This is an open access article distributed under the terms of the Creative Commons
Attribution License (Labeling individuals based on performance potential is a fact of organizational life and many companies attempt to identify high potential employees early in their employees’ careers (Cappelli, 2008; Collings & Mellahi, 2009; Karakowsky & Kotlyar, 2011; Slan-Jerusalim & Hausdorf, 2007). In this paper, we investigate whether the initial expectations reflected in the high potential label influence reactions to performance feedback. Specifically, we are concerned with the effect of labeling individuals in terms of potential performance capabilities on their subsequent task commitment and performance satisfaction following positive and negative feedback. This study makes a contribution by demonstrating that a high potential label can act as a double edge sword, whereby, on the one hand, it can boost performance expectations, while, on the other, it can contribute to a sharper drop in performance satisfaction and task commitment, potentially reducing engagement and retention of the very employees organizations consider most valuable for their future success.
Literature Review and Hypotheses
Typically the label of high potential refers to those employees who are perceived as possessing the talents and abilities necessary for advancement in the organization (Cappelli, 2008; Collings & Mellahi, 2009; Iles, 1997; Slan-Jerusalim & Hausdorf, 2007). Having identified their high potential employees, organizations tend to inform these individuals of their status to facilitate their career development, advancement and retention (Ready, Conger, Hill, & Stecker, 2010). While some organizations are secretive about their high potential list, even they acknowledge the challenges of protecting such information, since employees can usually figure out who the high potentials are by the type of assignments people receive, who they have lunch with, who sponsors or mentors them, and their level of visibility in the organization (Derr, Jones, & Toomey, 1988, p. 276).
Communicating high expectations to subordinates, which may include labeling them as high potentials, can have beneficial effects (e.g., Eden, 1984; Tierney & Farmer, 2004). However, not much is known about the role of high expectations when performance feedback repeatedly falls short of such expectations. In organizations, evaluative feedback, including performance reviews (Ashford, 1986), is considered essential for employee development and improvement (Halpern, 2004; Kluger & DeNisi, 1996; Maki, 1998; Morrison & Brantner, 1992). In particular, negative feedback is seen as helpful in this regard, as it creates awareness of discrepancies and motivates individuals to either work harder or to modify their behavioral strategies in order to reduce these discrepancies (e.g., Bandura, 1997; Klein, 1989; Locke & Latham, 1990). However, people tend to feel dissatisfied and discouraged in response to negative feedback (Brett & Atwater, 2001; Kluger, Lewinsohn, & Aiello, 1994), and, in particular, individuals labeled as high potentials may be especially sensitive to feedback, because the gap between high expectations and their actual performance challenges their alleged high potential. Thus, we propose that individuals labeled as high potentials are likely to experience larger decrements in task commitment and performance satisfaction as compared to their counterparts who are not labeled as high potentials (i.e., non-high potentials) following negative performance feedback.
People have a natural tendency to evaluate their relative status (Wood, 1996) and pursue information for making comparative judgments. This is accomplished through a variety of channels including observations of conversations or interactions between the leader and peers or active information seeking (Wood, 1996). In addition, research indicates that it is other individuals in one’s social group that provide the reference point and context for social comparison (Turner, Brown, & Tajfel, 1979). According to cognitive-based research, individuals gather information through both conscious (controlled) and unconscious (automatic) processes (Lord & Mayer, 1991). Such information seeking processes are aimed at detecting both similarities and differences between oneself and another for the purpose of judging one’s relative status on some attribute (Wood, 1996).
In line with this reasoning, it is likely that receiving the label high potential provides the recipient with information regarding their capabilities. Thus employees labeled as high potentials would view their capabilities differently than the employees not granted such preferential status and would hold a different set of expectations prior to performance. There is research evidence to suggest that expressed expectations of high performance can have a positive impact on the recipient of such expectations.
Among the research attesting to the positive impact of high expectations are the studies demonstrating the Pygmalion effect, whereby leader’s expectations of subordinate’s performance can impact subordinate’s self-expectations and, ultimately, behavior (Eden, 1984, 1990a, 1990b, 1992; Eden & Shani, 1982). Tierney and Farmer (2004) observed that according to the Pygmalion model, leader behavior boosts self-efficacy because of the performance expectations it signals. All this is consistent with the notion that publicly labeling an individual as a high potential performer can have a positive impact on their performance expectations. This can be summarized in the following hypothesis.
Hypothesis 1: Individuals who are labeled as high potentials will have higher expectations with regard to their future performance compared to those individuals not receiving such labels.
The question remains – what happens to recipients of the high potential label following performance feedback? Research suggests that people interpret feedback by comparing it to expectations (or, standards) in order to evaluate its meaning (Swann, 1990). The direction and the extent of the gap between feedback and expectations impact an individual’s interpretations, attitudes and responses (Shrauger, 1975; Swann, 1990; Taylor, Fisher, & Ilgen, 1984). Performance feedback that meets or exceeds the initial expectations of performance tends to be viewed as desirable and serves to reaffirm one’s self-image, thus generating favourable reactions (Jussim, Yen, & Aiello, 1995). This type of feedback tends to result in a more pleasant mood and a greater degree of satisfaction (Blakely, 1993; Swann, Griffin, Predmore, & Gaines, 1987; Taylor et al., 1984).
On the other hand, feedback indicating that the level of performance is below a certain standard tends to be viewed as undesirable and results in negative affect and dissatisfaction (Taylor et al., 1984). A discrepancy between the expectations and the outcome can also be viewed as threatening to self-image (Jussim, Yen, & Aiello, 1995) and create psychological discomfort that individuals seek to reduce (Bandura, 1989; Higgins, 1987; Mento, Locke, & Klein, 1992). Therefore, performance feedback perceived as falling below a standard could lower a person’s level of commitment to the task at hand (e.g., Belschak & Den Hartog, 2009).
The above-described tendency applies to both types of individuals, those labeled as high potential and those not labeled as high potential. However, high potentials and non-high potentials are likely to differ in terms of their expectations (e.g., Eden, 1984, 1990a, 1990b, 1992; Eden & Shani, 1982; Tierney & Farmer, 2004) and, thus, in terms of their standards for interpreting feedback. Therefore, high potentials are likely to interpret the same level of feedback (e.g., “above average”) differently than those not labeled as high potentials. In turn, the difference in their perceptions of feedback would generate differences in their attitudes – their level of satisfaction and commitment.
For those labeled as high potentials, the highest expectations of performance (e.g., “outstanding”) would likely be used as the standard for evaluating feedback. Given the high level of expectations set by the high potential label, any feedback indicating an exceptional or outstanding level of performance would further reinforce the positive expectations. However, any positive feedback (e.g., “above average”) that fails to meet such high standards would likely be perceived as unfavourable or unsatisfactory and would create a negative discrepancy. Research supports this logic by demonstrating that individuals are less satisfied with feedback when their expectations (self-assessment) are extremely favorable (Blakely, 1993; Swann et al., 1987). Discrepancies between expectations and outcomes create psychological discomfort that individuals seek to reduce (Bandura, 1989; Higgins, 1987; Mento et al., 1992), which may take the form of reducing their commitment to the task at hand (e.g., Belschak & Den Hartog, 2009). Overall, given their initial expectations of performance well above the norm, individuals labeled as high potentials would be negatively affected by positive feedback framed as “above average”, especially compared to those individuals who were not given any initial expectations regarding their performance potential.
For those not regarded as having high potential (i.e., non-high potentials), the expected performance standard would likely be lower. Knowing that they are not among those who were identified as high potentials, the non-high potentials would likely be somewhat uncertain about their capabilities with respect to the task, other than realizing that their performance expectations should not be set as high as those of high potentials. Nevertheless, all individuals prefer to view themselves as “above average” as they seek to maintain a favorable self-concept (Steele, 1988; Taylor & Brown, 1988). In work environments, most employees consider themselves to be “above-average” performers (Meyer et al., 1979) and tend to perceive “average” (or “satisfactory”) performance ratings as negative feedback (e.g., Pearce & Porter, 1986). Therefore, it is reasonable to expect that in the absence of a high potential label, the “above average” level would likely serve as the standard for evaluating feedback. In other words, for non-high potentials, any feedback that is viewed as “above average” would serve to validate their capabilities and, therefore, would be viewed as image enhancing and favourable.
Based on our theorizing above, the following series of hypotheses can be generated.
Hypothesis 2: After receipt of positive feedback (framed as "above average" performance), individuals who were not labeled as high potentials will exhibit higher levels of performance satisfaction compared to those individuals labeled as high potentials.
Hypothesis 3: After receipt of positive feedback (framed as "above average" performance), individuals who were not labeled as high potentials will exhibit higher levels of task commitment compared to those individuals labeled as high potentials.
Our discussion above implies that the potential for discrepancies between positive feedback and expectations are greater for recipients of the high potential label compared to those without such a label. That is, for any given level of feedback, the discrepancy will likely be higher for those with raised expectations.
In the case of negative feedback, both kinds of individuals will experience discrepancies upon the receipt of such feedback. There is research to suggest that negative feedback can result in decrements in commitment (e.g., Belschak & Den Hartog, 2009) and satisfaction (Stake, 1982). While this suggests that all individuals may respond unfavorably to negative feedback, this pattern will be more pronounced among individuals who have been primed to expect very high performance feedback. Recipients of the high potential label, as indicated earlier, feel highly visible and under public scrutiny to “live up to expectations”. Consequently, negative feedback will likely unsettle those labeled as high potentials much more since their publicly known status is now being questioned. This is not the case for non-labeled individuals who have no perceived public expectations to “live up to”. Consequently, negative feedback will have a significantly greater adverse effect on recipients of the high potential label compared to their non-labeled counterparts in terms of performance satisfaction and task commitment.
Based on the above theorizing, the following series of hypotheses can be generated.
Hypothesis 4: After receiving negative feedback, individuals labeled as high potentials will experience a significantly greater drop in performance satisfaction compared to their non-labeled counterparts.
Hypothesis 5: After receiving negative feedback, individuals labeled as high potentials will experience a significantly greater drop in task commitment compared to their non-labeled counterparts.
Hypothesis 6: After receiving negative feedback individuals labeled as high potentials exhibit significantly lower performance satisfaction compared to their non-labeled counterparts.
Hypothesis 7: After receiving negative feedback individuals labeled as high potentials exhibit significantly lower task commitment compared to their non-labeled counterparts.
Method
Participants
Subjects included 477 undergraduate students participating in exchange for a partial course credit. The average age of the participants was 20.5 years (SD = 3.99), and 53.5 percent of subjects were female. The study was approved by the university research ethics board and was deemed not to involve any risks to participants from partaking in the study.
Task
As part of the experiment, participants were required to complete several cognitive tasks. Subjects were presented with a long list of four-digit numbers and instructed to derive a single-digit answer by following specific rules (similar to Vancouver & Tischner, 2004, p. 1094). For example, one rule involved identifying the absolute difference between the first two digits as well as the difference between the latter two digits and ultimately finding the difference between the two differences. A different rule was used in each round. Participants were given 4 minutes to complete as many calculations as possible. Upon completion of each task, the facilitator collected the booklets and his two assistants, stationed at the back of the room, quickly tabulated scores for each participant. Participants were then privately provided with bogus "feedback" (with the exception of the Practice Task) and instructed to fill out a questionnaire containing measures (see Measures below). The first task was considered to be a Practice Task, and subjects were told that this practice round would give them an idea of how the exercise works and that their performance in this round would not be considered as part of their overall performance in the actual exercise.
Procedure
The study employed a 2 potential labeling (high-potential, non-high potential) x 2 performance feedback (positive, negative) between-subjects experimental design. Participants were randomly assigned to experimental conditions prior to the beginning of the experiment.
Sessions were conducted in groups of 6 to 12 individuals. The number of subjects in the high-expectation condition varied based on the total number of participants, such that in groups of 9-12 participants, 3 subjects were assigned to the high-expectations condition, and in groups of 6-8 participants, 2 subjects were assigned to the high-expectations condition.
Participants arrived and were seated in a large classroom, where they were briefed about the nature of the study and provided their informed consent. Subjects were told that the study involved completing a series of cognitive tasks (they were not told how many) and that their performance would be evaluated in relation to other participants in the study to date by comparing the total number of correct responses. In order to further increase their level of commitment to the task, participants were told that the task was a measure of mental processing speed and ability, which was developed by social scientists and was highly predictive of intelligence. Participants were instructed to give their best effort so the researchers could make an accurate assessment of how well they could perform the task.
Participants were then provided with a verbal explanation of how their performance would be evaluated and shown a large poster at the front of the room displaying the five levels of performance (Table 1). A smaller version of the display was also taped to the right top corner of each subject’s table.
Table 1
Performance Levels
| Performance Description | Ranking |
|---|---|
| Outstanding | You are in the Top 10%. This means that you are performing better than 90% of the participants. |
| Very Good | You are in the Top 20%. This means that you are performing better than 80% of the participants. |
| Above Average | You are in the Top 35%. This means that you are performing better than 65% of the participants. |
| Average | You are performing on par with 50% of the participants. |
| Below Average | You are performing in the bottom 35% of the participants. |
The facilitator told participants that “those students who do well on this task, especially those who perform in the top 10% - or in other words perform better than 90% of their peers – tend to exhibit significantly higher mental processing ability.” He then asked the participants to indicate their desire to perform well on the ensuing tasks using a Likert-type scale, anchored by “not at all” (1) and “very much” (7), which was used as one of our manipulation checks.
Next, the facilitator explained that before proceeding to the actual main task, he would like to give the participants a chance to practice the exercise in order to provide them with an idea of how it works and to make sure everyone properly understands it (the Practice Task). The facilitator distributed the practice task and instructed participants to complete as many items as possible in 4 minutes. Upon completion of the practice task, the facilitator collected the response forms and had his two assistants, positioned at the back of the room, “evaluate” the results and “make sure everyone is on track with this exercise.”
Once the “evaluation” was completed, the facilitator informed participants that everyone had understood the task. He then publically informed them that several individuals performed exceptionally well in the practice round and he expected these individuals would perform in the Outstanding category in this exercise (see Manipulations below). Participants then completed the questionnaire containing measures of dependent variables (see Measures below).
Next, the facilitator distributed the First Task and instructed participants to complete as many items as possible in 4 minutes. Upon completion of the task, he again had the results “evaluated” with the help of his two assistants. This time, he privately provided participants with a form containing bogus feedback on their personal performance according to the condition. Following that, the facilitator distributed the Second Task and instructed participants to complete as many items as possible in 4 minutes. Upon completion of the task, he again had the results “evaluated” with the help of his two assistants and privately provided participants with a form containing bogus feedback according to the condition. The Third Task was then distributed and the participants were instructed to complete as many items as possible in 4 minutes. Upon completion of the task, the facilitator again had the results “evaluated” with the help of his two assistants and privately provided participants with a form containing bogus feedback according to the condition. He then asked participants to fill out a questionnaire.
Finally, after the participants had completed the third task and filled out a questionnaire containing the measures of dependence variables, the facilitator announced that the last task was actually the final task in the experiment. He then asked participants to complete another short exit questionnaire containing questions designed as manipulation checks and demographic variables (age and gender). The manipulation check included the following questions with responses on a 7-point Likert scale: “Do you think the experimenter had any expectations of how you would perform on these tasks?”, anchored by “No, not at all” (1) and “Yes, definitely” (7), and “How would you rate the quality of your overall performance on these tasks?” anchored by “Very poor” (1) and “Outstanding” (7). Once participants submitted their forms, the facilitator provided a debriefing, explaining the nature of the experiment and reassuring them that these tasks do not necessarily accurately reflect mental processing capabilities.
Manipulations
Labeling Based on Performance Potential. We created two labeling conditions: high-potential and non-high potential. The manipulation was carried out immediately after the completion of the practice round. Participants assigned to the high-potential condition were publically told by the facilitator that they performed exceptionally well in the practice round and that, in this exercise, he expected them to perform in the top 10% of all participants. Upon receiving the “results” of the Practice Task, the facilitator informed participants that everyone has understood the task, and then exclaimed in a delighted tone, as if in passing:
“Interesting, some of you have actually scored in the top 10% in this practice round (pointed to the Outstanding category in the table). This doesn’t happen often in a group of this size. Students with numbers X, Y, Z… you, you and you (pointing and smiling at each student one by one): Congratulations on your outstanding ability! Based on my experience… I expect you will perform in the Outstanding category on the upcoming tasks. Keep up the good work!”
Participants assigned to the non-high potential condition were not specifically told anything, but the researcher made sure that these subjects clearly heard his praise of the high-potential participants.
The experiment was structured in line with how organizations commonly implement their high-potential programs in real life, whereby they identify certain individuals as high potentials early in their employees’ careers – and sometimes even at the point of hiring them right out of university – on the basis of their assessment of these individuals’ ability to perform in future roles and/or future tasks. Organizations often make such judgment calls about their employees’ potential to perform in future roles and/or future tasks by observing their performance in their current roles and/or tasks. Similarly, in this experiment, participants in the high potential label condition were told that based on their performance on a practice task they were considered to have an outstanding ability and were expected to perform at the highest level on a series of the upcoming cognitive tasks which were created specifically for the purpose of this experiment and were unfamiliar to the participants. We employed the claims about the subjects’ alleged performance on a practice task in order to administer a convincing manipulation by creating a credible rationale for assigning a high potential label.
Performance Feedback. Two performance feedback conditions were created: Positive feedback and negative feedback. Participants received bogus normative feedback from the facilitator about their performance in relation to other students (i.e., subjects received predetermined feedback regardless of their actual performance). In the positive feedback condition, subjects were told that their performance on the task was “Above Average”; in the negative feedback condition, subjects were told that their performance on the task was “Below Average.” In order to make the manipulation believable, the facilitator and his two assistants, positioned at the back of the room, pretended to analyze the results after each task. The facilitator then handed each subject a pre-printed piece of paper with that participant’s ID number a checkmark next to a particular performance category, indicating how he or she “performed” on the previous task in comparison with other participants. The facilitator had no knowledge of the participants’ actual performance during the session.
Measures
Dependent Variables
Performance Expectations. Participants were asked to indicate their personal expectations of their performance on a 7-point Likert scale: “How confident are you in your ability to perform this task effectively?”; “How well do you think you will perform on the next task?”; “How confident are you that you can perform either in the Outstanding (top 10%) or the Very Good (top 20%) range on the next task?”; “How well do you think you will perform on the next task in comparison to other students?”. The Cronbach's Alpha for this 4-item measure was .92 following the Practice Round.
Task Commitment. This construct was used to refer to the probability that a participant would persevere with the task, as opposed to leave the task (in line with the definition of Job Commitment; Farrell, & Rusbult, 1981, p. 79). We measured task commitment after each task (i.e., practice task, first task, second task, and third task) by asking participants to respond to the following two items on a 7-point Likert-type scale: “Would you continue with this task regardless of how many more tasks are remaining?” and “Have you thought about quitting this activity?”. The reliability of this 2-item scale was α = .76 following the Third Task.
Performance Satisfaction. We measured participants’ satisfaction with their performance on the task after each task (i.e., practice task, first task, second task, and third task) by asking participants to respond to the following two items on a 7-point Likert-type scale: “How satisfied are you with your performance on the task?” and “Are you at all disappointed with your performance?”. The second item was reversed in order to create a composite measure. The Alpha coefficient of this 2-item measure was .72 following the Third Task.
Control Variable. In our analysis we controlled for gender (F = 0; M = 1) given that women may react differently than men to raised expectation (Dvir, Eden, & Banjo, 1995), experience emotions more intensely (e.g., Barrett, Robin, Pietromonaco, & Eyssell, 1998; Fujita, Diener, & Sandvik, 1991; Kring & Gordon, 1998), and have been shown to respond differently to stress elicited by using mental arithmetic tasks (Wang et al., 2007).
Results
Manipulation Check
We used three manipulation checks in this study. First, in order to check whether subjects took the study seriously, we asked participants at the beginning of the experiment to indicate their desire to perform well on the tasks using a Likert-type scale, anchored by “Not at all” (1) and “Very much” (7). The mean response of 5.72 (SD = 1.26) indicated that the participants took this exercise seriously and were committed to doing well.
Second, to check the effectiveness of the manipulation of labeling, we asked participants at the end of the study to respond to this question “Do you think the experimenter had any expectations of how you would perform on these tasks?” on a Likert scale, anchored by “No, not at all” (1) and “Yes, definitely” (7). We compared responses of high-potentials and non-high potentials using one-way ANOVA. The results confirmed that subjects in the high-potential condition (M = 4.09, SD = 1.88) were significantly more likely to think that the experimenter had expectations regarding their success than those in the non-high potential condition (M = 3.25, SD = 1.90), F(1,475) = 18.99, p < .0001.
Third, to check whether the participants believed our manipulation of feedback, we asked participants at the end of the study to respond to this question “How would you rate the quality of your overall performance on these tasks?” on a Likert scale, anchored by “Very poor” (1) and “Outstanding” (7). We compared responses of subjects in the positive feedback and negative feedback conditions using one-way ANOVA. The results confirmed that subjects in the positive feedback condition (M = 4.49, SD = 1.18) were significantly more likely to think that they had performed better than those in the negative feedback condition (M = 2.79, SD = 1.62), F(1,474) = 170.64, p < .0001.
Test of Hypotheses
The means, standard deviations and correlations among the variables are presented in Table 2.
Table 2
Means, Standard Deviations and Correlations Among Variables
| Spearman Correlation Coefficients
|
|||||||
|---|---|---|---|---|---|---|---|
| M | SD | Labeling | Feedback | 1 | 2 | 3 | |
| 1. Performance Expectations (Practice Task) | 4.94 | .91 | .32** | .04 | (.92) | ||
| 2. Performance Satisfaction (Task 3) | 3.48 | 1.72 | -.13** | .44** | -.11* | (.72) | |
| 3. Task Commitment (Task 3) | 5.10 | 1.63 | -.09* | .10* | .21** | .19** | (.76) |
| 4. Gender (F=0; M=1) | -.03 | .07 | .16** | .01 | -.02 | ||
Note. N = 477. Coefficient alpha is shown in brackets along the diagonal.
*p < .05. **p < .01.
Prior to testing our hypotheses, we checked that our experimental conditions did not significantly differ by gender. No significant difference was found, Chi-Squared χ2 (1, N = 477) = .33, p = .56, indicating adequate randomization by gender in creating pre-experimental equivalence among treatment groups.
In our Hypothesis 1, we predicted that individuals who were labeled as high-potentials would have higher expectations of their future performance compared to those individuals not receiving such labels. We used a general linear model to compare performance expectations among subjects immediately following the manipulation of labeling (i.e., after the Practice Task), while controlling for gender. In the model, labeling was entered as a factor and gender was entered as a covariate. The results revealed a main effect for labeling F (1,474) = 55.24, p < .0001, partial η2 =.11. The average performance expectations reported by high-potentials (M = 5.38, SD = .81) were significantly higher than those reported by non-high potentials (M = 4.75, SD = .89), thus providing support for Hypothesis 1. The main effect of gender was also found to be significant, F (1,474) = 16.12, p < .0001, partial η2 = .03, whereby men tended to have somewhat higher expectations than women.
In Hypothesis 2, we predicted that after receipt of positive feedback, non-high potentials would exhibit higher levels of performance satisfaction compared to those individuals labeled as high potentials. To test this hypothesis, we excluded all data involving negative performance feedback, and applied a general linear model to only the data where subjects received positive feedback. We entered labeling as a factor and gender as a covariate. The results showed that, following the final (i.e., third) task, the level of performance satisfaction was higher for non-high potentials (M = 4.42, SD = 1.46) than high potentials (M = 3.74, SD = 1.45), F (1,236) = 10.95, p < .001, partial η2 = .05. Thus, Hypothesis 2 was confirmed: After receiving positive performance feedback, high-potentials reported lower satisfaction with their performance than non-high potentials.
In Hypothesis 3, we proposed that after receipt of positive feedback, individuals who were not labeled as high potentials would exhibit higher levels of task commitment compared to those individuals labeled as high potentials. To test this hypothesis, we repeated the process described immediately above with respect to the task commitment. The results revealed that, following the final task, the level of task commitment was higher for non-high potentials (M = 5.33, SD = 1.51) than high potentials (M = 5.13, SD = 1.65), but failed to reach a level of statistical significance, F (1,236) = .94, n.s.
These results indicate that, in the positive feedback condition, labeling subjects as high-potential and, thus, assigning to them high performance expectations increased their expectations of their future performance, but also reduced their feelings of satisfaction with their performance, as compared to non-high potentials.
In Hypothesis 4, we stated that after receiving negative feedback high potentials would experience a significantly greater drop in performance satisfaction compared to their non-labeled counterparts. For the purpose of our analysis we excluded all data involving positive feedback, and ran a general linear model using only the data involving negative feedback. Our dependent variable was computed as a difference in performance satisfaction measured after the practice task and measured after the final (i.e., third) task. We entered labeling as a factor and gender as a covariate. The results revealed a main effect for labeling F (1,225) = 86.81, p < .0001, partial η2 = .28, such that the average decline in performance satisfaction reported by high-potentials (M = -3.87, SD = 2.12) was significantly higher than that reported by non-high potentials (M = -1.41, SD = 1.66), thus providing support for Hypothesis 4.
In our Hypothesis 5, we predicted that after receiving negative feedback, high potentials would experience a significantly greater drop in task commitment compared to their non-labeled counterparts. To test this hypothesis we repeated the above procedure with respect to task commitment. The results revealed a significant main effect for labeling F (1,235) = 15.33, p < .0001, partial η2 = .06, whereby the average decline in task commitment (between the practice round and the final task) was greater for high-potentials (M = -1.70, SD = 1.67) than non-high potentials (M = -.82, SD = 1.49), thus providing support for Hypothesis 5.
In other words, the results showed that following negative feedback, high-potentials experience a more rapid decline in their performance satisfaction and task commitment than non-high potentials.
In Hypothesis 6, we predicted that after receiving negative feedback individuals labeled as high-potentials would exhibit significantly lower performance satisfaction compared to their non-labeled counterparts. We used a general linear model to compare the level of performance satisfaction among high-potentials and non-high potentials in the negative feedback (i.e., below average) condition only while controlling for gender. The results revealed that in line with our prediction, subjects in the high-potential group (M = 2.49, SD = 1.68) reported lower levels of satisfaction following the final task than subjects in the non-high potential group (M = 2.82, SD = 1.59), however the difference failed to reach a level of significance at p < .05, F (1,232) = 1.93, n.s. Therefore, Hypothesis 6 was not supported.
Finally, in our Hypothesis 7, we predicted that after receiving negative feedback individuals labeled as high potentials would exhibit significantly lower task commitment compared to their non-labeled counterparts. We again used a general linear model to compare task commitment among subjects after the final task, while controlling for gender. In the model, labeling was entered as a factor and gender was entered as a covariate. The results confirmed that, in the condition of negative feedback, high-potentials exhibited lower levels of task commitment (M = 4.51, SD = 1.83) than non-high potentials (M = 5.08, SD = 1.61), F (1,235) = 5.41, p < .05, partial η2 = .03. Therefore, Hypothesis 7 was supported.
In summary, in the negative feedback condition, both high potentials and non-high potentials reported a decline in performance satisfaction and commitment between the Practice Task and the Final Task. However this deterioration was more pronounced for high potentials, and after the Final (i.e., Third) Task their level of commitment was actually lower than that of non-high potentials.
Discussion
This study attempts to shed more light on a very topical, critically important and under-researched management issue. High potential programs are increasingly popular (e.g., Ready et al., 2010) and many organizations regard the practice of identifying high potential employees as an important component of building a long-term competitive advantage (Cappelli, 2008; Corporate Leadership Council, 2005). However the results of this study indicate that labeling individuals in terms of their performance capabilities could represent a double edge sword. On the one hand, as has been demonstrated by earlier research, it can boost individuals’ expectations and, in turn, their self-efficacy and performance (e.g., Tierney & Farmer, 2004). On the other, as has been demonstrated by the results of this study, labeling individuals as high potentials could impact how they interpret performance feedback, potentially lowering their level of satisfaction and commitment.
The extant research has provided ample evidence of the positive effects of communicating high expectations to subordinates that create a self-fulfilling prophecy and can result in higher performance (e.g., Eden, 1984). For example, numerous studies have demonstrated the occurrence of the Pygmalion effect, whereby leader’s expectations of employee’s superior performance have been shown to positively impact that individual’s self-expectations and behavior (Eden, 1984, 1990a, 1990b, 1992; Eden & Shani, 1982). In fact, the findings of this study also show that labeling individuals as high potentials can indeed raise their performance expectations (Hypothesis 1).
However, much less is known about the role these expectations play following performance feedback that does not meet the performer’s expectations. That is, what happens if the individual feels that he or she has not lived up to these high expectations? What is the impact of performance feedback when it fails to affirm those initially high expectations? The extant research has not satisfactorily addressed these questions. This study specifically attempts to address the role of labeling and its influence on responses to performance feedback.
The results of this study draw attention to a possible downside of high potential labeling. The practice of high potential labeling may create additional vulnerabilities to feedback. While people in general tend to feel discouraged by feedback that falls short of their expectations (Brett & Atwater, 2001; Kluger et al., 1994; Taylor et al., 1984), individuals labeled as high potentials could be particularly likely to be affected by feedback because of the heightened level of their initial expectations. In other words, labeling individuals as high potentials may cause them to perceive a greater gap between expectations and performance and possibly even a challenge to their alleged high potential status, leading to deterioration in satisfaction and commitment.
In this study, we compared how individuals in the high potential and the non-high potential conditions reacted to positive (“Above average”) and negative (“Below average”) feedback. In the case of positive feedback, those labelled as high potentials reacted less favourably than those who did not receive such a label. By the end of the Third Task, individuals in the high potential condition, whose initial performance expectations were boosted through labeling, were less satisfied with their performance than non-high potentials when feedback was framed as “above average”.
The consequences of negative feedback were adverse for both high potentials and their non-labeled counterparts. However, the decrements in performance satisfaction and task commitment were greater for high potentials, whereby the non-high potentials were not as adversely affected by negative feedback. Thus, our results suggest that while personal performance expectations are positively related to task commitment, building up performance expectations through labeling can create a false promise of commitment. When faced with “below average” feedback, both high-potentials and non-high potentials reported a decline in performance satisfaction and commitment; however high-potentials experienced a more rapid decline compared to non-high potentials. By the end of Third Task, high potentials exhibited lower task commitment than non-high potentials.
Thus, the double edge sword of high potential expectations: Labeling individuals as high-potentials can increase their performance expectations, but may also reduce their satisfaction and commitment.
Implications
The results of this study add value to a consideration of the role of labeling high potentials and responses to performance feedback. We believe this study is among the first to examine this issue. This research has important practical implications, as organizations are increasingly implementing high potential programs, in large part, to improve engagement and retention of their high potential employees. Ironically, high potential programs may actually reduce satisfaction and commitment under certain conditions, thus potentially reducing engagement and retention of the very employees organizations consider most valuable for their future success. The findings of this study suggest that organizations should carefully consider how they communicate high expectations to their employees and how they manage expectations and attitudes when performance feedback falls short of the initial expectations.
Limitations and Future Research
While this study has important implications for understanding how labeling individuals as high potentials can influence responses to feedback, there are a number of limitations in this study as well as questions raised that merit future research attention. First, the study employed students within a laboratory based setting, which limits the generalizability of our findings. Although a laboratory study was used, it is promising that the present research demonstrated an effect of labeling in a situation with minimal consequences; in a field setting it is possible that such effects may be more pronounced. Field based studies that monitor employees who have actually been targeted as high potentials would be of great interest for future research. Second, it is desirable to examine the interaction of the factors identified in this study in a “real” work context – among individuals within traditional organizational settings. Our manipulation has likely created a more specific expectancy (performance in the Outstanding category) on specific tasks, as compared to a “real” work context, where a high potential label doesn’t always specify either how good “high” is or how soon that “potential” is likely to manifest. This may have resulted in a more focused contrast between expectations and performance, and thus may have more explicitly activated processes related to expectancy disconfirmation and adaptation level theory when the task feedback contradicts the initial label, and, therefore, might have produced stronger results than a manipulation of perceived potential would produce. Future field research should also examine how labeling an employee as a high potential might influence their attitudes toward performance feedback; and to what extent non recipients of the high potential label are motivated to improve their status and thereby might be more disappointed with feedback that does not reflect an improvement in their capabilities. Third, future research should consider strategies for informing individuals of their potential and for communicating higher performance expectations that do not lead to a drop in satisfaction and commitment following poor performance feedback. We hope this study encourages further exploration of the impact of labeling on employee behavior and organizational outcomes.