RUNNING HEAD: ASSESSMENT PREDICTING ALCOHOL USE 1
Experienced-Based Versus Scenario-Based Assessments of Shame and Guilt and Their
Relationship to Alcohol Consumption and Problems
Jason Luoma, Ph.D.a Paul Guinther, Ph.D.a Jacqueline Potter,a & Megan Cheslock, B.A.a
aPortland Psychotherapy Clinic, Research, and Training Center
Luoma, J. B., Guinther, P., Potter, J., & Cheslock, M. (in press). Experienced-Based Versus
Scenario-Based Assessments of Shame and Guilt and Their Relationship to Alcohol
Consumption and Problems. Substance Use and Misuse."
This is an Author’s Accepted Manuscript of an article published in Substance Use and Misuse.
The final version of the article as published in Substance Use and Misuse (copyright Taylor &
Francis) is available online at: http://dx.doi.org/10.1080/10826084.2017.1305416.”
Address correspondence to:
Jason Luoma, Ph.D.
ASSESSMENT PREDICTING ALCOHOL MISUSE 2
Background: Empirical studies of the relationships between shame, guilt, and drinking
are sparse and sometimes appear contradictory. However, a more coherent picture emerges
when researchers differentiate between measures of experienced guilt and shame (i.e.,
questionnaires that ask how often people experience thoughts, feelings, and sensations
associated with the emotion) versus proneness to guilt and shame (i.e., self-reports of likely
responses to imagined situations). Objectives: Assess the extent to which experiential versus
proneness measures of shame and guilt are associated with alcohol consumption and alcohol-
related problems. Methods: Between 2012 to 2013, 89 community-dwelling non-abstaining
adults were interviewed on a single occasion about their drinking and completed self-report
measures of shame, guilt, and drinking-related behaviors. Results: Overall, shame and guilt
were most strongly related to alcohol-related problems and not drinking amount per se, and
shame was more strongly related to alcohol-related problems than was guilt. A measure of
experienced shame, the Internalized Shame Scale, was the strongest predictor of drinking-
related problems and predicted problems above and beyond other measures of shame and guilt.
While guilt proneness was related to less problematic drinking, guilt experienced at the time of
the assessment was related to more problematic drinking. Conclusions: Shame appears to be
more central to the experience of problematic drinking than guilt. Results also support the idea
that guilt/shame proneness is distinct from experienced shame and guilt. Incorporating this
distinction appears to account for the inconsistencies in the literature regarding how shame and
guilt are related to drinking.
Key words: shame, guilt, measurement, alcohol use, alcohol problems
ASSESSMENT PREDICTING ALCOHOL MISUSE 3
Shame: Cognitions, emotions, and other behaviors related to the perception that the self
is flawed in the eyes of oneself or others.
Guilt: Cognitions, emotions, and other behaviors related to the perception that the self
has engaged in a negative action that caused harm.
Experiential measure: Assesses whether an individual consistently experiences or is
currently experiencing target (e.g., shame-related) cognitions, feelings, and behaviors.
Proneness measure: Assesses likely responses to hypothetical scenarios where target
(e.g., shame-related) cognitions, feelings, and behaviors could potentially be elicited.
Alcohol use: The rate of alcohol consumption in terms of frequency and quantity.
Alcohol problems: Negative consequences resulting from drinking alcohol.
ASSESSMENT PREDICTING ALCOHOL MISUSE 4
Both shame and guilt are so-called “self-conscious” emotions, as they require the ability
to reflect on oneself and the relationship between oneself and others (Tangney, & Dearing,
2002). Shame relates to a perception that the self is flawed in the eyes of oneself or others,
whereas guilt relates to a perception that the self has engaged in a negative action that caused
harm (Dearing, Stuewig, & Tangney, 2005; Gilbert, 1998; Lewis, 1971). Although shame and
guilt have been widely discussed as important variables in the maintenance of problematic
drinking (Randles & Tracy, 2013; Treeby & Bruno, 2012), empirical study of these relationships
is limited and seemingly conflicting findings are commonplace. Shame is related to a wide
variety of problems in social and psychological functioning (Schmader & Lickel, 2006; Tangney,
Stuewig, & Mashek, 2007), including addictive behavior (Luoma & Kohlenberg, 2012), though
shame may sometimes play an adaptive role in certain contexts (de Hooge, Nelissen,
Breugelmans, & Zeelenberg, 2011; Luoma & Kohlenberg, 2012). In contrast, guilt is more
variably reported as being either adaptive or maladaptive across a variety of contexts (Meehan et
al., 1996; Stuewig et al., 2014). Before reviewing some apparently conflicting findings, it will be
useful to have a framework for interpretation, to which we now turn.
At least part of the apparent inconsistency in the literature linking shame and guilt to
substance use is due to differing measurement approaches (Cook & Nathanson, 1996; Fontaine,
Luyten, Estas, & Corveleyn, 2004). Two kinds of self-report measures of shame and guilt have
been identified: experience-based measures (i.e. experiential measures) and scenario-based
measures (i.e. proneness measures). Experience-based measures, such as the Internalized Shame
Scale (ISS; Cook & Nathanson, 1996) and State Shame and Guilt Scale (SSGS; Marschall,
Sanftner, & Tangney, 1994), ask participants whether they consistently experience or are
ASSESSMENT PREDICTING ALCOHOL MISUSE 5
currently experiencing cognitions, feelings, and other behaviors associated with shame or guilt.
In contrast, scenario-based measures, such as the Test of Self-Conscious Affect (TOSCA;
Tangney, Wagner, & Gramzow, 1989), ask participants to indicate their likely response to
hypothetical scenarios where shame and/or guilt could be potentially elicited. As opposed to
experienced shame/guilt, the TOSCA therefore measures a global disposition or proneness
Many studies of the relationship between shame/guilt and substance use treat experience
and proneness measures as if they were equivalent. However, experienced shame/guilt may
differ from shame/guilt proneness in important ways. When examining cross-sectional between-
subjects effects, this distinction between reported experiences of shame or guilt (whether
prolonged or immediate) and a global response disposition becomes extremely important. This
distinction parallels the well-known distinction between situation and trait influences on
responding. Traits describe patterns of relatively stable responding across situations, but traits
can be overridden by situations that elicit relatively similar responding across most people. For
example, even people who are very extraverted can act introverted in certain contexts. Due to
measurement approach, shame and guilt proneness measures should be more trait-like, and less
affected by the person’s current situation or life context. In contrast, experience measures should
be more responsive to life events and should fluctuate as the person encounters situations that
tend to elicit shame or guilt. For example, even a person who is not very prone to shame might
experience increased shame when encountering events that tend to elicit shame in most
individuals. Similarly, a person who is not very prone to guilt might experience increased guilt
when encountering events that tend to elicit guilt in most individuals.
ASSESSMENT PREDICTING ALCOHOL MISUSE 6
The importance of this distinction between experience and dispositional measures is
perhaps most clear when considering the example of guilt. A global, dispositional tendency
toward guilt reflects a tendency toward taking reparative action when a person engages in
behavior they perceive as harming others, as can happen in problem drinking (Tracy et al.,
2007). This is consistent with recent data from Giguere, Lalonde, and Taylor (2014) who found
that guilt in response to exceeding group norms predicted reduced drinking at a future occasion.
Thus, a dispositional tendency toward guilt is likely to be protective against the development of
alcohol use problems for those who are not currently problem drinkers. On the other hand, it is
likely that people who are experiencing elevated negative consequences associated with problem
drinking may experience elevated levels of both guilt and shame when reflecting on their alcohol
use. This idea fits with the findings of Meehan et al. (1996) who found that people in residential
addictions treatment had higher scores on measures of experienced guilt compared to non-
addicted controls, while simultaneously scoring lower on guilt proneness measures. Since guilt is
a relatively universal human emotion (at least in European-American populations), then guilt
would be a natural reaction for many individuals when reflecting on the harms created by
drinking. Based on these ideas, we would expect that, in a cross-sectional study, experienced
guilt might be positively correlated with drinking-related problems, while dispositional guilt
proneness might be negatively correlated (i.e., protective).
A different pattern of relationships is likely between shame and problematic drinking.
Shame proneness appears to be a risk factor for the development of problematic substance use
(Stuewig et al., 2014). The link between shame and the development of problematic drinking is
based on the observation that shame tends to lead to problematic avoidance behaviors such as
social withdrawal and substance use (Dearing et al., 2005). In this view, shame serves as an
ASSESSMENT PREDICTING ALCOHOL MISUSE 7
antecedent for problematic drinking. At the same time, shame could also be a consequence of the
problematic drinking, regardless of a strong dispositional tendency toward shame. Shame as a
consequence of problematic drinking would be expected when drinking results in failures to
meet role expectations (e.g., unemployment or failed relationships) or violations of important
standards of behavior (Tracy, Robins, & Tangney, 2007). These violations of standards and
norms are increasingly likely as drinking becomes more problematic, which should result in
more and more shame as drinking increases. While these temporal considerations cannot be fully
disentangled in a cross-sectional study, we would expect that both experienced shame and shame
proneness would be positively correlated with drinking-related problems and that experienced
shame would be more strongly associated with drinking-related problems than dispositional
To review, guilt motivates people to discontinue behaviors that harm relationships and
make reparations if damage has occurred (Tangney & Dearing, 2002), such that a single occasion
of overdrinking followed by guilt would be expected to lead to subsequent drinking constraint
for a period of time (Giguère, Lalonde, & Taylor, 2014). Because guilt prone individuals would
likely make reparations after overdrinking, they would be less likely to accrue problems related
to alcohol consumption. Therefore, one would expect a negative relationship between guilt
proneness and alcohol-related problems. However, we propose that over extended periods of
time, there may be factors more powerful than guilt that could lead to problematic drinking and
subsequent alcohol-related problems. These negative alcohol-related consequences would give
people ample material to feel guilty about, and thereby increase the frequency of personal guilt
experiences among problem drinkers (even for those without a high proneness toward guilt). As
such, one would expect a positive relationship between experienced guilt and global alcohol-
ASSESSMENT PREDICTING ALCOHOL MISUSE 8
related problems. In contrast, shame proneness and shame experiences would not be expected to
motivate reparations, instead potentially motivating behaviors such as social withdrawal. As
such, one would expect both shame proneness and experienced shame to be positively related to
global alcohol-related problems, though proneness-based measures may be less strongly
predictive than experience-based measures.
With this shame/guilt experience versus proneness framework in mind, we can now turn
to, and make better sense, of the literature on shame, guilt, and addictive behaviors. We begin
with cross-sectional studies. A study of college students and jail inmates (Dearing et al., 2005)
showed that shame proneness was related to drug and alcohol use problems, but guilt proneness
was either unrelated or protective. A second study with a student sample demonstrated that
shame proneness was positively correlated with alcohol-related problems, but guilt proneness
was negatively correlated (Treeby & Bruno, 2012). A third study found that people attending
peer support groups or in treatment for addiction had higher mean scores on shame proneness
compared to a non-addicted control sample; the people with substance use problems evidenced
lower scores on guilt proneness compared to controls (O’Connor, Berry, Inaba, Weiss, &
Morrison, 1994). Similarly, a cross sectional study of college students (Ianni, Hart, Hibbard, &
Carroll, 2010) found that experienced shame was positively correlated with severity of drinking.
A study that combined measurement frameworks (Meehan et al., 1996) found that people in
residential addictions treatment had higher scores on measures of shame proneness, experienced
shame, and experienced guilt compared to non-addicted controls. In contrast, the residential
addictions group had lower scores on guilt proneness. In general, correlations between
shame/guilt proneness and measures of addictive behavior were smaller than correlations
between experienced shame/guilt and addictive behavior. Finally, using a college student sample,
ASSESSMENT PREDICTING ALCOHOL MISUSE 9
Giguère, et al. (2014) found that both experienced guilt and experienced shame were elevated
after a drinking experience that was higher than normative levels.
Four studies have examined self-reported shame or guilt as prospective predictors of
substance use. Stuewig et al. (2014) reported that children who were more shame-prone in the
fifth grade started drinking at an earlier age and used a greater variety of drugs by the time they
were 18 years of age. In contrast, guilt proneness in fifth grade was related to delayed age of
onset for drinking and less drug use by age 18. Among people entering a smoking cessation
program, an experienced shame measure predicted earlier smoking relapse and higher rates of
smoking at follow up (Boudrez, 2009). A daily diary study with college students (Mohr,
Brannan, Mohr, Armeli, & Tennen, 2008) used a single-item measure of experienced shame and
guilt and found that daytime reports of experienced ashamed mood predicted drinking at home
that evening better than reports of guilt or other reported emotions. Giguère, et al. (2014) found
that shame following drinking that exceeded group norms predicted more drinking over the
following week, while guilt experienced after drinking that exceeded group norms led to reduced
drinking over the following week.
We have not provided an exhaustive literature review on the relationship between
shame/guilt and substance misuse, however, our review does indicate the relationships between
shame, guilt, and drinking is complex. The pattern of relationships between shame/guilt and
substance misuse across studies can appear contradictory without an explicit appreciation for the
distinction between experience and proneness measures and without consideration of the
temporal trajectory involved in the development of problem drinking. Additional research
advancements would likely be facilitated by a closer examination of such measures in a single
study, thereby permitting direct comparison of their utility in predicting substance misuse.
ASSESSMENT PREDICTING ALCOHOL MISUSE 10
The present study was, therefore, designed to examine the relative utility of three
commonly used measures of shame and guilt in predicting drinking-related behaviors in a
community sample of drinkers: the ISS measure of experienced shame, the SSGS measure of
experienced shame and guilt, and the TOSCA measure of shame and guilt proneness. The ISS
was specifically created for use with alcohol-misusing populations, though no studies have
empirically examined the relative utility of the ISS versus the SSGS and TOSCA in predicting
drinking or drinking-related problems. The SSGS has been found to be correlated with drinking-
related problems in one prior study (Ianni et al., 2010), and the TOSCA has been weakly but
consistently associated with drinking-related problems (Dearing et al., 2005; Meehan et al.,
1996; O’Connor et al., 1994; Treeby & Bruno, 2012). We hypothesized that both shame and
guilt would be more highly related to drinking-related problems than drinking per se, and that
experience measures would be more highly related to levels of alcohol use and drinking-related
problems than proneness measures. We also expected that experience-based measures of shame
and guilt and also shame proneness would be positively associated with drinking-related
problems, whereas guilt proneness would be negatively associated with drinking-related
problems. Finally, we hypothesized that measures of shame, as a whole, would be more
predictive of drinking-related problems than measures of guilt, as a whole.
Data for the present study were collected during the intake of a larger longitudinal
investigation examining psychological predictors of drinking behavior; the protocol was
approved by an independent ethics committee. A community sample of drinkers was recruited
through online ads and posted flyers. Eligibility criteria included internet access, ability to travel
ASSESSMENT PREDICTING ALCOHOL MISUSE 11
to the research center, and having at least one drink of alcohol in the two weeks prior to
screening. Exclusion criteria included inability to read English, pregnancy, or inability to provide
informed consent. One participant was an extreme outlier in regression analyses but his removal
did not substantially alter significance values of parameters; we nevertheless removed the case
from all analyses because his especially high level of drinking relative to the rest of the sample
(i.e., 428 drinks over the past 30 days) produced what appeared to be artificially inflated zero-
order relationships between that variable and self-reported shame. After removing that case, the
final sample consisted of 89 adults (57 female, 32 male; years of age M = 33.75, SD = 12.83;
years of education M = 15.11, SD = 2.71). Eighteen percent (n = 16) of the participants identified
themselves as non-Caucasian (Black or African-American, n = 5; Asian, n = 1; multiracial, n =
10), and 7% (n = 3) identified as Hispanic; one participant identified his race as Middle Eastern.
Just over half were employed (n = 49), some unemployed (n = 31), and others not in the labor
force (n = 9). Independent from employment status, a minority of participants identified as being
current students (n = 19). At the time of the study, the majority of participants were not receiving
any kind of mental health treatment (n = 70); 20% were receiving mental health treatment (n =
17), and 2 participants (2%) were being treated for alcohol misuse.
2.2.1 Measures of shame and guilt.
2.2.2. Internalized Shame Scale. The Internalized Shame Scale (ISS; Cook, 1987;
current study α = .96, M = 1.13, SD = 0.77) is a 24-item self-report questionnaire measuring
internalized shame. The measure has previously shown construct validity and reliability in both
clinical and nonclinical populations and is considered particularly appropriate for examination of
ASSESSMENT PREDICTING ALCOHOL MISUSE 12
shame in substance dependent populations (Rybak & Brown, 1996). An example item is “I feel
intensely inadequate and full of self doubt,” with scale anchors “Never” and “Almost Always.”
2.2.3. State Shame and Guilt Scale. The State Shame and Guilt Scale (SSGS; Marschall
et al., 1994) is a 15-item self-report measure with subscales measuring state shame, guilt, and
pride; only the state shame (SSGS-S; current study α = .80, M = 1.41, SD = 0.61) and state guilt
(SSGS-G; current study α = .87, M = 1.82, SD = 1.01) subscales were examined in the present
study. The measure is intended to assess shame in terms of present-moment global negative
feelings about the self, and assess guilt in terms of present-moment negative feelings about some
specific event. The SSGS has previously been shown to be sensitive to experimental shame
inductions (Marschall et al., 1994). An example SSGS-S item is “I feel like I am a bad person”
and an example SSGS-G items is “I feel bad about something bad that I have done,” with scale
anchors “Not feeling this way at all” and “Feeling this way very strongly.”
2.2.4. Test of Self-Conscious Affect. The Test of Self-Conscious Affect (TOSCA-3;
Tangney et al., 1989) is a 16-item measure of self-reported shame and guilt proneness in
response to shame or guilt potentiating scenarios. Following the scoring procedure recommended
by Tangney and Dearing (2002) we regressed raw shame and guilt scores on each other to
produce residual scores of guilt-free shame (TOSCA-GFS; current study raw-score α = .75, M =
2.76, SD = 0.74) and shame-free guilt (TOSCA-SFG; current study raw-score α = .65, M = 4.04,
SD = 0.50). The measure has generally shown good psychometrics across studies (Tangney &
Dearing, 2002). An example scenario is “You make plans to meet a friend for lunch. At 5
o'clock, you realize you stood him up.” With scales anchored by “Not Likely” and “Very
Likely,” participants then rated the extent to which they would have a shame reaction (e.g., “You
ASSESSMENT PREDICTING ALCOHOL MISUSE 13
would think ‘I’m inconsiderate’”) and also the extent to which they would have a guilt reaction
(e.g., “You’d think you should make it up to him as soon as possible”).
2.2.5 Measures of alcohol consumption and problematic drinking.
2.2.6. Time Line Follow Back. The Time Line Follow Back (TLFB; Sobell & Sobell,
1992) involves retrospectively reporting the number of drinks participants consumed on each of
the last 30 days. With the assistance of a calendar and prompts from the research assistant,
participants were interviewed about their drinking in order to determine the number of drinking
Days (TLFB-Days; current study M = 10.71, SD= 7.26), the total number of standard Drinks
(TLFB-Drinks; current study M = 33.98, SD = 33.03) they consumed, and the number of Binges
(TLFB-Binges; current study M = 2.78, SD = 4.56) in which they engaged. Binges were defined
as 4 or more drinks in a day for women and 5 or more drinks in a day for men (c.f. Courtney &
Polich, 2009); sixty-four percent of participants had binged at least once within the last month.
Sobell, Brown, Leo, and Sobell (1996) found the TLFB to have an alternate-forms test-retest
reliability of .90.
2.2.7. Alcohol Use Disorders Identification Test. The Alcohol Use Disorders
Identification Test (AUDIT; Saunders, Aasland, Babor, de la Fuente, & Grant, 1993; current
study α = .85, current study M = 9.44, SD = 6.28) is a 10-item self-report measure of intake,
dependence, and problems related to the consumption of alcohol. Using an AUDIT raw-score
cutoff of 8 or more, 53% of the sample was identified as engaging in “problematic” use of
alcohol (Babor, Higgins-Biddle, Saunders, & Monteiro, 2001), whereas 35% of participants
exceeded a cutoff of 10 or more and were therefore identified as engaging in “hazardous or
harmful” levels of alcohol misuse (Saunders et al., 1993). The AUDIT is a reliable and valid
ASSESSMENT PREDICTING ALCOHOL MISUSE 14
measure of alcohol abuse, dependence, and harmful use (de Meneses-Gaya, Zuardi, Loureiro, &
3.1. Checking assumptions. Where indicated, variables were root or logarithmically
transformed prior to regression analyses to correct for skewness (Tabachnick & Fidell, 2007); the
positively-skewed SSGS-S distribution was corrected by recoding raw scores into a binary
variable (i.e., absence or presence of any shame coded as 0 and 1, respectively). A negative skew
in the raw TOSCA-G distribution was corrected via logarithmic transformation prior to
calculation of residual TOSCA-SFG and TOSCA-GFS scores.
3.2. Correlations. Spearman correlational analyses (see Table 1) showed that measures
of shame and guilt did not predict alcohol drinking days, total drinks, nor number of binges, with
the exception that ISS scores were positively correlated with number of binges. Levels of
alcohol-related problems (i.e., AUDIT scores) were significantly related to all measures of
shame and guilt; relationships were positive with the exception of a significant negative
relationship between problematic drinking and scenario-based guilt proneness (i.e., higher
TOSCA-SFG was protective against drinking problems). The strongest associations were with
the ISS, followed by the other shame and then guilt measures. It is also worth noting that zero-
order correlations between shame measures and state guilt were all positive, whereas the
correlation between guilt proneness and state guilt was negative.
ASSESSMENT PREDICTING ALCOHOL MISUSE 15
Table 1. Spearman Correlations between Shame, Guilt, and Drinking Measures
6. TLFB Days
7. TLFB Drinks
8. TLFB Binges
*p < .05, ** p < .01
3.3. Regression analyses. All predictor variables were entered into two-step regression analyses
to assess their capacity to predict drinking outcomes, with shame variables entered in the first
step and guilt variables entered in the second step (see Tables 2 and 3). We compared model fit
for overdispersed Poisson regression and negative binomial regression in predicting TLFB
Binges, a count variable. A Vuong test revealed that the negative binomial model was a better fit
to the data (Z = 2.56, p > .05) and thus the negative binomial model is reported. All other
regressions were conducted using ordinary least squares regressions. TLFB Days and Drinks
were not significantly predicted in overall models at either step; the addition of guilt variables in
the second step did not improve prediction, and no variable emerged as a unique predictor. TLFB
Binges was significantly predicted in overall models when only shame variables were included,
but this relationship weakened to be only marginally significant once guilt variables were
included in the second step. No variable emerged as a unique predictor of binges at either step.
Finally, AUDIT scores were significantly predicted at both steps, though the addition of guilt
variables in the second step did not significantly improve prediction. The ISS was a significant
unique predictor of AUDIT scores at both steps, whereas no other variable was a significant
unique predictor at either step.
ASSESSMENT PREDICTING ALCOHOL MISUSE 16
We ran additional models to address potential confounds. As the AUDIT includes items
measuring both alcohol consumption and alcohol-related problems, we ran models that
statistically controlled for levels of drinking in predicting AUDIT scores. In models predicting
AUDIT scores that controlled for TLFB total drinks and TLFB drinking days, results were
essentially unchanged from those reported in Table 2. Exact parameters changed slightly in the
models, but all results were identical in terms of whether parameters were statistically significant
or not. We do not present these results in detail since they did not affect results appreciably. We
also ran stepwise regressions on AUDIT scores entering guilt variables in the first step (F =
6.68, p = .002, R2 = .13) and found the addition of shame variables significantly increased
prediction in the second step (F change = 4.46, p = .006, R2 change = .12); the SSGS-G was a
significant unique predictor in the first step but not in the second.
ASSESSMENT PREDICTING ALCOHOL MISUSE 17
Table 2. Ordinary Least Squares Stepwise Regressions of Shame and Guilt scores on Days,
Drinks, and AUDIT scores.
** p < .01
Note: Sr = semipartial correlation. F and adjusted R2 are reported for Overall models; F Change
and ΔR2 are reported with respect to model Changes resulting from the addition of guilt variables
in Step 2. No individual variables were unique contributors in days and drinks models and
therefore they were not reported for space purposes.
ASSESSMENT PREDICTING ALCOHOL MISUSE 18
Table 3. Negative Binomial Stepwise Regressions of Shame and Guilt Scores on Number of
OLS adj. R2
* p < .05
Note: Likelihood Ratio χ2 is shown for tests of Overall model effects and Wald χ2 is shown for
tests of parameter estimates. Semipartial correlations (Sr) and adjusted R2 were calculated using
Ordinary Least Squares (OLS) multiple regressions on the transformed Binges variable, and
therefore represent only approximate effect sizes.
Overall, findings were consistent with previous studies showing shame to be predictive
of alcohol consumption and alcohol-related problems (Meehan et al., 1996; O’Connor et al.,
1994; Randles & Tracy, 2013; Treeby & Bruno, 2012). However, this study provides
additional information about what shame and guilt measurement methods may be most
important in relation to alcohol misuse. In general, shame and guilt measures were predictive
of alcohol-related problems and not drinking per se. None of the shame or guilt variables,
either alone or in combination, were related to days of drinking or number of drinks over the
ASSESSMENT PREDICTING ALCOHOL MISUSE 19
last 30 days; only the ISS was predictive of binges. In contrast, all the shame and guilt
variables correlated with problematic drinking as measured by the AUDIT. When entered into
simultaneous regression, shame and guilt measures, as a whole, predicted little variance in
number of drinking days, drinks, or binges over the last 30 days (R2’s ≤ .02 ). In contrast, when
shame and guilt measures were entered into a regression predicting problematic drinking,
variance accounted for was much larger (R2 = .21). Together, these results suggest that shame
and guilt are more highly related to problematic drinking or drinking consequences than
amount of drinking per se. This makes sense in that shame and guilt are both emotions that
relate to the violation of social standards or harms to relationships, which are likely
consequences of problematic drinking.
Shame was clearly a stronger predictor of problematic drinking behavior than was guilt.
Zero-order correlations showed that shame measures were more strongly associated with
problematic drinking than were guilt measures. Regression analyses showed that shame
measures, as a whole, were able to predict approximately 22% of the variance in problematic
drinking. In contrast, guilt measures did not contribute any additional variance in prediction
once shame measures were already taken into consideration. Overall, these patterns suggest
that shame is likely to be more influential in the maintenance of problematic drinking
compared to guilt.
In terms of individual predictors, internalized shame stood out as the most robust and
unique predictor. Internalized shame was the only measure able to independently predict
variance in problematic drinking after controlling for other shame and guilt variables. One
explanation could relate to the fact that the ISS was specifically created for alcohol misusing
populations; perhaps the content of the items more specifically reflects the shameful experiences
ASSESSMENT PREDICTING ALCOHOL MISUSE 20
of people struggling with alcohol misuse. Another possible explanation is that experienced
shame may be more strongly associated with drinking problems because repeated drinking-
related problems may tend to lead to an experience of shame even in the absence of a strong
dispositional tendency toward shame.
Results confirmed our expectations that shame proneness and experience-based measures
of shame and guilt would be positively associated with drinking-related problems, whereas guilt
proneness would be negatively associated with drinking-related problems. Shame proneness,
experienced shame, and experienced guilt were all positively correlated with each other, but
negatively correlated with guilt proneness. This pattern is consistent with previous findings of
guilt proneness being protective against problematic substance use (Dearing et al., 2005; Treeby
& Bruno, 2012). These findings highlight the importance of being precise in one’s definition of
shame and guilt, and measuring accordingly. We speculate that a disposition toward guilt
proneness (i.e., as measured by the TOSCA-SFG) may be protective against people developing
problematic drinking in that they would be likely to cut back their drinking if it results in
relationship harm to others. However, when a person who has had episodes of embarrassing or
problematic drinking behavior is asked to recall their drinking experiences, they may tend to
experience guilt over the harms that have occurred as a result of their drinking. Longitudinal
studies and studies including other assessment methods are needed in order to further disentangle
these possible explanations.
A major strength of this study was the use of a community sample of drinkers (versus
an only college student sample), most of whom were not receiving any treatment, which may
make the results more generalizable to drinkers as a whole. Our sample was also representative
in terms of overall levels of shame in that mean levels of shame were similar to the only
ASSESSMENT PREDICTING ALCOHOL MISUSE 21
previous study to publish ISS means, which used a college student sample (del Rosario &
White, 2006). We also included individuals exhibiting a wide range of drinking behavior,
thereby avoiding restricted ranges that can occur when only including problematic drinkers.
While a relatively small sample size is a potential limitation of the study, the sample
size exceeds published recommendations of at least 10 participants per predictor (Vanvoorhis
& Morgan, 2007). Another potential limitation is the use of the AUDIT as a measure of
drinking-related problems; it includes items related to drinking frequency and therefore is not a
pure measure of negative drinking consequences. This weakness of the AUDIT is mitigated
somewhat given the minimal relations between shame/guilt and drinking frequency that were
observed. In addition, models statistically controlling for drinking days and drinking amount
did not alter results involving the AUDIT. Finally, the cross-sectional nature of this study
limits our ability to determine the temporal pattern of associations. On one hand, problematic
drinking involving norm violations, role violations, and/or harms caused to the self or others
(Klingemann & Gmel, 2001) could cause heightened experiences of shame and guilt. On the
other hand, shame and/or guilt could serve as antecedents for problematic drinking, since these
emotional states are typically aversive and may lead to avoidant coping through drinking
(Cooper, Frone, Russell, & Mudar, 1995). Current evidence that shame serves as a direct
antecedent to drinking is limited. Future studies, such as ecological momentary assessment
approaches or daily diary studies, are needed to understand temporal patterns. The
investigation of relevant measures in the present study is an important first step in that
While there is increasing consensus in the literature regarding conceptual distinctions
between shame and guilt, researchers are often not clear about the distinctions between
ASSESSMENT PREDICTING ALCOHOL MISUSE 22
experience and proneness measures. In our study, a proneness measure of guilt actually
correlated in the opposite direction with problematic drinking compared to a measure of
experienced guilt, suggesting that this distinction is important. The development of new and
improved temporally and contextually-sensitive measures is needed to accurately capture the
full range of guilt and shame experiences in relation to drinking.
Our findings support the conclusions of other researchers who have called for a
reevaluation of confrontational interventions for substance abuse (Meehan et al., 1996;
O’Connor et al., 1994). Our data indicate that confrontational approaches that increase shame
are likely to be iatrogenic or at least not helpful. This conforms with other research on the often
problematic nature of shame (Wong & Tsai, 2007). As such, treatments that specifically target
the elicitation of adaptive guilt, are non-shaming, reduce maladaptive shame, or alter shame to
function adaptively could prove to be more effective than shame-inducing interventions. To
expand on these possibilities, future research should focus on determining temporal and causal
relations between interventions, shame and guilt responses, and problematic alcohol use.
ASSESSMENT PREDICTING ALCOHOL MISUSE 23
The authors would like to thank Dalai Luoma and James Guinther for their assistance in
preparing this article.
Jason Luoma, Ph.D. served as principal investigator and contributed to experimental design, data
analyses, and manuscript preparation. Paul Guinther, Ph.D. served as laboratory coordinator and
contributed to experimental design, data collection, database management, data analyses, and
manuscript prepraration. Jacqueline Potter and Megan Cheslock served as research assistants and
contributed to data collection, preliminary data analyses, and manuscript preparation.
This research was fully supported with internal private funding through the clinical-research
social business model of Portland Psychotherapy Clinic, Research, and Training Center.
ASSESSMENT PREDICTING ALCOHOL MISUSE 24
Babor, T. F., Higgins-Biddle, J. C., Saunders, J. B., & Monteiro, M. G. (2001). AUDIT-The
Alcohol Use Disorders Identification Test-Guidelines for Use in Primary Care. World
Health Organization (2nd ed.). Geneva, Switzerland: World Health Organization.
Boudrez, H. (2009). Psychological factors and long-term abstinence after smoking cessation
treatment. Journal of Smoking Cessation, 4(1), 10–17. http://doi.org/10.1375/jsc.4.1.10
Cook, D. R. (1987). Measuring shame: The internalized shame scale. Alcoholism Treatment
Quarterly, 4(2), 197–215. http://doi.org/10.1300/J020v04n02_12
Cook, D. R., & Nathanson, D. L. (1996). Empirical studies of shame and guilt: The Internalized
Shame Scale. In D. L. Nathanson (Ed.), Knowing feeling: Affect, script, and psychotherapy
(132–165). New York, NY: W W Norton & Co.
Cooper, M. L., Frone, M. R., Russell, M., & Mudar, P. (1995). Drinking to regulate positive and
negative emotions: A motivational model of alcohol use. Journal of Personality and Social
Psychology, 69(5), 990–1005. http://doi.org/10.1037/0022-3522.214.171.1240
Courtney, K. E., & Polich, J. (2009). Binge drinking in young adults: Data, definitions, and
determinants. Psychological Bulletin, 135(1), 142–56. http://doi.org/10.1037/a0014414
Coxe, S., West, S. G., & Aiken, L. S. (2009). The analysis of count data: A gentle introduction to
poisson regression and its alternatives. Journal of Personality Assessment, 91(2), 121–136.
de Hooge, I. E., Nelissen, R. M., Breugelmans, S. M., & Zeelenberg, M. (2011). What is moral
about guilt? Acting “prosocially” at the disadvantage of others. Journal of Personality and
Social Psychology, 100(3), 462–473. http://doi.org/10.1037/a0021459
ASSESSMENT PREDICTING ALCOHOL MISUSE 25
de Meneses-Gaya, C., Zuardi, A. W., Loureiro, S. R., & Crippa, J. A. S. (2009). Alcohol Use
Disorders Identification Test (AUDIT): An updated systematic review of psychometric
properties. Psychology & Neuroscience, 2(1), 83–97.
Dearing, R. L., Stuewig, J., & Tangney, J. P. (2005). On the importance of distinguishing shame
from guilt: Relations to problematic alcohol and drug use. Addictive Behaviors, 30(7),
del Rosario, P. M., & White, R. M. (2006). The Internalized Shame Scale: Temporal stability,
internal consistency, and principal components analysis. Personality and Individual
Differences, 41(1), 95–103. http://doi.org/10.1016/j.paid.2005.10.026
Fontaine, J. R. J., Luyten, P., Estas, C., & Corveleyn, J. (2004). Scenario-based and frequency-
based approaches towards the measurement of guilt and shame: Empirical evidence for
unique contributions. In S. P. Shohov (Ed.), Advances in psychology research (pp. 141–
154). Hauppauge, NY: Nova Science Publishers.
Giguère, B., Lalonde, R. N., & Taylor, D. M. (2014). Drinking too much and feeling bad about
it? How group identification moderates experiences of guilt and shame following norm
transgressions. Personality and Social Psychology Bulletin, 40(5), 617-632.
Gilbert, P. (1998). What is shame? Some core issues and controversies. In P. Gilbert & B.
Andrews (Eds.), Shame: Interpersonal behavior, psychopathology, and culture (pp. 3–38).
New York, NY: Oxford University Press.
Ianni, P. A., Hart, K. E., Hibbard, S., & Carroll, M. (2010). The association between self-
forgiveness and alcohol misuse depends on the severity of drinker’s shame: Toward a
buffering model. Addictive Disorders & Their Treatment, 9(3), 106–111.
ASSESSMENT PREDICTING ALCOHOL MISUSE 26
Klingemann, H., & Gmel, G. (Eds.). (2001). Mapping the social consequences of alcohol
consumption. Dordrecht, Netherlands: Kluwer Academic Publishers.
Lewis, H. (1971). Shame and guilt in neurosis (1st ed.). New York, NY: International University
Luoma, J. B., & Kohlenberg, B. S. (2012). Self-stigma and shame in addictions. In S. C. Hayes
& M. Levin (Eds.), Mindfulness and acceptance for addictive behaviors (pp. 216-238).
Oakland, CA: New Harbinger.
Marschall, D., Sanftner, J., & Tangney, J. P. (1994). The State Shame and Guilt Scale. Fairfax,
VA: George Mason University.
Meehan, W., O’Connor, L. E., Berry, J. W., Weiss, J., Morrison, A., & Acampora, A. (1996).
Guilt, shame, and depression in clients in recovery from addiction. Journal of Psychoactive
Drugs, 28(2), 125–134. http://doi.org/10.1080/02791072.1996.10524385
Mohr, C., Brannan, D., Mohr, J., Armeli, S., & Tennen, H. (2008). Evidence for positive mood
buffering among college student drinkers. Personality and Social Psychology Bulletin,
34(9), 1249–1259. http://doi.org/10.1177/0146167208319385
O’Connor, L. E., Berry, J. W., Inaba, D., Weiss, J., & Morrison, A. (1994). Shame, guilt, and
depression in men and women in recovery from addiction. Journal of Substance Abuse
Treatment, 11(6), 503–510. http://doi.org/10.1016/0740-5472(94)90001-9
Randles, D., & Tracy, J. L. (2013). Nonverbal displays of shame predict relapse and declining
health in recovering alcoholics. Clinical Psychological Science, 1(2), 149–155.
Rybak, C. J., & Brown, B. (1996). Assessment of internalized shame: Validity and reliability of
ASSESSMENT PREDICTING ALCOHOL MISUSE 27
the Internalized Shame Scale. Alcoholism Treatment Quarterly, 14(1), 71–83.
Saunders, J. B., Aasland, O. G., Babor, T. F., de la Fuente, J. R., & Grant, M. (1993).
Development of the Alcohol Use Disorders Identification Test (AUDIT): WHO
collaborative project on early detection of persons with harmful alcohol consumption--II.
Addiction, 88(6), 791–804. http://doi.org/10.1111/j.1360-0443.1993.tb02093.x
Schmader, T., & Lickel, B. (2006). The approach and avoidance function of guilt and shame
emotions: Comparing reactions to self-caused and other-caused wrongdoing. Motivation
and Emotion, 30(1), 42-55. http://doi.org/10.1007/s11031-006-9006-0
Sobell, L. C., Brown, J., Leo, G. I., & Sobell, M. B. (1996). The reliability of the Alcohol
Timeline Followback when administered by telephone and by computer. Drug and Alcohol
Dependence, 42(1), 49–54. http://doi.org/10.1016/0376-8716(96)01263-X
Sobell, L. C., & Sobell, M. B. (1992). Timeline follow-back: A technique for assessing self-
reported alcohol consumption. In R. Z. Litten & J. P. Allen (Eds.), Measuring alcohol
consumption: Psychosocial and biological methods (pp. 41–72). Totowa, NJ: Humana
Stuewig, J., Tangney, J. P., Kendall, S., Folk, J. B., Meyer, C. R., & Dearing, R. L. (2014).
Children’s proneness to shame and guilt predict risky and illegal behaviors in young
adulthood. Child Psychiatry & Human Development, 46(2), 217–227.
Tabachnick, B. G., & Fidell, L. S. (2007). Using multivariate statistics (5th ed.). Boston, MA:
Tangney, J. P., & Dearing, R. L. (2002). Shame and guilt. P. Salovey (Ed.). New York, NY: The
ASSESSMENT PREDICTING ALCOHOL MISUSE 28
Tangney, J. P., Stuewig, J., & Mashek, D. J. (2007). Moral emotions and moral behavior. Annual
Review of Psychology, 58, 345–372.
Tangney, J. P., Wagner, P., & Gramzow, R. (1989). The test of self-conscious affect. Fairfax,
VA: George Mason University.
Tracy, J. L., Robins, R. W., & Tangney, J. P. (2007). The self-conscious emotions: Theory and
research. New York, NY: Guilford Press.
Treeby, M., & Bruno, R. (2012). Shame and guilt proneness: Divergent implications for
problematic alcohol use and drinking to cope with anxiety and depression symptomatology.
Personality and Individual Differences, 53(5), 613–617.
Vanvoorhis, C. R. W., & Morgan, B. L. (2007). Understanding power and rules of thumb for
determining sample sizes. Tutorials in Quantitative Methods, 3(2), 43–50.
Wong, Y., & Tsai, J. L. (2007). Cultural models of shame and guilt. In J. Tracy, R. Robins, & J.
Tangney (Eds.). Handbook of self-conscious emotions (pp. 210-223). New York, NY: