the worsening of their symptoms and to ultimately allow partners to support and help cope with mental illness.

Footnotes

Funding

The author(s) received no financial support for the research, authorship, and/or publication of this article.

ORCID iD

Allison A. Vaughn

Open research statement

As part of IARR’s encouragement of open research practices, the authors have provided the following information: This research was not pre-registered. The data used in the research are available. The data can be obtained by emailing: avaughn@sdsu.edu. The materials used in the research are available. The materials can be obtained by emailing: avaughn@sdsu.edu.

References

Agnew

C. R.

Loving

T. J.

Drigotas

S. M.

(2001). Substituting the forest for the trees: Social networks and the predictions of romantic relationship state and fate. Journal of Personality and Social Psychology, 81, 1042–1057.

Anxiety and Depression Association of America. (2014). Retrieved from http://www.adaa.org/about-adaa/press-room/facts-statistics

Benazon

N. R.

(2000). Predicting negative spousal attitudes toward depressed persons: A test of Coyne’s interpersonal model. Journal of Abnormal Psychology, 109, 550–554. doi:10.1037//0021-843X.109.3.550

Blais

Renshaw

(2012). The association of biological and psychological attributions for depression with social support seeking intentions in individuals with depressive symptoms. Behavioural and Cognitive Psychotherapy, 40, 605–617. doi:10.1017/S1352465812000355

Blais

Renshaw

(2014). Perceptions of partners’ attributions for depression in relation to perceptions of support and conflict in romantic relationships. Journal of Marital and Family Therapy, 40, 498–508. doi:10.1111/jmft.12055

Corrigan

P. W.

Watson

A. C.

(2002). Understanding the impact of stigma on people with mental illness. World Psychiatry, 1, 16–20.

Coyne

J. C.

(1976). Toward an interactional description of depression. Psychiatry: Journal for the Study of Interpersonal Processes, 39, 28–40.

Cronan

S. B.

Key

K. D.

Vaughn

A. A.

(2016). Beyond the dichotomy: Modernizing stigma categorization. Stigma and Health, 1, 225–243. doi:10.1037/sah0000031

Doyle

D. M.

Molix

(2014). Love on the margins: The effects of social stigma and relationship length on romantic relationship quality. Social Psychological and Personality Science, 5, 102–110. doi:10.1177/1948550613486677

10.

Furman

C. R.

Luo

Pond

R. S.

Jr. (2017). A perfect blame: Conflict-promoting attributions mediate the association between perfectionism and forgiveness in romantic relationships. Personality and Individual Differences, 111, 178–186. doi:10.1016/j.paid.2017.01.052

11.

Hettema

Neale

Myers

Prescott

Kendler

K. S.

(2006). A population-based twin study of the relationship between neuroticism and internalizing disorders. American Journal of Psychiatry, 163, 857–864. doi:10.1176/ajp.2006.163.5.857

12.

Hofmann

. (2014). Interpersonal emotion regulation model of mood and anxiety disorders. Cognitive Therapy and Research, 38, 483–492. doi:10.1007/s10608-014-9620-1

13.

Joiner

Alfano

Metalsky

(1993). Caught in the crossfire: Depression, self-consistency, self-enhancement, and the response of others. Journal of Social and Clinical Psychology, 12, 113–134. doi:10.1521/jscp.1993.12.2.113

14.

Katz

Beach

Joiner

(1999). Contagious depression in dating couples. Journal of Social and Clinical Psychology, 18, 1–13. doi:10.1521/jscp.1999.18.1.1

15.

Kenny

D. A.

Acitelli

L. K.

(2001). Accuracy and bias in the perception of the partner in a close relationship. Journal of Personality and Social Psychology, 80, 439–448.

16.

Kenny

D. A.

Kashy

D. A.

Cook

W. L.

(2006). Dyadic data analysis. New York, NY: Guilford Press.

17.

Kessler

Walters

Forthofer

(1998). The social consequences of psychiatric disorders, III: Probability of marital stability. American Journal of Psychiatry, 155, 1092–1096. doi:10.1176/ajp.155.8.1092

18.

Kim

J. E.

Saw

Zane

(2015). The influence of psychological symptoms on mental health literacy of college students. American Journal of Orthopsychchiatry, 85, 620–630. doi:10.1037/ort0000074

19.

Leff

Vearnals

Brewin

C. R.

Wolff

Alexander

Asen

… Everitt

. (2000). The London depression intervention trial: Randomised controlled trial of antidepressants v. couple therapy in the treatment and maintenance of people with depression living with a partner: Clinical outcome and costs. The British Journal of Psychiatry, 177, 95–100. doi:10.1192/bjp.177.2.95

20.

Lovibond

S. H.

Lovibond

P. F.

(1995). Manual for the depression anxiety stress scales (2nd Ed.). New South Wales: Psychology Foundation Monograph.

21.

MacDonald

T. K.

Ross

(1999). Assessing the accuracy of predictions about dating relationships: How and why do lovers’ predictions differ from those made by observers? Personality and Social Psychology Bulletin, 25, 1417–1429.

22.

Martinez

A. G.

Piff

P. K.

Mendoza-Denton

Hinshaw

S. P.

(2011). The power of a label: Mental illness diagnoses, ascribed humanity, and social rejection. Journal of Social and Clinical Psychology, 30, 1–23. doi:10.1521/jscp.2011.30.1.1

23.

Masuda

Price

Anderson

Schmertz

Calamaras

(2009). The role of psychological flexibility in mental health stigma and psychological distress for the stigmatizer. Journal of Social and Clinical Psychology, 28, 1244–1262. doi:10.1521/jscp.2009.28.10.1244

24.

McLeod

J. D.

(1994). Anxiety disorders and marital quality. Journal of Abnormal Psychology, 103, 767–776. doi:10.1037/0021-843X.103.4.767

25.

Muthén

L. K.

Muthén

B. O.

(2017). Mplus user’s guide (8th Ed.). Los Angeles, CA: Muthén & Muthén.

26.

Pierce

Sarason

(1991). General and relationship-based perceptions of social support: Are two constructs better than one? Journal of Personality and Social Psychology, 61, 1028–1039. doi:10.1037/0022-3514.61.6.1028

27.

Porter

Chambless

(2014). Shying away from a good thing: Social anxiety in romantic relationships. Journal of Clinical Psychology, 70, 546–561. doi:10.1002/jclp.22048

28.

Qualtrics (2018, July). Computer Software. Provo, UT: Qualtrics.

29.

Segrin

Powell

Givertz

Brackin

(2003). Symptoms of depression, relational quality, and loneliness in dating relationships. Personal Relationships, 10, 25–36. doi:0.1111/1475-6811.00034

30.

Starr

Davila

(2008). Excessive reassurance seeking, depression, and interpersonal rejection: A meta-analytic review. Journal of Abnormal Psychology, 117, 762–775. doi:10.1037/a0013866

31.

Stewart

J. G.

Harkness

K. L.

(2015). The interpersonal toxicity of excessive reassurance-seeking: Evidence from a longitudinal study of romantic relationships. Journal of Social and Clinical Psychology, 34, 392–410. doi:10.1521/jscp.2015.34.5.392

32.

Swann

W. B.

Griffin

J. J.

Predmore

Gaines

(1987). The cognitive-affective crossfire: When self-consistency confronts self-enhancement. Journal of Personality and Social Psychology, 52, 881–889. doi:10.1037/0022-3514.52.5.881

33.

Weiner

(1985). An attributional theory of achievement motivation and emotion. Psychological Review, 92, 548–573.

34.

Weiner

Perry

R. P.

Magnusson

(1988). An attributional analysis of reactions to stigmas. Journal of Personality and Social Psychology, 55, 738–748. doi:10.1037/0022-3514.55.5.738

35.

West

T. V.

Kenny

D. A.

(2011). The truth and bias model of judgement. Psychological Review, 118, 357–378. doi:10.1037/a0022936

36.

Whisman

M. A.

(2001). Marital adjustment and outcome following treatments for depression. Journal of Consulting and Clinical Psychology, 69, 125–129. doi:10.1037/0022-006X.69.1.125

37.

Zaider

T. I.

Heimberg

R. G.

Iida

(2010). Anxiety disorders and intimate relationships: A study of daily processes in couples. Journal of Abnormal Psychology, 119, 163–173. doi:10.1037/a0018473

38.

Zinbarg

R. E.

Mineka

Bobova

Craske

M. G.

Vrshek-Schallhorn

Griffith

J.W.

… Anand

(2016). Testing a hierarchical model of neuroticism and its cognitive facets: Latent structure and prospective prediction of first onsets of anxiety and unipolar mood disorders during 3 years in late adolescence. Clinical Psychological Science, 4, 805–824. doi:10.1177/2167702615618162