Sage Journals: Discover world-class research

Abstract

Why do people change their self-perceptions (i.e., self-concept) after receiving external, self-discrepant information? While recent theorizing posits that reflection is a prerequisite for such change, empirical evidence supporting this is sparse. Here, we investigated the role of reflection in self-concept change after feedback in six studies. In Studies 1a–e (total N = 2,422), participants received feedback regarding a self-concept domain and we assessed reflection as well as self-concept change. Across studies, the amount of reflection was positively correlated with self-concept change. In Study 2 (N = 1,149), we experimentally varied whether participants were encouraged to reflect on the feedback and their self-concept or prevented from doing so via a distractor task. We found that self-concept change was larger after inducing (vs. suppressing) reflection. Exploratory analyses showed that the association between reflection and self-concept change was more pronounced for negative than for positive feedback in Studies 1a–e, but not Study 2.

Keywords

reflection self-concept self-concept change feedback

People’s self-concept is generally stable. Yet, external information that is discrepant from their self-concept can sometimes lead to self-concept change (Brotzeller & Gollwitzer, 2024; Elder et al., 2022; Müller-Pinzler et al., 2019). Such self-concept change can occur on different dimensions of the self-concept, which encompasses all self-perceptions a person holds (Shavelson et al., 1976). A person might, for example, think of themselves as a hard worker. But when their boss says that they are not investing enough time and energy into work, this discrepancy may make them re-think their self-concept of being a hard worker. While receiving feedback only once or observing just one single instance of discrepant behavior might lead to small changes in a person’s self-concept, receiving discrepant feedback repeatedly may alter one’s self-concept more profoundly and sustainably (Bleidorn et al., 2018). But what is the psychological mechanism linking everyday experiences to self-concept change?

Theoretical accounts (e.g., the TESSERA framework or the literature on wisdom and narrative identity; see below) posit that long-term change occurs when people reflect on self-relevant experiences as this helps them gain self-knowledge and integrate their experiences into their self-concept (Glück & Weststrate, 2022; McAdams & McLean, 2013; Wrzus & Roberts, 2017). However, there is little empirical evidence for an association between reflection and self-concept change so far (Quintus et al., 2021; Wrzus, 2021). Even less is known about the factors that might impact this association. The present research aims to fill these gaps by investigating the association between reflection and self-concept change as well as exploring one potential moderator: the direction of discrepancy, that is, whether the feedback negatively or positively deviates from a person’s self-concept.

Conceptualizing Reflection

There are different conceptualizations of reflection about self-relevant experiences. An early conceptualization defines reflection as the “process of creating and clarifying the meaning of experience (present or past) in terms of self” that results in a changed conceptual perspective (Boyd & Fales, 1983, p. 101). In a literature review, Rogers (2001) finds that reflection is often conceptualized as an active cognitive and affective process that is triggered by certain situations and during which new information can be contrasted with and later integrated into existing beliefs. Both definitions imply that reflection has important functions for the self. When a person encounters new and potentially discrepant information about themselves, this might trigger aversive emotions (e.g., shame, in case of negative information). Reflecting on the new information can then help process the information, regulate oneself, and potentially reevaluate one’s self-concept. This is consistent with the literature on self-reflection and self-awareness, which highlights a link between these constructs and self-regulation as well as self-evaluation (Morin, 2011; Tangney & Tracy, 2012). More recent conceptualizations of reflection build upon these earlier works; yet, the definitions of reflection are often less detailed. For instance, Wrzus and Roberts (2017) describe reflection as “consciously thinking about one’s past experiences, behavior, thoughts, and feelings” (p. 261).

In an attempt to create a comprehensive, yet concise definition of reflection based on existing definitions and the elements highlighted by Rogers (2001), we define reflection as a process during which individuals cognitively engage with self-relevant experiences and potentially integrate them into their existing beliefs. In the present research, we are especially interested in self-concept change resulting from an integration of new information into the existing self-concept (Brotzeller & Gollwitzer, 2024).

The Effect of Reflection on Self-Concept Change

One prominent theoretical model highlighting the role of reflection for changes in people’s self-concept regarding their personality is the TESSERA framework (Wrzus & Roberts, 2017). The authors propose that reoccurring sequences of self-relevant experiences lead to changes in self-perceptions via reflective processes. They specify that these self-relevant experiences consist of several elements: First, a triggering situation (e.g., a person being asked to take over a task at work) creates an expectation regarding a thought, feeling, or behavior that a person should display (the expectation can be either internal or external, e.g., the boss expecting a conscientiously completed task). In response to the triggering situation and the expectation, the person displays a state or state expression (e.g., the person works on the task conscientiously), followed by a reaction to the state or state expression (either by oneself or by someone else, e.g., the boss gives positive feedback on the task result). One or several such sequences make the person reflect on the reactions they have received. Such reflection about the reactions and their implications for one’s self-concept is assumed to be the pathway to changes in the explicit self-concept. Importantly, while the TESSERA framework also covers changes in implicit personality or in personality observed by others, our research specifically focuses on self-concept change, which refers to changes in self-perceptions people hold.

While the framework focuses on changes in the personality self-concept, Wrzus and Roberts (2017) argue that the framework applies to other domains of the self-concept as well. This is in line with theorizing in other areas, suggesting that reflection is decisive for changes regarding all domains of the self-concept. The literature on narrative identity, for example, proposes that people actively reflect on difficult experiences to increase their self-understanding and integrate these experiences into their identity, which can contain all kinds of self-related information (McAdams & McLean, 2013; Pals, 2006). Moreover, the wisdom literature regards reflection on one’s life experiences, mistakes, and successes as decisive in gaining self-insight as well as questioning and adapting one’s self-views (Glück & Weststrate, 2022).

Despite these theoretical arguments for a link between reflection and self-concept change, little empirical research has directly tested this link. Previous research that is relevant in that regard provides merely indirect evidence for the role of reflection in self-concept change after discrepant feedback. For instance, Sedikides and Skowronski (1995) asked participants to name and rate the importance of sources of self-knowledge and found that participants rated activities related to self-reflection as highly important (e.g., remembering past interactions or relationships). But this study did not look at self-concept change. In a more recent study, Miller (2020) examined how engaging in reflective activities helped healthcare leaders gain self-knowledge after stressful situations. They report responses from 17 participants rating the degree to which the reflective activities increased their self-knowledge, finding that those who more often engaged in reflective activities also reported larger gains in self-knowledge. Yet, they examined only self-reported gains in self-knowledge and also did not measure actual changes in participants’ self-concepts. This, in addition to the very small sample size, limits the conclusions that can be drawn from this study regarding our research question: whether reflection after self-discrepant feedback positively predicts self-concept change. The most direct evidence for a link between reflection and self-concept change comes from a study by Quintus et al. (2021). In several waves of daily diary assessments over a period of 2 years, the authors assessed participants’ daily experiences, states, and their reflection on these experiences. Furthermore, they measured participants’ explicit personality self-concept regarding the Big Five traits at several measurement occasions throughout the 2-year period. For conscientiousness, more extensive reflection in combination with more trait-relevant (i.e., conscientious) behavior led to larger increases in trait conscientiousness (Quintus et al., 2021). However, such an effect was not observed for the other Big Five traits. Thus, there was no strong evidence for reflection as the pathway to self-concept change in their study.

Despite theoretical models highlighting their importance, assuming that reflection in response to self-relevant experiences leads to self-concept change is not trivial. While self-concept change is one way of dealing with a self-relevant experience that is discrepant from one’s self-concept, there are other options that do not require adapting one’s self-concept. One option could be to dismiss the self-discrepant experience as irrelevant (“immunization”; see Brandtstädter & Greve, 1994). When a person receives negative feedback regarding their work ethic from someone who dislikes them, for example, they can easily interpret the feedback as maliciously intended, invalid, or unreliable. Another option for dealing with self-discrepant experiences could be to adapt one’s behavior in order to prevent such discrepancies in the future (“assimilation”; see Brandtstädter & Greve, 1994). A person receiving negative feedback about their work ethic could, for example, plan to invest more time and energy into their work in the future. Importantly, either of the two options can render self-concept change unnecessary: After dismissing feedback as irrelevant or adapting one’s behavior to avoid similar feedback in the future, self-concept change is no longer necessary. Processing self-discrepant experiences during reflection, therefore, does not necessarily lead to self-concept change, but might also even reinforce one’s initial self-concept (Quintus et al., 2021; Wrzus & Roberts, 2017). Previous research supports the assumption that reflecting on a self-discrepant experience can result in outcomes other than self-concept change: Several qualitative studies have shown that reflecting on feedback can produce (plans for) behavioral changes aimed at improving and, therefore, preventing discrepant feedback in the future (Boyd & Fales, 1983; Overeem et al., 2009; Sargeant et al., 2009). Moreover, a quantitative study on leadership development showed that structured reflection about relevant experiences (compared to a control group in which no structured reflection was induced) leads to more changes in leadership behaviors (DeRue et al., 2012).

In summary, while theoretical models such as the TESSERA framework as well as the literature on wisdom and narrative identity posit that reflection after self-relevant experiences leads to self-concept change, there are plausible theoretical arguments that speak against such an effect and few empirical studies supporting either position. To gain a better understanding of the cognitive processes that shape how the self-concept changes in light of external feedback, we think it is relevant to start with the broad question of whether and how reflection is involved in such change. This is what we aim to do in the present research. More specifically, our first goal is to study whether reflecting on a discrepant, self-relevant information leads to self-concept change. In doing so, we build upon and extend previous research by examining how single instances of receiving self-relevant information contribute to self-concept change regarding specific aspects of the self-concept.

The Direction of Discrepancy as a Potential Moderator

While reflection is regarded as an important process that might lead to self-concept change, it is plausible to assume that several person and situation characteristics impact this association. The authors of the TESSERA framework propose, for example, that individual differences in self-evaluation motives or different ways of remembering and processing positive and negative information about the self might play a role (Wrzus, 2021; Wrzus & Roberts, 2017). In the present research, we focus on the latter aspect and, in addition to the main effect of reflection on self-concept change, explore the impact of negative versus positive information on that association. More specifically, we are interested in the effect of the direction of discrepancy, which describes whether the self-relevant information a person encounters is more negative or positive than their initial self-concept. There is a third case in which the feedback received does not deviate from one’s self-concept. This case, however, is not interesting for the current research question. It will therefore be neglected in our theoretical argument as well as our analyses.

The comparison between negative and positive information is interesting because we know that both types of information are processed differently (Unkelbach et al., 2020). In the case of self-discrepant feedback, this implies that the process and the outcomes of reflecting about negative versus positive information differ. Negative information—feedback according to which one’s abilities, performance, etc., is worse than expected—is arguably more threatening (Hakmiller, 1966) and may invoke defensive strategies that render self-concept change less likely than positive information—feedback suggesting that one is better than expected. In that regard, reflecting thoroughly, elaborately, and open-mindedly might be necessary to counteract such defensive reactions to negative feedback. It is plausible, therefore, that the association between reflection and self-concept change is more pronounced after having received negative compared to positive self-discrepant information. Investigating this effect is a second goal of the present research.

The Present Research

We present six studies investigating the role of reflection in self-concept change after receiving self-relevant information. Furthermore, we explore the effects of the direction of discrepancy on this association. To do so, we applied a quantitative approach that enabled us to examine self-concept change at a fine-grained level. In all studies, we measured people’s self-concept regarding a specific domain, asked them to complete a task or scale measuring this domain, and, subsequently, gave them feedback on their results. Afterwards, we measured participants’ self-concept a second time. Applying a correlational approach in Studies 1a–e, we additionally assessed self-reported reflection after the feedback to examine whether it is associated with self-concept change. In Study 2, we experimentally induced or suppressed reflection and tested whether subsequent self-concept change differed between conditions. All studies were conducted online and with German-speaking participants.

Data, R codes for primary analyses, and supplemental materials are available online at https://osf.io/cfvjs/ for all studies. We report how we determined our sample size, all data exclusions, all manipulations, and all measures for all studies. All studies except for Study 1c were preregistered. Preregistrations can be found at https://aspredicted.org/31S_M1M for Study 1a, https://aspredicted.org/VYH_M31 for Study 1b, https://aspredicted.org/3TQ_Y5S for Study 1d, https://aspredicted.org/R33_23D for Study 1e. These preregistrations detail the study designs, pre-planned stopping rules, and exclusion criteria. As these studies had originally been designed to test other research questions, the hypotheses tested here differ from those that had been preregistered. The data from Studies 1d and 1e were used in a previous publication (Brotzeller & Gollwitzer, 2024). Study 2 was designed to replicate the (correlational) findings in Studies 1a–e with an experimental design. The preregistration for Study 2 can be found at https://aspredicted.org/12F_H31.

Studies 1a–e

In Studies 1a–e, the association between reflection after self-relevant feedback and self-concept change was investigated using a correlational approach. To examine this association across a broad range of domains of the self-concept, we focused on a different self-concept domain in each study. We gave participants feedback regarding the respective self-concept domain and measured the extent of their reflection as well as self-concept change.

Method

Procedures and Measures

As the procedures and measures for Studies 1a–e are similar, an overview with examples from Study 1a is given in the following. Details on the procedures and measures for all studies can be found in Table 1. The full study materials for all studies are available at https://osf.io/cfvjs/. In all five studies, participants were first informed which domain of the self-concept the study would focus on (e.g., weight estimation abilities in Study 1a). Participants were given a description of and indicated their self-perception regarding this self-concept domain (e.g., via one item in Study 1a: “How good do you think you are at estimating the weight of objects or living beings?,” scale from 0% = very bad to 100% = very good). In case the self-perception was not already assessed in a percentage format (cf. Table 1), participants were informed which percentage their self-rating corresponded to. This was done to ensure that self-perceptions were directly comparable to the feedback participants received later in the study. After indicating their self-perceptions, participants completed either a task with several subtasks or a scale measuring their level on the respective self-concept domain (e.g., a weight estimation task in Study 1a). They then received feedback regarding their result on this task or scale (e.g., “Your answers on the estimation test have been evaluated. On a scale from 0% [very bad estimation ability] to 100% [very good estimation ability] you are at: 75%” in Study 1a). The feedback was always presented to them in the same percentage format used for their self-perceptions.

Table 1.

Overview of Studies 1a–e.

Study	SCD	Measurement of SCD	Task/scale	Feedback	Measurement of reflection
1a	Weight estimation	“How good do you think you are at estimating the weight of objects or living beings?,” scale from 0% = very bad to 100% = very good; M_t1 = 46.21, SD_t1 = 21.09; M_t2 = 33.64, SD_t2 = 17.73; M_SCC = 15.41, SD_SCC = 14.39	Weight estimation task with 12 subtasks in which participants were presented an object or living being and had to choose the correct of four options for its weight; the task was self-created	“Your answers on the estimation test have been evaluated. On a scale from 0% (very bad estimation ability) to 100% (very good estimation ability), you are at: [RESULT]%”; real feedback representing the percentage of correctly solved subtasks	“During the course of the study, I thought a lot about my self-assessment of my ability to estimate weights and about my feedback on the estimation test,” scale from 1 = strongly disagree to 6 = strongly agree; M = 3.74, SD = 1.20
1b	Health-conscious lifestyle	“How health-conscious do you consider your lifestyle to be? When answering, please refer to the last 6 months,” scale from 0% = not at all heath-conscious to 100% = very health-conscious; M_t1 = 63.74, SD_t1 = 19.55; M_t2 = 65.86, SD_t2 = 16.84; M_SCC = 5.40, SD_SCC = 6.60	Scale assessing how health-conscious a person’s lifestyle is via items from five categories (e.g., nutrition, sports, sleep, and stress; adapted from Rodriguez Añez et al., 2008; Wilson et al., 1984)	“Your answers on the lifestyle scale were evaluated. On a scale from 0% (not at all health-conscious) to 100% (very health-conscious lifestyle) you are at: [RESULT]%”; real feedback representing a point score that sums up responses to the items and is converted to a percentage score for the feedback (point scoring adapted from Rodriguez Añez et al., 2008)	“During the course of the study, I thought a lot about my self-assessment and my feedback on the lifestyle scale,” scale from 1 = strongly disagree to 6 = strongly agree; M = 3.72, SD = 1.23
1c	General knowledge	“My general knowledge is as good as or better than that of ____ of the German-speaking population,” scale from 0% to 100%; M_t1 = 59.07, SD_t1 = 17.55; M_t2 = 59.90, SD_t2 = 19.43; M_SCC = 10.19, SD_SCC = 11.10	General knowledge test with 10 questions (adapted from Trepte & Verbeet, 2010)	“Based on your answers in the general knowledge test, your general knowledge is as good as or better than that of [RESULT]% of the reference group,” real feedback representing the participant’s percentile score compared to a reference group of >160,000 German speakers (Trepte & Verbeet, 2021)	“During the course of the study, I thought a lot about my self-assessment and my feedback on the general knowledge test,” scale from 1 = strongly disagree to 6 = strongly agree; M = 3.59, SD = 1.13
1d	Emotion recognition	Four items, for example, “It’s very easy for me to read a person’s emotions from their eyes,” scale from 1 = strongly disagree to 9 = strongly agree; α_t1 = .94, M_t1 = 5.10, SD_t1 = 1.73; α_t2 = .94, M_t2 = 4.84, SD_t2 = 1.59; M_SCC = 0.90, SD_SCC = 0.93	Emotion recognition task with 10 subtasks in which participants were presented photos of human eye areas and had to choose the correct one out of four options given for the emotion the person might be feeling (adapted from Bölte, 2005)	“Your percentage of correct answers in the photo task was: [RESULT]%”; real feedback representing the percentage of correctly solved subtasks	“How much did you think about your feedback on your self-assessment and the photo task during the course of the study?,” scale from 1 = very little to 9 = very much; M = 5.22, SD = 1.82
1e	Visual-spatial ability	Five items, for example, “I find it very easy to examine objects from different spatial perspectives in my imagination,” scale from 1 = strongly disagree to 9 = strongly agree; α_t1 = .82, M_t1 = 5.65, SD_t1 = 1.29; α_t2 = .90, M_t2 = 5.34, SD_t2 = 1.65; M_SCC = 0.76, SD_SCC = 0.71	Visual-spatial ability tasks with subtasks in which participants had to mentally manipulate objects to solve problems (adapted from Formann et al., 2011; Kersting et al., 2008); due to an experimental manipulation, participants completed either one or three tasks with 18 or 20 subtasks each and received feedback on each completed task (for more information, see https://osf.io/cfvjs/)	“You have solved [RESULT]% of the tasks correctly”; fake feedback supposedly representing the percentage of correctly solved subtasks; feedback was either positive or negative and differed from participants’ self-perceptions by a smaller or larger amount (for more information, see https://osf.io/cfvjs/)	“I reflected a lot on the feedback I received about my ability for spatio-visual thinking during the study,” scale from 1 = strongly disagree to 5 = strongly agree; M = 2.93, SD = 1.15

Note. SCD = self-concept domain, SCC = self-concept change. [RESULT] stands for the result a participant achieved in the respective task or scale.

In Studies 1a–d, the feedback was real and reflected participants’ actual task performance. In Study 1e, participants received fake feedback as we manipulated the direction and the size of the discrepancy: Participants’ feedback was randomly assigned to be either more negative or more positive than their self-perceptions and to deviate by either a small amount (i.e., around 5%) or a large amount (i.e., around 20%) from their self-perceptions.¹ After having received feedback, participants again indicated their self-perceptions regarding the self-concept domain on the same item(s) used at the beginning of the study. Self-concept change scores were created by calculating the absolute value of the difference between the first and second self-concept measurement.

Participants further rated how much they had reflected on their level on the self-concept domain and the feedback during the study (e.g., via one item in Study 1a: “During the course of the study, I thought a lot about my self-assessment of my ability to estimate weights and about my feedback on the estimation test,” scale from 1 = strongly disagree to 6 = strongly agree). The specific items measuring reflection in each of the five studies, respectively, are reported in the last column of Table 1.²

Samples

For all studies, participants were recruited through (university) mailing lists and, in some cases, via social media postings and personal networks. In return for their participation, participants were given the opportunity to participate in a raffle for vouchers or to receive course credit. For all preregistered studies, we collected data and applied exclusion criteria as delineated in the respective preregistration, except for one additional exclusion criterion applied in Study 1d (see below for an explanation). In Study 1c, which was not preregistered, we collected data for one month and applied the same exclusion criteria as in Studies 1a and 1b.³

Study 1a

In total, 667 participants completed the study within the preregistered time frame. After applying the preregistered exclusion criteria, the final sample consisted of N = 592 participants (M_age = 31.05, SD_age = 14.49; 412 female, 170 male, 10 “other”).

Study 1b

Within the preregistered time frame, 654 participants completed the study. Applying the preregistered exclusion criteria resulted in a final sample of N = 544 participants (M_age = 36.10, SD_age = 15.16; 401 female, 133 male, 10 “other”).

Study 1c

A total of 587 participants completed the study within a time frame of one month. After applying the same exclusion criteria as in Studies 1a and 1b, the final sample consisted of N = 460 participants (M_age = 34.75, SD_age = 15.47; 313 female, 140 male, seven “other”).

Study 1d

In total, 548 participants completed the study within the preregistered time frame. After applying the preregistered exclusion criteria, our sample consisted of n = 459 participants. We applied one additional exclusion criterion: As our goal was to examine responses to discrepant feedback, we excluded n = 6 further participants with no discrepancy between feedback and initial self-concept.⁴ Our final sample thus comprised of N = 453 (M_age = 32.61, SD_age = 14.03; 331 female, 114 male, eight “other”).

Study 1e

A total of 627 participants completed the study within the preregistered time frame. We applied the preregistered exclusion criteria, which led to a final sample of N = 373 participants (M_age = 29.33, SD_age = 12.50; 290 female, 77 male, five “other,” one did not respond).

Results

We used R to conduct all analyses (R Core Team, 2023). In preparation for our analyses, we rescaled the reflection and self-concept change variables to be consistent between studies. Rescaled reflection values ranged from 0 = low to 5 = high reflection for all studies, while rescaled self-concept change values ranged from 0 = no to 8 = maximum change.⁵

Main Analysis

For our main analysis, we used the combined data from all five studies with a total N = 2,422. This also meant that our data had a multilevel structure as participants were nested within studies. Therefore, we used multilevel modeling to analyze our data, which was proposed by Raudenbush and Bryk (1985) for meta-analytic procedures in which results are analyzed across studies. Such approaches, in which the original data for each participant are included in the analyses instead of using aggregate data, are also called individual participant data (IPD) meta-analyses. Due to their many advantages over meta-analytic approaches that merely take the study-level effect size estimators (and their standard errors) into account, IPD meta-analyses are currently considered the gold standard (Rakshasbhuvankar, 2021).

Based on the combined collected data, we conducted simulations to estimate the statistical power to detect the main effect of interest in our multilevel model using the simr package (Green & MacLeod, 2016). Assuming α = .05 and 1 − β = .80, our sample size was sufficient to detect significant effects as small as B = 0.04. Further details are provided in the supplementary materials on OSF.

To analyze our data, we specified a random intercept model including absolute self-concept change as the dependent variable, reflection as a fixed effect, and study as a random factor.⁶ To do so, we used the lme4 package in R and parameters were estimated via restricted maximum likelihood estimation (Bates et al., 2015). Our results are summarized in Table 2. Reflection emerged as a significant predictor of self-concept change, B = 0.15, p < .001.

Table 2.

Multilevel Regression Model Results for Self-Concept Change.

Model parameters and variance components	Main analysis	Exploratory analysis
(Intercept)	0.44* (0.13)	0.45** (0.11)
Reflection	0.15*** (0.01)	0.15*** (0.01)
DoD		0.02 (0.04)
Reflection × DoD		−0.07*** (0.01)
SD _{Random Intercept}	0.28	0.23
SD _{Level-1-Residual}	0.86	0.84
R ²	.133	.148

Note. Standard errors are displayed within parentheses. N = 2,422. DoD = direction of discrepancy. DoD: negative = −1, positive = 1.

p < .05. **p < .01. ***p < .001.

Exploratory Analyses

Next, we explored the effect of the direction of discrepancy on the association between reflection and self-concept change. In preparation for this analysis, we created the scores for the direction of discrepancy. In Studies 1a–d, we compared participants’ initial self-concept with their task scores (both variables are comparable because we used percentage values; see above). Cases in which task scores were lower than the self-concept were coded as a “negative discrepancy,” cases in which task scores exceeded the self-concept were coded as a “positive discrepancy.” Cases in which the discrepancy was 0 were excluded from our analyses as outlined in the sample description. In Study 1e, direction of discrepancy was manipulated in the feedback participants received. For our analyses, we effect-coded the direction of the discrepancy variable for all studies (i.e., “negative” = −1, “positive” = 1).

To investigate the effect of the direction of discrepancy on the association between reflection and self-concept change, we again calculated a random intercept model with the combined data set, this time including the direction of discrepancy as well as its interaction with reflection as additional fixed effects.⁷ The results are displayed in Table 2. The interaction effect between reflection and the direction of discrepancy was significant, B = −0.07, p < .001.⁸ To explore this effect further, we conducted simple slopes analyses using the reghelper package (Hughes & Beiner, 2021). We found that even though the association between reflection and self-concept change was significant in both groups, it was stronger for negative, B = 0.22, SE = 0.02, t(2414.55) = 10.58, p < .001, than for positive discrepancies, B = 0.07, SE = 0.02, t(2,414.68) = 3.87, p < .001.

Discussion

Our results provide correlational evidence for a positive association between reflection and self-concept change, which is in line with the literature on personality change, wisdom, and narrative identity. Across a variety of self-relevant domains, we found that the more a person reflects about discrepant, self-relevant feedback and its implications for the self-concept, the larger the changes in their self-concept. In exploratory analyses, we furthermore found that this association is stronger after receiving negative feedback (i.e., worse than expected) compared to positive feedback (i.e., better than expected). One drawback of all of these studies is that they only provide correlational evidence for a link between reflection and self-concept change. Yet, in the respective literatures, such as the TESSERA framework, reflection is considered a necessary condition for self-concept change after receiving self-relevant feedback. The first five studies reported here are unable to test this causality in a rigorous fashion. One alternative explanation for the findings we reported here could be that the correlation between reflection and self-concept change was produced by a common causal factor (such as self-concept clarity: lower self-concept clarity is associated with more self-reflection; Campbell et al., 1996; lower self-concept clarity is also associated with more self-concept change; Carter & Bruene, 2019). To test the causal effect of reflection on self-concept change after receiving discrepant feedback more rigorously, we decided to pursue an experimental approach in which we manipulated reflection experimentally.

Study 2

Study 2 was designed to test the effect of reflection on self-concept change after discrepant feedback in an experimental fashion. To do so, we used a similar design as in the previous studies but included an experimental manipulation geared towards either inducing or suppressing reflection.

Reflection on self-relevant information is a complex cognitive process that might include, for example, evaluating the relevance of the feedback or integrating it into one’s existing beliefs. As such, reflection requires working memory capacity, a limited resource needed for temporarily storing and processing information (Baddeley, 1992; Hitch & Baddeley, 1976). In this study, we capitalized on the fact that working memory capacity is limited: By creating cognitive load through a distractor task in one of two experimental conditions, we aimed at occupying the available working memory capacity and, thus, at suppressing reflection about the feedback. Such an approach has been successfully applied to limit the cognitive resources available for a specific task (Dijkstra & Hong, 2019; Johnson et al., 2014). In the other experimental condition, we aimed at inducing reflection by encouraging participants to reflect about the feedback and its implications via several prompts. Previous research has demonstrated that structured reflection via prompts can increase the depth as well as quality of reflection and is beneficial for learning outcomes (Cengiz, 2020; Chen et al., 2009; DeRue et al., 2012). We hypothesized that self-concept change would be larger after reflection was induced than after it was suppressed.⁹

Method

Procedures and Measures

Participants were informed that the study would consist of two blocks. They were told that the first block would be about procedural thinking, that is, the ability to grasp complex problems and solve them efficiently by imagining processes and/or spatial motion. Participants received a detailed definition of procedural thinking and were then asked to indicate their self-concept by rating their self-perceived ability for procedural thinking via four items (e.g., “I am able to quickly find effective solutions to procedural problems”; scale from 1 = strongly disagree to 6 = strongly agree; α_t1 = .88; M_t1 = 4.18, SD_t1 = 0.83; adapted from Heppner & Petersen, 1982).

Next, they were asked to work on two sets of tasks measuring their ability for procedural thinking. The two sets were presented in random order. Each set of tasks consisted of five subtasks. Participants were tasked with solving as many subtasks as they could within a given time limit. In the first set of tasks, participants were shown several interconnected gear wheels. A starting cue indicated in which direction the first wheel turned. They were then asked to determine in which direction one or several of the other wheels turned. Each subtask displayed a different arrangement of gear wheels and was considered correctly solved only if participants indicated the correct direction for all relevant gear wheels (tasks based on similar tasks by Bennett & Fry, 1969). The time limit for this set of tasks was 120 s. The second set of tasks consisted of subtasks showing several spatial geometric figures. Participants had to mentally combine these figures. They were presented with four options of what the combined figure looked like and had to choose the correct one (tasks adapted from Berkowitz et al., 2020). The time limit for this set of tasks was 160 s.

For all subtasks, participants received feedback on whether they had correctly solved the subtask directly after submitting their response. The feedback either read “Your answer is correct” or “Your answer is incorrect.” If participants did not manage to respond to all subtasks within the time limit, the remaining subtasks were considered incorrectly solved; participants were informed accordingly for each unsolved task. In total, participants therefore received feedback ten times, once for each subtask.

After completing the two sets of tasks, the second block of the study began. At this point, participants were randomly assigned to one of the two experimental conditions. In the inducing reflection condition, participants were asked to take three to five minutes to reflect on their ability for procedural thinking. They were instructed to respond to several open-ended questions aimed at prompting in-depth reflection during this time. The questions were “After having indicated your self-perception and worked on the tasks on procedural thinking, what are you thinking right now?,” “How did you perceive your ability for procedural thinking while working on the tasks?,” “Regarding your self-perception and the tasks on procedural thinking, how are you feeling right now?,” “Which insights did you gain after indicating your self-perception and working on the tasks on procedural thinking?,” and “Do you have any further thoughts?.” In the suppressing reflection condition, participants worked on a cognitively demanding distractor task: They were asked to repeatedly memorize lists of 10 sequentially presented nouns and reproduce them in a free recall format (words were taken from Kroneisen et al., 2014). Similar tasks have been shown to demand cognitive resources (Kroneisen et al., 2014) and have been used to induce cognitive load in previous research (Kidder et al., 1997; Shears et al., 2007). After three trials of the memory task, participants were presented with a fourth list of nouns but were not immediately asked to reproduce them. Instead, they were asked to remember the nouns until they had to reproduce them at a later point in time.

All participants were then asked to again rate their self-perceived ability for procedural thinking via the same four items used at the beginning of the study (α_t2 = .90; M_t2 = 3.65, SD_t2 = 1.06). Following the procedure applied in Studies 1a–e, we calculated the self-concept change score by subtracting participants mean self-rated ability at the first from that at the second measurement occasion. Absolute values of this score served as the final self-concept change score used in our analyses (M_SCC = 0.70, SD_SCC = 0.69). After their second self-rating, participants in the suppressing reflection condition were asked to reproduce the words they had memorized in the fourth trial of the memory task. In the end, we asked all participants to indicate how much they had reflected on their ability for procedural thinking (i.e., “After completing the procedural thinking tasks, I reflected extensively on my ability for procedural thinking,” scale from 1 = strongly disagree to 6 = strongly agree; M = 3.31, SD = 1.15). This was used as a manipulation check.¹⁰

Sample

Participants were recruited through a university mailing list, personal networks, and the PsyWeb panel (see https://psyweb.uni-muenster.de/). In return for their participation, participants could take part in a raffle for vouchers. We collected data until the date specified in our preregistration, which resulted in 1,399 data sets. We applied the preregistered exclusion criteria and excluded data from n = 145 participants. In addition, we again excluded participants with no discrepancy between feedback and initial self-concept as they were not relevant to our research question, resulting in n = 99 further exclusions.¹¹ Furthermore, we excluded the data from n = 6 participants who had participated more than once (i.e., their participant IDs occurred more than once in the dataset). Thus, the final sample consisted of N = 1,149 participants (M_age = 43.90 years, SD_age = 15.36 years; 732 female, 404 male, 12 “other,” one non-response). According to a sensitivity power analysis conducted in G*Power (α = .05, 1 − β = .80, sample size group 1 = 564, sample size group 2 = 585; Faul et al., 2007), we could detect an effect of Cohen’s d = 0.15 for our main analysis with this sample.

Results

Manipulation Check

First, we conducted a manipulation check to determine whether our experimental manipulation was successful. Calculating a Welch’s two-sample t-test, we found that self-reported reflection was higher in the inducing reflection condition (M = 3.46, SD = 1.17) than in the suppressing reflection condition (M = 3.16, SD = 1.10), t(1135.70) = 4.44, p < .001, Hedges’ g = 0.26, 95% CI [0.15, 0.38] (effsize package; Torchiano, 2020).¹² Thus, our manipulation produced a small to medium sized effect on self-reported reflection.

Main Analysis

To test whether self-concept change was larger after inducing than after suppressing reflection, we calculated a Welch’s two-sample t-test. We found the expected effects, with self-concept change being larger in the inducing (M = 0.76, SD = 0.74) than in the suppressing reflection condition (M = 0.64, SD = 0.63), t(1100.20) = 2.79, p = .005, Hedges’ g = .17, 95% CI [0.05, 0.28] (effsize package; Torchiano, 2020).

Exploratory Analyses

Like in Studies 1a–e, we explored whether the effect of reflection on self-concept change was moderated by the direction of the discrepancy—that is, whether the effect was larger for negative compared to positive discrepancies. First, we computed the difference between self-perceptions at the first measurement occasion and task feedback scores; second, we dichotomized this variable into negative and positive discrepancies.

Next, we specified a multiple regression model with reflection (“suppressing reflection” = −1, “inducing reflection” = 1), direction of discrepancy (“negative” = −1 and “positive” = 1), and the interaction of the two as predictor terms and self-concept change as the dependent variable. Overall, the model explained 4% of the variance in self-concept change, F(3, 1145) = 15.11, p < .001, R² = .04, 95% CI [0.02, 0.06]. Our analysis revealed significant main effects for reflection, B = 0.06, t(1145) = 2.49, p = .013, 95% CI for B [0.01, 0.11], sr² = .01, and direction of discrepancy, B = −0.15, t(1145) = −6.10, p < .001, 95% CI for B [−0.19, −0.10], sr² = .03. The interaction term was not significant, B = −0.00, t(1145) = −0.09, p = .928, 95% CI for B [−0.05, 0.05], sr² = .00.^13,14

Discussion

Consistent with the findings from Studies 1a–e, Study 2 provides experimental evidence that reflection leads to self-concept change after discrepant feedback. The small effect we observed is consistent with theoretical accounts as single instances of reflecting on self-relevant experiences should produce only small changes in the self-concept that should, over time and after repeated reflection on self-relevant experiences, condense into larger self-concept changes (e.g., Wrzus & Roberts, 2017). While we found that the association between reflection and self-concept change was larger for negative compared to positive discrepancies in Studies 1a–e, we did not find such a moderating effect in Study 2.

General Discussion

In six studies, we examined the association between reflection after self-relevant feedback and self-concept change. We further explored the effect of the direction of discrepancy on this association. Across different self-concept domains, we found that reflection was positively associated with self-concept change: The more participants reflected about the feedback and its implications for the self, the more they changed their self-concept. This is in line with recent theorizing on the mechanism behind changes in the self-concept (Wrzus & Roberts, 2017) and supports previous, tentative evidence for a link between reflection and self-concept change (Quintus et al., 2021). Regarding the direction of discrepancy as a moderator, the aggregated results from Studies 1a–e suggest that the association between reflection and self-concept change was more pronounced for negative than for positive discrepancies. In Study 2, the reflection × direction interaction was not significant. One potential issue with the design of Study 2 is that there were more participants receiving negative compared to positive feedback (i.e., task scores were more often worse [as compared to better] than participants’ expectations indicated in their self-concept self-reports). This imbalance may have made it more difficult to detect small differences in the reflection effects between the groups. Future research should therefore examine the interaction between reflection and the direction of discrepancy in an experimental design that aims for a more balanced distribution of positive and negative deviations from one’s prior expectations (i.e., self-concept).

Theoretical Implications

The present research adds to the existing literature by investigating the role of reflection after receiving self-discrepant feedback systematically, rigorously, and with considerable contextual breadth (i.e., for different self-concept domains, with different approaches, etc.). Theoretical work such as the TESSERA framework and research on narrative identities or wisdom have suggested reflection to be a decisive factor for self-concept change. Yet, little quantitative empirical research has tested this assumption. One initial test of the TESSERA framework examined the role of reflection in a longitudinal design (Quintus et al., 2021). Yet, this study did not look at single instances of receiving and reflecting on self-relevant feedback. Our own research allowed for a more fine-grained investigation of the link between reflection and self-concept change. We found that reflection was associated with self-concept change even after single instances of self-relevant feedback. In our studies, we examined a variety of self-concept dimensions, from intelligence subfacets (Studies 1c and 1e) to abilities (Studies 1a, 1d, and 2). Furthermore, we did not only use correlational designs, but also experimentally induced or suppressed reflection in Study 2 and found that inducing reflection led to more self-concept change than suppressing reflection.

Our findings contribute to the literature on self-concept change in several regards. First, our research provides a very conservative test of the theoretical idea that reflection is the pathway to self-concept change: In each study, we presented participants with self-relevant feedback and assessed or manipulated reflection as well as self-concept change. In doing so, we examined isolated instances of receiving and reflecting on self-relevant information. In line with theory, the effects we found in our studies are small. Over time and across several instances, these small effects should accumulate into larger changes in people’s self-concept (Wrzus & Roberts, 2017).

Second, it is by no means trivial to assume that reflection is positively associated with self-concept change: Reflecting about self-discrepant feedback may also trigger cognitive processes that eventually stabilize one’s self-concept. For instance, reflection might enable individuals to strategically shield their self-concept against self-discrepant feedback (“immunization” strategies; Brandtstädter & Greve, 1994). This might include denying the relevance of a discrepant information (e.g., “When my boss tells me I work too little this is irrelevant because she doesn’t know how many hours I really work”; “data-oriented immunization”) or redefining whether the discrepant information is self-relevant regarding the self-concept domain in question (e.g., “When my boss tells me I work too little, this doesn’t impact my self-concept as a hard worker—how many hours I work is not relevant here, it’s about how much I get done in the hours I do work”; “concept-oriented immunization”). In addition, reflection may motivate people to reduce self-concept discrepancies by trying to manage the impression they make on other people, or to work harder so that, next time, their actual abilities are more in line with their self-concept (“assimilation” strategies; Brandtstädter & Greve, 1994; Pinquart et al., 2021). Single instances of feedback—such as the ones we used here—might be particularly likely to invite immunization or assimilation strategies: It is easier to discredit a single feedback or avoid similar feedback in the future than to adapt one’s self-concept. Supporting this notion, previous research in the context of expectation violations has shown that when expectation violations happen once or only under specific circumstances, the inconsistent information is discredited or behavior changes are made rather than adapting one’s expectations (Pietzsch & Pinquart, 2023). In light of these findings, it is particularly noteworthy that we consistently found a positive association between reflection and self-concept change even after single instances of feedback.¹⁵

As an additional contribution to the existing literature, we produce first evidence on the impact of the direction of discrepancy on the association between reflection and self-concept change. Previous research suggests that positive and negative information are processed differently, and that this might affect processes of self-concept change (Unkelbach et al., 2020; Wrzus & Roberts, 2017). In the present research, we found correlational evidence for a stronger link between reflection and self-concept change after negative than after positive feedback. In Study 2, however, we could not replicate these results in an experimental design. Therefore, while we find some evidence that reflection might produce more self-concept change after negative than after positive feedback, this effect should be further investigated in future studies.

Limitations and Future Research Directions

Despite its contributions to the literature on self-concept change, there are some limitations to the present research and avenues for future research arising from our results. One such limitation lies in our operationalization of reflection as well as self-concept change. When measuring or manipulating reflection, we targeted reflection in general without restricting its content. Furthermore, we measured and manipulated reflection shortly after participants had received self-relevant feedback. Yet, reflective processes can manifest in many different forms and range from briefly focusing one’s attention on the feedback to engaging deeply with the implication of such feedback for one’s self-concept and one’s social interactions. There is currently no theoretical framework that helps inform us how long such processes last and which types of reflection should be most impactful in influencing self-concept change. Furthermore, it is plausible that people differ in how they process self-relevant information (Wrzus, 2021) and that situational factors impact how much and about what people reflect (e.g., people might reflect more in an artificial study design even after mildly discrepant feedback in comparison to a more naturalistic setting).

Similar points apply to self-concept change: While we measured self-concept change shortly after participants had received their feedback, it is plausible that such processes can last past this point and differ between individuals. Moreover, it is plausible that some aspects of the self might be more difficult to change than others (e.g., aspects that are central to a person’s identity might be more difficult to change than peripheral aspects¹⁶). While we studied changes in different aspects of the self-concept in all studies, we were only able to examine a selection of potentially interesting aspects. Furthermore, the way we measured self-concept change enabled us to indirectly assess changes in self-perceptions. However, our measure was a self-report measure and suffered from shortcomings such as being influenced by motivated information processing, cognitive biases, or idiosyncratic motives (such as self-enhancement). That being said, self-reports are useful because they indicate the declarative aspects of a person’s self-concept in the most direct fashion.

One might additionally argue that the self-concept changes we measured in our studies may not necessarily reflect “true” changes in one’s self-concept, but rather participants’ responses to the experimental demand that was created as part of our feedback paradigm. While this might, in principle, be a valid concern (Coles et al., 2023), it remains an empirical question to what extent our measured self-concept changes are influenced by demand effects. One way to look at this question empirically is to revisit the data collected in Studies 1a and 2. In Study 1a, the final part of the survey contained one question in which participants were asked to report their perception of the researchers’ assumptions on whether they expected to find self-concept change after discrepant feedback or not. Adding participants’ responses to that question as a moderator of the effect of reflection on self-concept change in Study 1a did not change the pattern of results (i.e., reflection was a significant predictor, and the interaction effect was not significant). More details on the results regarding the perceived researcher assumption, together with the results of an additional study on the role of demand effects in self-concept change after discrepant feedback, are reported in Brotzeller & Gollwitzer (2025). In Study 2, apart from directly measuring self-concept change, we also assessed participants’ self-reported willingness to adapt their self-concept in response to the feedback. If the differences in self-concept change between reflection conditions were due to experimental demand, we would expect to find the same pattern of results for willingness to adapt the self-concept. This is not the case, however: An additional analysis comparing the mean willingness to adapt the self-concept in the two reflection conditions did not result in significant differences between groups. Based on these analyses, the assumption that our effects may have been merely the result of experimental demand seems unlikely.

Nevertheless, we acknowledge the limitations in our operationalization of reflection and self-concept change and call for future research to measure and manipulate reflection as well as self-concept change more comprehensively. This should be done in different settings as well as for a variety of self-concept aspects not yet studied in the present research to expand the generalizability of the findings. One focus for future research could lie in measuring the content and extent of reflection in more detail and over a longer period to investigate its progression and gain deeper insights into its association with self-concept change. When manipulating reflection, future research might develop stronger experimental manipulations than the one we applied in Study 2. While we employed prompts as well as a distractor task, which had been found to be effective techniques for inducing and suppressing reflection (Chen et al., 2009; Shears et al., 2007), our manipulation check showed that this only produced a small to medium effect. To increase the effectiveness of such a manipulation, future research could, for example, improve the prompts that participants receive or increase the time spent on reflecting about self-discrepant feedback (and, in the contrasting condition, time spent working on the distractor task in order to suppress reflection). Furthermore, including a control condition in which reflection is neither suppressed nor induced would allow to determine whether the positive effect of inducing reflection on self-concept change is smaller or larger than the negative effect of suppressing reflection on self-concept change. This could help develop more effective experimental manipulations for future studies.

Moreover, while we only examined the main effect of reflection on self-concept change and merely looked at one potential moderator for this association (i.e., direction of discrepancy), future research should take into account further factors that might play a role in this context. Person characteristics of interest might, for example, be self-evaluation motives such as self-enhancement and self-verification (i.e., the motivations to view oneself as positively as possible and to confirm one’s existing self-concept, respectively; Wrzus, 2021). Previous research has shown that how people perceive themselves does not merely reflect their past thoughts, feelings, and behaviors, but is also influenced by inaccurate information processing and motivational processes (Schriber & Robins, 2012). It seems plausible to assume that self-evaluation motives might also impact how one processes and reflects on self-discrepant information and subsequently adapts one’s self-concept. Future research should examine these and other person characteristics that might be of interest here, such as intellectual performance or personality traits (e.g., openness or need for cognition¹⁷). Another factor potentially playing a role might be the source of the self-discrepant information. Previous research has demonstrated, for example, that a higher source credibility is linked with a higher perceived feedback accuracy and feedback acceptance (Lechermeier & Fassnacht, 2018). It is plausible to assume that the source credibility might also impact the duration and content of reflection on discrepant feedback, with feedback from a more credible source leading to more in-depth reflection, which, in turn, is more likely to result in self-concept change. Feedback based on a task in a scientific study, for example, might be perceived as more credible than some forms of feedback a person receives in daily life, while it is likely less credible than others (e.g., it might be more credible than feedback from a stranger on the street but less credible than feedback from a family member who has known the person for years and has much more information to base the feedback on). Future research should further examine the role of such factors. Furthermore, future studies could expand on our research by examining the interaction between reflection and the direction of discrepancy in a more rigid fashion.

Lastly, starting from our findings on the effects of reflecting about single self-relevant experiences, future research should aim to gain a better understanding of long-term processes of reflection and self-concept change. One approach could, for example, be to expand on our research by examining reflection on not one but several instances of self-relevant experiences and by measuring subsequent reflection as well as self-concept change. Producing several such experiences in a controlled setting and assessing reflection as well as self-concept change over longer periods would help us understand conditions for sustainable changes in one’s self-concept.

Conclusion

While reflection on self-relevant experiences has been theorized to lead to self-concept change, there is little research examining their association. The present research provides correlational and experimental evidence that reflection is positively associated with self-concept change. We further find some evidence that this association might be more pronounced for reflection after negative than after positive feedback.

Footnotes

We would like to thank several people for their help in creating study materials and collecting data: Elena Cadeggianini for her help with Study 1a,Nia Pavlova for her help with Study 1b,Elèna Lombardo for her help with Study 1c,Veronika Fischer,Alina Kolb,Emma Runge,Vivian Swoboda,and Leonie Thürmer for their help with Study 1e,and Benedikt Ruhdorfer as well as David Wetterich for their help with Study 2.

Author Note

Study materials,data sets,and analysis scripts are openly available at the project’s Open Science Framework page ( https://osf.io/cfvjs/ ). Studies 1a–e were originally designed to test other research questions (see preregistrations at https://aspredicted.org/31S_M1M for Study 1a;https://aspredicted.org/VYH_M31 for Study 1b;https://aspredicted.org/3TQ_Y5S for Study 1d;https://aspredicted.org/R33_23D for Study 1e;Study 1c was not preregistered). Thus,the hypotheses tested here differ from those that were preregistered for those studies. Study 2 is reported here exactly as it was preregistered (

Declaration of Conflicting Interests

The author(s) declared no potential conflicts of interest with respect to the research,authorship,and/or publication of this article.

Funding

The author(s) received no financial support for the research,authorship,and/or publication of this article.

ORCID iDs

Franziska Brotzeller

Gideon L. Spägele

Mario Gollwitzer

Supplemental Material

Supplemental material is available online at

References

Baddeley

(1992). Working memory. Science, 255(5044), 556–559. https://doi.org/10.1126/science.1736359

Bates

Mächler

Bolker

B. M.

Walker

S. C.

(2015). Fitting linear mixed-effects models using lme4. Journal of Statistical Software, 67(1), 1–48. https://doi.org/10.18637/jss.v067.i01

Beißert

Köhler

Rempel

Beierlein

(2014). Eine deutschsprachige Kurzskala zur Messung des Konstrukts Need for Cognition: Die Need for Cognition Kurzskala (NFC-K) (GESIS-Working Papers, 2014/32). GESIS—Leibniz-Institut für Sozialwissenschaften. https://nbn-resolving.org/urn:nbn:de:0168-ssoar-403157

Bennett

G. K.

Fry

D. E.

(1969). Bennett mechanical comprehension test. Psychological Corporation.

Berkowitz

Gerber

Thurn

C. M.

Emo

Hoelscher

Stern

(2020). Novel spatial ability tests. osf.io/jf5mx

Bleidorn

Hopwood

C. J.

Lucas

R. E.

(2018). Life events and personality trait change. Journal of Personality, 86(1), 83–96. https://doi.org/10.1111/jopy.12286

Bölte

(2005). Reading the mind in the eyes test für Erwachsene (dt. Fassung) von S. Baron-Cohen. J.W. Goethe-Universität Frankfurt/M.

Boyd

E. M.

Fales

A. W.

(1983). Reflective learning: Key to learning from experience. Journal of Humanistic Psychology, 23(2), 99–117. https://doi.org/10.1177/0022167883232011

Brandtstädter

Greve

(1994). The aging self: Stabilizing and protective processes. Developmental Review, 14(1), 52–80. https://doi.org/10.1006/drev.1994.1003

10.

Brotzeller

Gollwitzer

(2024). Exploring asymmetries in self-concept change after discrepant feedback. Personality and Social Psychology Bulletin, 01461672241232738. https://doi.org/10.1177/01461672241232738

11.

Brotzeller

Gollwitzer

(2025). The role of demand effects in studying self-concept change after discrepant feedback. Manuscript in preparation.

12.

Campbell

J. D.

Trapnell

P. D.

Heine

S. J.

Katz

I. M.

Lavallee

L. F.

Lehman

D. R.

(1996). Self-concept clarity: Measurement, personality correlates, and cultural boundaries. Journal of Personality and Social Psychology, 70, 141–156. https://doi.org/10.1037/0022-3514.70.1.141

13.

Carter

M. J.

Bruene

(2019). Examining the relationship between self-perceptions of person, role, and social identity change and self-concept clarity. Imagination, Cognition and Personality, 38(4), 425–451. https://doi.org/10.1177/0276236618792267

14.

Cengiz

(2020). The effect of structured journals on reflection levels: With or without question prompts? Australian Journal of Teacher Education, 45(2), 23–43. https://doi.org/10.3316/ielapa.192236383097301

15.

Chen

N.-S.

Wei

C.-W.

K.-T.

Uden

(2009). Effects of high level prompts and peer assessment on online learners’ reflection levels. Computers & Education, 52(2), 283–291. https://doi.org/10.1016/j.compedu.2008.08.007

16.

Coles

N. A.

Wyatt

Frank

M. C.

(2023). A quantitative review of demand characteristics and their underlying mechanisms. PsyArXiv. https://doi.org/10.31234/osf.io/uw85a

17.

DeRue

D. S.

Nahrgang

J. D.

Hollenbeck

J. R.

Workman

(2012). A quasi-experimental study of after-event reviews and leadership development. Journal of Applied Psychology, 97(5), 997–1015. https://doi.org/10.1037/a0028244

18.

Dijkstra

K. A.

Hong

(2019). The feeling of throwing good money after bad: The role of affective reaction in the sunk-cost fallacy. PLoS ONE, 14(1), e0209900. https://doi.org/10.1371/journal.pone.0209900

19.

Elder

Davis

Hughes

B. L.

(2022). Learning about the self: Motives for coherence and positivity constrain learning from self-relevant social feedback. Psychological Science, 33(4), 629–647. https://doi.org/10.1177/09567976211045934

20.

Faul

Erdfelder

Lang

A.-G.

Buchner

(2007). G*Power 3: A flexible statistical power analysis program for the social, behavioral, and biomedical sciences. Behavior Research Methods, 39(2), 175–191. https://doi.org/10.3758/BF03193146

21.

Formann

A. K.

Waldherr

Piswanger

(2011). Wiener Matrizen-Test 2: Ein Rasch-skaldierter sprachfreier Kurztest zu Erfassung der Intelligenz; WMT-2. Beltz Test, Hogrefe.

22.

Glück

Weststrate

N. M.

(2022). The wisdom researchers and the elephant: An integrative model of wise behavior. Personality and Social Psychology Review, 26(4), 342–374. https://doi.org/10.1177/10888683221094650

23.

Green

MacLeod

C. J.

(2016). SIMR: An R package for power analysis of generalized linear mixed models by simulation. Methods in Ecology and Evolution, 7(4), 493–498. https://doi.org/10.1111/2041-210X.12504

24.

Hakmiller

K. L.

(1966). Threat as a determinant of downward comparison. Journal of Experimental Social Psychology, 1, 32–39. https://doi.org/10.1016/0022-1031(66)90063-1

25.

Heppner

P. P.

Petersen

C. H.

(1982). The development and implications of a personal problem-solving inventory. Journal of Counseling Psychology, 29(1), 66–75. https://doi.org/10.1037/0022-0167.29.1.66

26.

Hitch

G. J.

Baddeley

A. D.

(1976). Verbal reasoning and working memory. Quarterly Journal of Experimental Psychology, 28(4), 603–621. https://doi.org/10.1080/14640747608400587

27.

Hughes

Beiner

(2021). Reghelper: Helper functions for regression analysis. R package version 1.1.0. https://CRAN.R-project.org/package=reghelper

28.

Johnson

E. D.

Tubau

De Neys

(2014). The unbearable burden of executive load on cognitive reflection: A validation of dual process theory. Proceedings of the Annual Meeting of the Cognitive Science Society, 36(36), 2441–2446.

29.

Kersting

Althoff

Jäger

A. O.

(2008). Wilde-Intelligenz-Test 2: WIT-2. Hogrefe.

30.

Kidder

D. P.

Park

D. C.

Hertzog

Morrell

R. W.

(1997). Prospective memory and aging: The effects of working memory and prospective memory task load. Aging, Neuropsychology, and Cognition, 4(2), 93–112. https://doi.org/10.1080/13825589708256639

31.

Kroneisen

Rummel

Erdfelder

(2014). Working memory load eliminates the survival processing effect. Memory, 22(1), 92–102. https://doi.org/10.1080/09658211.2013.815217

32.

Lechermeier

Fassnacht

(2018). How do performance feedback characteristics influence recipients’ reactions? A state-of-the-art review on feedback source, timing, and valence effects. Management Review Quarterly, 68(2), 145–193. https://doi.org/10.1007/s11301-018-0136-8

33.

Leiner

(2021). SoSci survey (Version 3.2.29) [Computer software]. https://www.soscisurvey.de

34.

McAdams

D. P.

McLean

K. C.

(2013). Narrative identity. Current Directions in Psychological Science, 22(3), 233–238.

35.

Miller

J. P.

(2020). Finding truth within: Exploring the importance of reflective practice in deepening self-knowledge. Integral Review, 16(1), 357–390.

36.

Morin

(2011). Self-awareness part 1: Definition, measures, effects, functions, and antecedents. Social and Personality Psychology Compass, 5(10), 807–823. https://doi.org/10.1111/j.1751-9004.2011.00387.x

37.

Müller-Pinzler

Czekalla

Mayer

A. V.

Stolz

D. S.

Gazzola

Keysers

Paulus

F. M.

Krach

(2019). Negativity-bias in forming beliefs about own abilities. Scientific Reports, 9(1), Article 1. https://doi.org/10.1038/s41598-019-50821-w

38.

Overeem

Wollersheim

Driessen

Lombarts

Van De Ven

Grol

Arah

(2009). Doctors’ perceptions of why 360-degree feedback does (not) work: A qualitative study. Medical Education, 43(9), 874–882. https://doi.org/10.1111/j.1365-2923.2009.03439.x

39.

Pals

J. L.

(2006). Narrative identity processing of difficult life experiences: Pathways of personality development and positive self-transformation in adulthood. Journal of Personality, 74(4), 1079–1110. https://doi.org/10.1111/j.1467-6494.2006.00403.x

40.

Pietzsch

M. C.

Pinquart

(2023). Predicting coping with expectation violations: Combining the ViolEx model and the covariation principle. Frontiers in Psychology, 14, 1152261. https://doi.org/10.3389/fpsyg.2023.1152261

41.

Pinquart

Rothers

Gollwitzer

Khosrowtaj

Pietzsch

Panitz

(2021). Predictors of coping with expectation violation: An integrative review. Review of General Psychology, 25(3), 321–333.

42.

Quintus

Egloff

Wrzus

(2021). Daily life processes predict long-term development in explicit and implicit representations of Big Five traits: Testing predictions from the TESSERA (triggering situations, expectancies, states and state expressions, and reactions) framework. Journal of Personality and Social Psychology, 120(4), 1049–1073. https://doi.org/10.1037/pspp0000361

43.

R Core Team. (2023). R: A language and environment for statistical computing. R Foundation for Statistical Computing. http://www.r-project.org/

44.

Rakshasbhuvankar

(2021). Individual participant data (IPD) meta-analysis. In Patole

(Ed.), Principles and practice of systematic reviews and meta-analysis (pp. 147–155). Springer International Publishing. https://doi.org/10.1007/978-3-030-71921-0_14

45.

Raudenbush

S. W.

Bryk

A. S.

(1985). Empirical Bayes meta-analysis. Journal of Educational Statistics, 10(2), 75–98. https://doi.org/10.3102/10769986010002075

46.

Rodriguez Añez

C. R.

Reis

R. S.

Petroski

E. L

. (2008). Brazilian version of a lifestyle questionnaire: Translation and validation for young adults. Arquivos Brasileiros de Cardiologia, 91, 102–109. https://doi.org/10.1590/S0066-782X2008001400006

47.

Rogers

R. R.

(2001). Reflection in higher education: A concept analysis. Innovative Higher Education, 26(1), 37–57. https://doi.org/10.1023/A:1010986404527

48.

Sargeant

J. M.

Mann

K. V.

van der Vleuten

C. P.

Metsemakers

J. F.

(2009). Reflection: A link between receiving and using assessment feedback. Advances in Health Sciences Education, 14(3), 399–410. https://doi.org/10.1007/s10459-008-9124-4

49.

Schriber

R. A.

Robins

R. W.

(2012). Self-knowledge: An individual-differences perspective. In Vazire

Wilson

T. D.

(Eds.), Handbook of self-knowledge (pp. 105–127). The Guilford Press.

50.

Sedikides

Skowronski

J. J.

(1995). On the sources of self-knowledge: The perceived primacy of self-reflection. Journal of Social and Clinical Psychology, 14(3), 244–270. https://doi.org/10.1521/jscp.1995.14.3.244

51.

Shavelson

R. J.

Hubner

J. J.

Stanton

G. C.

(1976). Self-concept: Validation of construct interpretations. Review of Educational Research, 46(3), 407–441.

52.

Shears

Miller

Ball

Hawkins

Griggs

Varner

(2007). Cognitive demand differences in causal inferences: Characters’ plans are more difficult to comprehend than physical causation. Discourse Processes, 43(3), 255–278. https://doi.org/10.1080/01638530701226238

53.

Tangney

J. P.

Tracy

J. L.

(2012). Self-conscious emotions. In Handbook of self and identity (2nd ed., pp. 446–478). The Guilford Press.

54.

Torchiano

(2020). _effsize: Efficient effect size computation_. https://doi.org/10.5281/zenodo.1480624

55.

Trepte

Verbeet

(2010). Allgemeinbildung in Deutschland: Erkenntnisse aus dem SPIEGEL-Studentenpisa-Test. Springer-Verlag.

56.

Trepte

Verbeet

(2021). Spiegel-Studentenpisa-Test. GESIS Datenarchiv, Köln. ZA6268 Datenfile Version 1.0.0. https://doi.org/10.4232/1.13780

57.

Unkelbach

Alves

Koch

. (2020). Chapter Three—Negativity bias, positivity bias, and valence asymmetries: Explaining the differential processing of positive and negative information. In Gawronski

(Ed.), Advances in experimental social psychology (Vol. 62, pp. 115–187). Academic Press. https://doi.org/10.1016/bs.aesp.2020.04.005

58.

Wilson

D. M. C.

Nielsen

Ciliska

(1984). Lifestyle assessment: Testing the FANTASTIC instrument. Canadian Family Physician, 30, 1863–1866.

59.

Wrzus

(2021). Processes of personality development: An update of the TESSERA framework. In Rauthmann

J. F.

(Ed.), The Handbook of Personality Dynamics and Processes (pp. 101–123). Elsevier. https://doi.org/10.1016/B978-0-12-813995-0.00005-4

60.

Wrzus

Roberts

B. W.

(2017). Processes of personality development in adulthood: The TESSERA framework. Personality and Social Psychology Review, 21(3), 253–277. https://doi.org/10.1177/1088868316652279

I Think,Therefore I Change: Reflection Leads to Self-Concept Change After Discrepant Feedback

Abstract

Keywords

Conceptualizing Reflection

The Effect of Reflection on Self-Concept Change

The Direction of Discrepancy as a Potential Moderator

The Present Research

Studies 1a–e

Method

Procedures and Measures

Samples

Study 1a

Study 1b

Study 1c

Study 1d

Study 1e

Results

Main Analysis

Exploratory Analyses

Discussion

Study 2

Method

Procedures and Measures

Sample

Results

Manipulation Check

Main Analysis

Exploratory Analyses

Discussion

General Discussion

Theoretical Implications

Limitations and Future Research Directions

Conclusion

Footnotes

Author Note

Declaration of Conflicting Interests

Funding

ORCID iDs

Supplemental Material

References