An Investigation Of The Assumptions Of Self-handicapping

Table of contents xiii Evidence for the Two Formulations for the Uncertainty Set for Selfhandicapping 261 Support for the Assumptions for Individual Difference Variables in the use of Self-Handicapping. 262 Evidence for the Onset of Self-Handicapping 262 Evidence for Achievement Goal Orientation and Self-Handicapping 263 Evidence for the Relationship Between Self-Esteem and Self-Handicapping 265 Conclusion 267 Appendix A Survey of Frequency of Use of Physical Skills Test and Suitability of Physical Skills Tests 269 Appendix B Survey of the Frequency of Use and Suitability of Physical Skills Tests 274 Appendix C Face Validity Scale for Overarm Throwing Test 275 Appendix D Information Letter to Principles and HPE Teachers 276 Appendix E Information Letter to Students 277 Appendix F Parental Information Letter 278 Appendix G Consent Forms 280 Appendix H Measure of Ego-relevance for the Physical Domain 282 Appendix I Measure of Perceived Control Over Test Outcome 283 Appendix J Situational Impediments Checklist 284 Appendix K Information Letter to Principles and HPE Teachers for the Project: Self-Handicapping by Australian 286 Appendix L Information Letter for the Project: Self-Handicapping by Australian Youth in the Physical Domain 287

Table of contents xiv Appendix M Thank You Letter 289 Appendix N Feelings About Sport Questionnaire 290 Appendix O Self-handicapping Scale (Jones and Rhodewalt, 1986) 291 Appendix P Task / Ego Orientation Scale (Duda & Nicholls, 1992) 293 Appendix Q Global Measure of Self-Esteem (Coopersmith, 1990) 295 Appendix R The Physical Self-Perception Profile: Physical Self-worth Subscale (Fox, 1989) 298 Appendix S The Physical Self-Perception Profile: Physical Sport Competence Sub-scale (Fox 1989) 300 Appendix T Task Specific Self-Efficacy 302 REFERENCES 303

Table of tables xv TABLE OF TABLES Table 1 Sunamary of Frequency Analysis of Common Tests of Physical Ability Used in Tasmanian High and Primary Schools 87 Table 2 Summary of Teachers' Perceptions of Suitability of Five Physical Skill Tests 88 Table 3 Summary of Post Hoc Analysis of Mean Differences in Teachers' Perceptions of Suitability For Tests of Physical Skill 89 Table 4 Summary of Descriptive Statistics for Age Groups' Performance Scores on the Sham Test of Overarm Throwing 94 Table 5 Mean and Percentiles for the Sham Test of Overarm Throwing Ability 97 Table 6 Summary of Frequency Statistics for Perceived Impediments to Performance 120 Table 7 Descriptive Statistics for Perceptions of Control 122 Table 8 Summary of ANCOVA Results for Perception of Control 123 Table 9 Mean and Standard Deviation Scores for Self-handicaps 125 Table 10 Summary of ANOVA Results for Feedback Contingency Condition, Gender, and their Interaction 126 Table 11 ANOVA Results for Feedback Contingency Condition, Gender, and their Interaction 127 Table 12 Summary of Mean Differences in Self-reported Performance Disruption by Feedback Contingency Conditions 128 Table 13 Summary of Frequency of Reported Perceived Impediments to Performance Table 14 Descriptive Statistics for Perceptions of Control 154 155

Table of tables xvi Table 15 ANCOVA for Perception of Control 156 Table 16 Descriptive Statistics for Self-reported Disruption Scores 158 Table 17 ANOVA for Self-reported Performance Disruption 159 Table 18 Summary of Post Hoc Comparisons for Self-reported Performance Disruption 160 Table 19 Suncunary of Two-Way Table for Gender for Reported Perceived Impediments to Performance 184 Table 20 Summary of Two-Way Table for Evaluative Conditions for Perceived Impediments to Performance 185 Table 21 Descriptive Statistics for Perceptions 186 Table 22 Summary of ANCOVA Results for Perception of Control 187 Table 23 Summary of Descriptive Statistics for all Variables used in the Regression Analysis 189 Table 24 Summary of Standard Analysis for Variables Predicting of Self-reported Performance Disruption (N 160) Table 25 Correlation Matrix for Self-report Measures (N 160) 190 219 Table 26 Summary of Frequency Analysis for Impediments to Performance 221 Table 27 Descriptive Statistics for Perceptions of Control 222 Table 28 ANCOVA for Post Experimental Perception of Control 224 Table 29 Summary of Descriptive Statistics For Variables Used to Predict Self-reported Performance Disruption 225 Table 30 Summary of Multiple Regression Analysis for Variables Predicting of Self-reported Performance Disruption (N 160) 227 Table 31 Summary of Frequency Analysis for impediments to Performance 248 Table 32 Descriptive Statistics for Perceptions of Control 249

Table of tables Table 33 Sunamary of ANCOVA Results for Perception of Control xvii 250 Table 34 Descriptive Statistics for Variables Used to Predict Situational Self-Handicapping 251 Table 35 Summary of Standard Analysis for Variables Predicting Self-reported Performance Disruption (N 196) 253

Introduction 1 CHAPTER 1 Introduction The purpose of this thesis was to test some central assumptions of selfhandicapping theory. Jones and Berglas (1978) originally described selfhandicapping as the process whereby individuals actively plan and use excuses before possible flawed performances. For example, athletes may claim that their injuries will possibly interfere with their upcoming performances. Why would individuals do such things? There are two different motivations for self-handicapping. The motivation to self-handicap may lie in either a self-presentational emphasis (i.e., impression management) or the maintenance of personal beliefs about competence and control (Arkin & Baumgardner, 1985; Higgins, 1990, Snyder, 1990). Selfpresentational concerns seem to be the main motivation for self-handicapping when evaluative conditions are public whereas maintenance of personal beliefs are prime motivation when evaluative conditions are more private. Within self-handicapping theory is the proposition that individuals use selfhandicaps when they experience performance uncertainty on evaluative tasks (Arkin & Baumgardner, 1985). Snyder (1990) highlighted that there is agreement on the proposition that performance uncertainty is an underpinning factor in selfhandicapping, but there is disagreement regarding the antecedents to performance uncertainty. The disagreement regarding the antecedents to performance uncertainty has lead to two divergent streams of research. Jones and Berglas (1978) hypothesised a history of non-contingent positive reinforcement leads to self-handicapping.

Introduction 2 Alternatively, Snyder and Smith (1982) argued that a history of simple concerns about performance were the antecedent to self-handicapping. Tests of the Jones and Berglas hypothesis have used deceptive non-contingent success performance data to manipulate performance certainty. Investigations of the Snyder and Smith hypothesis have used the implication of failure as the mechanism to manipulate performance certainty. There is also a difference in the definition of the type of self-handicap in the different streams of research. Jones and Berglas (1978) maintained that people self-handicap by manipulating their behaviours. For example, in one study (Berglas & Jones, 1978), participants manipulated their own behaviour by choosing higher dosages of a drug known to inhibit performance. Snyder and Smith (1982) argued that people self-handicap by using self-avowals of impediments. For example. Smith, Snyder, and Handelsman (1982) documented how individuals self-reported high levels of test anxiety to try to offset the implications of failure. Subsequent investigations using both protocols have produced evidence for some assumptions for self-handicapping. Nonetheless, there are a number of shortcomings to this evidence. There is no published confirmation for either hypothesis for the physical domain. It is unclear if individuals who are exposed to either protocol will self-handicap when they experience performance uncertainty about outcomes associated with physical ability. It is also unclear what antecedent conditions are associated with the two hypothesised types of self-handicaps for the physical domain. For example, it is unknown if individuals exposed to failure will self-handicap by manipulating their behaviour. Finally, one line of research has shown that individual difference variables are associated

Introduction 3 with differences in the use of self-handicaps. Higgins (1990) defined individual difference variables as characteristics normally associated with personality. There is documented evidence that supports assumptions for a variety of individual difference variables, such as self-esteem (Harris, Snyder, Higgins, & Schrag, 1986), self-esteem certainty (Harris & Snyder, 1986), gender (Berglas & Jones, 1978), trait self-handicapping (Carron, Prapavessis, & Grove, 1994; Hausenblas & Carron, 1996; Prapavessis & Grove, 1998; Rhodewalt, Saltzman, & Wittmer, 1984), and achievement goal orientation (Ryska, Yin, & Boyd, 1999). Nonetheless, evidence for the assumptions holding for the physical domain is limited because there is no published evidence, and where there is evidence, the protocols have not included an experimental manipulation of performance certainty. Also, tests of the assumptions are limited in terms of the sample demographics. The overarching aim of the studies in this thesis was to investigate how young people responded to evaluative threat that was specific to the physical domain. Young people were the target population for this thesis for three reasons. First, there is little literature that describes the self-handicapping concept as it relates to young people. Second, where there is literature, the evidence shows young people use self-handicaps in different ways to adults. Finally, as selfhandicapping represents somewhat health ego-centrism, but if used continually, pathological intemalisation is a possibility, it is necessary to understand how young people respond to evaluative threat across domains to help educators and coaches avoid dysfunctional instructional strategies.

Introduction 4 The review of literature structure is in terms of the key assumptions tested in this thesis. The assumptions that formed the basis of the research questions that guided the investigations are as follows: 1. Some individuals will respond to evaluative threat by externalising the cause of possible flawed performance to factors other than ability (Higgins, 1990). 2. The motive to self-handicap is concerns about performance uncertainty. There are two hypotheses for the antecedents of performance uncertainty with each hypothesis predicting differing types of self-handicaps (Berglas «& Jones, 1978; Snyder & Smith, 1982). 3. Individual differences in personality and gender are assumed to cause differences in self-handicapping (Higgins, 1990; Rhodewalt, 1990; Self, 1990). In the review of literature, I have documented the evidence for each of the aforementioned assumptions. These limitations involved methodological concerns, limited testing of the assumptions in the physical achievement domain, and a lack of testing with differing populations. Addressing these limitations formed the research aims. The research aims were as follows: 1. The first research aim was to develop a sham test of physical ability and associated performance data to use in the remaining studies (Study 1). 2. The second research aim was to test the two hypotheses for the antecedent of performance uncertainty (Study 2). 3. An associated aim was to establish if specific types of handicapping behaviour

Introduction 5 were associated with specific types of antecedents and gender to performance uncertainty. 4. A third research aim was to examine the onset of self-handicapping behaviour (Study 3). 5. The fourth research aim was to determine the relationship between achievement goal orientation, trait self-handicapping, and self-handicapping (Study 4). 6. The final aim was to examine the relationship between self-esteem and selfhandicapping (Studies 5 & 6). To fulfil the aforementioned aims, the research replicated previous experimental protocols (Berglas & Jones, 1978; Smith et al., 1982). The protocol involved a random sample of young people (aged between 13 and 16) who believed that they were to complete two parallel tests of physical ability. The young people did not know that the test was a sham. After the first test, the young people were exposed to the experimental manipulation. The experimental manipulation involved giving participants deceptive first test performance scores. Regardless of their actual performance scores, participants received scores that represented either non-contingent success or non-contingent failure. A nonevaluative control group allowed for comparisons in self-handicapping responses between evaluative groups. After the experimental manipulation, the young people had an opportunity to use mitigating excuses before the second test. Previous protocols using experimental designs have modified an existing test of ability (i.e., intelligence tests) to enhance the effect of the manipulation. Modifications have normally centred on changing the items of tests in such a

Introduction 6 way that they remain realistic, yet unsolvable. There has been no rephcation of the protocol for the physical domain, so it was necessary to select and modify an existing test of physical skill. Given criticisms (Thompson, 1993) in terms of the threat to external validity of protocols that use sham deceptive performance scores, performance data on the sham test of physical skill was obtained from a population with similar characteristics as the participants in this thesis. The development of the sham physical skill test and the associated performance data that helped determine non-contingent failure and non-contingent success fulfilled the first research aim. To test the second research aim, I used an experimental manipulation that involved both types of performance feedback (non-contingent success and noncontingent failure) to test the two formulations for the antecedent of performance uncertainty. Moreover, participants were able to select two types of handicaps as mitigating excuses. The outcomes of Study 2 limited the remaining studies to one antecedent condition for performance uncertainty and one type of self-handicap as a mitigating excuse. In Study 3,1 investigated the onset of the use of selfreports as self-handicaps using young children (aged 10) and young people (aged 13). The results hmited the scope of the remaining studies to young people aged between 13 and 16. In the remaining studies (Studies 4, 5, & 6), I assessed the assumptions related to individual difference variables by having young people complete various self-report measures (self-esteem, trait self-handicapping, selfefficacy, and achievement orientation) and then randomly allocating them to a control or evaluative condition. Self-handicapping is an effective short-term strategy, but left unchecked it may lead people to be ineffectual in certain performance domains. Classrooms

Introduction 7 constitute an obvious potential source of non-contingent feedback and, inevitably, self-handicapping will occur despite the best efforts of teachers to create climates that minimise evaluative threat. This study, by examining the veracity of some assumptions for the climate and personal dispositions that cause young people to self-handicap in the physical domain will enable educators to continue to use or develop specific therapeutic and counselling guidelines to help young people deal with failure.

Literature 7 CHAPTER 2 Literature Snyder (1990) defined the term uncertain anticipatory set as the antecedent conditions that cause people to become so uncertain about the outcomes of their performances on evaluative tasks that they self-handicap. Two hypotheses outline the antecedent conditions for performance uncertainty. Both hold in common the proposition that uncertainty about performance outcomes in evaluative situations elicits self-handicapping. The non-contingent success hypothesis holds that performance uncertainty arises out of an inability to discern what behaviour will replicate previous success (Berglas & Jones, 1978). The self-esteem protection hypothesis holds that self-handicapping arises from concerns over the maintenance of the self-image (Snyder & Smith, 1982). The two formulations do refer to the same attribution strategy. The similarities are the prerequisite components of self-handicapping: the evaluative threat, the protagonist, the self-handicap, and the bad act. Obviously, there must be a protagonist who anticipates uncertainty about success in an upcoming evaluative performance. In the temporal progression of events, individuals are the authors of the bad acts, and such acts can be linked to sense of self Selfhandicaps operate to obscure and weaken the causal Hnkage between people's core sense of self and possible flawed performances. For example, the citing of injuries before events provides athletes with ready-made excuses for flawed performances. That is, the athletes can claim that the injuries hindered them from giving performances that reflected their true abilities. This proposition implies that self-handicapping strategy fits under the rubric of attribution theory. There are a variety of models associated with attribution theory and to review each is

Literature 8 beyond the scope of this thesis. In the next section, I have limited the overview of attribution theory to Heider's (1958) original concepts to give a background to attribution theory. Contained within this overview are comments on the contributions of Kelley (1971) and Weiner (1986), specifically in terms of their augmentation and discounting principles and the development of causal dimensions, respectively. What then follows is how self-handicapping fits into the overall framework of attribution theory. Attribution Theory and the Concept of Self-handicapping Attribution theory concerns the processes involved in the perception of causation and the consequences of such perception. There is not one but many attribution theories, and the term refers to several different kinds of problems. Heider's (1958) two concepts of personal perception and attributional processes are common to attribution theories. He suggested a useful fr

associated with higher level osf self-handicapping i n young people. Moreover, certainty of self-esteem and the trait of self-handicapping wer noe t associated with self-handicapping. Stud 6 explorey d the relationship between self-esteem and self-handicapping using domain-specific measure of self-esteems an, d task specific self-efficacy.