The Affective Style Questionnaire: Development and Psychometric Properties

Abstract

Affective style is an individual difference variable that refers to tendencies for regulating emotions. The emotion research literature has consistently identified three general strategies to handle emotional reactions: some strategies are aimed at re-adjusting affect to adapt successfully to situational demands; other strategies intend to conceal or suppress affect; and a third approach is to tolerate and accept emotions including unwanted and aversive reactions. We conducted two studies to develop a self-report measure to assess these affective styles. In the first study (n = 434), a list of 127 items related to this construct was administered. A factor analysis supported three factors: habitual attempts to conceal or suppress affect (Concealing subscale; 8 items), a general ability to manage, adjust, and work with emotions as needed (Adjusting subscale; 7 items), and an accepting and tolerant attitude toward emotions (Tolerating subscale; 5 items). The scale showed satisfactory internal consistency. Furthermore, the respective subscales showed different pattern of relations with existing instruments measuring similar constructs. Findings were cross-validated in an independent sample (n = 495). The factor structure and results of psychometric analyses were replicated. The final 20-item Affective Style Questionnaire is a brief instrument to measure individual differences in emotion regulation.

One of the most remarkable features of humans is the capacity to regulate and adjust their emotions depending on particular situational demands. It is likely that this capacity is evolutionarily adaptive (e.g., Davidson, 2003; Ekman, 2003; Izard, 1992; Lazarus, 1991) and closely connected to cognitive appraisal processes that distinguish humans from non-humans (e.g., Frijida, 1986; Gross, 2002; Lazarus, 1991; Scherer & Ellgring, 2007).

Emotion regulation refers to the process by which people influence which emotions they have, when they have them, and how they experience and express these emotions (Gross, 1998). Consistent with previous authors, we define affective style as inter-individual differences in the sensitivity to and regulation of emotions (Davidson, 1998). Some affective styles effectively regulate the experience and expression of emotions in ways that increase progress toward valued aims, whereas other strategies have apparently unintended, counterproductive effects. For example, attempts to suppress emotions increase physiological arousal (Gross, 1998; Gross & Levenson, 1997), and rumination over negative emotional events prolongs angry and depressed affective states (Nolen-Hoeksema & Morrow, 1993; Rusting & Nolen-Hoeksema, 1998). In contrast, an accepting stance toward arousing emotional experiences without unnecessary attempts to change or avoid them has been linked to increased persistence in challenging situations and reductions in subjective distress (Hayes, Luoma, Bond, Masuda, & Lillis, 2006).

Emotion regulation strategies can be classified based on the time point when people engage in these activities during the emotion generation process and the efficacy of these efforts (Gross, 2002; Gross & John, 1997; Gross & Levenson, 1997). Antecedent-focused regulation occurs before the emotional response has been fully activated. This includes tactics such as attention deployment, situation modification, cognitive reframing of a situation, and any preparatory action (e.g., listening to particular energizing music before a work-out routine). Response-focused regulation reflects attempts to alter the experience or expression of ongoing emotions. This includes tactics such as suppression and acceptance-based attitudes.

Laboratory studies suggest that antecedent-focused strategies such as reappraisal are relatively effective for regulating emotions in the short-term, whereas suppression based response-focused strategies are often unintentionally counterproductive (e.g., Gross, 1998; Gross & Levenson, 1997). It has further been shown that people differ in their habitual use of antecedent- and response-focused emotion regulation strategies, and that these individual differences are meaningfully associated with emotional experiences and psychosocial functioning (Gross & John, 2003). For instance, people who are more reliant on reappraisal as a regulatory strategy experience better social functioning and greater well-being, as assessed by self-reports, peer reports, and reactivity to laboratory stimuli. In contrast, people who are more reliant on suppression as a regulatory strategy experience worse social functioning and well-being (Gross & John, 2003).

An important question pertaining to emotion regulation concerns the variation between people in their habitual tendency to use some regulatory strategies over others, especially if the preferred strategy has undesirable outcomes. A potential explanation for the persistence of ineffective emotion regulation is people’s acceptability and tolerance of particular emotional experiences (Salovey, Mayer, Goldman, Turvey, & Palfai, 1995). Some people respond to the onset of emotions by appraising them as intolerable and subsequently engage in avoidance, concealment, or other counterproductive response-focused interventions. Recently developed treatments for emotional disorders employ techniques that target such negative judgments of emotions and maladaptive emotional control efforts (Hayes, Strosahl, & Wilson, 1999; Segal, Williams, & Teasdale, 2002).

In sum, the emotion literature consistently identifies different affective styles for regulating emotions. The first style includes suppression and other response-focused strategies aimed at concealing and avoiding emotions after they arise. We refer to this affective style as concealing strategies. A second style characterizes people who are “more able to access and utilize emotional information in adaptive problem solving, and better able to modulate emotional experience and expression according to contextual demands” (Mennin, Heimberg, Turk, & Fresco, 2002, p. 88). These individuals possess the tools to readjust or balance emotions as needed to successfully navigate the rewards and punishments of everyday life. We refer to this affective style as adjusting. Finally, a third style reflects comfort and non-defensiveness in response to arousing emotional experiences as they exist in the present moment. This includes a strong tolerance of distress. We refer to this affective style as tolerating strategies.

To our knowledge no instrument exists that measures these three broad affective styles. The most relevant instruments are the 10-item Emotion Regulation Questionnaire (ERQ; Gross & John, 2003) measuring individual differences in expressive suppression and cognitive reappraisal; the 10-item Acceptance and Action Questionnaire-II (Bond et al., submitted; Hayes et al., 2004) measuring individual differences in the willingness to accept and work with private thoughts and feelings in the pursuit of valued goals (an aggregation of multiple facets); and the 36-item Difficulties in Emotion Regulation Scale (DERS: Gratz & Roemer, 2004) measuring various ways that people habitually find themselves unable to successfully regulate difficult, aversive emotional experiences. Although useful, the existing instruments are limited by the relatively restricted scope of possible affect regulation strategies and the constraints of a particular theoretical orientation. The goal was to develop a brief but psychometrically sound scale to measure the general affective styles of concealing, adjusting, and tolerating.

Study 1

Subject Sample

A total of 457 undergraduate students of Boston University (BU), 18 years of age or older, participated in this study. Due to missing data, the following analyses are based on 434 participants. The mean age of participants was 19.15 (SD = 2.61). The majority of the sample was female (67%) and Caucasian (68.1%). Remaining participants self-identified as being Asian or Asian-American (18.40%), Hispanic, Latino, or Mexican-American (3.7%), and African American (2.6%).

Procedure

Participants completed a web-based survey (PsychData) that included demographic questions, a pool of emotion regulation items for the development of our new scale, and several published self-report instruments. Participants received research credits toward psychology classes.

Item Pool Generation

The two authors generated 127 items assessing different ways of dealing with emotions. Many items derived were based on the work by Gross and John (1997; 2003) who distinguish antecedent- and response-focused strategies, and the acceptance and mindfulness based literature (e.g., Hayes et al., 1999). The complete item pool can be obtained from the authors.

Self-Report Measures

In addition to the 127 author-generated emotion items, several validated self-report measures were administered.

Brief COPE (Carver, 1997). The Brief COPE, an abbreviated version of the COPE (Carver, Scheier, & Weintraub, 1989), is a 28-item inventory consisting of 14 subscales. The instrument assesses individual differences in the use of effective and non-effective coping strategies. Each subscale has two items (α = .50 to .90). Examples of coping scales include Denial, Active Coping and Behavioral Disengagement.

Toronto Alexithymia Scale (TAS-20; Bagby, Parker, & Taylor, 1994). The TAS-20 is a 20-item scale to measure alexithymia, a construct reflecting difficulty identifying, describing, and being aware of emotions. The scale has become the most widely used measure of this construct with adequate psychometric properties (Parker et al., 2003).

Difficulties in Emotion Regulation Scale (DERS: Gratz & Roemer, 2004) contains 36 items to assess six dimensions of self-regulatory difficulties: nonacceptance of emotional responses (accept), difficulties engaging in goal-directed behavior (when upset; goals), impulse control difficulties (when upset; impulse), lack of emotional awareness (aware), limited access to effective emotion regulation strategies (strategies), and lack of emotional clarity (clarity). The scale shows adequate reliability (Gratz & Roemer, 2004) and some evidence of validity (e.g., Gratz, Rosenthal, Tull, Lejuez, & Gunderson, 2006).

Berkeley Expressivity Questionnaire (BEQ; Gross & John, 1995). The Berkeley Expressivity Questionnaire consists of 16 items to assess three facets of emotional expressivity: negative expressivity, positive expressivity, and impulse strength. The measure shows adequate psychometric properties (Gross & John, 1997).

Emotion Regulation Questionnaire (ERQ; Gross & John, 2003). This scale consists of 10 items and assesses individual differences in two emotion regulation strategies: expressive suppression and cognitive reappraisal. The scale shows good psychometric properties (Gross & John, 2003).

Acceptance and Action Questionnaire-II (AAQ-II; Bond et al., submitted). The AAQ-II is a 10-item measure, a revised refinement of the original scale (Hayes et al., 2004), to assess individual differences in acceptance and experiential avoidance. The psychometric characteristics are adequate.

Results

Factor Structure

The principal component analysis of the entire item pool resulted in 30 factors with Eigenvalues greater than 1 accounting for 67.48% of the total variance. A close inspection of the factor structure suggested that items loading on the first three factors describe three distinct affective styles, concealing, adjusting, or tolerating affect. The remaining factors could not be easily interpreted. In the next step, we selected items loading highly on these factors and eliminated items with poor item validity.

Psychometric Data

The first factor (Concealing), consisting of 8 items, accounted for 22.18% of the variance. The second factor consisting of 7 items and accounting for 15.81% of the variance was interpreted as the Adjusting subscale. The third factor, the Tolerating subscale, consisting of 5 items, accounted for 10.09% of the variance. These three factors were the only ones with Eigenvalues above 1. The scree plot further confirmed the 3-factor solution.

The correlation matrix was subjected to a varimax rotation. The resulting 3-factor structure is presented in Table 1 (Study 1 column). Concealing showed a mean of 23.94 (SD = 6.94, median = 24, mode = 24, range: 8–40); Adjusting showed a mean of 20.36 (SD = 5.05, median = 20, range: 7–34), and Tolerating showed a mean of 15.23 (SD = 3.42, media = 15, range: 6–25).

Table 1.

Factor Structure from Study 1 (Boston University; N = 434) and Study 2 (George Mason University; N = 495).

Items	Study 1			Study 2
	Concealing	Adjusting	Tolerating	Concealing	Adjusting	Tolerating
People usually can’t tell how I am feeling inside.	.65	−.02	−.23	.64	−.02	−.06
I often suppress my emotional reactions to things.	.64	−.03	−.03	.66	−.07	−.12
I am good at hiding my feelings.	.81	.06	−.03	.78	.05	−.02
People usually can’t tell when I am upset.	.72	.19	−.09	.77	.10	−.04
People usually can’t tell when I am sad.	.73	.20	−.09	.76	.10	−.02
I can act in a way that people don’t see me being upset.	.72	.19	.06	.73	.24	.03
I could easily fake emotions.	.60	.05	.14	.58	−.13	.09
I can hide my anger well if I have to.	.56	.33	.18	.58	.36	.04
I have my emotions well under control	.21	.63	−.10	.00	.62	.04
I can avoid getting upset by taking a different perspective on things.	.17	.63	.05	.10	.65	.06
I am able to let go of my feelings.	.13	.66	−.03	.05	.60	−.10
I can calm down very quickly	.14	.71	.05	.22	.72	.04
I can get out of a bad mood very quickly	.07	.68	.15	.01	.69	.07
I know exactly what to do to get myself into a better mood.	−.05	.70	.12	.00	.63	.11
I can get into a better mood quite easily.	−.00	.80	.08	−.05	.77	.03
I can tolerate having strong emotions.	.05	.20	.66	.00	.06	.72
It’s ok if people see me being upset.	−.25	.06	.61	−.27	.10	.61
It’s ok to feel negative emotions at times.	.04	−.07	.67	.03	−.07	.63
I can tolerate being upset.	.15	.27	.63	.13	.22	.53
There is nothing wrong with feeling very emotional.	−.05	−.10	.70	−.05	.04	.74

Note: The Table shows the factor scores (varimax rotation) of the items of the Affective Style Questionnaire items. High-loading items are printed in bold.

Internal consistency was acceptable for the Concealing (α = .84), Adjusting (α = .80), and Tolerating (α = .66) subscales. The item-total correlation coefficients were rs >.59 for Concealing, rs >.61 for Adjusting, and rs >.57 for Tolerating subscales (ps < .0001). The Concealing and Adjusting subscales showed a correlation of r = .17, the Adjusting and Tolerating subscales correlated at r = .13, and the Concealing and Tolerating subscales showed a correlation of r = −.08.

Construct Validity

Table 2 (Study 1 column) shows the correlations between the three subscales of the ASQ and related instruments. As expected, large correlations were observed between the ASQ-Adjusting subscale and the AAQ-II (r = .47) and the Reappraisal subscale of the ERQ (r = .54), and between the ASQ-Concealing subscale and the ERQ suppression subscale (r = .60) and BEQ-Negative Expressivity (r = −.68). Interestingly, we also found a moderate negative correlation between the ASQ-Tolerating subscale and the ERQ suppression subscale (r = −.34), suggesting that suppressing one’s emotions requires concealment and low distress tolerance. Also, the ASQ-Concealing and Adjusting subscales were differentially linked to difficulties describing and identifying feelings, and only the ASQ-Tolerating subscale was linked to being emotionally aware (r = −.34 with externally oriented subscale of TAS-20).

Table 2.

Correlations between Affective Style Questionnaire and Other Instruments in Study 1 (Boston University) and Study 2 (George Mason University).

Questionnaires	Study 1			Study 2
	Concealing	Adjusting	Tolerating	Concealing	Adjusting	Tolerating
ERQ
Reappraisal	.13*	.54**	.09	.14*	.57**	.14**
Suppression	.60**	−.03	−.34**	.52**	−.05	−.32**
BEQ
Negative Expressivity	−.68**	−.16*	.16*	−.70**	−.27**	.10*
Positive Expressivity	−.23**	.07	.28**	−.33**	.07	.28**
Impulse Strength	−.15*	−.39**	.19**	−.30**	−.38**	.21**
AAQ-II	−.05	.47**	.22**	−.03	.48**	.18**
Brief Cope
Self-Distraction	.10*	.12*	.09	.17**	.19**	.07
Active Coping	−.11	.08	.29**	−.12	.05	.27**
Substance Use	.08	.10*	.22**	.01	.14*	.21**
Emotional Support	.17**	.16**	.08	.20**	.24**	.13**
Instrumental Support	.04	−.15**	−.06	.06	−.07	−.01
Behavioral Disengagement	−.04	.20**	.03	−.01	.14**	.10*
Venting	−.19	.01	.18**	−.10*	.03	.07
Positive framing	.04	−.04	.14**	.04	−.02	.07
Planning	−.02	−.32**	.04	−.07	−.32**	.11*
Humor	.08	.06	.02	.12**	.05	.14**
Acceptances	−.02	−.02	−.05	−.07	−.01	−.01
Religion	−.19**	−.10*	.05	−.14**	−.14**	.06
Self-Blame	.05	−.24**	−.09	.01	−.20**	−.20**
TAS-20
Difficulty Identifying Feelings	.18**	−.27**	−.16**	.11*	−.25**	−.12**
Difficulty Describing Feelings	.38**	−.18**	−.28**	.29**	−.17**	−.25**
Externally Oriented	.13	−.02	−.34**	.16**	−.03	−.28**
DERS
Total	.11*	−.48**	−.32**	.02	−.51**	−.23**
Clarity	.18**	−.30**	−.28**	.09*	−.30**	−.24**
Aware	.15**	−.08	−.46**	.20**	−.13	−.39**
Impulse	−.04	−.44**	−.13**	−.09*	−.44**	−.08
Accept	.10*	−.27**	−.31**	.12**	−.27**	−.15**
Goals	−.06	−.40**	−.20	−.08	−.43**	−.08
Strategies	.06	−.54**	−.19**	−.04	−.54**	−.08**

Notes:

^*p < .05;

^**: p < .001. The Table shows product moment correlations between the ASQ subscales and other questionnaires. AAQ-II: Acceptance and Action Questionnaire; BEQ: Berkeley Expressivity Questionnaire; ERQ: Emotion Regulation Questionnaire; DERS: Difficulties in Emotion Regulation Scale; TAS-20: Toronto Alexithymia Scale.

As for the DERS subscales, the highest correlations were observed between the Adjusting subscale and the DERS strategies subscale (limited access to effective emotion regulation strategies), goals subscale (difficulties engaging in goal-directed behavior), and the impulse subscale (impulse control difficulties). Moderate correlations were also observed between the Tolerating subscale and the DERS accept subscale (nonacceptance of emotional response) and the aware subscale (lack of emotional awareness). The three ASQ subscales showed consistently low correlations with the Brief Cope Subscales, supporting their discriminant validity.

Discussion

With the goal of developing a short and psychometrically sound measure of affective style, our results were promising. We found evidence for 3 meaningful and interpretable factors leading to Concealing, Adjusting, and Tolerating affect subscales. Each of these subscales showed high item validity and internal consistency. Inter-correlations among the ASQ subscales with other measures of emotion regulation, psychological flexibility, and other personality traits provided evidence of convergent and discriminant validity. To further evaluate the structure and psychometric properties of our scale, we conducted a second study with an independent sample.

Study 2

Subject Sample

A total of 528 undergraduate students of George Mason University (GMU) in Virginia participated in this study. Due to missing data, the following analyses are based on 495 participants.

Participants were at least 18 years old with a mean age of 22.02 (SD = 5.23). The majority of the sample was female (78%) and Caucasian (54.5%). Remaining participants self-identified as being Asian or Asian-American (17.8%), Hispanic, Latino, or Mexican-American (7.4%), and African American (8.4%).

Compared with the Study 1 sample, the GMU sample was significantly older, t (797.57) = −11.23, p < .0001, and there was a greater proportion of women, χ² (1) = 15.20, p < .0001, and racial and ethnic diversity, χ² (1) = 18.01, p < .0001.

Procedure

The procedure and instruments were identical to Study 1.

Results

Factor Structure

The principal component analysis with the 20-item ASQ resulted again in 3 factors with Eigenvalues greater than 1 accounting for 49.6% of the total variance. As shown in Table 1 (Study 2 column), the factor structure of Study 1 was replicated perfectly. The Concealing, Adjusting, and Tolerating factors accounted for 19.70%, 18.34, and 11.54% of the variance, respectively. The scree plot confirmed the 3-factor solution.

Psychometric Data

Concealing showed a mean of 22.60 (SD = 6.31, median = 22, mode = 21, range: 8–40); Adjusting showed a mean of 20.96 (SD = 5.15, median = 21, range: 7–35), and Tolerating showed a mean of 15.47 (SD = 3.43, media = 15, range: 6–25).

Internal consistency was acceptable for Concealing, α = .84, Adjusting, α = .82, and Tolerating, α = 68, subscales. The item-total correlation coefficients were high for Concealing, rs >.61, Adjusting, rs >.66, and Tolerating, rs >.64, subscales (ps < .0001). Concealing and Adjusting subscales showed a correlation of r = .30, the Adjusting and Tolerating subscales correlated at r = .17, and the Concealing and Tolerating subscales showed a correlation of r = −.03.

Construct Validity

Table 2 (Study 2 column) shows the correlations between the ASQ subscales and other instruments, with patterns similar to Study 1. The strongest relations with the ASQ-Concealing subscale were the BEQ-Negative Expressivity (r = −.70), ERQ-Suppression (r = .52), and BEQ-Positive Expressivity (r = −.33) subscales. The strongest relations with the ASQ- Adjusting subscale were ERQ-Reappraisal (r = .57), DERS Strategies (r = −.54), and DERS Goals (r = −.43) subscales and the AAQ-II (r = .48); indicative of adaptive regulation and psychological flexibility. The strongest relations with the ASQ-Tolerating subscale were DERS Aware (r = −.39), ERQ-Suppression (r = −.32), BEQ-Positive Expressivity (r = .28), and TAS-20 Externally Oriented (r = −.28) subscales; indicative of effective use of acceptance and mindfulness strategies.

Discussion

The goal of this study was to cross-validate the findings of Study 1 and further validate the ASQ in a sample of college students from a different university. Although there were differences between the two samples in terms of gender, racial, and ethnic diversity, the results were replicated. The factor analysis again revealed the identical 3-factor solution with the same items loading on subscales interpreted as concealing, adjusting or tolerating affect. Based on the pattern of correlations with other instruments measuring emotion regulation, coping and personality, the results support the uniqueness of each ASQ subscale and their convergent and discriminant validity.

General Discussion

Our objective was to develop a short, reliable, and valid measure of affective style. Based on a thorough review of the emotion and clinical literature, a large pool of items was generated. Using two large samples of college students at separate universities, we were able to create a 20-item scale consisting of 3 subscales: Concealing, Adjusting, and Tolerating affect. Given the brevity of the measure, the scale shows excellent psychometric properties. The subscales map onto existing measures of emotion regulation and also include additional facets of affective style.

An important limitation of our work is the sole reliance on undergraduate student samples. Although we found evidence for the separability of each affective style, we did not evaluate whether the ASQ subscales differentially predict how people respond to aversive and rewarding events in the laboratory and naturalistic environment. It will be important to employ multi-method assessments and investigate the temporal course of how people regulate their emotions in future research. Despite the limitations of our two initial studies, this measure may become a potentially useful tool in basic and clinical research.

Appendix ASQ

Instructions

We are interested in how you experience and manage your emotions. Obviously, different situations bring out somewhat different responses, but think about what you usually do. Please try to respond to each item separately in your mind from each other item. Do not indicate agreement with things that you think you should do or wish you do. Instead, choose your answers thoughtfully, and make your answers about what is true FOR YOU. Please answer every item. There are no “right” or “wrong” answers, so choose the most accurate answer for YOU-- not what you think “most people” would say or do. Use the scale below to answer each item.

1	2	3	4	5
not true of me at all	a little bit	moderately	quite a bit	extremely true of me

1.	People usually can’t tell how I am feeling inside.	1—2—3—4—5
2.	I have my emotions well under control	1—2—3—4—5
3.	I can tolerate having strong emotions.	1—2—3—4—5
4.	I can avoid getting upset by taking a different perspective on things.	1—2—3—4—5
5.	I often suppress my emotional reactions to things.	1—2—3—4—5
6.	It’s ok if people see me being upset.	1—2—3—4—5
7.	I can calm down very quickly	1—2—3—4—5
8.	I am able to let go of my feelings.	1—2—3—4—5
9.	I am good at hiding my feelings.	1—2—3—4—5
10.	People usually can’t tell when I am upset.	1—2—3—4—5
11.	It’s ok to feel negative emotions at times.	1—2—3—4—5
12.	I can get out of a bad mood very quickly	1—2—3—4—5
13.	People usually can’t tell when I am sad.	1—2—3—4—5
14.	I can tolerate being upset.	1—2—3—4—5
15.	I can act in a way that people don’t see me being upset.	1—2—3—4—5
16.	I know exactly what to do to get myself into a better mood.	1—2—3—4—5
17.	There is nothing wrong with feeling very emotional.	1—2—3—4—5
18.	I could easily fake emotions.	1—2—3—4—5
19.	I can get into a better mood quite easily.	1—2—3—4—5
20.	I can hide my anger well if I have to.	1—2—3—4—5