Insight’s Insecurities: The Impact of Self-Esteem on Insight Problem Solving

Insight’s Insecurities: The Impact of Self-Esteem on Insight Problem Solving

Abstract
Varying levels of self-esteem (SE) foster the ability to think about problems in different ways but little is known about how this affects insight problems. This paper investigates the hypothesis that individuals experiencing high SE will solve more insight problems correctly than those experiencing low SE. It was also expected that manipulating contingency would mediate the relationship between SE and problem solving. To investigate this, participants in four conditions completed a dingbat problem-solving task. The results of the study support the hypothesis that high SE persons would solve more dingbat problems correctly than low SE persons. This is consistent with the idea that high SE individuals are more likely to abandon the mental set whereas low SE individuals will rely on their mental set until they receive evidence that it may be problematic. The interaction hypothesis was not supported suggesting that contingency does not mediate the relationship between SE and insight problem solving.

Introduction
Variations in self-esteem (SE) have long been the focus of studies relating to problem solving. SE is literally defined by how much value people place on themselves. High SE refers to a highly favourable, global evaluation of the self whereas low SE, by definition refers to an unfavourable self-evaluation. SE does not require accuracy in regards to self-evaluation (Baumeister, Campbell, Krueger & Vohs, 2003). Recent research has indicated that when defining SE it may be more accurate to differentiate between global SE and SE contingencies. SE contingencies represent the domain to which each individual’s global SE is most reactive (Crocker & Wolfe, 2001). This suggests that the exploration of the relationship between SE and behaviour may need to focus on global SE as well as SE contingent on the specific behaviour being evaluated.
Problem solving can be defined as a mental process and is part of the larger problem process that includes problem finding and representation of the problem based on what seems to be the crucial aspects. Many different types of problem solving have been identified by literature, such as heuristic and algorithmic problem solving. However, one particular form of problem solving that still remains under-explored with regards to SE is insight problem solving. Insight problem solving, the focus of this study, can be defined as the ‘eureka’ moment (can occur with or without incubation) whereby the solution of a problem instantly becomes clear. Although much research has been done in the area of SE, there remains an uncertainty whether high or low SE is more adaptive to problem solving. The goal of this study is to thoroughly explore these issues and determine a more exacting link between SE and insight-related problem solving.

SE and problem solving has been the focus of a substantial amount of research. SE has been studied in connection with social problem solving. Davila, Hammen, Burge, Daley, & Paley (1996) found that women with low global self-worth and insecure attachment style had poorer social problem solving skills. Additionally, previous research has linked cognitive processes such as heuristic and algorithmic problem solving with SE. In particular, low SE was found to be associated with negative problem orientation, as well as impulsive/careless and avoidant orientation (D’Zurilla, Chang & Sanna, 2003). The most significant association was with negative problem orientation, defined as, doubting one’s ability to solve and use solution strategies (D’Zurilla et al. 2003). In a similar vein, Tang & Reynolds (2006) found that for participants with low SE, perceived task difficulty adversely affected certainty, perceived ability and task satisfaction. This was not the case for participants with high SE. With regards to high SE, Hoffman & Spatariu (2007) found that self-efficacy and meta-cognitive prompting influenced better mathematic problem-solving performance suggesting that strategy knowledge improved problem solving. Similarly, Whitesell, Mitchell & Spicer (2009) indicated that factors associated with SE played an important role in academic success. The above illustrates the link research has found between SE and various kinds of problem solving, however, it is as yet unclear whether SE has a similar effect on insight problems.

One aspect of SE and problem solving that may be important for insight was identified by Weiss & Knight (2004). Participants were requested to search for the commonality between three numbers. It was found that low SE participants were more likely to search for longer and for more information. Since searching for more information was functional for this task, low SE participants performed better (Weiss & Knight, 2004). In other words, adherence to the original mental set resulted in better performance. A mental set can be defined as the inability to think about an event in a different way (Öllinger, Jones, & Knoblich, 2008). When solving insight problems it would be counter-productive to adhere to the mental set as the solution is intuitive and therefore requires looking at the problem from a new perspective. Gasper (2003) suggests that sticking to a mental set is un-adaptive to problem solving. Similarly, Jaquish & Ripple (1981) found that high SE participants irrespective of age were more likely to form creative ideas based on the examination of all possible solutions than those with low SE. The research explored above suggests that those with high SE will perform better on insight problems because they are better able to break out of an inadequate mental set and form creative solutions to a novel problem.

The relationship between SE and insight problem solving may change when SE contingency is made relevant to the task. SE contingency is the area or domain within the individual’s life to which his global SE is most likely to react. For example, if an individual’s self-evaluation is dependent on his cognitive ability, then cognitive ability is one of his SE contingencies; his SE will fluctuate when he is successful on a cognitive task and decrease when he fails at a similar task. Support for the existence of contingent SE was provided by Park & Crocker (2008) who found that when SE is contingent on others’ approval, negative feedback resulted in lower state SE and greater negative affect than those whose SE was less contingent on the approval of others. Those participants with low SE contingent on others’ approval made greater effort to appear attractive after receiving negative feedback. In contrast, participants with high SE that received negative feedback indicated an increased wish to seem warm and caring the more they founded their SE on the approval of others. Evidence for the contribution of performance contingent SE as a better predictor for job performance than global SE was provided by Ferris, Brown, Pang and Keeping (2010). In particular, a stronger positive correlation between SE and job performance was found for individuals whose SE was not contingent on performance compared to participants whose SE was highly contingent on performance. Those participants whose SE was highly contingent on performance were motivated to perform better because any failure in the domain of performance would have negative consequences on their self-worth, regardless of their level of global SE, high or low. However, for those whose global SE was not contingent on performance, bad performance would be of no particular threat to their self-worth and they would therefore be able to perform in a way that was consistent with their SE level. Those with low SE performed significantly worse than those with high SE who then performed almost as well as those with high performance-contingent SE.

Bearing in mind the relationship between SE and creative thinking, the goal of the current study is to infer a relationship between SE and insight problem solving. To this end, we will examine the effect that levels of SE have on insight problem solving. Based on previous findings, we expect that participants experiencing high SE will solve more insight problems within a predetermined time frame, compared to participants experiencing low SE. Furthermore, we expect that manipulating contingency will moderate the relationship between SE and problem solving. This prediction is derived from research into relevance of emotion to prospective behaviour. Tuan Pham (1998) found that affective evaluation played a key role in consummatory judgements if the emotion was evaluated as being contingent and representative of the decisions at hand. If the emotion could be attributed to some other external cause and therefore not contingent on the target, the participant’s current emotion did not have an effect on the decision making process. Similarly, Berkowitz and Troccoli (1990) revealed that negative affect results in a negative evaluation of the subject when participants do not pay attention to the source of their discomfort, thereby attributing their negative feeling to the subject. However, those participants that engaged in an evaluation of the source of their current affectual state were less likely to judge the subject negatively. Research on the experience of emotion contingent on the decision-making target provides a theoretical basis for the notion that SE in the domain of insight will affect participants’ ability to solve insight problems more so than global SE alone.

In order to test for these effects, SE and SE contingency on insight problem solving, a 4 groups 2 x 2 design was applied, with SE and SE contingency as manipulated variables. In group 1 (high SE/Contingent), participants will be asked to write about a positive creative problem solving related experience; group 2 (high SE/non-Contingent), will write about a great compliment they received; group 3 (low SE/Contingent) will write about a negative creative problem solving experience; group 4 (low SE/non-Contingent) will write about a hurtful criticism they received. Participants in all conditions will then have to complete a series of insight related dingbat problems. We predict that high SE will increase performance in insight problem solving, so that participants in the high SE condition will solve more dingbat problems correctly within the given time limit than those in the low SE condition. Since both high global SE and contingency are important to insight we predict that there will be a SE X contingency interaction: participants manipulated to experience SE contingent on insight performance will solve more insight problems regardless level of SE, whereas those with non-contingent SE will perform in a way consistent with their level of global SE; those with low SE will perform worse than those with high SE.

Method
Participants: 60 participants (38 women and 22 men ranging in age from 18 to 53, median – 22) volunteered to participate in the study. All of the participants provided informed consent.
Materials: SE manipulation: In the present study a SE manipulation measure was adapted from Mikulincer and Sheffi (2000) who induced participants’ affect. In the original study participants were asked to remember and write down an emotionally positive/negative/neutral experience. This manipulation of emotion was found to be effective. The present study carried out a SE manipulation using similar categories. In all conditions, participants were asked to recollect an event, visualise it and then provide a written description of the event. Since insight and creativity are of similar constructs in that they both require a break in original mental set, the contingency manipulation will require recall of a creative experience. The manipulations per condition are as follows: those participants in; 1) High SE / Contingent wrote about a significant creative success. 2) SE / non-Contingent wrote about a meaningful compliment they had received. 3) Low SE / Contingent condition wrote about a creative failure, 4) SE/ non-Contingent condition recalled an experience of hurtful criticism. All participants in the four conditions completed the task successfully. Following the recall task, participants were asked to rate their SE using the Rosenberg Scale (Rosenberg, 1965) (Appendix 1). The scale contains ten questions to which participants can strongly agree, agree, disagree and strongly disagree. Scores are calculated for items 1, 2, 4, 6 and 7 with strongly agree equalling 3 points, agree 2 points, disagree 1 point and strongly disagree 0 points. Items 3, 5, 8, 9 and 10 are reversed in valence. The scale ranges from 0-30. Scores between 15 and 25 are within normal range; scores below 15 suggest low SE.
Insight: Insight problem solving was evaluated using a test comprised of 15 dingbat problems. Dingbats are picture puzzles in which the images are symbolic and need to be organized and interpreted to obtain a logical structure. We define dingbat problems as insight because the viewer is required to break the accustomed mental set of the image or word presented, in order to see the problem from a novel perspective. The dingbats that were presented had varying difficulty levels. The dingbats were ordered randomly in three different fixed presentations and each participant was assigned to one of the orders randomly. The dingbats were presented using Microsoft PowerPoint 98' on a 14” color monitor with a 600 x 800 resolution at a 60 cm distance. Each slide contained one dingbat for 15 seconds followed by a blank slide for 5 seconds during which time participants were required to write the answer on a sheet of paper that was previously provided by the experimenter. The answer sheet was numbered 1 – 15, with space left for a handwritten answer. This way, every participant spent the same amount of time on each problem because the slides continued automatically. The amount of correct answers was measured in order to establish the effectiveness of the participant in solving dingbat insight problems. The average amount of correct answers attained by participants in each SE condition was then compared.
Procedure: On agreement to participate in the study, volunteers were randomly assigned to one of the four conditions (15 per condition). They were given an explanation and instructions for the experimental task after which they provided informed consent. Participants were then given the SE induction task and the Rosenberg SE measure to complete. When this initial stage of the experiment was concluded, participants were given instructions to complete the dingbats task. It was explained that a total of 20 seconds would be given to solve each dingbat problem (15 seconds for presentation of each dingbat followed by blank screen for 5 seconds) on the computer screen in front of them after which another dingbat would appear until all 15 were shown. Participants were requested to write their answers on the paper provided; the paper was numbered 1 – 15 with space left for the hand written answer. Following the completion of the experiment participants were debriefed regarding the goals of the study.

Results
In order to explore the study’s hypotheses, a 2 x 2 two-way ANOVA was applied to the data, with total correct dingbat solutions as dependent variable and manipulated SE and Contingency as independent variables.

Manipulation check: In order to check the effectiveness of SE manipulation, an independent samples t-test was used to compare the SE of participants manipulated to experience high SE with that of participants manipulated to experience low SE. The analysis showed that participants in the high SE condition reported higher scores on the Rosenberg SE scale (M = 18.37, SD = 5.37) compared to those in the low SE condition (M = 21.3, SD = 4.94), t (58) = -2.2, p < .05.

Demographic variables: Gender: In order to test whether there was a significant difference in gender distribution between the conditions a chi square analysis was performed. No significant difference between the groups was found, ²(3) = .861, N.S.

Age: In order to examine whether there were age differences between experimental groups a two way ANOVA was performed. The analysis revealed a significant main effect for SE, F(1,56) = 4.528, p < .05, partial η² = .075. The analysis revealed a near significant effect for Contingency, F(1,56) = 3.031, p = .09, partial η²=.051. These results indicate that exploring the research hypotheses may need to control for age related effects.

Education: In order to examine whether there were education differences between experimental groups a two way ANOVA was performed. The analysis revealed a near significant main effect for SE, F(1,56) = 2.899, p = .09, partial η² = .049. These results indicate that exploring the research hypothesis may need to control for education related effects.

Hypothesis testing: In order to examine the effect of SE and SE contingency on dingbat solution performance, a 2 (SE: high vs. low) x 2 (Contingent, non-Contingent) two-way ANOVA was conducted. The analysis did not reveal a significant main effect for SE, F(1,58) = 1.96, N.S, partial η² = .034. However, when controlling for the effects of education, the analysis revealed a near significant result, F(1,55) = 3.033, p = .09, partial η² = .052. Participants in the high SE condition (M = 8.133, SD = 3.181) solved significantly more dingbat problems than those in the low SE condition (M = 6.933, SD = 3.383). Controlling for age had no effect on the results of the original analysis. The analysis did not reveal a significant main effect for Contingency, F(1,58) = .604, N.S. partial η² = .011. These results remained when controlling for either education or age. In addition, the analysis did not reveal a significant SE x Contingency interaction, F(1,58) = .024, N.S, partial η² = .000. These results remained when controlling for either education or age.

Discussion
In order to examine the relationship between SE and Contingency on insight problem solving participants were randomly assigned to one of four SE conditions (high SE/Contingent, low SE/Contingent, high SE/non-Contingent, low SE/non-Contingent). Participants were then required to complete a series of 15 insight problems. The results of the study support the hypothesis that those participants manipulated to experience high SE would solve more insight problems than those experiencing low SE. Participants in the high SE condition solved more dingbat problems correctly than participants in the low SE condition. However, the interaction hypothesis was not supported. Contingent SE in conjunction with global SE did not have an effect on insight problem solving ability.

The demonstration of the importance of SE for insight problem solving has important implications for established theories of cognitive problem solving. Past research has ascertained that SE can influence the ability to solve social and cognitive problems in the domain of algorithmic and heuristic problem solving. In such cases high SE positively influenced the ability to solve problems and low SE negatively influenced problem solving ability. The present study suggests that there may be a similar link between SE and insight problem solving. We therefore suggest that insight problem solving may require similar cognitive processes to algorithmic, heuristic and social problem solving. Consequently, rather than there being a number of cognitive processing systems there may be one global system with many subdivisions. Support for this theory requires further research.

While an effect for SE was found, the data suggests that education acts as a confounding variable in the relationship between SE and insight. Highly educated individuals have more knowledge about the world than those with less education, therefore one might think that a highly educated individual is more capable of solving insight problems because he is better able to restructure his mental set to account for more relevant knowledge stored in memory than someone with less education. However, a highly educated individual with low SE may be less capable of flexibility and therefore may not be able take full advantage of his large knowledge store, when required to restructure the mental set of an insight problem. On the other hand, individuals with less education have a smaller knowledge base than highly educated individuals. Consequently, the ability to be flexible which is necessary for insight problem solving (DeYoung, Flanders & Peterson, 2008) facilitated by SE level (Weiss & Knight, 2004) may not be helpful since the relevant knowledge is not stored in memory. Additional research could aim to clarify the interaction between education and SE and their combined influence on insight problem solving ability. Additional research may also uncover if the interplay of SE and education has a significant effect on other kinds of problem solving.

The failure to support the interaction hypothesis could have significant implications for theories of contingency. Despite the fact that contingency mediates the relationship between SE and certain kinds of problem solving, it could be that it does not play a role in the relationship between SE and insight. An alternative way to view the failure to confirm the interaction hypothesis could be an issue of construct validity. It was theorised that participants with no psychological background would not fully understand the term insight. Therefore, creativity was used as an operational variable for insight. This is because creativity, flexibility and the ability to abandon irrelevant mental sets are often theorised as necessary skills for insight problem solving (DeYoung, Flanders & Peterson, 2008). However, if creativity is in fact a completely independent process from insight, it would follow that priming for creativity contingent SE would have no effect on the ability to solve insight problems. Additional research could aim to operationalise insight contingent SE in a different way. If additional operationalisations of insight replicate the null results obtained in the present study, the role of contingent SE on behaviour and problem solving may have to be reconsidered. It is possible that contingent SE is relevant only to behaviours that are directly relevant to self-identity. In other words, contingent SE may be highly individualised and only effect efficacy in domains that are uniquely important to the individual’s self-worth. For example, creativity contingent SE may affect an artist’s ability but not a mathematician’s ability to create something truly unique. Further empirical research is necessary in order to explore this suggestion for the role of contingent SE on behaviour.

In spite of the null results obtained for the interaction hypothesis, the hypothesis that SE affects our ability to think insightfully was supported. When addressing the first hypothesis, it is notable that despite the small sample (n=60) used in this study the results were significant, suggesting that were further research to be done with larger samples drawn from a more varied population, similar significant results would be found. Furthermore, given the establishment of an initial link between insight and SE it would be prudent to further substantiate the relationship through further research using standard insight tests. Even though we theorise that dingbats do require insight this conclusion has as of yet not been scientifically proven.

The knowledge that high SE has a positive influence on insight ability could be integrated and used in institutions that would like to construct creativity or insight-building workshops. School and office administrations that wish to revolutionise the way that the institution is run by for example, increasing the motivation to learn or work on difficult tasks, could implement training, and establish working environments that positively influence employees’ SE. In work environments that particularly require insight and innovation e.g. advertising, positively influencing employees’ SE could encourage both a healthier working environment as well as higher productivity. Organisations that wish to implement a SE-building intervention with the aim of increasing creativity and insight, should note that these interventions might not be effective for all employees. Once further research is done into the exact role of education on insight ability, SE interventions could be implemented for the appropriate target populations.

References
Baumeister, R. F., Campbell, J. D., Krueger, J. I., & Vohs, K. D. (2003). Does high self-esteem cause better performance, interpersonal success, happiness, or healthier lifestyles? Psychological Science in the Public Interest, 4(1), 1-44.
Berkowitz, L. & Troccoli, B. T. (1990). Feelings, direction of attention, and expressed evaluations of others. Cognition and Emotion, 4(4), 305-325.
Crocker, J., Wolfe, C. T. (2001). Contingencies of self-worth. Psychological Review, 108, 593-623.
Davila, J., Hammen, C., Burge, D., Daley, S. E. & Paley, B. (1996). Cognitive / interpersonal correlates of adult interpersonal problem-solving strategies. Cognitive Therapy and Research, 20(5), 465-480.
DeYoung, C. G., Flanders, J. L. & Peterson, J. B. (2008). Cognitive abilities involved in insight problem solving: An individual differences model. Creativity Research Journal, 20(3), 278-290.
D’Zurilla, T. J., Chang, E. C. & Sanna, L. J. (2003). Self-esteem and social problem solving as predictors of aggression in college students. Journal of Social and Clinical Psychology, 22(4), 424-440.
Ferris, D. L., Brown, D. J., Pang, F. X. J., Keeping, L. M. (2010). Self-esteem and job performance: The moderating role of self-esteem contingencies. Personnel Psychology, 63, 561-593.
Gasper, K. (2003). When necessity is the mother of invention: Mood and problem solving. Journal of Experimental Social Psychology, 39(3), 248-262. doi:10.1016/S0022-1031(03)00023-4
Hoffman, B. & Spatariu, A. (2008). The influence of self-efficacy and metacognitive prompting on math problem-solving efficiency. Contemporary Educational Psychology, 33(4), 875-893.
Jaquish, G. A. & Ripple, R. E. (1981). Cognitive creative abilities and self-esteem across the adult life-span. Human Development, 24(2), 110-119.
Mikulincer, M. & Sheffi, Y. (2000). Adult attachment style and cognitive reactions to positive affect: A test of mental categorization and creative problem solving. Motivation and Emotion, 24(3), 149-174.
Ollinger, M., Jones, G. & Knoblich, G. (2008). Investigating the effect of mental set on insight problem solving. Experimental Psychology, 55(4), 270-282.
Park, L. E., Crocker, J. (2008). Contingencies of self-worth and responses to negative interpersonal feedback. Self and Identity, 7(2), 184-203.
Rosenberg, M. (1965). Society and the adolescent self-image. Princeton, NJ: Princeton University Press.
Tang, T. L. & Reynolds, D. B. (2006). Effects of self esteem and perceived goal difficulty, certainty, task performance and attributions. Human Resources and Development Quarterly, 4(2), 153-170.
Tuan Pham, M. (1998). Representativeness, Relevance, and the use of feelings in Decision Making. The Journal of Consumer Research, 25(2), 144-159.
Weiss, H. M., & Knight, P. A. (1980). The utility of humility: Self-esteem, information search, and problem-solving efficiency. Organizational Behavior and Human Performance, 25(2), 216-223. doi:10.1016/0030-5073(80)900641
Whitesell, N. R., Mitchell, C. M., Spicer, P. & The Voices of Indian Teens Project Team (2009). A longitudinal study of Self-esteem, cultural identity, and academic success among American Indian adolescents. Cultural Diversity and Ethnic Minority Psychology, 15(1), 38-50. DOI: 10.1037/a0013456