How Do People Use Reappraisal? An Investigation of Selection Frequency and Affective Outcomes of Reappraisal Tactics

Abstract

Although the effects of different emotion regulation strategies are well-documented, most studies to date have focused on the selection and implementation of broad strategies, while overlooking the selection and implementation of specific tactics to enact those strategies. The present research investigated the strategy of cognitive reappraisal and the differences in selection frequency and affective outcomes that are associated with the implementation of different reappraisal tactics to enact that strategy. Participants completed a laboratory task in which they were instructed to reappraise or not to reappraise negative images and reported on their use of specific reappraisal tactics for every trial. Using established reappraisal tactic coding, we assessed how people selected from among common tactics for each image (Study 1) and all tactics (Study 2) and implemented those tactics to reappraise negative images. We compared reappraisal tactic selection and implementation when used during instructed reappraisal vs. during spontaneous reappraisal, in the non-reappraise condition. Results of both studies indicate that tactics were used more often when instructed to reappraise vs. when spontaneously reappraising. Participants used some tactics (e.g., reality challenge) more frequently compared to the rest of the tactics in both conditions. Negative affect was lower following instructed vs. spontaneous reappraisal. Some tactics (e.g., change current circumstances) were more effective at decreasing negative affect in both conditions. Knowing which reappraisal tactics are most frequently selected, and their affective outcomes when used when prompted or spontaneously, may help us better understand how to improve people’s ability to use reappraisal to achieve their emotional goals.

Imagine you have been interviewed for a highly-sought-out position at a prestigious company. Maybe the interview went not as you had expected, and you are feeling sad, embarrassed, and unmotivated. There are many opportunities to think about the situation differently to feel better. You could re-frame the interview as not as bad as it seemed from the candidate’s point of view and consider that the offer could still be on the table. You could also accept that rejection is a common part of career progression. There are multiple ways to reconsider your emotions with the same goal of feeling less negative. Knowing how often these various types of reinterpretations are used, as well as their affective outcomes (i.e., efficacy), may help us better understand how to improve people’s ability to achieve their emotional goals.

Emotion Regulation, Cognitive Reappraisal, and Tactics

Emotions can be helpful in achieving goals, but when they do not match with our context and hinder our ability to achieve our goals, people often seek to change them. Emotion regulation is the process by which individuals influence their experience and expression of emotions (Gross, 1998). There are many strategies that can be used to effectively downregulate negative emotions, but reappraisal (sometimes also referred to as cognitive reappraisal) is largely accepted as one of the most adaptive and common strategies (Gross, 1998; Gross & John, 2003; McRae & Gross, 2020). Reappraisal is an emotion regulation strategy that involves reconsidering, reevaluating, or reframing an emotional situation to change its meaning (Gross & John, 2003). For example, someone who just had a lackluster job interview could tell themselves that perhaps it is an opportunity to get professional feedback to improve their interview skills, or that perhaps in a few years this will be seen as a pivotal change in their career trajectory. The new narrative changes the initial reaction of sadness and disappointment, and these negative emotions may become more manageable.

Reappraisal is often treated as a unitary concept when its use is directly compared to other emotion regulation strategies such as expressive suppression and distraction (Hayes et al., 2010; Sheppes et al., 2011), but it can be implemented using a number of different tactics. For instance, prior research distinguished between situation-focused and self-focused reappraisals (Ochsner et al., 2004), compassion-focused and benefit-focused reappraisals (Witvliet et al., 2010), and between four distancing forms of cognitive reappraisals (i.e., spatial, temporal, objective, and hypothetical; Powers & LaBar, 2019). In an attempt to form a broader tactic categorization scheme, McRae et al. (2012) differentiated at least six distinct reappraisal tactic categories (i.e., change current circumstances, change future consequences, reality challenge, acceptance, distancing, and agency) in response to picture-based reappraisal task. According to McRae et al. (2012), a change current circumstances tactic alters the interpretation of the present situation as being much better than it first seemed. A change future consequences tactic also involves changing interpretations of the situation for the better, but it requires invoking a future improvement rather than reinterpreting the present. A reality challenge tactic contests the authenticity of the situation (e.g., “it isn’t real” or “it’s fake/staged”). An acceptance tactic normalizes the situation, focusing on the fact that “sometimes things like this just happen” and “in the grand scheme of things, this is a normal or frequent event.” In using a distancing tactic, individuals attempt to reconsider the situation from an objective or psychologically removed perspective thereby minimizing the event’s salience or personal relevance (e.g., “I don’t know these people” or “this does not impact me”; Powers & LaBar, 2019). An agency tactic focuses upon one’s ability to handle the problem, and one might affirm that “I have the skills to change things for the better.” Lastly, a technical–analytic–problem solving tactic is practical tactic of reappraisal, focusing on what steps can be taken to solve a problem (i.e., “this is a problem to be solved by focusing on specific steps to be taken”).

Frequency of Tactic Selection

When extended to reappraisal, the process model of emotion regulation (Gross, 2015) distinguishes between how often people select reappraisal (or reappraisal tactics; selection frequency), and how well reappraisal is implemented to achieve one’s emotional goal (affective outcomes, McRae & Gross, 2020). Previous studies, focused on reappraisal strategy selection frequency in picture-based (or film-based) reappraisal tasks, have allowed us to estimate how often people use reappraisal when it is an option (e.g., Suri et al., 2015; Nook et al., 2020). Compared to other strategies, reappraisal tends to be selected more frequently than expressive suppression (Opitz et al., 2015; Volokhov & Demaree, 2010) and less frequently than distraction (Scheibe et al., 2015). In laboratory studies comparing selection frequency of reappraisal to no strategy at all, Suri et al. (2015) found that participants chose to reappraise on fewer than 50% of trials. In these paradigms, researchers have also begun to estimate the frequency of spontaneous reappraisal strategy use (how often people reappraise in the absence of the instruction to downregulate; Volokhov & Demaree, 2010). Surprisingly, it seems to occur infrequently, on only 16% of trials (Suri et al., 2015). In contrast, during paradigms where reappraisal is prompted, compliance with the reappraisal instruction occurs relatively frequently, on over 88% of trials (Nook et al., 2020).

In addition to reports of how often people use reappraisal broadly, in recent years a small number of studies have measured the frequency of selection of specific reappraisal tactics (McRae et al., 2012; Daros et al., 2020; Nook et al., 2020). With the goal of decreasing negative emotion to negative images, McRae et al. (2012) and Nook et al. (2020) identified change current circumstances, reality challenge, and change future consequences to be the three most frequently selected tactics, while distancing and acceptance were less frequently selected. Similarly, Daros and colleagues (2020) observed that both clinical populations (i.e., individuals with borderline personality disorder and individuals with mixed anxiety and/or depressive disorders) and healthy controls consistently chose to reappraise using tactics like reality challenge and change future consequences. These studies provide initial indication that selection frequency may be tactic-specific, yet they did not assess variation in selection frequency when more than one tactic option was available. Moreover, all reappraisals for a classic reappraisal task (changing emotional responses to negative images) were coded retroactively by trained research assistants. To better understand the specificity in selection frequency of tactics, it is necessary to measure the self-reported use of multiple reappraisal tactics on a trial-by-trial basis.

Affective Outcomes of Tactic Implementation

Across several studies, reappraisal has been demonstrated to be a successful strategy when implemented (Waugh et al., 2022), in that individuals demonstrate changes in emotion in accordance with their goals (Webb et al., 2012). The differential affective outcomes of tactic implementation are less understood. One study employing instructed use of two different tactics (i.e., situation-focused and self-focused) found that the tactics were equally effective (Ochsner et al., 2004). Other researchers have observed superiority of distancing (Powers & LaBar, 2019; Shiota & Levenson, 2012) or change current circumstances and change future consequences (Martin & Dahlen, 2005; Waugh et al., 2016) in achieving the goal of downregulating negative emotion. These studies experimentally manipulated which tactics were implemented, thus were unable to simultaneously measure the frequency of tactic selection with the effectiveness of their implementation. McRae et al. (2012) investigated selection frequency and affective outcomes and found reality challenge to be relatively less effective than the other tactics at changing affect. However, because tactic use was measured retrospectively and not measured on each trial, these reports of affective outcomes should be considered preliminary. Additionally, because selection frequency and affective outcomes were examined on a participant level rather than a trial level, conclusions from this investigation only provide support for understanding how people who use a particular tactic fare, and not for understanding how using a particular tactic affects affective outcomes. A study measuring selection frequency and affective outcomes of tactics on a trial-by-trial basis would paint a clearer picture of the relationship between the two.

Prompted vs. Spontaneous Reappraisal

Nearly all studies to date that manipulated reappraisal by instruction compared affective outcomes of reappraisal to a non-reappraisal condition. Despite strong evidence that the reappraisal instruction produces meaningful changes in use of regulation and affective outcomes, few studies have measured trial-level reappraisal both when prompted and spontaneously. Prompted selection refers to the use of reappraisal following the instruction to decrease negative emotion (regulation condition), while spontaneous selection refers to the use of reappraisal in the absence of the instruction to use any emotion regulation or decrease negative emotion (non-regulation condition). Suri et al. (2015) indicated that while individuals select reappraisal more often when prompted compared to spontaneously, selection frequency of prompted reappraisal is not 100%, and spontaneous reappraisal does occur. Furthermore, there is little data on spontaneous tactic frequency, because participants are typically only instructed to report on the reappraisal tactics they used when instructed to reappraise. Therefore, our goal was to measure trial-by-trial tactic selection when individuals were instructed to reappraise, and also when they were not, providing a novel estimate of spontaneous reappraisal tactic frequency selection, as well as tactic frequency when prompted to reappraise.

Additionally, the effect of instruction on the affective outcomes of tactic implementation remains unknown. Some studies that focus on the independent effects of goals and means in emotion regulation suggest that activating a goal (e.g., to downregulate negative emotion) results in desirable affective outcomes even without instructing specific means of doing so (i.e., a given reappraisal tactic; Tamir et al., 2019). According to this formulation, when reappraisal is not instructed, individuals may not have an explicit goal to downregulate, and therefore, spontaneous tactic implementation might be less successful at decreasing negative affect. By contrast, instructed reappraisal might result in significantly greater changes in affective outcomes. However, to our knowledge, no study to date has compared affective outcomes of tactics used following prompted vs. spontaneous reappraisal.

Current Studies

Extant research on reappraisal tactics does not address the following gaps: tactic selection frequency in response to individual emotional stimuli, affective outcomes associated with tactic implementation in response to individual stimuli, and differences in selection frequency and affective outcomes of tactics when reappraisal is prompted vs. spontaneous. In order to address this, the present set of studies investigated the selection frequency and affective outcomes of reappraisal tactics using an experimental reappraisal manipulation that enables measuring the tactic used and its affective outcome in each trial. Further, participants reported the tactics they used following instructions to either regulate or to respond naturally, thereby allowing for both prompted and spontaneous tactic selection to be measured independently. In our research, we used established, but arguably not exhaustive, set of tactics formed by McRae et al. (2012). In the first study, we investigated the selection frequency and affective outcomes for the most common tactics for each image. In the second study, we aimed to replicate the results of Study 1 and assess the selection frequency and affective outcomes for all reappraisal tactics specified in McRae et al.’s (2012) categorization scheme.

Study 1

Method

Participants

Our previous investigation (McRae et al., 2012) used between-subjects methods with a total N = 58. Given trends for increasingly well-powered studies, we used N = 58 as a minimum recruitment goal for our within-subjects design but instructed experimenters to continue to enroll participants beyond this goal until the end of the academic term during which data were collected to increase power and to allow for any necessary exclusions based on non-compliance with the task. During the academic term designated for data collection, 88 undergraduate students from the University of Denver participated in the study. Data from three students were excluded from the sample due to an apparent lack of understanding of the task as indicated by post-task debriefing questionnaires and low response variability. The final sample included 85 participants (64 females and 21 males). We also collected information about participants’ gender identities. At the time of the experiment, 62 participants identified as women, 21 participants identified as men, one participant chose “other” as a response, and one participant chose not to answer. The mean age of the final sample included in analyses was 19.08 years old with a range of 18 to 23 years old. The racial composition was 82.4% Caucasian, 14.1% Asian, 1.2% biracial and 2.4% missing. In addition, 4.7% of participants identified their ethnicity as Hispanic/Latino. Students were recruited through SONA systems (a cloud-based subject pool software for experiment management; Research Participation System, 2018) and received course credit for their participation. Inclusion criteria were: enrollment at the University of Denver, working proficiency in English, 18 years of age or older, documented consent to participate. Exclusion criteria were: age less than 18 years, individuals with severe cognitive disability, noncompliance with task instructions as indicated by post-task debriefing questionnaires.

Procedure

Participants engaged in a computerized emotion regulation task similar to those used by our group and others previously (Jackson et al., 2000; McRae et al., 2012). The experiment was conducted online through Qualtrics (Qualtrics XM, 2005) in 2018 – 2019. Forty-eight negative images were selected from the International Affective Picture System (Lang et al., 1997). Images were counterbalanced into two equal groups based on normative ratings of valence and arousal as well as on the average decrease in negative emotion that occurs while reappraising, extracted from an in-house database of previous reappraisal studies. Images were selected based on moderate levels of intensity so that they would elicit negative emotion without subjecting participants to unethical levels of distress. Two versions of the task were created in order to counterbalance the images such that each image was assigned to a different experimental instruction (24 “look” / 24 “decrease”) in each version. The images were then pseudorandomized so that the same instruction was shown no more than three times in a row. The two resulting versions of the task did not significantly differ on levels of affect, reappraisability, or tactic usage. Task version was counterbalanced across participants.

Participants were trained in reappraisal by reading instructions, answering multiple-choice questions about reappraisal correctly, and completing several practice trials before starting the computer task. During each trial of the task, participants viewed an instruction (2 seconds) to either “look” or “decrease” which was followed by a negative image (7 seconds) (Figure 1). When instructed to “look,” participants were told to keep their eyes on the image the entire time and feel whatever emotions surfaced naturally. When instructed to “decrease,” they were asked to “tell yourself something about the image that lessens your negative emotional response or makes you feel better. In other words, reinterpret, reevaluate, or gain a new perspective that changes your emotional response.” During training, participants were provided with a practice image and subsequent examples of how the image could be reinterpreted.

Figure 1. Trial Structure of Experimental Task.

Note: Participants viewed an instruction to “look” (spontaneous) or “decrease” (prompted) for two seconds and a negative image for seven seconds. Reporting of negative affect, tactic use, and free response when appropriate was self-paced.

When the negative image disappeared from the screen, participants reported how negative they felt about the image on a scale of 1 (not at all negative) to 7 (extremely negative). The negative affect rating scale was followed by a multiple-choice question. Participants were presented with 3 of 6 tactics (Table 1) and indicated whether their reinterpretation of the image fit into one of those tactics.

Table 1.

Reappraisal Tactic Options and Choice Text for Multiple-choice Options

Tactic Name	Multiple-Choice Option
Change current circumstances	It is not as bad as it first seemed.
Reality challenge	It is not real / It is fake or staged.
Change future consequences	It will get better soon / in the future.
Agency	Someone has the ability to change things for the better.
Distancing	I don’t know these people / this does not relate to me.
Acceptance	This is normal or natural, and sometimes things like this happen.

Note: Only multiple-choice options were presented to participants during the task (tactic names were not presented)

The most frequently used tactic categories were calculated for each image based on coded free responses from a previous reappraisal study (Davis et al., 2014; Waugh et al., 2016; see supplemental materials). Since a technical–analytic–problem solving tactic was not among the most frequently used tactics for any of the images, we did not include it in our task. Tactic choices appeared in all conditions to match trial structure across conditions and to capture spontaneous, uninstructed use of tactics. Since the “look” instruction specifically discouraged reinterpretation of the image, a response option of “I did not or could not change my emotional response to the image” was always presented. In addition, an “Other” option was available for every trial, and participants were asked to elaborate in an open text box if they selected this option. Results presented here are summaries of trials with these multiple-choice options. To incentivize generation of reappraisals for each trial rather than selecting a multiple-choice response post-hoc, attention check trials were used. For these, on randomly selected trials, participants were asked to describe what they were telling themselves when viewing the image by typing in a text box rather than responding via multiple-choice. Written responses from both the “other” and attention check trials were coded into tactic categories, and patterns of tactic selection were compared against the multiple-choice data to ensure that participants were appropriately generating reappraisals in-line with task instructions (see supplemental materials).

Following completion of the main task, participants completed a battery of surveys and questionnaires about their experience and emotional state (see supplemental materials). Average completion time for the experiment – including the instructions, main task, and questionnaires – was 60 minutes. The procedure was approved by the University of Denver Institutional Review Board. The materials and data for the study can be found on the Open Science Framework website (https://osf.io/3ugk7/).

Analytic Strategy

Selection frequency.

The frequency of tactic selection was calculated as a proportion of the number of tactic selections per participant out of the number of tactic presentations. To evaluate selection frequency, we used a 2 (instruction: “look” vs. “decrease”) × 8 (tactic type)¹ repeated measures analysis of variance (ANOVA) model with tactic type and instruction as within-subjects factors. To follow-up, paired-samples t tests were used to further identify which pairwise differences in frequency of tactic selection met our significance threshold. Because Mauchly’s Test of Sphericity indicated that the assumption of sphericity had been violated for tactic type effect, χ²(27) = 258.95, p < .001 and interaction of instruction and tactic type, χ²(27) = 356.61, p < .001, Greenhouse-Geisser corrected results are reported. To conduct these analyses, we used SPSS software.

Affective outcomes.

The affective outcomes of tactic implementation were analyzed using a linear mixed effect model with a random intercept for each participant and a random intercept for stimulus that had negative affect as the dependent variable and instruction, tactics, and their interaction as predictors. To follow up, paired-samples t tests were used to further identify pairwise differences in affective outcomes met our significance threshold. To conduct these analyses, we used ‘lme4,’ ‘lmerTest,’ and ‘emmeans’ packages in R software.

Transparency and Openness

In our manuscript, we report how we calculated our sample sizes, data exclusions, and analytic strategies. The data, analysis code and instruction materials can be found at https://osf.io/3ugk7/.

Results

Selection Frequency

There was a significant main effect of tactic type, F(4.02, 337.96) = 39.81, p < .001, η_ρ² = 0.32, on selection frequency, qualified by a significant interaction between instruction and tactic type, F(3.32, 279.25) = 43.89, p < .001, η_ρ² = 0.34. One way of characterizing this interaction is to examine the effects of instruction on each tactic’s frequency of selection. Though patterns of relative selection frequency of tactics were largely similar between the “look” and “decrease” conditions, prompting with the reappraisal instruction was associated with a significant increase in selection frequency across nearly all tactics. Paired t tests revealed that participants selected the following six tactics more frequently following the “decrease” as compared to “look” instruction: agency (t(84) = 4.43, p < .001, 95% CI of difference [0.08, 0.23]), change current circumstances (t(84) = 4.18, p < .001, 95% CI of difference [0.03, 0.07]), change future consequences (t(84) = 4.00, p < .001, 95% CI of difference [0.03, 0.09]), distancing (t(84) = 5.26, p < .001, 95% CI of difference [0.14, 0.31]), reality challenge (t(84) = 9.06, p < .001, 95% CI of difference [0.22, 0.34]), and other (t(84) = 5.45, p < .001, 95% CI of difference [0.04, 0.08]). Conversely, the did not / could not reappraise option was selected less frequently following the “decrease” compared with the “look” instruction (t(84) = 8.84, p < .001, 95% CI of difference [−0.48, −0.30]) (Figure 2).

Figure 2. Selection frequency of prompted versus spontaneous tactics in Study 1 and Study 2.

Note: Significant pairwise differences (p < .05) between prompted and spontaneous conditions are marked with a “*.” Note that proportions were adjusted for presentation frequency in Study 1, but did not need this adjustment in Study 2. Error bars represent the standard error of the mean.

Another way to characterize the interaction between instruction and tactic type on selection frequency is to report relative tactic use separately for the prompted and spontaneous conditions. When instructed to respond naturally and not decrease negative emotion, participants reported that they did not / could not reappraise on 51.56% of trials on average. The most frequently selected spontaneous tactics were acceptance (M = 0.22, SE = 0.02), agency (M = 0.18, SE = 0.02), and reality challenge (M = 0.21, SE = 0.02); the least frequently selected spontaneous tactics were change current circumstances (M = 0.06, SE = 0.01), distancing (M = 0.05, SE = 0.02), and other (M = 0.05, SE = 0.01) (see Figure 2).

When instructed to decrease negative emotion (prompted), participants reported that they did not / could not reappraise on only 12.69% of trials on average. The most frequently selected prompted tactics were agency (M = 0.34, SE = 0.03), distancing (M = 0.28, SE = 0.04), and reality challenge (M = 0.49, SE = 0.02); the least frequently selected prompted tactics were change current circumstances (M = 0.10, SE = 0.01), change future consequences (M = 0.16, SE = 0.01), and other (M = 0.11, SE = 0.01) (see Figure 2).

Affective Outcomes

There was a significant main effect of instruction on affective outcomes, F(1, 3247.8) = 7.11, p = .008, and a significant main effect of tactic type, F(7, 3315.8) = 30.83, p < .001, qualified by a significant interaction between instruction and tactic type, F(7, 3307.3) = 2.66, p = .01. (Figure 3). Paired t tests revealed that participants reported higher negative affect following the “look” as compared to “decrease” instruction for two tactics: agency (t(3316) = 2.83, p = .005, 95% CI of difference [0.19, 1.03]), and other (t(3308) = 3.54, p < .001, 95% CI of difference [0.27, 0.93]).

Figure 3. Affective outcomes of prompted versus spontaneous tactics in Study 1 and Study 2.

Note: Significant pairwise differences (p < .05) between prompted and spontaneous conditions are marked with a “*.” Negative affect was generally lower following the instruction to decrease, but patterns of affective outcomes associated with tactic implementation were largely similar across conditions. Error bars represent standard error of the mean.

One way to characterize the main effect of tactic type and the tactic type x instruction interaction is to examine affective outcomes by tactic for the prompted and spontaneous conditions. The tactics associated with the lowest negative affect ratings (i.e., best affective outcomes) in the spontaneous condition were acceptance (M = 3.69, SE = 0.18), change current circumstances (M = 3.58, SE = 0.18), and distancing (M = 3.72, SE = 0.42); the tactics associated with the highest negative affect ratings (i.e., worst affective outcomes) were agency (M = 4.30, SE = 0.21), change future consequences (M = 4.14, SE = 0.17), and other (M = 4.08, SE = 0.19 ) (see Figure 3).

The tactics associated with the lowest negative affect ratings (i.e., best affective outcomes) in the prompted condition were, as in spontaneous condition, change current circumstances (M = 3.38, SE = 0.16), and distancing (M = 3.56, SE = 0.25), but also other (M = 3.48, SE = 0.16); the tactics associated with the highest negative affect ratings (i.e., worst affective outcomes) were agency (M = 3.69, SE = 0.19), change future consequences (M = 3.98, SE = 0.16), and reality challenge (M = 3.70, SE = 0.15) (see Figure 3).

Discussion

Study 1 investigated the selection frequency and affective outcomes of six specific reappraisal tactics using an experimental emotion regulation task. Participants rated negative emotion and reported tactic usage on a trial-by-trial basis following the instruction to either decrease emotion (prompted) or to respond naturally (spontaneous) to series of negative images. We observed differences in selection frequency and affective outcomes among the tactics. Agency and reality challenge were among the most frequently selected tactics in both conditions while change current circumstances was among those that were associated with the lowest level of negative affect. Differences in tactic frequency and effectiveness were clear across instruction condition, but the instruction to reappraise also changed the use and effectiveness of some tactics. Although tactics were used more during the prompted reappraisal condition, a few of them were more effective when prompted (e.g., agency) while most did not vary in effectiveness depending on condition (e.g., reality challenge). These results might help us better understand how to improve people’s ability to use reappraisal in order to achieve their emotional goals. While the results of Study 1 were informative, some aspects of our design limited broader interpretation of our findings. One major limitation of our design, especially for estimating frequency, is the uneven presentation of reappraisal tactic across the experiment. To resolve this issue and replicate our findings, we conducted Study 2, with some critical modifications to our experimental design.

Study 2

While our previous study was useful in identifying the different frequencies of selecting reappraisal tactics, a few of our design choices limited our interpretations of the results. Specifically, by presenting only a subset of the most popular tactics for each image, we might have primed participants to use some tactics more frequently than others. Additionally, by combining did not and could not reappraise response options, we were not able to separate failed attempts to reappraise from decisions to not reappraise at all, which are quite different. In Study 2, we address these issues by presenting all tactic options at each trial.

Method

Participants

During the academic term designated for data collection, fifty-four undergraduate students from the University of Denver participated in the study. Data from two students were excluded from the sample due to an apparent lack of understanding of the task as indicated by post-task debriefing questionnaires. The final sample included 52 participants (40 females, 10 males, 1 participant who chose not to share sex assigned at birth, and 1 participant who did not report any demographic data). We also collected information about participants’ gender identities. At the time of the experiment, 41 participants identified as women, and 10 participants identified as men. The mean age of the final sample included in analyses was 19.39 years old with a range of 18 to 28 years old. The racial/ethnic composition was 73.1% Caucasian, 9.6% Asian, 5.8% African American, 3.8% Biracial, 1.9% American Indian or Alaskan Native, and 3.8% rather not answer. In addition, 17.3% of participants identified as Hispanic/Latino. One person did not provide any demographic data about themselves. The recruitment procedures and inclusion/exclusion criteria were the same as in Study 1.

To calculate an appropriate sample size, we used the effect size η_p² = .343 for an interaction between instruction and tactic type for frequency selection from Study 1. Aiming to achieve 95% power at α = .05, we calculated a sample size equal to 6 people in G*Power software. We also calculated sample size estimations for the smallest significant pair-wise comparison in the same interaction (i.e., 0.046 mean difference between “decrease” and “look” conditions for change current circumstances tactic). To achieve 90% power at α = .05 would require 53 people. Since we hoped to replicate the results of Study 1 and pilot the study design before moving to fMRI environment, we aimed to recruit between 40 and 60 people. The final sample size of 52 people was sufficient to detect significant differences in our design.

Procedure

In our follow-up study, we aimed to replicate the results of Study 1. While the design of Study 2 was similar to Study 1, we added a few novel features. In Study 2, immediately following picture presentation, participants chose from all reappraisal tactics (i.e., acceptance, agency, change current circumstances, change future consequences, distancing, reality challenge, technical–analytic–problem solving and other) instead of the three previously most used tactics for a particular image. Notably, such change allowed for addition of technical–analytic–problem solving tactic in our analysis, which focuses on what steps can be taken to solve a problem (i.e., “This is a problem to be solved by focusing on specific steps to be taken”). Furthermore, we separated the did not / could not reappraise option into did not reappraise and could not reappraise options. This allowed us to distinguish between trials in which participants either followed the look instruction with fidelity or opted out of reappraisal when instructed (did not reappraise) versus trials on which they attempted to reappraise but were not successful (could not reappraise). Finally, the order of these multiple-choice tactic options was randomized for each image. The procedure was approved by the University of Denver Institutional Review Board. The data were collected in 2021.

Analytic Strategy

Similar to Study 1, we conducted a 2 (instruction: “look” vs. “decrease”) × 10 (tactic type) repeated measures analysis of variance (ANOVA) for frequency of tactic selection with tactic type and instruction as within-subjects factors. Paired-samples t tests were used to further identify pairwise differences in frequency of tactic selection between tactic types across conditions, as well as within each condition. Mauchly’s Test of Sphericity indicated that the assumption of sphericity had been violated for tactic type effect, χ²(44) = 297.20, p < .001 and interaction of instruction and tactic type, χ²(44) = 408.67, p < .001. Therefore, Greenhouse-Geisser corrected results are reported. For affective outcomes, the linear mixed effect model was analogous to that used in Study 1.

Results

Selection Frequency

As in Study 1, there was a significant main effect of tactic type, F(3.60, 183.61) = 19.71, p < .001, η_ρ² = 0.28, qualified by a significant interaction of instruction and tactic type, F(2.41, 122.98) = 25.02, p < .001, η_ρ² = 0.33. Paired t-tests revealed that participants selected four tactics more frequently following the “decrease” as compared to “look” instruction: acceptance (t(51) = 2.73, p = .010, 95% CI of difference [0.01, 0.07]), change future circumstances (t(51) = 3.50, p = .001, 95% CI of difference [0.03, 0.09]), distancing (t(51) = 3.90, p < .001, 95% CI of difference [0.02, 0.06]), and reality challenge (t(51) = 6.55, p < .001, 95% CI of difference [0.10, 0.19]). Similar to Study 1, participants selected did not reappraise less frequently following the “decrease” as compared to “look” instruction (t(51) = 6.00, p < .001, 95% CI of difference [−0.42, −0.21]). Finally, the selection frequencies for the remaining options were not significantly different between “decrease” and “look” conditions: agency (t(51) = 0.02, p = .985, 95% CI of difference [−0.02, 0.02]), change current circumstances (t(51) = 1.14, p = .249, 95% CI of difference [−0.01, 0.04]), technical–analytic–problem solving (t(51) = 1.22, p = .248, 95% CI of difference [−0.01, 0.03]), other (t(51) = 0.40, p = .687, 95% CI of difference [−0.03, 0.02]), and could not reappraise (t(51) = 0.30, p = .773, 95% CI of difference [−0.05, 0.06]). See Figure 2.

When instructed to respond naturally, participants indicated they could not reappraise on 12.29% of the trials and did not reappraise on 36.29% of the trials on average. Apart from these options, the most frequently selected spontaneous tactics were acceptance (M = 0.09, SD = 0.08), change future circumstances (M = 0.07, SD = 0.08), and, similar to Study 1, reality challenge (M = 0.11, SD = 0.11). The least frequently selected spontaneous tactics were change current circumstances (M = 0.05, SD = 0.07), technical–analytic–problem solving (M = 0.04, SD = 0.06), and other (M = 0.05, SD = 0.06).

When instructed to decrease negative emotion (prompted), participants indicated they could not reappraise on 13.08% of the trials on average. They also indicated that they did not reappraise on 5.13% of the trials on average. Analogous to the spontaneous condition, the most frequently selected prompted tactics were acceptance (M = 0.13, SD = 0.08), change future circumstances (M = 0.13, SD = 0.07), and reality challenge (M = 0.25, SD = 0.16). The least frequently selected prompted tactics were agency (M = 0.06, SD = 0.06), technical–analytic–problem solving (M = 0.05, SD = 0.06), and other (M = 0.04, SD = 0.06).

Affective Outcomes

As in Study 1, there was a significant main effect of instruction on affective outcomes, F(1, 2195.8) = 10.32, p =.001, and a significant main effect of tactic type, F(9, 2197.4) = 30.75, p < .001, but no significant interaction of instruction and tactic type, F(9, 2183.8) = 0.95, p = .484 (see Figure 3).

The tactics associated with the lowest negative affect ratings (i.e., best affective outcomes) collapsing across both conditions were acceptance (M = 3.41, SE = 0.19), change current circumstances (M = 3.28, SE = 0.20), and distancing (M = 3.38, SE = 0.20); the tactics associated with the highest negative affect ratings (i.e., worst affective outcomes) were could not reappraise (M = 4.93, SE = 0.19), agency (M = 4.04, SE = 0.20), and technical–analytic–problem solving (M = 4.17, SE = 0.21) (See Figure 3).

Discussion

Study 2 investigated the selection frequency and affective outcomes of seven specific reappraisal tactics that were presented on every trial of the experimental emotion regulation task. The most frequently selected tactics in both prompted and spontaneous reappraisal conditions were acceptance, change future circumstances, and reality challenge. The tactics associated with the best affective outcomes following both instructions were acceptance, change current circumstances, and distancing. While in Study 1 acceptance was among the most frequently chosen and the most effective tactics in the spontaneous condition, it had similar frequency and effect across conditions in Study 2. Notably, change current circumstances was among the most effective reappraisal tactics in the prompted and spontaneous conditions in Study 1 and across conditions in Study 2. Additionally, we were able to separate the decision to not reappraise from the failure to reappraise by using separate did not reappraise and could not reappraise options. Failure to reappraise (could not reappraise) was associated with worst affective outcomes across both conditions. With some variation, these results replicated the general pattern of tactic selection and affective outcomes we observed in Study 1, without the limitations of our initial design.

General Discussion

In the two current studies, we assessed selection frequency and affective outcomes for reappraisal tactics in a classic reappraisal task. We observed variation in tactic use across conditions, while the prompt to reappraise resulted in more frequent use of tactics. Reality challenge was among the most frequent tactics across both studies and conditions. Other commonly endorsed tactics included acceptance, agency and change future consequences. Interestingly, the most common tactics were not always associated with the lowest ratings of negative affect. Participants’ choice of the tactics might be potentially influenced either by characteristics of the tactic or the image, which we discuss in more detail in the Limitation section. In both studies, change current circumstances was among the tactics that were associated with lowest levels of negative affect across prompted and spontaneous conditions. In addition, distancing and acceptance were also commonly associated with lower levels of negative affect. Finally, for the first time we captured reports of attempted (but failed) reappraisal, as well as reports that participants chose not to reappraise, across prompted and spontaneous conditions. Together, these results deepen our understanding of how often and well specific tactics might be used and contribute to motivational and energetic accounts of emotion regulation.

Patterns of selection and implementation

Across both studies, participants engaged in reappraisal using reappraisal tactics on the majority of trials on which they were instructed to do so. This high rate of compliance following the instruction to reappraise is consistent with previous reports that reappraisal is widely endorsed (Opitz et al., 2015; Nook et al., 2020; Volokhov & Demaree, 2010). As an extension of prior work, the current studies also evaluated whether participants engaged in reappraisal spontaneously, on trials when they were not prompted to do so. In both studies, participants reported selecting tactics more frequently following prompting than spontaneously, but reappraisal tactics were still reported on approximately half of the spontaneous trials. Though previous studies have reported on the occurrence of spontaneous tactic selection, they report spontaneous tactic selection to occur on far fewer, only about 16%, of trials (Nook et al., 2020; Suri et al., 2015). High rates of both prompted and spontaneous tactic selection in the present context affirm the ubiquity of reappraisal as a frequently selected form of emotion regulation (John & Gross, 2007; Suri et al., 2015).

High selection frequency of reappraisal under both the prompted and spontaneous conditions could possibly have been present in this research because we may have established default preference towards reappraising in the task design. Defaults are important contextual variables for predicting choice (Suri et al., 2015). Though our data show that instruction certainly has a strong influence on tactic use, it is possible that the presentation of tactic response options on nearly every trial unintentionally set a default to reappraise for both prompted and spontaneous conditions. However, if participants were relying on the frequency of the presentation of the multiple-choice options as an implicit cue for how they should process the picture, they should have selected the did not / could not reappraise option on 25% of all spontaneous trials in Study 1 and 20% in Study 2. Rates of spontaneous endorsement of this option were found to be much higher, about 50%. Thus, our results suggest that instruction, more so than the mere presentation of certain response options, increases reappraisal frequency.

The propensity to reappraise in prompted and spontaneous conditions could also be explained by the apparent hedonic benefit of reappraising. Both prompted and spontaneous implementation of many tactics were associated with lower levels of negative emotion than the use of no reappraisal tactic (see supplemental materials). Therefore, reappraisal need not be instructed to be successful. These results align with previous findings which observe successful changes in emotion following the use of instructed and uninstructed strategies (Hagstrøm et el., 2020; Opitz et al., 2015). There is an intuitive benefit of using reappraisal to decrease negative emotion, which may motivate using it under both conditions. It is important to take these results into consideration for future research. While most studies on emotion regulation treat uninstructed/passive viewing condition as non-regulation condition, that might vary by individual motivation to regulate emotion, even when instructed not to. Future research should aim to better differentiate what contributes to the use of regulation in uninstructed conditions.

The importance of goals in selection and implementation

Previous work postulated that reappraisal use depends on the emotion goal pursued and whether that goal is externally prompted or internally motivated (Vishkin et al., 2020; Tamir et al., 2019). Our task design allowed for the establishment of different explicit goals for each instruction: the instruction to “decrease” established an externally prompted goal, and the instruction to “look” did not provide an explicit regulatory goal. It is possible, however, that participants spontaneously engaged an internally motivated goal to regulate. By recording tactic selection trial-by-trial, it was first demonstrated that providing goals via reappraisal instruction may be an important mechanism driving increased tactic selection frequency.

In terms of the success of tactic implementation, prompted reappraisal resulted in similar changes in affective outcomes compared to spontaneous reappraisal. Tamir et al. (2019) proposed that having a goal to downregulate is sufficient for regulatory success as it might trigger the use of familiar regulatory strategies, whereas activating the means of regulation (reappraisal) without a goal does not alone result in typical levels of emotional change .In our studies, the presence of the “decrease” instruction prompted an explicit goal to downregulate, which was associated with significant changes in affect. However, our results also provide novel information about the effect of tactics in the absence of an explicit goal – trials on which participants used some tactics (e.g., change current circumstances and acceptance were consistent across both studies), even when they were not given a goal to downregulate, resulted in significantly lower negative affect than trials on which no tactic was used (see supplemental materials). Our instructions explicitly discouraged participants from using reappraisal to change emotion during these trials, but it is unclear whether they activated an explicit or implicit goal to regulate, or engaged in tactic use without such a goal. Future work should investigate the degree to which a goal was unintentionally activated during our “look” trials, or whether some tactics are able to change affect when no such goal is activated at all. In Study 2, we also observed higher negative affect for trials in both conditions where participants reported that they could not reappraise. Future research might investigate the association between failure to meet the regulatory goal and one’s emotional state.

The importance of effort in selection and implementation

In addition to goals and means, Cognitive Energetics Theory (CET; Kruglanski et al., 2012) outlines factors driving the choice to engage in reappraisal. According to CET, the choice to reappraise is also determined by restraining forces such as the perceived difficulty of reappraising. Thus, tactics which are perceived as more difficult might be used less frequently because the effort required to engage with the particular stimuli is greater than the motivation to change negative emotion. This might be particularly pronounced in experimental conditions that include only short-term exposure to the stressors. Some tactics are perceived as more difficult to implement compared to others; reality challenge has been demonstrated to be a less cognitively demanding tactic than change current circumstances and change future consequences (Miliavsky et al., 2019). Consistent with this, reality challenge was one of the most frequently selected prompted tactics in both our studies. It is possible that this is because it was perceived as less effortful for the particular picture set we used. In addition, it is possible, and consistent with previous work, that change current circumstances was one of the least frequently selected prompted tactics because it was perceived as more effortful (Miliavsky et al., 2019). It is important to note that we observed an unpredicted separation between the frequency of tactic selection and the affective outcomes of using different tactics. Although change current circumstances was selected infrequently, perhaps due to the tactic being perceived as difficult, the implementation of change current circumstances was associated with one of the best affective outcomes.

Limitations and future directions

This study, like all studies, is not without its limitations. In the present study, reappraisal was evaluated with a laboratory experiment where participants were asked to reappraise or not reappraise negative images and record their experience by selecting from multiple-choice options (or typing free responses). Recording tactic use following most images created a novel opportunity for both prompted and spontaneous tactic selection to be recorded. However, because tactic options were presented following most images, it is possible that participants could have selected a tactic post-hoc rather than generating reappraisals on their own. In Study 1, our analysis of random free response questions and other free responses (see supplemental materials) render this relatively unlikely to be responsible for the results reported here. As we report in supplemental materials, the data trends in the randomly presented free response questions in Study 1 align with the multiple-choice data, suggesting that participants were not waiting to reappraise and only selecting tactics post-hoc. Furthermore, we addressed this concern by presenting all tactic options for each image in Study 2.

Though a primary focus of this study was the effect of instruction on reappraisal selection and implementation, it is possible that the specific images presented influenced tactic selection and implementation (Aldao, 2013). Reappraisal selection frequency has been shown to be lower for extremely high and extremely low valence images (Suri et al., 2015). For the current study, images were selected based on moderate levels of intensity and normed on valence, arousal, and reappraisability (from an in-house database of previous studies) for each condition. The presentation of moderately negative pictures may have motivated the high rate of reappraisal frequency we observed. Images may have been negative enough to motivate reappraisal, and not so negative that the effort of using a tactic offset the motivation to reappraise (Miliavsky et al., 2019). Additionally, we might suggest that the specific content of the images might have influenced tactic choice and effectiveness as well. While some images might allow for multiple interpretation, others might have a limited number of tactics that might apply to the depicted stressor. While we have included stimulus as a factor into our statistical models, future studies should understand the degree to which particular stressors influence the frequency and effectiveness of reappraisal tactics.

In focusing on matching groups of images between conditions, we did not consider the effect of specific emotion categories on tactic selection and implementation (images were, however, assigned to condition in a manner counterbalanced across participants). Previous work has demonstrated that tactic use may vary based on which emotions are being regulated (Vishkin et al., 2020). For example, change future consequences and reality challenge tend to be more frequently selected for the regulation of fear and sexual desire (Vishkin et al., 2020; Shafir et al., 2018) while acceptance and distancing tend to be more frequently selected and more effective for the regulation of sadness, regret, and disappointment (Vishkin et al., 2020). Manipulating the context of appraisal – the emotional intensity of the stimuli – and which emotions are being evoked to then examine its effect on selection frequency and affective outcomes is a promising avenue for future reappraisal research. It is also important to note that not only characteristics of stimuli but also the modality of stimuli might influence the choice of reappraisal tactics (e.g., reality challenge might be better suited for visual media than for stressful events experienced first-hand). Future studies might further explore how reappraisal tactics are used in response to various types of stressors.

Like all studies of instructed emotion regulation, it is possible that experimental demand influenced selection frequency of tactics and ratings of self-reported affect. It is also possible that other variables such as the demographic characteristics of the sample influenced experimental outcomes. This preliminary investigation recruited a relatively homogeneous sample of available college students at a traditional four-year private university. It is important to consider characteristics like age, gender, race, ethnicity, and clinical status when interpreting patterns of reappraisal use and ability. For example, Hagstrøm et al. (2020) found the ability to reappraise to be positively correlated with age, where older children tended to use a wider range of reappraisal tactics and regulate more effectively compared to younger children. Other studies found the use of change current circumstances and distancing to increase with age while the use of change future consequences and reality challenge to decrease (Nook et al., 2020). There also may be gender variation in cognitive engagement during reappraisal, wherein men may recruit neural systems associated with cognitive effort to a lesser degree than women when reappraising (McRae et al., 2008). Because the majority of our sample consisted of healthy white women in young adulthood, future work should aim to recruit a sample demographically representative of the local or national community.

Conclusion

There are a multitude of ways to change how to think about a situation to achieve the goal of feeling less negative. The present study demonstrated that different reappraisal tactics are selected at varying frequencies and are associated with diverse affective outcomes. Further investigations into reappraisal tactics can provide illuminating insight into the application of reappraisal for various emotional situations as well as the important differences in how and when individuals use reappraisal for their benefit.