Anxiety and Avoidance in Close Relationships Are Associated With Responses to Negative and Positive Stimuli

Introduction

Bowlby’s (1958; 1982) attachment theory explains the nature of the child’s ties to the caregiver primarily based on an evolutionary perspective, postulating that a child’s attachment behaviors represent activation of a species-wide behavioral system to enhance the likelihood of survival. Romantic partnerships between adults have also been studied through the lens of attachment theory (Guzman-Gonzalez et al., 2020; Hazan & Shaver, 2017; Shaver et al., 1988). A relevant question is whether infant-caregiver and romantic relationships share similar evolutionary functions. According to Zeifman and Hazan (2016), while attachment to another adult is typically not essential to survival as in infant attachment, romantic attachment can promote individuals’ chances of survival and the survival of joint offspring. Psychologists have studied romantic bonds between adults by focusing on three distinct behavioral systems: attachment, caregiving, and sex (Mikulincer & Shaver, 2019). Attachment and caregiving have a bidirectional relationship in romantic attachment; when a person feels threatened, the attachment system is activated (as in infant attachment), which triggers a need for closeness with a partner who can provide protection. In turn, when the other person senses their partner in distress, the caregiving system is activated so that they provide a safe, comforting environment for their partner (Guzman-Gonzalez et al., 2020).

According to attachment theory, early interactive experiences with caregivers lead to the formation of working representations of the self and others. Secure attachments are expected to form when the caregiver is reliable and responsive, particularly in times of need (e.g., Ainsworth & Bowlby, 1991). These early attachments come into play with other individuals later in life who may be outside the family (e.g., Ainsworth, 1991; Bowlby, 1982; Mikulincer & Shaver, 2019). If a sense of safety and security is not in place, negative representations of self and others can form, leading to attachments to others (e.g., romantic partners) that can be anxious or avoidant. Attachment anxiety is characterized by a fear of rejection or abandonment, whereas attachment avoidance is characterized by discomfort with emotional closeness (Mikulincer & Shaver, 2019).

To our knowledge, there has been very little research on how attachment in close relationships relates to the appraisal of emotional and neutral environmental stimuli. Our emotional experience with stimuli in our environment may prime our thinking and actions to be more (or less) positive or more (or less) negative (Elgendi et al., 2018; Gollan et al., 2016; Watkins & Roberts, 2020). This appraisal of the world around us may also explain how we behave in social contexts, such as in interactions with a romantic partner.

It is important to assign emotional valence to environmental stimuli, as it is essential to human survival. Stimuli that are associated with a negative valence (e.g., snakes, spiders) may be threatening, promoting an aversive response (e.g., fight or flight). Stimuli that are associated with a positive valence (e.g., palatable foods, smiling faces) may be beneficial for nourishment or reproduction, promoting an approach response. Despite the importance of interpreting emotional stimuli, people often vary in their emotional appraisals of an event (Lazarus, 1991). For example, people who are undergoing a risky surgical procedure may be more optimistic or pessimistic about the process or the outcome, and people who have experienced trauma-like exposure to combat may have different affective responses to the same event (Bremner et al., 1999; Taylor et al., 2000).

A number of studies have shown that emotional interpretation of information differs across individuals based on mental health factors or personality traits. These differences are often reflected in a bias in processing valenced information, in which positive and negative information are not attended to, remembered, or processed at the same rate. Positivity bias can refer to a processing advantage (i.e., faster reaction time or higher accuracy) for positive versus negative stimuli, and negativity bias, a processing advantage for negative (vs. positive) stimuli (Kauschke et al., 2019). Positivity bias can also refer to the tendency to give more weight to positive stimuli, whereas negativity bias is the tendency to give more weight to negative stimuli.

In addition to the focus on negative and positive stimuli, recent research has examined how a neutral stimulus is evaluated. A growing body of empirical evidence suggests that “neutral” stimuli may not actually be neutral, and rather, that humans tend to (consciously or unconsciously) assign a valence to these “emotionless” stimuli (e.g., Ito & Cacioppo, 2005). Several studies demonstrate a positivity offset to neutral stimuli, a greater positive affect (vs. negative affect) when the environment is neutral (e.g., Norris et al., 2011). It has been proposed that this offset occurs because of an imbalance in activation systems, in that the appetitive/approach system is more active than the aversive/avoidance system to neutral stimuli (Koh et al., 2017).

People who rate higher on depression or anxiety scales are less likely to demonstrate this positivity offset. Gollan and colleagues (2016) found that people higher in depression show a weaker positivity offset, in that they have a lesser tendency to perceive neutral stimuli as positive. Winton et al. (1995) found that people higher in social anxiety (specifically, fear of negative evaluation) were more likely to identify neutral faces as negative. Therefore, prior research has demonstrated individual differences in the positivity offset, which will be further explored in this study.

There is a vast literature that supports a negativity bias in depression and anxiety (Bradley et al., 1995; Mogg et al., 1995). Abler et al. (2007) found that depressed patients had more amygdala activation in anticipation of a negative stimulus relative to neurotypical controls. Hamilton and Gotlib (2008) found better memory recollection for negative stimuli in depressed individuals relative to nondepressed controls. They also found an overrecruitment of the neural network involved in memory for affective stimuli in depressed individuals. Williams et al. (2007) used electroencephalography (EEG) to examine responses to fear stimuli. Participants higher in self-reported anxiety had greater early stage neural activation to fearful face stimuli that was presented for 10 s, indicative of an automatic response to fear stimuli in anxious individuals.

In the present exploratory study, we focus on a relational component to better understand individual differences in response to affective and neutral stimuli. For example, if neutral stimuli in the environment (e.g., a table, a book) are perceived as more negative to someone who experiences more anxiety about rejection or abandonment from their partner, then perhaps this negative appraisal primes a negative mindset, which leads to heightened anxiety about the partnership (even in emotionally “neutral” contexts). In turn, having more anxiety about rejection or abandonment in a partnership could prime a negative appraisal of objects in the environment. For someone who avoids intimacy in relationships, perhaps emotional stimuli are perceived as less positive or less negative, which may explain a general blunted emotional response. Although examining causation is beyond the scope of the current paper, we aim to understand how anxiety and avoidance in close relationships are associated with the appraisal of valenced and neutral stimuli.

Method

Results

Response to Images

The rmANOVA on emotional appraisal by stimulus type revealed a main effect of stimulus type, F (1, 92) = 1905.60, p < .001, η _p ² = .95. Participants gave the lowest ratings to negative stimuli (M = 1.28) and the highest ratings to positive stimuli (M = 3.56). Neutral ratings fell in the middle (M = 2.93), and ps < .001 for all pairwise differences (see Table 1).

OPEN IN VIEWER

Table 1.

Mean Evaluation of Valence With Standard Deviation in Parentheses by Stimulus Type in the Image Classification Task.

Negative	Neutral	Positive
1.28 (.22)	2.93 (.33)	3.56 (.30)

Note. Participants rated the emotional valence of each stimulus on a 4-point scale: 1 = “very negative,” 2 = “slightly negative,” 3 = “slightly positive,” and 4 = “very positive.”

We hypothesized that we would find a positive bias of neutral stimuli. We found that neutral stimuli were most commonly evaluated as “slightly positive,” with “very positive” and “slightly negative” as the second most common appraisals. A Wilcoxin signed rank test comparing the number of “3” (“slightly positive”) versus “2” (“slightly negative”) responses for neutral stimuli revealed that participants categorized neutral stimuli as “slightly positive” more often than “slightly negative” (Mdn = 30.0 vs. Mdn = 8.5), z = −6.42, p < .001. Participants categorized neutral stimuli as “slightly positive” more often than “very positive” (Mdn = 30.0 vs. Mdn = 9.0), z = −6.85, p < .001. The average number of “very positive” and “slightly negative” evaluations did not differ, z =−5.02, p = .62.

Individual Differences Measures

Pearson product-moment correlations of note are reported below. All correlations between individual difference measures and RT, accuracy, and ratings on trials are shown in Table 2.

OPEN IN VIEWER

Table 2.

Zero-Order Pearson Correlations Between the Individual Difference Measures and Evaluation of Valence in the Image Classification Task.

	1	2	3	4	5	6	7	8	9	10	11
1. ECR-R anxiety	—
2. ECR-R avoidance	.317**	—
3. RT_Nega	−.076	.117	—
4. RT_Neut	.029	−.074	.726**	—
5. RT_Posi	.055	.041	.802**	.830**	—
6. Acc_Nega	−.142	−.219*	−.120	−.114	.000	—
7. Acc_Neut	−.109	.141	.004	−.258*	.092	.496**	—
8. Acc_Posi	−.180	−.122	.077	−.022	−.047	−.016	−.001	—
9. Rating_Nega	.006	.316**	.365**	.035	.144	−.563**	−.086	.147	—
10. Rating_Neut	−.077	−.031	−.036	−.037	−.173	−.285**	−.474**	.531**	.215*	—
11. Rating_Posi	.051	−.219*	−.048	.203	−.116	−.085	−.505**	.441**	−.284**	.424**	—

Note. *p < .05, two-tailed. **p < .01, two-tailed. ECR-R = Experience in Close Relationships. RT = reaction time. Acc = accuracy. Nega = negative. Neut = neutral. Posi = positive.

Experiences in Close Relationships (ECR-R; Anxiety and Avoidance)

We also predicted that participants who had higher scores on the ECR-R Avoidance and ECR-R Anxiety would show a negativity bias (the processing advantage of negative over positive stimuli).

We found that individuals who scored higher on ECR-R Anxiety had a greater negativity bias, r = .209, p < .05. This indicates that RTs were substantially faster for negative than for positive stimuli, indicating a bias towards this type of stimuli. All other individual difference measures did not correlate with the negativity bias, rs < ±.130, ps > .22.

Correlations between accuracy, RTs, and mean ratings of positive or negative stimuli were not significant for ECR-R Anxiety, rs < ±.143, ps > .175. There is a rather interesting pattern that emerged with ECR-R Avoidance, however. Participants higher in Avoidance had lower accuracy for negative stimuli, r = −.219, p < .05. Participants higher in Avoidance also gave higher ratings for negative stimuli, r = .316, p < .005. Both of these findings suggest that participants higher in Avoidance were more likely to identify negative stimuli as “slightly positive” or “very positive”. In other words, these participants perceived negative stimuli as less negative.

In addition to differences in identifying negative stimuli, there were also differences in ratings of positive stimuli by ECR-R Avoidance. Participants higher in ECR-R Avoidance gave lower ratings to positive stimuli, r = −.219, p < .05, which indicates that they perceived positive stimuli as less positive. The correlation between ECR-R Anxiety and accuracy for positive stimuli was negative and marginally significant, suggesting that participants higher in Anxiety gave positive stimuli a “slightly negative” or “very negative” label more often, r = −.180, p = .084.

We also predicted that participants who had higher scores on the ECR-R Avoidance and ECR-R Anxiety would be more likely to rate neutral stimuli as negative. These measures did not correlate with neutral stimuli ratings, rs < ±.078, ps > .45.

Discussion

The main purpose of the current exploratory study was to examine how anxiety and avoidance in close relationships are associated with emotional appraisal of a stimulus (positive, negative, or neutral). Our first hypothesis was that there would be an overall positive bias of neutral stimuli, reflecting a positivity offset that has been demonstrated in previous studies. Our results support this hypothesis, in that neutral stimuli were rated as more positive than negative, on average. These data are consistent with a number of previous findings that support a positivity offset, in that people tend to experience more positive (than negative) affect when the environment is neutral (Ito & Cacioppo, 2005; Koh et al., 2017; Norris et al., 2011).

Although we predicted that we would find an overall positivity offset, we expected to find individual differences in the response to negative, neutral, and positive stimuli. Our second hypothesis was that we would find a negativity bias (the processing advantage of negative over positive stimuli) for people who have more anxiety and avoidance tendencies in close relationships. This hypothesis was partially supported. Individuals who scored higher on ECR-R Anxiety had a greater response time negativity bias, in that RT were substantially faster for negative than for positive stimuli, indicating a bias towards this type of stimuli. Furthermore, participants higher in anxiety in close relationships assigned positive stimuli a negative (either “slightly negative” or “very negative”) marginally more often than people lower in anxiety. These findings are consistent with previous reports that people higher in anxiety have a negativity bias (Bradley et al., 1995; Mogg et al., 1995; Williams et al., 2007). These findings suggest that anxiety in close relationships may operate similarly to trait anxiety. One limitation of the current study is that we did not measure trait anxiety. Thus, further research is needed to disentangle the differences between these variables. All other individual difference measures did not correlate with the negativity bias.

We also predicted that people higher in anxiety and avoidance in close relationships would have a greater tendency to rate neutral stimuli as negative. Neither of these measures (ECR-R Anxiety or ECR-R Avoidance) correlated with ratings of neutral stimuli, and thus, this hypothesis was not supported by the data. Previous studies have found a negative bias of neutral stimuli for people higher in social anxiety (e.g., Winton et al., 1995). Furthermore, people higher in depression are less likely to demonstrate a positivity offset (Gollan et al., 2016). Our study differed from previous studies in multiple ways. Winton’s study was conducted with strictly faces, whereas our study was conducted with both social (e.g., faces) and non-social (e.g., scenes, objects) stimuli. Gollan included a response scale in the form of a 5 x 5 grid in which there were 10 total points (0-5 for both positive and negative evaluations, from “not at all” to “extremely”). It may be worth increasing the points on our response scale in future studies, as 10 points allows for more statistical sensitivity in detecting differences in participant evaluations.

An interesting pattern emerged with ECR-R Avoidance. We found that people higher in ECR-R Avoidance rated negative stimuli as less negative (i.e., more positive) and rated positive stimuli as less positive (i.e., more negative). These findings may suggest a blunted response to emotional stimuli. This is consistent with findings that individuals who score higher on Behavioral Inhibition System (BIS) show less neuronal activation in response to a positive reward stimulus (Simon et al., 2010). BIS has, however, been correlated with stronger physiological responses to negative stimuli, which contradicts our findings here (e.g., Balconi et al., 2012). One major limitation of the current study is that demographic information was not obtained for this set of participants due to a procedural error. The estimated age of the participants in this study, however, should reflect college-wide demographics (median age is likely approximately 20 years old), but future research can offer more insight into the relationship between these individual differences and response to valenced and neutral stimuli as a function of age and other demographic variables.

To our knowledge, our study is the first investigation into how avoidance in close relationships relates to the emotional appraisal of stimuli. More research is needed to understand how individual differences relate to the appraisal of emotional and neutral stimuli in both neurotypical and clinical populations. Future research should also examine causal links. For example, experimental work can address the question about whether appraisal of environmental stimuli leads to differences in anxiety or avoidance in close relationships. If so, this can inform the design of interventions that focus on changing appraisal of stimuli in the environment in order to decrease avoidance or anxiety in a relationship with a partner.

Here, in an exploratory study, we address a novel research question regarding how anxiety and avoidance in close relationships relates to the emotional evaluation of objects, events, or people in the environment. Our results show that individuals higher in anxiety in close relationships show a negativity bias in two measures: (1) faster RT to negative than to positive stimuli, and (2) more frequently evaluating positive stimuli as “negative.” Individuals higher in avoidance in close relationships show a different pattern of emotional appraisal: more positive ratings of negative stimuli and more negative ratings of positive stimuli, suggesting a general blunted response to emotional stimuli. This area of research lends important insight into partner dynamics in close relationships.