Overcoming fixed mindsets: The role of affect

Abstract

When there is an established strategy to solve a problem, we often approach the problem with a mindset that makes us blind for more efficient solutions. We examined the role of affect in overcoming such blinding effects of mindsets. As positive affect is known to broaden and negative affect to narrow thought–action repertoires, we speculated that positive affect facilitates and negative affect impedes the overcoming of a current mindset. To induce a mindset, participants initially solved 60 similar problems which were only solvable using the same complex strategy. After a short break in which positive or negative affect was induced, participants continued to work on the problems. Critically, there now was an additional simple way to solve the problems. Participants experiencing positive affect were more likely to detect the simple solution than participants experiencing negative affect. These findings reveal that affect modulates how much we are constrained by current mindsets.

Keywords:

• Emotions
• Insight
• Mental set
• Mood
• Problem solving
• Mental fixation

Notes

¹ Sample size was not determined before data collection. We did a first data analysis after data from 50 participants had been collected, showing that happy participants experiencing positive affect were more likely to detect the simple solution than participants experiencing negative affect, χ²(1, N = 50) = 4.67, p = .031. We then decided to check the robustness of this finding by collecting data from 30 additional participants. The results for the additional 30 participants fully replicated the results for the first 50 participants, χ²(1, N = 30) = 4.66, p = .031.

² Because only extreme outliers are excluded with such a criterion when the sample size is small (e.g., Shiffler, 1988), we reanalysed our data using a number of other correction methods (i.e., different general cut-offs, log-transformation, median). None of the analyses showed a significant main effect of affect, or a significant interaction between affect and block, all ps > .121.

³ All participants showing a behavioural change in performance were able to provide explicit knowledge about the easy solution. One possibility why detection of the easy strategy was not evident in solution times of the four additional participants identified as having detected the easy strategy by the post-experimental questionnaire may have been that these participants continued to use the instructed rules although they had detected the easy strategy. However, given that the benefit from using the easy strategy was extremely large (i.e., a benefit of an 84%-decrease in solution times), it seems unlikely that this was the case. More likely, detection of the easy strategy was not evident in solution times of these participants because they had detected the easy strategy so late that a sustained decrease of solution times was not detectable.

⁴ The increased rate of detecting the easy strategy in the positive affect condition was also evident in mean solution times in the critical test. Averaging solution times across all 120 problems in the critical test (incorrect responses and solution times above 15 sec were removed) revealed that mean solution times were faster in the positive affect condition (M = 4.44 sec, SD = 2.70) than in the negative affect condition, (M = 5.71 sec, SD = 2.74), t(78) = −2.08, p = .041.

⁵ We additionally analysed our data using a number of other criterions (i.e., checks within only the next 20 trials after detection of the easy strategy, coding a check if solution times were two or four seconds above the median). None of the analyses showed a significant difference in the number of checks between affective groups, all ps > .181.