Experiential avoidance in idiographic, autobiographical memories: Construct validity and links to social anxiety, depressive, and anger symptoms

Abstract

Experiential avoidance, or attempts to alter or avoid undesirable thoughts and feelings, has been theorized to be relevant to the development of emotional disturbances, particularly anxiety problems. Prior work has relied on two methodologies: global self-report measures or laboratory manipulations. To better understand links between experiential avoidance and emotional disturbances, we measured experiential avoidance in the context of prominent anxious autobiographical events. Trained raters coded events for emotionality and reliance on experiential avoidance. Our interest was whether experiential avoidance could be measured as a memory characteristic and how it relates to social anxiety, depressive, and anger symptoms. As evidence of construct validity, experiential avoidance ratings were related to more intense negative emotions and coping difficulties during anxious events, memory vividness, and emotion suppression tendencies. Experiential avoidance was positively related to social anxiety and depressive symptoms and predicted an increase in social anxiety over a 3-month period; findings could not be attributed to the emotionality of memories. In contrast, no relations were found with inward or outward expressions of anger, or longitudinal change in depressive or anger symptoms. Results suggest that experiential avoidance is an important dimension of people’s life narratives and particularly relevant to social anxiety problems.

Recent theoretical models suggest trans-diagnostic processes that might explain the transition from negative emotions to disorder. For instance, Ingram (1990) provided a comprehensive review of the literature on self-focused attention and how it influences the development and maintenance of anxiety, mood, eating, and psychotic disorders. A few years later, researchers suggested that experiential avoidance as another candidate (Hayes, Wilson, Gifford, Follette, & Strosahl, 1996). Experiential avoidance can be defined as the tendency to negatively evaluate feelings and thoughts and spend inordinate effort to avoid or alter these private events, as well as situations that led to them in the past and situations expected to lead to them in the future. An over-reliance on this regulatory strategy has been shown to be associated with a wide array of dysfunctional outcomes (e.g., Hayes, Luoma, Bond, Masuda, & Lillis, 2006; Kashdan, Barrios, Forsyth, & Steger, 2006). In this study, we sought to extend this line of research by using an idiographic, autobiographical methodological approach to measure experiential avoidance.

Feeling anxious about giving a talk to an audience, being sad due to a lack of contact with other people, and feeling angry when a boss prematurely dismisses your ideas are all normal reactions. But if you avoid giving talks, confiding to other people, and attending business meetings because you do not want to experience anxiety, sadness, or anger, problems ensue. Response costs include the ironic rebound effect such that trying to conceal thoughts and feelings increase their frequency and intensity (Wegner, 1994). An excessive focus on controlling negative experiences drain emotional and physical resources leaving limited attention, stamina, and self-control for other demanding tasks (Baumeister, Galliot, DeWall, & Oaten, 2006).

To date, most of the work on experiential avoidance has focused on anxiety conditions (Eifert & Forsyth, 2005), in particular, social anxiety. Over the course of 21 days, people with excessive social anxiety reported fewer positive events on days when they experienced greater social anxiety and tended to avoid or suppress undesirable emotions (Kashdan & Steger, 2006). Other studies find that socially anxious people tend to conceal or avoid negatively valenced thoughts and feelings (e.g., Kashdan & Breen, 2007; Turk, Heimberg, Luterek, Mennin, & Fresco, 2005); similar findings have been shown when focusing on people with diagnoses of social anxiety disorder (Kashdan, Morina, & Priebe, 2009).

At least two studies have shown that a willingness to pursue valued aims despite the presence of anxious feelings is a useful target when treating people with social anxiety difficulties. In a small study of 39 college students with public speaking anxiety, being randomly assigned to a 6 week cognitive-behavioral or acceptance and mindfulness based workshop led to similar reductions in social anxiety (Block, 2002). However, only students assigned to the acceptance and mindfulness based workshop showed significant reductions in the behavioral avoidance of public speaking situations. Building on these findings, in a 12-week acceptance and mindfulness based intervention for clients with social anxiety disorder, researchers found that changes in experiential avoidance during the initial phase of treatment was associated with changes in behavior during later phases of treatment (Dalrymple & Herbert, 2007). That is, socially anxious clients showed a greater willingness to enter and remain in social situations despite anxious thoughts, feelings, and physiological sensations. Although preliminary, these results suggest experiential avoidance is a potential mechanism accounting for the impairment associated with excessive social anxiety.

Much less is known about the contributions of experiential avoidance to depressive symptoms (Tull & Gratz, 2008) and anger symptoms (Eifert, McKay, & Forsyth, 2006). To further illustrate this point, the only existing work on experiential avoidance and depressive symptoms has been cross-sectional with tests of correlations between global self-report questionnaires and other global self-report questionnaires. Meanwhile studies showing that suppression of anger often has deleterious effects (e.g., DiGiuseppe, & Tafrate, 2006; Quartana & Burns, 2007) suggest a link between experiential avoidance and anger, but this relationship requires further inspection. The small body of research on anger issues compared with anxiety and depression is likely the result of a number of factors including beliefs that anger is the manifestation of “softer” emotions such as anxiety and depression, and lack of consensus in defining and operationalizing the phenomenology of anger experience and expression (e.g., Carver & Harmon-Jones, 2009; Eckhardt, Norlander, & Deffenbacher, 2004; Russell & Fehr, 1994).

Experiential avoidance occurs in the context of life events that are perceived as threatening, yet, most work on experiential avoidance has relied on context-free global self-report questionnaires such as the 9-item Acceptance and Action Questionnaire (Hayes et al., 2004). To add to this literature, we created an idiographic measure of experiential avoidance. Specifically, we asked people to recall a memory of when they felt particularly strong emotions and write about everything that happened from situational details to their reactions. From this rich, idiographic data, we created a system to code the intensity of their emotions and their tendency to engage in experiential avoidance efforts.

The current study is the first to use an autobiographical memory paradigm to measure experiential avoidance. Given that experiential avoidance is theoretically linked to a broad spectrum of psychopathology, we examined relations with social anxiety, depressive, and anger symptoms. Besides concurrent relations, we examined the effects of our idiographic measure of experiential avoidance on changes in social anxiety, depressive symptoms, and anger symptoms over a 3-month period.

We hypothesized that experiential avoidance could be reliably and validly measured in the context of idiographic memories. By using this alternative measurement strategy, relations with trait measures of social anxiety, depressive, and anger symptoms cannot be attributed to method variance effects. As the first attempt to measure experiential avoidance with this approach, we conducted tests of construct validity. For instance, it was important to show that experiential avoidance could not be explained by the intensity of negative affect and anxiety in the same narratives in predicting social anxiety, depressive, and anger symptoms. Based on prior theory and research, we hypothesized that experiential avoidance would show stronger concurrent and prospective relations with social anxiety compared with depressive and anger symptoms. Despite publication of a book on acceptance and mindfulness strategies to target experiential avoidance in the treatment of anger problems (Eifert et al., 2006), few studies exist to support this framework. Thus, we consider tests of how experiential avoidance relates to anger issues to be exploratory. In sum, we adopted a novel approach to studying experiential avoidance embedded in the context of personal memories to understand concurrent and temporal relations with various emotional disturbances.

Method

Participants

148 students in psychology at a large, public university in the Mid-Atlantic received course credit for completing an initial survey and a 3-month follow-up survey. We had a 98% retention rate from Time 1 (T1) to Time 2 (T2). Of the T2 participants, 109 were women (75.2%) and 36 were men (24.8%); 69% of participants identified themselves as Caucasian, 6.3% as Hispanic/Hispanic-American, 5.6% as Asian/Asian-American, 5.6% as Middle-Eastern, 3.5% as African-American, and 7.7% as Mixed or other.

Procedure

Participants completed an anonymous internet-based survey. The only personal identifier collected was email addresses. Participants completed several different questionnaires at T1 (see Measures), wrote autobiographical narratives, and gave consent to be contacted for future studies. Three months later, at T2, they were sent a web-link to their email address to answer the same measures completed at T1.

PsychData, the web-based company used for data collection, ensures high security through encrypted data transfer, password-required data access, and an environment where answers are written on a secure server with no traces on individual computers. Individual IP addresses were not collected.

Autobiographical Narratives

Participants completed three separate autobiographical narratives. Before writing, each participant was prompted to recall a prominent memory when they experienced a particular emotion―curiosity, gratitude, and anxiety, respectively. They were asked to write about the easiest memory to recall (accessibility being considered an index of prominence). For each narrative, participants rated the vividness of their remembered experience from 1 (not at all) to 5 (extremely). Due to the present interest in experiential avoidance, analyses are limited to anxiety narratives.

Two trained coders independently rated each narrative for the following constructs: global positive affect, global negative affect, anxiety, coping ability, and experiential avoidance. A brief primer on our coding system is provided in Appendix A. Presence of each construct in the narrative was rated on a Likert scale from 1 (not at all) to 7 (extremely) with the exception of experiential avoidance, which was rated from 1 (not at all) to 5 (extremely). Global positive affect was defined as a generic, pleasant state of well-being and ratings encompassed both high arousal states (e.g., joy, excitement) and low arousal states (e.g., contentment, serenity). Global negative affect was defined as feelings of distress or discomfort and ratings encompassed both high arousal states (e.g., scared, angry) and low arousal states (e.g., bored, sad). Anxiety was defined as experience of any of the following: the anticipation of feared stimuli, anxiety or fear about internal or external stimuli, anxiety in the aftermath of an event over perceived consequences, and acts of worry. Coping ability was defined as the person’s ability to effectively understand, work, cope, or thrive in the particular situation. That is, how well did they adapt and respond to the demands of the situation?

Experiential avoidance was defined as a process involving excessive evaluations of private thoughts, feelings, and sensations, and an unwillingness to experience these unwanted private events. Deliberate efforts to control or escape these events ensured higher ratings of experiential avoidance. Here is an example of a narrative with explicit markers of experiential avoidance attempts:

I constantly feel very anxious. Every single time I am put in a social situation I get pretty anxious. At parties, or gatherings, or meeting people and having to interact with people I'm not very familiar with makes me very nervous. It make my hands and body tense, it makes my heart race, it makes my face flush, it makes my stomach hurt. It makes me feel self-conscious and awkward and it makes me isolate myself or leave.

To develop our coding system, raters practiced on unrelated narrative datasets collected by other researchers. When raters reached agreement on more than 80% of narratives (± 1 point on the Likert scale), ratings began on the current dataset. Independent coding was conducted by clinical psychology Ph.D. students. To minimize behavioral drift, only five participants were coded at a time prior to meetings to record ratings, resolve discrepancies, and then continue to the next set. Both raters coded all narratives. Reliability was calculated using a two-way random intraclass correlation coefficient (ICC) for the absolute agreement of the single measure. By modeling random error, this is a conservative estimation. Reliability was excellent for experiential avoidance (ICC = .90), negative affect (ICC = .77), anxiety (ICC = .81), and coping ability (ICC = .75).

Trait Questionnaires

Social anxiety

The 19-item Social Interaction Anxiety Scale (SIAS; Mattick & Clarke, 1998) assesses the tendency to fear and avoid evaluation in social situations. Responses are provided using a 5-point Likert scale; rated from 0 (not at all) to 4 (extremely). The SIAS demonstrates stability across a four week period (r = .92; Mattick & Clarke, 1998) and reliably differentiates between people diagnosed with social anxiety disorder and other conditions (E. Brown et al., 1997).

Depressive symptoms

The 21-item Beck Depression Inventory-II (BDI-II; Beck, Steer, & Brown, 1996) assesses the severity of depressive symptoms, with higher scores indicating greater severity. In previous research, the BDI-II demonstrated acceptable psychometric properties and reliably distinguished between clinical and non-clinical samples (e.g., Sprinkle et al., 2002).

Anger symptoms

The 38-item Multidimensional Anger Inventory (MAI, Siegel, 1986) measures the frequency, duration, magnitude, and mode of expressed anger, as well as a range of anger-eliciting situations. Two subscales of the MAI, anger-in and anger-out, assess a person’s mode of anger expression. The 5-item anger-in subscale assesses the tendency to hold in anger and brood over it (Example items include: “I try to get even when I am angry with someone”; “Even after I have expressed my anger, I have trouble forgetting about it”) The 2-item anger-out subscale assesses the tendency to express anger to others (Items included: “When I am angry with someone, I let that person know” and “It’s difficult for me to let people know I’m angry” which is reverse-scored). In samples of college students and factory workers, there is evidence of acceptable reliability and validity (Siegel, 1986).

Experiential avoidance

The most widely used measure of experiential avoidance is the 9-item Acceptance and Action Questionnaire (AAQ; Hayes, Strosahl, & Wilson 2004). High scores correspond to greater reliance on experiential avoidance. Responses are provided using a 7-point Likert scale; rated from 1 (never) to 7 (always). Reliability including four-month stability has been shown to be acceptable (.64) (Hayes et al., 2004) and evidence for acceptable validity has been shown in experimental studies and examinations of treatment sensitivity (Hayes et al., 2006).

Emotion regulation strategies

The 10-item Emotion Regulation Questionnaire (ERQ; Gross & John, 2003) measures different ways that people tend to manage their emotions. The 6-item cognitive reappraisal subscale assesses a person’s tendency to change the way they think (Example items include: “I control my emotions by changing the way I think about the situation I’m in”; “When I’m faced with a stressful situation, I make myself think about it in a way that helps me stay calm”) The 4-item emotion suppression subscale assesses tendencies to inhibit or conceal the expression of experienced emotions (Example items include: “I control my emotions by not expressing them”; “When I am feeling negative emotions, I make sure not to express them”). Responses are provided using a 7-point Likert scale; rated from 1 (strongly disagree) to 7 (strongly agree). In prior studies, both subscales showed acceptable psychometric properties including three-month stability (Gross & John, 2003).

Results

Preliminary analyses

Descriptive data and inter-correlations for primary measures at T1 and T2 are reported in Table 1. Mean social anxiety scores at T1 (M = 23.62; SD = 13.80) were comparable to other non-clinical samples (Heimberg, Mueller, Holt, Hope, & Liebowitz, 1992; Mattick & Clarke, 1998) and scores at least one standard deviation above the mean were comparable to clinical samples (E. Brown et al., 1997).

Table 1.

Means, Standard Deviations, and Internal Consistency Coefficients for, and Zero-Order Relations among Variables

	1	2	3	4	5	6	7	8	9	10	11
1. T1 EA	---
2. T1 NA	.22*	---
3. T1 Anxiety	.11	.82**	---
4. T1 SIAS	.25*	.16	.16	---
5. T1 BDI-II	.24*	.11	.08	.48**	---
6. T1 Ang-In	.13	.13	.16	.39**	.43**	---
7. T1 Ang-Out	−.05	.08	.14	.08	.35**	.32**	---
8. T2 SIAS	.30**	.14	.14	.83**	.42**	.28**	.01	---
9. T2 BDI-II	.15	−.01	.03	.41**	.73**	.23*	.11	.46**	---
10. T2 Ang-In	−.02	−.04	.07	.21*	.30**	.44**	.20*	.29**	.26*	---
11. T2 Ang-Out	−.21*	−.02	.14	−.05	.31**	.14	.49**	.03	.28**	.38**	---
M	1.80	4.95	4.71	23.62	11.81	14.96	12.72	21.57	11.02	14.42	12.94
SD	1.56	1.39	1.37	13.80	9.46	3.33	2.79	12.33	10.42	3.19	3.01

Notes.

^*p < .05.

^**p < .001.

All p-values were two-tailed. T1 = Time 1; T2 = Time 2; EA = Experiential avoidance dimension of narratives; NA = Negative affect dimension of narratives; Anxiety = anxiety dimension of narratives; SIAS = Social Interaction Anxiety Scale; BDI-II = Beck Depression Inventory-II; Ang-In = Anger-In subscale of the Multidimensional Anger Inventory; Ang-Out = Anger-Out subscale of the Multidimensional Anger Inventory.

For anxiety narratives, average word count was 1830.45 (SD = 369.03). Coding of experiential avoidance had no relation with word count, r = .06; a similar near-zero correlation was found between word count and coding of anxiety in narratives, r = -.06. We also failed to find significant relations between narrative word count and outcome measures of social anxiety, r = −.04, depressive symptoms, r = −.02, and anger-in and anger-out symptoms, rs = .03 and .11, respectively. Thus, narrative length had a negligible influence on how experiential avoidance was coded and associations with distress.

Construct validity of experiential avoidance

We calculated correlations between our autobiographical memory measure of experiential avoidance and relevant scales to determine convergent and discriminant validity. As reported in Table 1, experiential avoidance was positively related to negative affect in narratives, trait social anxiety at T1 and T2, depressive symptoms at T1, and negatively related to the outward expression of anger at T2. In addition, experiential avoidance was positively related to tendencies to suppress emotions, r = .23, p = .01, and a marginal negative relation to cognitive reappraisal strategies, r = −.15, p = .10. Due to an administrative error, the most widely used self-report measure of experiential avoidance, the AAQ, was left out of the initial and follow-up survey, and only added mid-way into follow-up assessments. Thus, we only have data for 41 people. As expected, our narrative measure was positively related to 3-month follow-up global ratings on the AAQ, r = .24.

For anxiety narratives, experiential avoidance was negatively related to abilities to cope with threat and stress, r = −.36, p < .001, and positively related to participant ratings of memory vividness, r = .19, p = .03. The latter finding converges with prior work suggesting that engaging in experiential avoidance is an ironic process leading to more vivid, intrusive, and frequent unwanted experiences.

Overview of Primary Data Analyses

We examined how experiential avoidance was related to concurrent and prospective ratings of social anxiety (SIAS), depressive symptoms (BDI-II), and anger symptoms (anger-in, anger-out). We constructed a series of hierarchical regression models to examine whether our T1 autobiographical memory measure of experiential avoidance predicted T1 symptoms. Autobiographical memory measures of negative affect and anxiety were included as covariates to examine the construct specificity of experiential avoidance. Additionally, we used the same predictors with residual changes in symptoms from T1 to T2 as the dependent variable. To reduce collinearity, predictor variables were centered.

Predicting concurrent and prospective social anxiety

As shown in Table 2, experiential avoidance in autobiographical memories was positively associated with trait social anxiety. Despite stability of social anxiety scores over the 3-month assessment period, greater experiential avoidance was associated with greater social anxiety over time. These effects could not be attributed to the intensity of negative affect and anxiety in the same autobiographical memories.

Table 2.

Hierarchical Regression Models of Experiential Avoidance Predicting Concurrent and Prospective Social Anxiety, Depressive Symptoms, and Anger

Outcome: Social Anxiety at T1
Step		b	SEb	pr	t	ΔR2	ΔF
1	T1 NA	.10	.16	.06	.67	.03	1.72
	T1 Anxiety	.07	.16	.04	.46
2	T1 EA	.23	.09	.23	2.55*	.05	6.50*

Outcome: Social Anxiety at T2
Step		b	SEb	pr	t	ΔR2	ΔF
1	T1 SIAS	.89	.05	.86	18.35**	.74	336.72**
2	T1 NA	−.01	.08	−.01	−.08	.00	1.72
	T1 Anxiety	.01	.08	.01	.12
3	T1 EA	.10	.05	.18	1.98*	.01	3.92*

Outcome: Depressive Symptoms at T1
Step		b	SEb	pr	t	ΔR2	ΔF
1	T1 NA	.09	.12	.07	.79	.01	.72
	T1 Anxiety	−.01	.12	−.01	−.12
2	T1 EA	.15	.06	.23	2.53*	.05	6.40*

Outcome: Depressive Symptoms at T2
Step		b	SEb	pr	t	ΔR2	ΔF
1	T1 BDI-II	.76	.06	.76	12.58**	.57	158.21**
2	T1 NA	−.14	.08	−.16	−1.79	.01	1.79
	T1 Anxiety	.09	.08	.10	1.09
3	T1 EA	−.01	.04	−.02	−.22	.00	.05

Outcome: Anger-In at T1
Step		b	SEb	pr	t	ΔR2	ΔF
1	T1 NA	.01	.11	−.01	−.06	.03	1.59
	T1 Anxiety	.12	.11	.10	1.08
2	T1 EA	.08	.06	.12	1.26	.01	1.59

Outcome: Anger-In at T2
Step		b	SEb	pr	t	ΔR2	ΔF
1	T1 Anger-In	.44	.09	.43	5.14**	.19	26.43**
2	T1 NA	−.21	.10	−.19	−2.10*	.03	2.20
	T1 Anxiety	.17	.10	.15	1.62
3	T1 EA	−.01	.06	−.01	−.12	.00	.01

Outcome: Anger-Out at T1
Step		b	SEb	pr	t	ΔR2	ΔF
1	T1 NA	−.07	.11	−.06	−.64	.02	1.36
	T1 Anxiety	.15	.11	.13	1.40
2	T1 EA	−.03	.06	−.05	−.56	.00	.32

Outcome: Anger-Out at T2
Step		b	SEb	pr	t	ΔR2	ΔF
1	T1 Anger-Out	.50	.09	.47	5.76**	.22	33.17**
2	T1 NA	−.25	.10	−.23	−2.47*	.04	3.34*
	T1 Anxiety	.25	.10	.22	2.43*
3	T1 EA	−.09	.06	−.15	−1.63	.02	2.66

Notes.

^*p < .05.

^**p < .001.

All p-values were two-tailed. T1 = Time 1; T2 = Time 2; NA = Negative affect dimension of autobiographical anxiety memory; Anxiety = anxiety dimension of autobiographical anxiety memory; EA = experiential avoidance dimension of autobiographical anxiety memory; SIAS = Social Interaction Anxiety Scale; BDI-II = Beck Depression Inventory-II; Anger-In = subscale of the Multidimensional Anger Inventory; Anger-Out = subscale of the Multidimensional Anger Inventory.

Predicting concurrent and prospective depressive symptoms

As shown in Table 2, experiential avoidance in autobiographical memories was positively associated with depressive symptoms. However, experiential avoidance did not significantly predict changes in depressive symptoms over time.

Predicting concurrent and prospective anger symptoms

As shown in Table 2, experiential avoidance in autobiographical memories failed to show significant relations with anger symptoms.

Exploratory analyses

In prior studies, we found support for a “joint vulnerability” model with higher emotional disturbances and greater experiential avoidance linked with particularly problematic outcomes and a “joint resilience” model with the absence of emotional disturbances and an open, receptive attitude to negative emotions linked to adaptive outcomes (Kashdan & Breen, 2008; Kashdan, Morina, & Priebe, 2009; Kashdan & Steger, 2006). To test these models, we examined interactions between T1 symptoms and experiential avoidance in autobiographical memories on symptom changes. We failed to find empirical support for these interaction effects (from p =.19 in predicting changes in social anxiety scores to p = .42 in predicting changes in anger-out scores).

Discussion

We found that our idiographic approach to studying experiential avoidance to be reliable and valid. Compared with depressive and anger symptoms, our autobiographical memory measure of experiential avoidance showed robust relations with social anxiety. Of the three types of emotional disturbances, experiential avoidance only predicted increases in social anxiety symptoms over a 3-month period. The novel contribution of this study is that our measurement strategy embedded experiential avoidance into a personally meaningful context, allowing for objective evaluation by trained coders. Results from our idiographic, objective approach, converge with prior work illustrating how experiential avoidance can explain the transition from normal emotions into disorder (Eifert & Forsyth, 2005; Hayes et al., 1996).

People with excessive social anxiety are overly concerned about the possibility of social rejection and in response, engage in extensive efforts to control their anxiety and manage their impressions on other people (Clark & Wells, 1995; Rapee & Heimberg, 1997). Rigid attempts to regulate anxious thoughts and feelings, and manage impressions to avoid social rejection are a form of experiential avoidance. Attempts to regulate thoughts and feelings drain the limited supply of attention, stamina, and executive functioning capacity at any given point in time (Muraven & Baumeister, 2000). Even momentary engagement in experiential avoidance of this nature can impair functioning in subsequent activities by disrupting impulse control, lessening pain tolerance and intelligent thinking, and increasing reliance on rigid stereotypes (Baumeister et al., 2006). Regular tendencies to negatively evaluate, conceal, and avoid thoughts and feelings leads to a greater abundance of these deleterious outcomes as well as the selection of activities that prevent anxiety at the expense of strivings that are aligned with central values and interests (i.e., because they elicit some degree of anxiety and risk). More complex multi-method research designs can provide additional understanding of how self-regulatory processes and resource depletion factor into the etiology and maintenance of social anxiety problems (Kashdan & Steger, 2006).

Despite promising findings, it is important to recognize the small magnitude of experiential avoidance effects. The small magnitude of these effects is partially explained by the impressive stability of untreated social anxiety symptoms over time. Our multi-method approach most assuredly led to smaller effect sizes than previously found in studies of experiential avoidance and emotional disturbances. Most existing work has been narrowly focused on correlations between global self-report measures of emotional problems and global self-report measures of experiential avoidance that are completed during the same time period. We extended this work by asking people to recall prominent memories when they felt anxious and assessing acts of experiential avoidance in this context. We examined how objective coder ratings of experiential avoidance and affect embedded in these personal narratives relate to concurrent and prospective changes in self-report scales. To address the construct specificity of this measure, we contrasted experiential avoidance with objective coding of the intensity of negative affect and anxiety during the same narratives. Our results provided initial support that experiential avoidance efforts in a particular anxiety provoking situation are distinct from felt emotions. Despite our use of alternative methodologies, the vast majority of relations fit with prior theory and research.

Although experiential avoidance was less relevant to depressive and anger symptoms, this might be a consequence of relying on memories of anxious episodes in people’s lives. However, theory and research suggest that pain and distress of any nature are often antecedents and maintenance factors of anger (Berkowitz, 1989, 1990). When studying anger, situational contexts may be particularly important. For instance, after feeling rejected, the use of experiential avoidance as a regulatory strategy may increase the likelihood of internalized and externalized anger (e.g., Leary, Twenge, & Quinlivan, 2006). Data on anger in this study should be considered preliminary, given the relative paucity of research on anger compared with anxiety and mood conditions. Consider that there is nothing in the DSM other than intermittent rage disorder and singular symptoms in the criteria sets for various disorders (American Psychiatric Association, 2000). Additional work is needed on how to best define and operationalize the multiple dimensions of anger, as no consensus exists (Eckhardt et al., 2004).

There are other possibilities as to why many of the reported correlations were small and there was no support for a joint vulnerability model with experiential avoidance moderating growth in emotional disturbances over time. First, we relied on a single memory for each person and reliability would be enhanced with the collection and coding of multiple memories. Coding qualitative data is a resource intensive endeavor but offers a valuable neglected level of analysis in understanding psychological problems. Second, the stability of emotional disturbances reflects a floor effect with little possibility for further erosion. Third, people with emotional disturbances also exhibit negative interpretation biases that can affect reporting accuracy and variability in our measure of experiential avoidance. Another reason to question validity of our measure of experiential avoidance is that we relied on objective raters and it remains to be seen how precise people are at describing the process of how private events are experienced (meta-emotions and meta-cognitions) and what factors influence insight into this process. Fourth, writing skills and vocabulary might be an important source of non-random error when coding narratives. Fifth, the longitudinal period under study was arbitrary. To understand the role of experiential avoidance as a cause, correlation, or consequence, future work should examine periods of time when anxiety, depressive, and anger symptoms are most vulnerable to change. This includes early adolescence when social anxiety problems often develop (Schneier, Johnson, Hornig, Liebowitz, & Weissman, 1992) and periods of transition such as starting college or a new occupation, marriage, relocation, pregnancies, birth, and retirement. Finally, despite the use of an idiographic, autobiographical memory approach with an objective coding system, our outcome measures were limited to self-report ratings. Thus, all of the interpretative caveats relevant to self-report technologies are relevant.

To understand risk and protective factors for emotional disorders, researchers cannot rely solely on personality traits. Other levels of analysis to understand the totality of a person include their goals and strivings in the short and long term and their self-constructed life narratives (McAdams, 1996; Sheldon, 2004). In this study, we sought to begin understanding experiential avoidance in the context of people’s prominent remembered experiences, serving as part of their life narrative. We view this as a preliminary attempt to begin integrating life narrative perspectives into the study of psychopathology. For complex constructs, such as experiential avoidance, multi-method designs are needed that can address multiple levels of personality. Experimental and time-lag designs can address issues related to causality. In particular, we believe there is merit in using the current personal narrative approach and system to objectively code affect and experiential avoidance during critical periods of stress that might activate predispositions to psychological disorder. Examples of naturalistic samples to study experiential avoidance and how it might contribute to the development of disorder include (1) pre- and postnatal periods in parents, (2) before, during, and after military or police training, and (3) people working for organizations that are at high-risk for downsizing or bankruptcy. Understanding the narratives that people wield about their changing lives, and the embedded emotions and self-regulatory strategies (such as experiential avoidance) provide an opportunity to contrast competing, complex models of resilience, recovery, and disordered pathways in the aftermath of stress and loss.

Results of this study have implications for better understanding the phenomenology of experiential avoidance and varied relations to social anxiety, depressive, and anger symptoms. Reliance on experiential avoidance during a prominent memory of feeling anxious was associated with greater social anxiety and depressive symptoms, and an increase in social anxiety during a 3-month interval.

Appendix A

NARRATIVE CODING SYSTEM

All dimensions rated on a 1 (not at all) to 7 (extremely) scale should use the following:

General Guidelines

1. The person’s experiences, life, and the situation being described should be used as the context to make each rating (e.g., doing well on an exam may be more or less meaningful to a person depending on what is known about the person, their efforts and goals, and the situation).

2. Remember that some people will not provide that much detail and thus, there will be less information to derive ratings. Always use the available information.

3. If both general and specific memories are mentioned, only the specific memory should be considered. If multiple memories are listed, make ratings on the most salient memory.

4. For each rating, determine whether it meets one of the criteria with great intensity or whether it meets multiple criteria. Each way could lead to strong ratings.

5. If more than one feeling is mentioned in describing several people’s feelings throughout a memory, code according to the participant’s feelings.

Affective Intensity

When conducting ratings, consider explicit mention to the frequency, intensity, and/or longevity of the target. Some memories will clearly describe positive or negative experiences with no overt references to emotion. In these situations, make your ratings based on the positive tone (e.g., people make a positive evaluation of the situation or describe a situation that most people would find pleasant) or negative tone (e.g., people make a positive evaluation of the situation or describe a situation that most people would find aversive), respectively. For example, references to money or friends would reflect a positive tone and references to problems in relationships or being hospitalized would reflect a negative tone.

1. Global Positive Affect- generic, agreeable state of well-being characterized by emotions ranging from contentment to intense joy.

   1. positive affect ratings can reflect high arousal states (e.g., joy, excitement) or low arousal states (e.g., contentment, serenity)
2. Global Negative Affect- feelings of distress or discomfort characterized by emotions ranging from sadness and boredom to nervousness, shame, jealousy, and anger (includes frustration and other mundane feelings). Remember that hurt feelings reflects negative affect (e.g., feeling betrayed or ignored).

   1. negative affect ratings can reflect high arousal states (e.g., scared, angry) or low arousal states (e.g., bored, tired, sad)

3. Anxiety: This includes anticipation of feared stimuli, anxiety or fear about internal or external stimuli, or anxiety during the aftermath about what something means (e.g., perceived blunder during a social interaction may mean that the person might not call again) or acts of worry.

Experiential Avoidance―rated on a scale from 1 (not at all) to 5 (extremely)

Process involving excessive evaluations of private thoughts, feelings, and sensations, an unwillingness to experience these unwanted private events, and/or deliberate efforts to control or escape these events.

1. Ratings greater than 1 reflect a qualitative difference. That is, there is explicit endorsement of actions taken to intentionally alter or avoid the experience of anxiety because it is viewed as undesirable.

2. Make sure to differentiate negative thoughts (which can be nothing more than the cognitive component of anxiety) from experiential avoidance. The reference to excessive evaluations refers to meta-cognitions or evaluations about emotional responses (whether it be feelings, thoughts, or bodily reactions; e.g., I was worried about feeling anxious; anxiety is bad and is unpleasant).

3. Look for difficulties that people have in being in contact with and confronting disturbing feelings, thoughts, or circumstances. Avoidance often involves attempts to conceal, hide, disrupt, or block disturbing experiences. It can include attempts to suppress internal events, drinking or using other techniques to try and feel less or be numb, discontinuing the writing task because it is disturbing, avoiding therapeutic tasks, and isolating oneself from external stimulation.

4. Evidence of avoidance, withdrawal, or escape behavior, including internal events (not visible to others) indicates the presence of experiential avoidance (e.g., person describes not wanting to think about something).

Coping Ability

Rate the person’s ability to understand or their competence to effectively work, cope, or thrive in the situation being described. For example, how well do they adapt and respond to a situation? Focus on the process and not just the final outcome. After all, they may do well in the end but it may have required an inefficient use of regulatory resources (e.g., depletion, lost time and resources to brooding and psychological paralysis, inability to be mindfully engaged).

1. Use information provided by people in terms of how they feel that they are able to effectively deal with the given situation (self-efficacy)

2. Some degree of challenge needs to be confronted (no matter how little). Without a challenge, the rating should be not applicable. This is the only rating where not applicable is possible. However, only use this when there is clearly no challenge being faced.