Similarly to earlier laboratory studies27,28, we aimed to collect data from about 110 students. In the present work, participants were male university students (N=105), as based on prior studies [e.g.,30] males use more behavioral self-handicapping than females. Four participants were dropped from the analyses due to expressing general suspicion of the deception utilized in the study, resulting in a final sample of 101 participants. Statistical analyses were conducted on this final sample. The participants were between 18 and 55 years of age (Mage=27.33 years, SDage=8.79 years). Participants received partial course credit for their participation. The study was conducted in accordance with the Declaration of Helsinki and with the approval of the local university’s ethical committee as well as the informed consents of the participants.
The research procedure is shown in Figure 1. Participants were told that they can contribute to the validation of an online IQ test based on its reliable offline version.
1. Preliminary survey: First, they completed an online survey with items related to their intelligence beliefs and socio-demographic characteristics (see below in details). Two or three days later they were invited to the laboratory to participate in an experiment. The behavioral data were collected individually in a small university laboratory in which participants were randomly assigned to one of the two experimental conditions: (a) entity, (b) incremental mindset.
2. Manipulation of intelligence mindsets: Upon arrival at the laboratory participants had to wait for a few minutes in the corridor, where scientific posters illustrated with pictures and diagrams were pinned on the wall either about the heritability of the intellectual skills (entity condition), or the malleability of the brain (incremental condition). Then participants were greeted by the experimenter wearing white lab coat who accompanied them to the laboratory room. Next, the experimenter delivered the implicit message manipulation. In the entity condition, participants were told the following:
In this research intellectual skills are investigated. As you must know, intelligence is defined by an IQ score. When a baby is born, they already possess a particular intellectual potential that is unfolding throughout their life. This means that our intellectual skills are barely changeable, instead they predominantly remain stable and fixed through the lifespan. Hence, we are not able to do anything to change our intellectual skills. This research (the experimenter points to the same scientific report pinned on the wall that they had seen before) for instance is significant for this reason. Thompson and his colleagues studied identical twins in 2002. Identical twins are known to be 100% identical in terms of their DNA. The authors found that intellectual skills are 88% identical in the two siblings. This means that intellectual skills are highly determined by our genes. This is the reason behind the intelligence not changing throughout our lifespan.
In the incremental condition, participants were told the following:
In this research intellectual skills are investigated. Numerous previous studies showed that our intellectual skills change through the lifespan. This process depends on how much we learn and the amount of effort we exert. So, the scientific reports from the last few years evidently showed us that the structure of the brain develops and changes when we work hard on a task or exert great intellectual effort. As a result, new brain connections form and the existing connections strengthen. As such, human brain and intellectual skills can be developed. This research (the experimenter points to the same scientific report pinned on the wall that they had seen before) for instance is significant for this reason. Zatorre and his colleagues studied ordinary people learning to juggle in 2013. They found that after just one week of learning participants’ gray matter density increased, demonstrating development of their brains. This is the reason behind the statement that intellectual skills can be cultivated.
Both conditions additionally heard some statements about the importance of the intelligence:
Nonetheless we all know that intelligence is a crucial attribute of ourselves. Intelligence affects the grades we receive, the university and the job we get into, so basically and ultimately how successful we will be in our lives. Thus, intelligence is really an important factor.
3. Intelligence test exercise: Subsequently, participants were conseed that they would be exposed to some “trial” items before the “real” IQ test in order to assure that they understood the task and were familiar with the type of it. Participants were told that confounding factors such as misunderstanding needs to be ruled out so the real IQ test score would reflect their real intellectual skills. Six items were selected from the Raven Progressive Matrices31 and the last four were modified to be unsolvable. Participants had six minutes to complete the tasks and they were informed about the status of the countdown. After the 6th minute the experimenter took the sheet, and visibly flipped through it.
4. Behavioral measure of self-handicapping: Then participants were informed that the validation process needs to be explored in various circumstances. One of these circumstances is related to the effect of music on their performance. Therefore, the experimenter asked the participants to choose one of the available CDs to listen to while completing the IQ test27,32. According to the cover story, the music on each CD has a different effect on their cognitive functions. The CDs were stacked horizontally (so that the colored labels could be easily seen) in a CD rack and labelled: with Three green dots “Highly Enhancing” (1), Two green dots “Moderately Enhancing” (2), One green dot “Mildly Enhancing” (3) without a sign “Neutral” (4), One red dot “Mildly Detracting” (5), Two red dots “Moderately Detracting” (6) and Three red dots “Highly Detracting” (7) effects. The experimenter emphasized that the data collection is about to finish so there were enough music choices from all types. Therefore, they are free to choose whatever they really wish to listen to. Participants were left alone for one minute, then handed the selected CD to the experimenter.
5. Manipulation and suspicious checks: While the experimenter allegedly started to mount the music device, participants were requested to answer some manipulation checks regarding their CD choice ("What kind of effect the red/green signed music has on the cognitive functions?") and intelligence mindset (two items, see below). The potential suspicious effects were also tested using a funnel method33. First, participants were asked, “How could you describe the purpose of this study in your own words?”. Then, they were asked whether they had noticed anything particular during the experiment and especially during the practice session.
6. Debriefing: Finally, all participants were thoroughly debriefed directly following the study completion and dismissed.
Pre- and post-manipulation intelligence mindset beliefs. Students’ implicit theories of intelligence were assessed by the Theories of Intelligence Scale 13. The scale is composed of two four-item subscales of the entity (e.g., “Your intelligence is something about you that you can’t change very much”; Cronbach’s α = 0.91) and incremental theories (e.g., “No matter who you are, you can significantly change your intelligence level”; Cronbach’s α = 0.92) resulting in a unidimensional measure (Cronbach’s α = 0.93). The two subscales were interrelated (r=-0.73, p<0.001). Respondents indicated their answers on a six-point Likert scale. All eight items were used pre-manipulation, and two seminal items (“You have a certain amount of intelligence, and you can’t really do much to change it” and “You can always substantially change how intelligent you are”) of the scale were used post-manipulation as manipulation checks (rinter-item=0.68; p<0.001). The items on the pre- and post-treatment measures were averaged (separately) such that higher values correspond to more growth mindsets.
Importance of intelligence. Participants were asked about the importance of the domain that was about to be threatened: “How important is the level of your intelligence to you?”. The statement was rated on a 4-point Likert scale from not at all (1) to very much (5).
Behavioral self-handicapping measure. Self-handicapping behavior was operationalized as choosing suboptimal circumstances for performing the target intelligence task. Therefore, prior to taking the second test of intellectual ability, each participant had the opportunity to select one sort of music they wanted to listen to during the IQ test. The music selection ranged on a scale with values from 1 (highly enhancing) to 7 (highly detracting). The middle of the scale was represented by neutral music. The CD choice constituted the self-handicapping behavior. Based on prior literature we used it as a continuous measure as not choosing the most optimal condition (1-highly enhancing) can be interpreted as self-handicapping. Demographic data. Finally, demographic characteristics of participants were asked: gender, age, and whether their intellectual skills had been measured before.
Statistical analyses and data visualization were performed with IBM SPSS software version 22.0 and R 4.0.334. Within R, the tidyverse35, the ggplot236, packages were used for data transformation and visualization. First, we checked potential differences between the control and experimental groups in terms of age, prior intelligence mindset, importance of intelligence, prior participation in intelligence testing and found no differences between the two groups (all ps>0.58). Second, we treated the CD choice behavioral self-handicapping variable as a continuous measure, ordinary least squares (OLS) regression analysis was performed to examine the effect of the experimental manipulation (fixed vs. growth mindset) as well as the interaction between the experimental manipulation and the prior intelligence mindsets. The pre-experimental growth intelligence mindset, the intelligence mindset manipulation and their interaction were included in the regression model.