Selfhandicapping is a selfprotective strategy employed by individuals anticipating poor performance

1
Self-Handicapping Occurs in the Real World Too
A Naturalistic Approach to the Study of
Self-Handicapping
J. Scott Hanson Michael J Strube
Participants Participants were students enrolled
in the Spring semester introductory psychology
course at Washington University in St. Louis.
Enrollment for the course was approximately 107
students (45 men, 62 women). Several measurements
were taken over the course of the semester. The
sample reported here consists of the 51 students,
15 men and 36 women, for whom we received scores
for at least two state self-esteem measurement
occasions and all other pertinent measures2.
Instruments Three measures were collected in
this study the State Self-Esteem Scale (SSES
Heatherton Polivy, 1991) a short form of the
Self-Handicapping Scale (SHS Jones Rhodewalt,
1982) and a modified version of the
Self-Handicapping Checklist (SHC Strube,
1986). The SSES is designed to assess state
levels of esteem. The measure requires
respondents to use a 5-point scale to rate their
current feelings (1 Not at all and 5
Extremely) about several factors such as
abilities, public regard, body image, and
confidence. Factor analysis indicates that groups
of items load primarily on one of the three
constructs (Performance, Social, and Appearance),
and each set of items assesses state self-esteem
levels for the specific domain (Heatherton
Polivy, 1991). The SHS is an individual
difference measure of the tendency to engage in
strategic self-esteem maintenance (Jones
Rhodewalt, 1982). The short-form version (Strube,
1986) is a 10-item, self-report questionnaire in
which the respondent uses a 6-point scale to rate
agreement with statements reflecting the use of
self-handicapping behaviors and concern about
achievement settings (1 Disagree very much and
6 Agree very much) (Rhodewalt et al., 1984
Strube, 1986). The SHC is a list of circumstances
that students might cite as having affected their
preparation for an exam. The checklist (Strube,
1986) included the following circumstances heavy
course load, other exams to study for, sickness,
other commitments that prevented sufficient study
time, missed an important lecture, heavy work
schedule, lack of sleep, noisy roommates
prevented studying, didn't keep up with material,
studied wrong material, and didn't take good
notes. In an effort to de-emphasize the negative
tone of the checklist, we modified it to include
some items that could have been construed as
self-enhancing behaviors. That is, we included
behaviors on the checklist that could reasonably
be expected to improve the individual's score on
the exam desire to perform well or better than
last exam, put forth greater effort than normal,
understood the material better than usual,
prepared for exam with other people (study
group), and recently improved study habits. All
analyses, however, center on the negative
behaviors comprising 70 of the
list. Procedure TIME 1 During the first week of
class participants completed the SSES, and the
SHS, in addition to several measures of
self-concept. TIME 2 During the week before
mid-term exams participants provided information
about their anticipated performance on the
upcoming exam certainty of that projection, and
completed the SSES a second time. TIME 3
Immediately prior to taking their mid-term exam
in the course, students completed the SHC
indicating factors and events that they believed
may have influenced their ability to adequately
prepare for the exam. Once they had finished the
exam participants again completed the SSES. TIME
4 Participants completed a fourth SSES about
three weeks after the exam, not long after
returning from Spring Break.
Hierarchical regression analyses were conducted
to examine the extent to which self-handicapping
trait, self-esteem stability, prior exam
performance, anticipated score, and certainty of
prediction predicted the number of self-handicaps
claimed on the SHC. In Step 1, the trait
variables (SH Trait and SE Stability) were
entered simultaneously. In Step 2, Test 1 scores
were entered as a measure of domain ability. Step
3 consisted of the ratings of anticipated
performance and situational uncertainty. F tests
of the partial regression coefficients were used
to assess the significance of each of these
predictors as main effects. In Step 4, all
two-way product terms (i.e., SH Trait x Test1, SE
Stability x Anticipate, etc.) were entered as a
block. F tests of the partial regression
coefficients were again used to assess the
significance of these predictors. Finally, in
Step 5, the three-way interaction product term
for Test1 x Anticipate x Certainty was entered as
a representation of the interaction between prior
performance and uncertainty. This three-way
interaction conceptually replicates the
experimental conditions of prior feedback (Test 1
performance) and uncertainty (ambiguity about
factors associated with success), however, the
current procedure is based on changes in the
definitions of these constructs as outlined in
the introduction.
The major finding was that individuals who
received success feedback in an important domain
and expressed high confidence that they would
fail to replicate that performance made more
self-handicap claims than those with any other
combination of Test 1 performance, anticipated
Test 2 performance, and certainty of prediction.
This finding emerged from a design that improves
on the traditional experimental approach in that
the naturalistic setting and redefinition of
uncertainty provide a more meaningful
representation of antecedents to
self-handicapping behavior. The current design
conceptually replicates and extends previous work
on self-handicapping, but offers some important
advances over the traditional experimental
approach. Although the naturalistic setting
denies manipulation of feedback, it does permit
the assessment of both success and failure
feedback on self-handicapping behavior, as well
as the use of the more appropriate
conceptualization of uncertainty about the
expectation of replicating a previous
performance. Past experimental work has provided
participants with only noncontingent success
feedback, followed by the opportunity to handicap
in a given situation, with no direct assessment
of uncertainty. The current approach derives from
the premise that uncertainty is a meaningful and
measurable component of the self-handicapping
process, if considered appropriately. Individuals
rarely have little or no knowledge about factors
leading to a previous performance. Even if they
are unsure about the steps necessary to succeed,
past outcomes are generally inferred to be
attributable to behaviors carried out prior to
their occurrence - this is how superstitions are
formed (Skinner, 1948). Realistic uncertainty
originates from concern over one's preparation
for an upcoming test of ability, and may only
arise if one knows what is necessary for success
and is sure that it is not present in the given
situation. In this case, one will look for
alternate explanations or preparations in
anticipation of a negative outcome. This
uncertainty will be manifest and assessable in
predictions about performance and the confidence
placed in those predictions. These findings
provide valuable information about the factors
contributing to self-handicap claims and suggest
some modifications to future experimental work.
First, the uncertainty variable should not be
implied, but operationally defined and measured.
This requires assessing the individuals level of
confidence and expected performance prior to both
the first and second trials on the task. Second,
certainty should be treated as a continuous
variable and evaluated as a factor in the
self-handicapping process rather than inferred as
a product of the manipulation. Finally, the "no
success" group should be considered as more than
a control. These individuals self-handicap
claims appear to be a function of anticipated
score and certainty of estimate just as much as
those of the prior success group.
Self-handicapping is a self-protective strategy
employed by individuals anticipating poor
performance on an important evaluation (Jones
Berglas, 1978). Most support for
self-handicapping comes from laboratory studies
using tasks of questionable relevance. In
contrast, the current study investigated
self-handicapping behavior in the context of
performance on a psychology course exam.
Participants completed measures assessing
self-handicapping, state self-esteem, and made
predictions about future performance. Students
with more variable state self-esteem made more
self-handicap claims than those with more stable
self-esteem. Additionally, Test 1 performance
interacted with certainty and level of
anticipated Test 2 performance to affect
self-handicap claims compared to all other
participants, individuals scoring high on Test 1
with high certainty of a poor performance on Test
2 claimed more impediments prior to Test 2. These
results differ from those obtained in laboratory
studies and suggest that, although some
individuals are undoubtedly more inclined to make
self-handicap claims, the context may play a
larger role than previously believed.
Results of hierarchical regression of
Self-Handicap Claims on Self-Handicap Trait,
Performance Self-Esteem Stability Predictor
B SE B b R2 DR2 Traits
Self-Handicap .049 .037 .184
SSES-Stability .415 .190 .303 .159 Dom
ain Ability Test 1 Score -.005 .015 -.04
7 .161 .002 Situational Factors
Anticipated Score -.045 .020 -.323
Certainty of Score -.003 .016 -.026 .242 .
081 2-Way Interactions SHS X SSE
.002 .031 .029 SHS X Test
1 -.004 .002 -.270 SHS X Ant.
Score -.001 .004 .044 SHS X
Certainty -.002 .003 -.153 SSE X Test 1
.020 .018 .756 SSE X Ant.
Score -.021 .016 -.653 SSE X Certainty
.019 .017 .620 Test 1 X Ant.
Score -.0003 .001 -.052 Test 1 X
Certainty -.001 .001 -.238 Ant. Score X
Certainty .0002 .002 .026 .385 .143 3-Way
Interactions Test 1 X Ant. Score X
Certainty -.0002 .00006 -.568 .479 .0
94 Significance of b values and increments in R2
noted with . p lt .05
Jones and Berglas (1978 Berglas Jones, 1978)
first proposed that individuals fearful of a poor
showing on some diagnostic measure may erect
barriers that actually detract from their ability
to perform well. Construction of these
self-handicaps prior to evaluation permits one to
discount ability attributions should failure
occur, but to enhance such attributions for
successful outcomes (Kelley, 1971). That is,
because the handicap decreases the likelihood of
optimal performance, the handicapper is relieved
of responsibility for a poor showing. However,
should the handicapper do well, ability
attributions will be augmented because the
performance occurs despite obstacles to obtaining
a high score. The self-protective nature of the
self-handicapping strategy makes it attractive to
individuals with uncertain self-image or to those
in threatening situations. The three critical
factors believed to determine the likelihood of
self-handicapping behavior are success,
uncertainty, and self-esteem, and they generally
operate together in promoting the use of
self-protective strategies. Previous success in
an important area creates a "fragile and
ambiguous. . . self-concept" (Jones Berglas,
1978 p. 205) when the cause for success is
uncertain. Self-protective strategies, such as
self-handicapping, may bolster a fragile and
uncertain self-esteem (Harris Snyder, 1986).
Past research provided much evidence for the
individual and combined roles of these factors in
self-handicapping behavior. The most common
experimental design entails delivering
noncontingent success feedback in an important
domain to the participant, followed by the offer
of a handicap before engaging in an anticipated
replication of the domain-related task
(Rhodewalt, Morf, Hazlett, Fairfield, 1991).
This procedure creates uncertainty about the
performance by denying the individual useful
information about factors contributing to the
initial successful outcome. Individuals
experiencing this noncontingent success find
themselves under pressure to repeat the
performance, but with no firm understanding of
how the initial success was achieved. The
potentially severe impact of failure to
self-esteem and the uncertain ability to
replicate the performance can combine to make
construction of a self-handicap in anticipation
of a potential performance deficit quite
attractive (Arkin Baumgardner, 1985). Although
this design has been effective in producing
self-handicapping behavior, there is reason to be
concerned about the generalizability of the
findings. The contrived nature of the laboratory
manipulation substantially limits the realism of
the situation (Christensen, 1997), possibly
prompting participants to discount the importance
of the task, thus violating one of the primary
conditions promoting the use of a defensive
strategy (Shepperd Arkin, 1989). Furthermore,
the established laboratory procedure has failed
to effectively assess the role of either
situational or trait uncertainty in
self-handicapping. The current study attempts to
correct for past shortcomings by assessing the
roles of success, uncertainty, and self-esteem
variability on self-handicapping behavior in a
natural setting involving a task important to
participants. The two primary concerns addressed
in the present study are past methodological
problems and the conceptual representation of
self-esteem effects. In addressing
methodological concerns we employed a natural,
and personally important, anticipated task in the
form of exams in an academic course for credit.
We used previous exam performance as an
indication of ability and representation of
initial feedback in the domain. Participants
made predictions about their anticipated scores
on an upcoming exam and provided certainty
ratings about those estimates. We assessed
self-esteem levels at different time periods over
the course of the semester and used the variance
estimates from those scores as an indication of
self-esteem variability. This information
permitted us to assess the role of prior failure
as well as prior success, anticipated level and
certainty of performance, and self-esteem
variability in the use of self-handicapping
claims to account for exam performance.
Arkin, R. M., Baumgardner, A. H. (1985).
Self-handicapping. In J. H. Harvey G. Weary
(Eds.), Attribution Basic issues and
applications (pp. 169-202). New York Academic
Press. Christensen, L. B. (1997). Experimental
Methodology (7th ed.). Boston, MA Allyn and
Bacon. Harris, R. N., Snyder, C. R. (1986). The
role of uncertain self-esteem in
self-handicapping. Journal of Personality and
Social Psychology, 51, 451-458. Heatherton, T.
F., Polivy, J. (1991). Development and
validation of a scale for measuring state
self-esteem. Journal of Personality and Social
Psychology, 60, 895-910. Jones, E. E., Berglas,
S. (1978). Control of attributions about the self
through self-handicapping strategies The appeal
of alcohol and the role of under-achievement.
Personality and Social Psychology Bulletin, 4,
200-206. Jones, E. E., Rhodewalt, F. (1982).
Self-handicapping scale. (Available from the
authors at Department of Psychology, Princeton
University, or the Department of Psychology,
University of Utah.) Kelley, H. H. (1971).
Attributioni in social interaction. In E. E.
Jones, D. E. Kanouse, H. H. Kelley, R. E.
Nisbett, S. Valins, B. Weiner (Eds).
Attributions Perceiving the causes of behavior
(pp. 1-26). Morristown, NJ General Learning
Press. Rhodewalt, F., Morf, C., Hazlett, S.,
Fairfield, M. (1991). Self-handicapping The role
of discounting and augmentation in the
preservation of self-esteem. Journal of
Personality and Social Psychology, 61,
122-131. Rhodewalt, F., Saltzman, A. T.,
Wittmer, J. (1984). Self-handicapping among
competitive athletes The role of practice in
self-esteem protection. Basic and Applied Social
Psychology, 5, 197-209. Shepperd, J. A., Arkin,
R. M. (1989). Determinants of self-handicapping
Task importance and the effects of preexisting
handicaps on self-generated handicaps.
Personality and Social Psychology Bulletin, 15,
101-112. Strube, M. J (1986). An analysis of the
self-handicapping scale. Basic and Applied Social
Psychology, 7, 211-224.
Figure 1 Performance handicap claims as a
function of anticipated mid-term performance and
certainty of prediction for High Test 1 Performers
Figure 2 Performance handicap claims as a
function of anticipated mid-term performance and
certainty of prediction for Low Test 1 Performers
Direct Correspondence to J. Scott
Hanson Department of Psychology Campus Box
1125 Washington University in St. Louis St.
Louis, MO 63130 email jshanson_at_artsci.wustl.edu