Servicios personalizados
Inglés (pdf)
Articulo en XML
Referencias del artículo
Traducción automática
Enviar articulo por email
Citado por SciELO 
Citado por Google 
Similares en
SciELO 
Similares en Google 
Permalink
Introduction
Due to its evaluative nature, the academic context can be threatening for many students. Considering that performing below certain achievement standards is an unequivocal symptom of inferiority and incompetence, the fear of failure leads some students to adopt complex strategies that allow them to redefine failure to protect their personal worth (De Castella, Byrne, & Covington, 2013).
Among the various self-protective strategies identified by motivational research, self-handicapping stands out for its strong prevalence in academic settings (Martin & Marsh, 2003). This anticipatory mechanism allows students to develop some excuse, real or imagined, that allows them to externalize the causes of a hypothetical failure, separating it from their personal competence (Midgley & Urdan, 2001). This strategy is functional in the short term by protecting self-esteem and alleviating anxiety (Berglas & Jones, 1978); however, it entails significant damage in the long run, notably affecting academic performance (Schwinger, Wirthwein, Lemmer, & Steinmayr, 2014).
The wide range of potential excuses that can be used for self-defeating purposes has motivated the distinction between behavioral self-handicapping and claimed self-handicapping (Leary & Shepperd, 1986). The first involves some type of direct action (e.g., reduce effort, procrastinate), whereas the second comprises the verbalization of some handicap (e.g., anxiety, fatigue) without necessarily implying an external behavior that compromises achievement (Hirt, Deppe, & Gordon, 1991). From this perspective, behavioral self-handicapping is more maladaptive than claimed self-handicapping, as some studies have demonstrated (Clarke & MacCann, 2016; Ferradás, Freire, Valle, & Núñez, 2016).
Despite its considerable long-term counterparts, the immediate benefits of self-handicapping make it quite difficult for students to be dissuaded from it (Snyder, Higgins, & Stucky, 1983). This has motivated a prolific investigation into the factors in the academic context that may be related to self-handicapping, with achievement goals and self-esteem being the most cited.
Self-handicapping and achievement goals
One of the most productive lines of work in the motivational field focuses on the students' reasons for becoming academically involved. The widely accepted 2x2 model of achievement goals (e.g., Korn & Elliot, 2016) postulates that students can adopt learning goals1 if their objective is to satisfy intrapersonal standards (develop interest, increase personal competence) and performance goals if their priority is comparison with others, either to excel and demonstrate superiority (performance-approach goals) or to avoid poor grades and negative social judgments (performance-avoidance goals).
Although these different types of achievement goals are not mutually exclusive, so that they are compatible in the same student (Inglés, Martínez-Monteagudo, García-Fernández, Valle, & Castejón, 2015), each has different academic implications. Thus, in general terms, the adoption of learning goals has been associated with highly adaptive academic motivation such as intrinsic interest, effort, persistence or deep processing of information (Harackiewicz, Durik, Barron, Linnenbrink-García, & Tauer, 2008; Senko, Hama, & Belmonte, 2013; Wolters, 2004). More controversy has arisen, however, over the performance goals. Thus, performance-approach goals seem to relate positively to effort, self-efficacy or academic performance (Hulleman, Schrager, Bodmann, & Harackiewicz, 2010; Liem, Lau, & Nie, 2008) but negatively with intrinsic interest (Harackiewicz, Barron, Carter, Lehto, & Elliot, 1997). For its part, the adoption of performance-avoidance goals is associated with a pattern of reduction of effort and persistence, high levels of anxiety and low academic achievement (Huang, 2011; Skaalvik, 1997; Van Yperen, Blaga, & Postmes, 2014). However, the concern about preserving one’s social image leads students with these types of goals to maintain a certain degree of connection with academic work (Midgley & Urdan, 1995). In addition to this taxonomy, there is evidence that some students lack interest in being involved in tasks (i.e., work-avoidance goals). This kind of motivation is more maladaptive than learning and performance goals because it entails a greater degree of behavioral disaffection with academic work as well as lower achievement and emotional well-being (Harackiewicz et al., 1997; King & McInerney, 2014).
Research on achievement goals and self-handicapping suggests that this strategy is more common among those students with performance goals. However, it is not clear whether the vulnerability to self-handicapping is greater in students with performance-avoidance goals (Akin, 2014; Ntoumanis, Thogersen-Ntoumani, & Smith, 2009), in those who adopt performance-approach goals (Valle et al., 2007), or in those who combine both performance tendencies (Ferradás, Freire, Núñez, Piñeiro, & Rosário, 2017; Midgley & Urdan, 2001). There are also discrepancies between the studies that distinguish between behavioral self-handicapping and claimed self-handicapping. For example, Lovejoy and Durik (2010) suggested that the positive relationship between performance-approach goals and self-handicapping involves only the behavioral type of self-handicapping, whereas Ferradás et al. (2016) argued that this relationship also encompasses claimed self-handicapping. Likewise, some studies positively relate the work-avoidance goals to self-handicapping (Miki & Yamauchi, 2005; (Valle et al., 2007), although this relation may only involve behavioral self-handicapping (Ferradás et al., 2016).
Regarding to learning goals, it appears that they are negatively related to self-handicapping (Akin, 2014; Schwinger et al., 2014), acting as a protective factor against these strategies even in students who also pursue performance goals (Ferradás et al., 2017; Schwinger & Stiensmeier-Pelster, 2011).
Self-handicapping and self-esteem
Self-esteem is a self-reported feeling that reflects the degree to which a person loves, respects and feels competent, valuable and satisfied with himself or herself (González & Tourón, 1992). In the specific case of self-handicapping, the investigations that relate these strategies to self-esteem engendered notable controversies. Because self-handicapping is more common among those individuals who are to some degree unsure of their personal competence (Berglas & Jones, 1978), some studies determined a positive relation between these strategies and low self-esteem (e.g., Chen, Sung, & Wan, 2017; Eronen, Nurmi, & Salmela-Aro, 1998). However, other studies indicated that self-handicapping is more frequent in people with high self-esteem (e.g., Kim, Lee, & Hong, 2012; Rappo, Alesi, & Pepi, 2017), to the extent that this strategy allows safeguarding the personal worth of the individual faced with a hypothetical failure.
The present study
According to the studies reviewed, there does not appear to be a single profile of self-handicappers because both students with low self-esteem and those with high self-esteem are vulnerable to these strategies. However, the studies conducted to date do not allow this question to be empirically determined because they have analyzed the relationship between self-esteem and self-handicapping by adopting a variable-centered approach (classifying the variable self-esteem at several levels). Consequently, the first objective of this study was the identification of differentiated profiles of students from the combination of self-esteem and self-handicapping using a person-centered approach. This approach is more reliable in the motivational reality of the students (Schwinger & Wild, 2012), in this case, of those eager to protect their personal worth by self-handicapping. Based on the reviewed studies that (from a variable-centered approach) have attempted to clarify the relationship between self-esteem and self-handicapping, we expected to find four profiles of self-handicapping students: (a) low self-esteem and high behavioral self-handicapping, (b) low self-esteem and high claimed self-handicapping, (c) high self-esteem and high behavioral self-handicapping and (d) high self-esteem and high claimed self-handicapping.
The second objective was to determine whether the identified profiles differ in their achievement goals. This will provide a deeper knowledge about the academic motivations underlying the different groups of self-handicappers observed. Based on the studies that relate self-handicapping to achievement goals, as well as on some studies (e.g., Chen et al., 2017; Gebka, 2014) that negatively relate self-esteem to performance-avoidance goals and positively relate self-esteem to learning and performance-approach goals, our hypotheses were as follows: (a) the two profiles of high self-esteem and high self-handicapping will demonstrate a stronger positive relationship with the performance-approach goals, (b) the two profiles of high self-handicapping and low self-esteem will demonstrate a greater positive relationship with the performance-avoidance goals, (c) the profile that combines low self-esteem and high behavioral self-handicapping will be the most prone to the work-avoidance goals, and (d) the four self-handicapping profiles will be negatively related to the learning goals, there being no significant differences among them.
In this work, we have sought to statistically control for the effect of gender and the course. There is abundant evidence that men resort more than women to behavioral self-handicapping strategies (McCrea, Hirt & Milner, 2008). Likewise, it appears that both types of self-handicapping are used more frequently as university students advance in their studies (Ferradás et al., 2015).
Method
Participants
The study was conducted with students from a Spanish university. Initially, 1087 (6.3%) of the 17,227 students studying at this university were selected through convenience sampling. The initial review of the data matrix indicated that some students did not respond to a sufficient number of items in the questionnaires and were thus eliminated from the study (56 cases, 5.15%). The cases that presented a data loss of less than 20% (27 cases, 2.62%) were replaced using the FIML (full information maximum likelihood) imputation method using the MPlus 7.11 program (Muthén and Muthén, 1998-2012). Three other cases were also eliminated for presenting outliers (Mahalanobis distance method) when exceeding the critical value χ² = 5.5 (gl = 7, p < .001). Consequently, 1028 students (M age = 21.36; SD age = 3.81) finally participated in the study. Of these, 69.9% were studying Health Sciences (Nursing, Physiotherapy and Podiatry), and 31.1% were studying Educational Sciences (Early Childhood Education, Primary Education, Social Education and Speech Therapy). Of the students, 37.2% were enrolled in first-year courses, 32.5% in second-year courses, and 30.3% in third-year courses. Regarding gender, 86.3% of the total participants were women.
Instruments
Self-Handicapping Scale
Self-handicapping was evaluated using the Spanish adaptation (Ferradás et al., 2016) of the Self-Handicapping Scale (Martin, 1998). The instrument considers two types of self-handicapping: behavioral self-handicapping (e.g., “I tend not to try the tasks so I have an excuse if I do not do as well as I expected”) and claimed self-handicapping (e.g., “I tell others that I am more exhausted than I really am when I have to do homework or exams, so if I do not do as well as I expected, I can say that's the reason”). The responses of the participants were recorded on a Likert scale (1 = never - 5 = always). The internal consistency of the scale was adequate in this study: behavioral self-handicapping (α = .84) and claimed self-handicapping (α = .90).
Rosenberg Self-Esteem Scale
The Spanish-adapted (Martín-Albo, Núñez, Navarro, & Grijalvo, 2007) version of the Rosenberg Self-Esteem Scale (Rosenberg, 1965) was used. This instrument (10 items, e.g., “In general, I am satisfied with myself”) demonstrated adequate reliability in our study (α = .88). Student responses were measured using a Likert scale (1 = in total disagreement - 5 = totally agree).
Goal Orientation Scale
The achievement goals were analyzed using the Goal Orientation Scale (Skaalvik, 1997) in its Spanish adaptation (Jover, Navas, & Holgado, 2014). This scale includes four types of goals: learning (six items, e.g., “It is important for me to learn new things in class”, α = .79), performance-approach (five items, e.g., “I try to get better grades than my classmates”, α = .85), performance-avoidance (six items, e.g., ”When I answer incorrectly in class, what worries me the most is what my classmates think of me”, α = .80), and work-avoidance (four items, e.g., “In class, I prefer to do as little as possible”, α = .76). The responses of the participants were recorded on a Likert scale (1 = never - 5 = always).
Design and procedure
In this study, a descriptive and comparative cross-sectional design was conducted. Regarding the procedure, prior to the data collection, the corresponding permits were requested from the university departments. The application of the instruments was conducted in the classrooms in which the students developed their lessons within the academic schedule and in a single session with no time limit. At the beginning of the session, participants were informed of the purposes of the study, the voluntary nature of their participation and the anonymity and confidentiality of the information collected.
Data analysis
To obtain the latent categorical variables (profiles) that allowed the participants to group according to their characteristics of self-handicapping and self-esteem, a Latent Profile Analysis -LPA- was conducted (Lanza, Flaherty, & Collins, 2003). The statistical program Mplus 7.11 (Muthén & Muthén, 1998-2012) was used to consider, from among the set of finite models, which model best fit the data, adding successive latent classes to the target model. The optimal number of classes was determined considering the formal test of the adjusted maximum likelihood ratio of Lo, Mendell, and Rubin (2001) -LMRT-, the Akaike information criterion (AIC), the Schwarz Bayesian information criterion (BIC), the BIC adjusted for the sample size (SSA-BIC) and the value of the entropy.
The p value associated with the LMRT indicates whether the solution with more (p < 0.05) or fewer classes (p > 0.05) best fit the data. The AIC, BIC and SSA-BIC indices have a descriptive character, with the lowest values indicating a better fit of the model. These criteria should complement the information provided by the LMRT but in no case should replace it, the latter being the final arbiter. Likewise, classes that contained less than 1% of the sample were considered spurious, a condition indicative of an excessive extraction of profiles (Hipp & Bauer, 2006). Finally, to evaluate the suitability of the potentially selectable model, using a MANOVA, the differences between the classes were analyzed with regard to the variables that were used to form them (self-esteem, behavioral self-handicapping and claimed self-handicapping). The effect size of these differences was determined by Cohen's d.
The relationship between the profiles of self-esteem and self-handicapping (latent classes, taken as independent variables) and achievement goals (dependent variables) was analyzed by means of a MANCOVA. As covariates, the gender and the course were incorporated into the model. The effect size was determined by the partial eta squared and Cohen’s d: small effect, between ηp 2 = 0.01 and ηp 2 = 0.059 (d = 0.20); medium, between ηp 2 = 0.059 and ηp 2 = 0.138 (d = 0.50); and large, when ηp 2 ≥ 0.138 (d = 0.80).
Results
Table 1 provides the descriptive statistics of the variables and the Pearson correlations. The asymmetry and kurtosis of the variables indicate that all of them presented normal distributions (see Finney & DiStefano, 2006).
Table 1: Matrix of correlations and descriptive statistics of the variables.
Note. BSH = Behavioral self-handicapping; CSH = Claimed self-handicapping; SE = Self-esteem; LG = Learning Goals; PApG = Performance-Approach Goals; PAvG = Performance-Avoidance Goals; WAG = Work-Avoidance Goals;*p <.01;**p <.001
Self-handicapping and self-esteem profiles
Several models of latent profiles have been adjusted to the data (models from two to six classes). Model fit stopped at six classes because a non-significant LMRT was obtained (LMRT = 387.103; p > .05). In addition, the LMRT indicated that the five-class model provided a better fit than the four-class model (LMRT = 838.482, p < .001; AICM4 > AICM5; BICM4 > BICM5; SSA-BICM4 > SSA-BICM5). The five-class model contains no class with a sample size of less than 1% and has higher entropy than the four-class model (entropy = .989).
Table 2 reports the accuracy of the classification in each class as well as the number of students in each of the five classes of the chosen model, both in absolute (n) and relative (%) terms. Two classes contained the majority of participants: Class 1 (54.67%) and Class 4 (32.88%). The other three classes contained less than 10%. The accuracy with which the subjects were classified within the classes (see diagonal of Table 2) was quite high, with values of 100% (Class 1) or quite close to this percentage (Classes 2 to 5).
Table 2: Characterization of the latent classes and accuracy of the classification of the individuals in each class.
Finally, statistically significant differences that were large in magnitude were observed among the five classes in the three variables that composed the profiles: self-esteem (F(4,1023) = 59.206; p < .001; ηp2 = 0.188), behavioral self-handicapping (F(4,1023) = 1751.689; p < .001; ηp2 = 0.873); and claimed self-handicapping (F(4,1023) = 2490.892; p < .001; ηp2 = 0.907).
The mean performances of the subjects (direct and standardized scores) belonging to the latent classes of the chosen model are presented in Table 3. Figure 1 presents the graphic representation of these profiles, using the standardized scores as a reference.
Table 3: Means, standard errors and confidence intervals of the solution of latent classes.
Note. Class 1 (moderate self-esteem/low self-handicapping); Class 2 (low self-esteem/high self-handicapping); Class 3 (low self-esteem/high claimed self-handicapping); Class 4 (moderate self-esteem/moderate selfhandicapping); Class 5 (low self-esteem/high behavioral self-handicapping); SE = Self-esteem; BSH = Behavioral self-handicapping; CSH = Claimed self-handicapping. Parentheses indicate standardized mean scores (z).
Note. MSE/LSH = moderate self-esteem/low self-handicapping; LSE/HSH = low self-esteem/high self-handicapping; LSE/HCSH = low self-esteem/high claimed self-handicapping; MSE/MSH = moderate self-esteem/moderate self-handicapping; LSE/HBSH = low self-esteem/high behavioral self-handicapping.
Figure 1: Graphic representation of self-esteem and self-handicapping profiles (Z scores).
In general, the two most quantitatively represented profiles (Classes 1 and 4) demonstrate moderately high self-esteem, differing from each other in self-handicapping levels. Thus, the subjects of Class 1 have low levels of behavioral and claimed self-handicapping (MSE/LSH profile), whereas the participants in Class 4 demonstrate moderate scores in both types of self-handicapping (MSE/MSH profile). The remaining three classes demonstrated low levels of self-esteem combined with a high use of self-handicapping, either behavioral only (Class 5: LSE/HBSH profile), claimed (Class 3: LSE/HCSH profile) or a combination of both (Class 2: LSE/HSH profile).
Relationship between self-esteem and self-handicapping profiles and achievement goals
With regard to the differences among the profiles in the achievement goals, once the effects of gender and the course were controlled for, MANCOVA results indicated the existence of statistically significant differences (λWilks = .502, F(4,1023) = 49.127, p < .001, ηp2 = 0.158) with a large effect size. The effect of the two covariates was also statistically significant: gender (F(4,1027) = 2.529, p = 0.039, ηp2 = 0.010) and course (F(4,1027) = 3.035, p = .017, ηp2 = 0.012).
Statistically significant differences were also observed between the profiles of self-esteem and self-handicapping in the four achievement goals (see Table 4): learning (F(4,1023) = 124.412, p < .001, ηp2 = 0.328), performance-approach (F(4,1023) = 18.723, p < .001, ηp2 = 0.068), performance-avoidance (F(4,1023) = 41.634, p < .001, ηp2 = 0.140), and work-avoidance (F(4,1023) = 15.762, p < .001, ηp2 = 0.058). The effect size was large in learning goals and performance-avoidance goals and medium in performance-approach goals and work-avoidance goals.
Table 4: Averages and standard deviations obtained for the profiles in the four achievement goals, together with the univariate tests for each of them.
Note. MSE/LSH = moderate self-esteem/low self-handicapping; LSE/HSH = low self-esteem/high self-handicapping; LSE/HCSH = low self-esteem/high claimed self-handicapping; MSE/MSH = moderate self-esteem/moderate self-handicapping; LSE/HBSH = low self-esteem/high behavioral selfhandicapping; Effect size (ηp2): small effect (< 0.059), medium effect (≥ 0.059 and < 0.138), large effect (≥ 0.138).
Table 5 presents the effect sizes of the differences among the five profiles in the four goals. In the learning goals, the post hoc contrasts revealed that the two groups that had moderately high self-esteem (MSE/LSH and MSE/MSH) adopted this goal significantly more often than did the other profiles. The effect size is large in all cases (d ≥ 1.99). In both performance goals, the profiles with low self-esteem and high self-handicapping (LSE/HBSH, LSE/HCSH, and LSE/HSH) scored significantly higher than profiles with moderate self-esteem and low/medium self-handicapping (MSE/LSH) and MSE/MSH). The differences between groups oscillated between medium and large in the performance-approach goals (d = 0.68 - d = 0.99), whereas in the performance-avoidance goals, these differences were large in all comparisons (d = 1.12 - d = 1.30).
Regarding work-avoidance goals, post hoc contrasts indicated that profiles with high behavioral self-handicapping (LSE/HBSH and LSE/HSH) scored significantly higher in this goal. In the LSE/HBSH profile, the differences between groups were statistically significant in all cases and large in magnitude (d ≥ 0.84). In the LSE/HSH profile, the differences were only statistically significant with regard to the LSE/HCSH group (large effect size: d = 1.33). In fact, the LSE/HCSH profile obtained the lowest scores in this achievement goal, differing widely (d ≥ 0.98) from the other four profiles.
Table 5: Effect size (Cohen's d) of the differences among classes in the four achievement goals.
Note. Class 1 (moderate self-esteem/low self-handicapping); Class 2 (low self-esteem/high self-handicapping); Class 3 (low self-esteem/high claimed selfhandicapping); Class 4 (moderate self-esteem/moderate self-handicapping); Class 5 (low self-esteem/high behavioral self-handicapping). Effect size (Cohen's d): small effect (< 0.50), medium effect (≥ 0.50 and < 0.80), large effect (≥ 0.80).
Discussion and conclusions
One of the main contributions of this study is the adoption of a person-centered approach to analyze the relationship between self-esteem and self-handicapping, an unpublished approach to date. The results obtained identified four profiles of self-handicappers: three with low self-esteem and one with moderately high self-esteem. The latter, the most numerous (nearly 33% of the students), was characterized by a moderate use of self-handicapping (behavioral and claimed). Regarding self-handicap profiles with low self-esteem, we identified a small percentage of students (just over 3%) with low self-esteem and high use of behavioral self-handicapping, another similar percentage of students with low self-esteem and high claimed self-handicapping, and a somewhat larger number of students (nearly 6%) with low self-esteem and a high utilization of both types of self-handicapping. To these four profiles was added a fifth, composing more than half of the sample, which presents moderate self-esteem and a scarce use of self-handicapping.
These results appear to support those obtained by other works that, by adopting a variable-centered approach, link self-handicapping with both high self-esteem (Kim et al., 2012; Rappo et al., 2017) and low self-esteem (Chen et al., 2017; Eronen et al., 1998). Our findings also deepen the understanding of the relationship between self-esteem and self-handicapping, to the extent that they suggest that the use of these strategies (whether behavioral, claimed, or both) is greater in profiles of students with low self-esteem, whereas in those with high self-esteem (at least, moderate), the use of self-handicapping is more modest.
However, our results did not allow us to determine the reasons that led students with low self-esteem and moderately high self-esteem to use self-handicapping, nor did these results explain why some students with moderately high self-esteem were self-handicappers (MSE/MSH profile) and others were not (MSE/LSH profile). Some studies (e.g., Tice, 1991) suggested that people with low self-esteem use self-handicapping to protect their self-esteem, whereas those with high self-esteem use self-handicapping to enhance self-esteem. Although the instruments applied in this work were not able to determine it, it cannot be ruled out that, unlike the students of the MSE/LSH profile, the self-esteem of the students of the MSE/MSH group is characterized by its instability (i.e., high at times and low at others), a characteristic that accentuates the tendency to self-handicapping (Newman & Wadas, 1997). These tentative explanations should be analyzed with more rigor in future studies.
In addition, the results of this study indicate that the profiles identified differ in the achievement goals adopted. It appears that the performance goals are the primary leitmotif of the three student profiles that most use self-handicapping (LSE/HSH, LSE/HBSH, and LSE/HCSH). Thus, these students, who are insecure about themselves (i.e., low self-esteem), face academic tasks with the double desire to demonstrate competence (performance-approach) and hide their limitations (performance-avoidance), using self-handicapping (behavioral, claimed or both) as a strategic alibi. This finding is consistent with the findings of other studies that, adopting a variable-centered approach, relate self-handicapping to both performance tendencies (Ferradás et al., 2017; Midgley & Urdan, 2001) rather than only one (e.g., Akin, 2014; Valle et al., 2007).
Conversely, the two profiles that demonstrate higher self-esteem and less self-handicapping (MSE/LSH and MSE/MSH) represent the students who are the most inclined to adopt learning goals. This finding offers a double interpretation. On the one hand, consistent with previous research (e.g., Chen et al., 2017; Gebka, 2014), it appears that self-esteem, even at moderately high levels, is positively related to an interest in learning and vice versa; this interest decreases when self-esteem is low. Conversely, the moderate use of self-handicapping does not appear to diminish the students’ interest in adopting learning goals. However, the high use of self-handicapping (profiles LSE/HSH, LSE/HBSH, and LSE/HCSH) leads to a significant decrease in the desire to learn, a fact that confirms the negative relationship between self-handicapping and learning goals evidenced by other works (Akin, 2014; Schwinger et al., 2014).
Finally, along the lines suggested by Ferradás et al. (2016), it is necessary to consider that the two profiles that stand out for the high utilization of behavioral self-handicapping (LSE/HBSH and LSE/HSH) are those that significantly adopt work-avoidance goals. These data confirm the importance of differentiating between claimed self-handicapping and behavioral self-handicapping (e.g., Clarke & MacCann, 2016) because the high use of the latter (either exclusively or in combination with the claimed self-handicapping) entails more dysfunctional achievement motivations in the academic context (King & McInerney, 2014).
Overall, the results obtained contribute to the understanding of academic self-handicapping. Our study allows the identification of unpublished profiles of self-handicapping students with low and moderately high self-esteem and their respective achievement goals.
These findings have important psychoeducational implications. Although self-handicapping appears to yield substantial short-term benefits for the student, its habitual use entails significant academic losses (Schwinger et al., 2014). Therefore, it is necessary to promote learning contexts that reduce the need to adopt these strategies. Assuming that anxiety and low perception of control increase vulnerability to self-handicapping (Cano, Martin, Ginns, & Berbén, 2017; Stewart & De George-Walker, 2014), teachers can reduce the incidence of these factors by providing clear, contingent and consistent feedback on students’ academic performance (Martin, Nejad, Colmar, Liem, & Collie, 2015). It would also be wise for teachers to promote learning environments that encourage students to trust in their own abilities and focus on personal improvement objectives, which would improve their self-determined motivation (Méndez-Giménez, Cecchini, Méndez-Alonso, Prieto, & Fernández-Río, 2018) and their academic involvement (Miñano, Gilar, & Castejón, 2012).
The contributions of this work, however, should be considered in light of the limitations of the study conducted and for its implications for future research. First, the correlational and cross-sectional nature of the proposed design does not allow causality relations to be inferred among the variables of the study. Therefore, it would be interesting if future studies adopted experimental designs that would allow the directionality of this relationship to be clarified. Second, the analyzed sample only includes students in educational and health fields, which restricts the generalizing of the results to the entire university population. Third, with respect to gender, there is a clear preponderance of female students (86.3%) among the participants of the present study. Although we have attempted to statistically control for the possible effect of gender by adopting it as a covariate, our results indicate that this variable had a statistically significant effect (medium effect size). Thus, another possible line of future research should be directed toward the study of the differences between women and men in the relationship between the profiles of self-handicapping and self-esteem and the adoption of achievement goals. This issue may be particularly important if one considers the solid evidence supporting gender differences in the use of self-handicapping (e.g., Ferradás et al., 2016; McCrea et al., 2008). Fourth, the data were collected using self-report instruments. Assuming the possibility that some students were unaware that they used self-handicapping strategies or were not willing to acknowledge their use (Martin, 2010), the results obtained from these instruments may be partially biased. Therefore, a more rigorous characterization of the motivational reality of the participants requires complementary evaluation tests such as in-depth interviews or classroom observation. This would also allow, for example, the analysis of differences in the use of self-handicapping depending on the task or subject, a relevant issue considering the high situational specificity of this strategy (Schwinger, 2013). Fifth, this study did not examine other achievement goals that are not strictly academic, such as the students’ desire to adapt socially and integrate with their peers (Esteban, Bernardo, Tuero, Cerezo, & Núñez, 2016). This complementarity between academic and extra-curricular goals would allow a more precise definition of student achievement behaviors (Wentzel, 2001), in this case, of the different profiles of self-handicappers with low and moderately high self-esteem identified in this work.
Finally, in the same manner in which the achievement goals considered in this study are not the only ones that can be adopted by students, self-handicapping is not the only strategy used for self-protection purposes (Martin, 2010). Therefore, it is possible that the profiles of identified self-handicappers may combine these self-defensive strategies with others such as defensive pessimism or helplessness, which could expand the spectrum of student profiles that, regardless of their self-esteem level, prioritize the protection of their personal worth. Future research should consider this question.
