Lack of guilt, shame, and remorse following weight stigma expression: a real-time assessment pilot study

Moral emotions

To understand why weight stigma is perpetrated and normalized, it may be helpful to examine how perpetrators of weight stigma feel, specifically with respect to moral emotions, after engaging in this harmful behavior. Moral emotions are emotions that are linked to the interest or welfare of society. They are thought to serve as an “emotional moral barometer,” providing people immediate feedback on the acceptability of behavior (Keyes & Haidt, 2003; Tangney, Stuewig & Mashek, 2007). Negatively valenced “self-conscious” emotions (e.g., shame, guilt, and remorse) often follow deviations from an individual’s moral compass (Scheff & Retzinger, 1991; Tangney, Stuewig & Mashek, 2007; Baumeister, Stillwell & Heatherton, 1994). In contrast, positively valenced moral emotions like pride are experienced following actions that conform to social standards or “doing the right thing” (Scheff & Retzinger, 1991; Tangney, Stuewig & Mashek, 2007).

Importantly, moral emotions, particularly shame and guilt, help to direct behavior following perceived or anticipated moral failure (Gausel & Leach, 2011). Although similar, there is some evidence that shame and guilt may differ in their foci, such that shame is associated with thinking that the self is defective or bad, while guilt is associated with believing that a specific behavior or thought was bad/undesirable (Tangney & Dearing, 2002). As a result, shame has been viewed as motivating avoidance tendencies compared with guilt, which may motivate approach tendencies and efforts to repair wrongdoing (Tangney et al., 1996). However, other research suggests that shame, too, can lead to personal change if one perceives that the moral failure is believed to result from a modifiable self-defect (Gausel & Leach, 2011; Gausel & Brown, 2012; Ferguson et al., 2007). Furthermore, shame and guilt are often highly correlated (Blum, 2008), especially in situations of self-caused wrongdoings, such as perpetrating weight stigma, compared with other-caused wrongdoings (Schmader & Lickel, 2006), and both can motivate behavior change (Lickel et al., 2014). Thus, it is important to understand the roles of shame and guilt in the expression of weight stigma.

Although not widely studied as intervention targets for weight stigma, some evidence from bullying literature implies that guilt and shame may play a role in bullying behavior and lack of bystander intervention (Rieffe et al., 2012; Mazzone, Camodeca & Salmivalli, 2018; Olthof, 2012). For example, similar to the lack of remorse and shame acknowledgement amongst bullies, perpetrators of weight stigma may not experience (or feel less) remorse for their behavior and may engage in shame displacement through other-directed blame (Ahmed & Braithwaite, 2004; Ahmed, 2006). Similarly amongst adults, workplace bullying has been attributed to low shame acknowledgement (Braithwaite, Ahmed & Braithwaite, 2008). Guilt-proneness has also been linked to decreased bullying behavior (Merkin, 2017; Mazzone, Camodeca & Salmivalli, 2016), potentially because of its association with feelings of personal responsibility in situations of offending others, which leads to greater repentance and efforts to make amends (Fisher & Exline, 2006; Merkin, 2017). It is possible that individuals who perpetrate weight stigma, like bullies, do not feel ashamed or guilty after their behavior (Menesini & Camodeca, 2008); however, the relationship between perpetrating weight stigma and the experience of moral emotions is yet unknown.

Current study

Studies on the expression of weight stigma have typically used self-reports, such as the Antifat Attitudes test (Crandall, 1994). These self-report questionnaires assess weight stigma attitudes but not actual weight stigma perpetration. Additionally, because weight stigma is so normalized, many perpetrators may fail to recall or recognize weight stigma transgressions, given the high degree of social acceptance around weight stigma. Experience sampling circumvents these limitations by collecting data in the natural environment. Only four studies to date have included weight stigma in experience sampling data collection (Carels et al., 2019; Vartanian, Pinkus & Smyth, 2018; Potter, Meadows & Smyth, 2020; Vartanian, Pinkus & Smyth, 2014). This pilot study aimed to explore daily weight stigma expression and trajectories of moral emotions following weight stigma expression in the perpetrators. The study explored the expression of various weight stigma thoughts and behaviors in everyday life, the frequency and types of these thoughts and behaviors, and perpetrators’ negative and positive moral emotions after they had engaged in a weight stigmatizing thought or behavior.

Hypotheses

The first aim of the present study was to simply describe the frequency and types of weight stigma perpetration from the perpetrators’ perspective as opposed to the recipients’ perspective. The second aim was to examine the trajectories and rates of change of negative (guilt, shame, remorse) and positive (pride) moral emotions after weight stigma perpetration. Due to the normative nature of weight stigma, it was hypothesized that negative moral emotions would not appreciably increase or decrease following weight stigma expression. Based on the limited work examining emotions surrounding expressions of weight stigma (Nguyen & Malti, 2014) and research suggesting that people believe that perpetrating weight stigma motivates people to lose weight (Nolan & Eshleman, 2016; Vartanian & Smyth, 2013), it was hypothesized that positive moral emotions (i.e., pride) would increase after weight stigma events.

Procedures

Participants from a large, public Mid-Atlantic university were recruited from an online undergraduate research pool and offered course credit for their participation. Participation in the study included an online screen, training phone calls, and answering the phone prompts via a smartphone. Participants did not attend any in-person visits. In order to conceal the true nature of the study, the research was described to prospective participants in an email as a study about thoughts and behaviors individuals may have toward a variety of different groups of people (e.g., the elderly, Mormons, physicians). Participants completed a screening online questionnaire that included weight stigma expression questions and filler questions asking about stigma expression towards the other groups of people. Participants were excluded from the study if they did not have a smartphone. Participants had to report at least 10 instances of weight stigma over the past month (indicating at least occasional recent weight stigma expression) and have a score of 70 or less on the Attitudes Toward Obese Persons questionnaire (ATOP). The ATOP ranges from 0 to 120, with higher scores indicating more positive attitudes towards heavy weight individuals. A score of 70 or less was chosen as a cutoff point to capture individuals with moderate to strong negative attitudes toward individuals who are overweight or obese (Allison, Basile & Yuker, 1991). The original development study for the ATOP demonstrated mean ATOP scores between 63 and 66. Choosing 70 as the cutoff helped to ensure that participants would express average to above-average weight stigma attitudes and engage in some level of weight stigma perpetration. Following the screener, a consent phone call was scheduled where participants were told that they had an equal chance of answering questions about the groups described in the screener during the remainder of the study (i.e., they would be randomly assigned to answer questions in the baseline survey and throughout the week about one of the groups).

Following the consent phone call, participants were sent an email with what they believed to be a randomly assigned link to the baseline questionnaire. In reality, all participants were sent a link to answer questions about their attitudes and behaviors toward higher-weight individuals. Participants were instructed that following completion of the online baseline questionnaire, they would be enrolled in a phone messaging system where they would receive text messages with links to complete a 2–3 min survey of their mood, behavior, and various questions related to the group to which they had been randomly assigned. Following completion of the baseline questionnaire, participants completed one-week of the experience sampling protocol, including a half day of practice, using their personal smartphone to answer questions during six time points between 10:30 AM and 10:00 PM. The six prompts were sent randomly throughout the day with an interval of at least one hour between prompts and an average of four hours between prompts, creating six approximately equal blocks of time. For this pilot study, all experience sampling data were collected using ESMcapture (http://esmcapture.com/). Participants were contacted 2–3 days after being enrolled to ensure they were receiving phone prompts and to answer any questions. Participants received class participation credits for completion of the study. All procedures were approved by the university’s Institutional Review Board (714290-1).

Materials

Data analysis

40 participants were recruited, and all 40 met the minimum weight stigma event requirement over the past month and endorsed moderate to strongly negative attitudes towards higher-weight individuals as measured by the ATOP. After the baseline questionnaire, five of the participants decided not to participate in the experience sampling protocol due to time constraints. A total of 35 participants completed the baseline questionnaire and one-week of phone prompts. Four participants, however, were excluded from the main analyses due to having 3 or fewer responses, leaving 31 participants. Results are based on data from the 31 participants.

Participants

The sample was primarily female (74.19% female) with a mean age of 19.74 years (SD = 2.35, range 18–28) and with a mean BMI, based on self-reported height and weight, of 25.45 kg/m² (SD = 5.81, range 17 - 41). 42% of participants had a BMI >25. The racial/ethnic background of the sample was as follows: 67.74% White/Caucasian, 9.68% Asian, 9.68% Hispanic, 6.45% Black/African American, and 6.45% Mixed heritage/unknown. The majority of participants were also born in the United States (87.1%, 27 students).

EMA compliance

The 31 participants who completed the experience sampling protocol responded for an average of 6.77 days (SD = 1.96) with an average of 2.88 responses per day (SD = 1.51). The compliance rate averaged 46.4% across participants.

Baseline weight stigma questionnaires and weight stigma events

Means and standard deviations of baseline weight stigma variables are displayed in Table 1. Correlations between anti-fat attitudes and total weight stigma expression measured via experience sampling also appear in Table 1. Importantly, total weight stigmatizing cognitions and behaviors were significantly correlated with baseline AFA total mean scores (r = 0.38, p = .04) and the AFA Dislike subscale (r = 0.40, p = .02), indicating that greater self-reported stigma expression was associated with greater self-reported anti-fat attitudes and particularly with dislike of individuals of higher weight (see Table 1). Total stigma events during EMA did not correlate with ATOP scores. Looking individually at weight stigma cognitions versus weight stigma behaviors, higher frequency of weight stigma cognitions was correlated with less positive attitudes toward higher-weight people (ATOP) (r = −.36, p = .048) at the trend level, while frequency of weight stigma behaviors was not. In contrast, only weight stigma behaviors had a significant positive correlation with AFA Dislike scores (r = .43, p = .02) and AFA Total scores (r = .39, p = .03). AFA Willpower was not significantly correlated with weight stigma perpetration cognitions or behaviors.

Frequency and types of weight stigmatizing perpetration

During the 7-day study period, the 31 participants reported perpetrating 1,008 weight stigma behaviors and cognitions. These weight stigma behaviors and cognitions were reported at 249 phone prompts (participants could report multiple weight stigma behaviors and cognitions at each time point). Additionally, the number of expressions reported at each phone prompt ranged from 0–18 with an average of 1.67 expressions (SD = 3.00) reported per time point across the 604 responses. Weight stigma events were recorded on all days of the study, and all participants reported perpetrating weight stigma at least once during the study period. Weight stigma events also occurred throughout all time periods of the day; no clear trends were evident. Of the 1,008 events, participants reported 590 weight stigma cognitions and 418 weight stigma behaviors. The frequencies of each type of weight stigma behavior and cognition are depicted in Fig. 1, which is color-coded to illustrate differences between cognitive (black) and behavioral (gray) weight stigma perpetration.

Trajectories following weight stigma events

GEE analyses indicated that the trajectory of guilt following expression of weight stigma significantly decreased following weight stigma expression (B = −.196, p = .021, Table 2, Fig. 2A). Similarly, participants reported a significant linear decrease in shame (B = −.153, p = .007, Table 2, Fig. 2B) and remorse (B = −.274, p = .002, Table 2, Fig. 2C) following weight stigma perpetration. For guilt, remorse, and shame, the quadratic and cubic components of the trajectory were non-significant, indicating that the rate at which the trajectory was changing over time, or the acceleration, did not change significantly over the 4 hour period following the event.

Similar to the trajectory of remorse, the trajectory of feeling proud after perpetrating weight stigma also significantly decreased following weight stigma events. Participants reported a significant linear decrease in pride (B = −.169, p = .033, Table 2, Fig. 2D). The quadratic component of the trajectory was also significant (B = −.037, p = .016, Table 2), which indicates that the rate at which pride was decreasing after the event accelerated over 4 h after the weight stigma event.

Analyzing the composite scores of the SSGS subscales for guilt, shame, and pride (tables provided in Supplemental Files), showed that guilt decreased significantly following weight stigma perpetration (B = −.247, p = .01), shame decreased following weight stigma perpetration (B = −.096, p = .05) at a trend level, and pride did not fluctuate significantly following weight stigma perpetration (B = −.059, p = .17). Negative affect, measured as a composite of PANAS items, decreased significantly following weight stigma perpetration (B = −.216, p = .01) as did positive affect (B = −.089, p = .01) (tables provided in Supplemental Files).

Limitations & future directions

This pilot study contributes to a limited body of work using real-time assessment of weight stigma, however, several limitations should be noted. Foremost, participant compliance rate was significantly lower than in other studies using experience sampling. The reason for the low compliance rate in the present study is unknown, but based on past publications from the authors’ own lab where compliance rates were higher (>80%) across participants (Fischer et al., 2017), differences in data collection are highlighted. Participants in this study were provided with class participation credit at one time point for study completion. In past studies from this lab (e.g., Fischer et al., 2017) where compliance rates have been higher, researchers provided incentives for completing additional prompts, in addition to completion incentives. In the current pilot study, participant compliance was not monitored routinely, and participants were not contacted when their response rates fell below ideal compliance rates (e.g., below 80%). Future studies may benefit from additional safeguards to ensure participant compliance, such as increasing communication frequency with participants, offering extra class participation credits based on the number of prompts answered, or decreasing participant burden by decreasing the number of prompts.

Experience sampling methodology has additional limitations that are inherent to correlational research. Due to the nature of experience sampling, in which participants are asked to report on multiple events and emotions periodically throughout the day, it is not possible to isolate weight stigmatizing events from other events that may have occurred during the same time period in which participants were responding to prompts. The only way to state that engaging in a weight stigma event caused a decrease in guilt and other emotions is to conduct an experimental study in which engagement in a weight stigma event is manipulated. These data show that the trajectory of guilt and other moral emotions statistically significantly decreased following these events, but it is always possible that this decrease may have been due to other behaviors or events that occurred at the same time. However, other behaviors that may cause decreases in moral emotions would likely need to occur at the same rate as weight stigmatizing behaviors during the same signaled prompt period in order for those behaviors to have a statistically significant impact on the emotion trajectory. It is possible that the use of signal-contingent prompting (e.g., asking participants to initiate data collection when they engage in an event) may increase confidence in concluding that moral emotions changed as a result of that specific event. However, given the pervasive nature of weight stigma documented in this study and normalization of weight stigma, it is also possible that participants would not have self-initiated data collection for a signal-contingent responding protocol. Last, it is possible that participants were over-reporting their weight stigma cognitions and behaviors due to reactivity to the experience sampling protocol and increased monitoring of thoughts and behaviors. Reported cognitions and behaviors could additionally be related to the same instance. For example, students may have reported judging someone due to their weight, finding them less sexually attractive, and excluding them from a picture due to their weight all at the same time. Future studies should ask participants to distinguish if reported weight stigma applied to one or multiple scenarios since their previous prompt.

Further, the data for the current study were collected from college students. While weight stigma is prevalent on college campuses (Puhl & Suh, 2015b), future studies should explore daily weight stigma expression in other environments. The sample size of the current study was also fairly small (although there were a large number of weight stigma events) and included very few males. The small sample size also precluded examining factors such as race, country of origin, gender, or BMI, or differences between weight stigma cognitions and behaviors as moderators of weight-stigma expression. Although the sample was on average higher-weight, BMI was not correlated with weight stigma perpetration or explicit anti-fat bias, suggesting that this sample of college students appeared to engage in similar patterns of weight stigma perpetration regardless of personal body size; however, future studies should examine weight stigma perpetration differences based on BMI in greater detail given the small sample size.

The present sample was fairly diverse with 34% of the sample being non-White. Cultural or racial differences in body ideals are widely recognized (Grogan, 2016) and may influence people’s attitudes toward people who are overweight or obese as well as their engagement in weight stigma. For example, White women have been found to express more stigmatizing attitudes toward other women, especially White women, compared with African American women. African American women also rated depictions of African American women who were larger bodied significantly less negatively (Hebl & Heatherton, 1998). Similar results have been found in adolescents (Gray et al., 2011). Replication in larger samples would allow for examination of the potential moderating effects of race, ethnicity, BMI, and gender on weight stigma expression.

Future research addressing these limitations should be conducted in order to improve confidence in the current results and better understand weight stigmatizing behavior across multiple contexts and in different cultures. Subsequent studies should also examine what characteristics make people most likely to exhibit weight stigma and less likely to experience guilt following its expression. Interestingly, some research suggests that weight status does not affect people’s stigmatization of obesity, suggesting that higher-weight individuals may be just as likely to engage in weight stigma as lower-weight individuals (Latner, Stunkard & Wilson, 2005). A better understanding of who expresses weight stigma, what motivates them to do so, and what maintains this behavior (e.g., lack of guilt following weight stigma expression) could help to inform intervention and policy development aimed at reducing this harmful and prevalent phenomenon.