Does Feedback Seeking Always Improve Performance? Investigating the Roles of Feedback Seeking Content and Frequency in Determining Goal Achievement and Behavior-Related Performance

FSB was defined and introduced as a proactive behavior (Ashford and Cummings, 1983). Various dimensions have been considered: frequency, method, source, timing, and content (Ashford, Blatt and VandeWalle, 2003). Feedback seeking content is an important aspect concerned with the questions feedback seekers pose and what they expect to gain from the process. For example, if the feedback seeker expects a positive feedback, raising the question differs from when the feedback seeker expects a negative feedback (Ashford and Tsui, 1991). Furthermore, the posed question can emphasize either the results or the process leading to those results (VandeWalle, 2003). The FSB sign (seeking positive or negative feedback) and whether the feedback is sought about work processes or the results of the processes are only two examples of how one can structure feedback seeking content.

Although researchers have investigated different dimensions of FSB, some have been emphasized more than others. For example, a meta-analysis conducted by Anseel, et al. (2015) looked into the relationship between over 20 phenomena and the feedback seeking frequency. However, feedback seeking content has not been studied extensively. The existing studies emphasize some aspects of the content, especially the sign of FSB, while overlooking others. To the best of our knowledge, for the first time, the present study utilizes three types of feedback preference proposed by Park, et al. (2007), namely diagnostic, normative, and assurance, for conceptualizing and measuring feedback seeking content. From this behavioral perspective (not as a preference), we define these three types of feedback seeking content as structuring the requested feedback. These types of content consist of analyzing individual performance outcomes and the diagnosis of any actions or behaviors required for the expected outcomes and areas for improvement (diagnostic); comparing individual performance with the performance of others (normative); and attempting to improve one's self-efficacy or self-esteem by having a positive feedback (assurance).

Studies have focused on formulating different aspects of individual performance for many decades (e.g., Campbell, 1990; Viswesvaran and Ones, 2000). Due to the multidimensional nature of performance, researchers have been unable to agree on a general definition, determine its dimensions, or introduce standard tools to assess it. Taking these different points of view into account will change the perception of the relationship between performance and other phenomena. Some researchers have proposed that a single definition of performance, applicable in all situations and professions, is impossible to formulate. However, other researchers have attempted such a formulation. Campbell (1990) defined individual work performance (IWP) as “behaviors or actions that are relevant to the goals of the organization”. As such, performance is defined based on behaviors rather than results. As Campbell proposed, we use IWP in this paper when referring to behavior-focused performance. Furthermore, Viswesvaran and Ones (2000) defined performance as “scalable actions, behavior, and outcomes that employees engage in or bring about that are linked with and contribute to organizational goals”. In addition to behavior, this definition recognizes the outcome type (e.g., goal achievement) as an integral part of the conceptualization of performance. Campbell (1990) has identified eight performance dimensions defined in terms of behavioral criteria about how an individual performs a task or plays administrative and teamwork roles. In addition to behavioral aspects, Viswesvaran, et al. (2005) have taken the outcome dimensions of quantity (such as lack of errors) and quality (productivity) into account for defining performance. In sum, it seems that “behavior-based” and “outcome-based” dimensions, such as achieving performance goals (Grant and Dweck, 2003), can be identified as two dimensions of individual performance conceptualized in the literature.

At first, researchers had assumed that FSB per se could lead to better performance (Ashford and Cummings, 1983). A common view was that the information gained through FSB is used to develop a clearer picture of how an organizational role can be performed. This information would ultimately reduce uncertainty around achieving higher performance. Furthermore, FSB provides individuals with insights into their performance, potentially improving their performance. Despite these assumptions, previous studies have failed to establish a statistically significant relationship between FSB and performance across all situations (Anseel, et al., 2015). This result may be explained in various ways (Audia and Locke, 2003) such as the unreliability of the feedback giver and the seeker's perceived incompetence in acting based on the provided feedback. In addition, individual factors such as psychological needs (Henry, et al., 2018), contextual factors such as the organization's feedback culture (Min and Ming, 2021), and task characteristics such as task complexity (VandeWalle, 2003) can also be considered.

As mentioned above, another possible explanation for these inconsistent results is that different studies might have used different types of performance. Both types of performance described above have been considered in the literature. While Luckoski, et al. (2022), for instance, investigated the relationship between FSB and outcome-based performance, Evans and Dobrosielska (2021) used task performance in their study, which is a behavior-based type of performance. Moreover, Lam, et al. (2015) have generally referred to work performance by using one question for evaluating teachers regarding how well they conducted instructional activities in a semester, providing no explicit observation of the difference between each type of performance.

Suppose a manager and his/her employees formulate specific goals and agree on them. In that case, the goal achievement can be discussed and evaluated during work processes more clearly than when there are no specific goals. In addition, even in a knowledge-based context with specific goals, performance issues and obstacles may occur due to the nonroutineness of tasks (Carneiro, 2000). This situation requires more performance conversations to acquire helpful information to manage performance. Thus, the more the employees seek feedback, the more likely they can resolve performance issues. Having specific goals can help such conversations because they know the target and what they want to accomplish, even when there are ambiguities regarding work processes. This is theoretically consistent with goal-setting theory, where gaining feedback determines strategies for achieving the goals more clearly, resulting in higher performance (Locke and Latham, 2002). This is also consistent with the expectancy theory that communication on goals can reduce ambiguities, increasing the possibility of goal achievement (Evans, 1996). Furthermore, due to the specificity of goals, the responses to FSB become more explicit. Thus, the likelihood of erroneous understanding of inquiries from the feedback provider and incorrect understanding of feedback from the feedback seeker is reduced as employees seek more feedback. Performance improves as uncertainties about goals and the required outcomes to achieve them subside (Ashford and Cummings, 1983).

Feedback seeking can be analyzed in terms of its content. As mentioned earlier, Park, et al. (2007) proposed three types of feedback content which feedback seekers prefer to receive: diagnostic, normative, and assurance. We propose that the same categorization can be adopted for conceptualizing the content of FSBs and investigating the relationship between feedback seeking and performance. As mentioned earlier, this paper's main theoretical contribution is to propose that examining the content of FSB adds to our understanding of how FSB may directly impact IWP.

The data were collected in a knowledge-based company in Tehran, which is active in research and high-tech equipment production. The structure of this company comprises three units: research, production, and staff (providing administrative services to the other two units). We invited every employee of these three organizational units engaged in nonroutine tasks to participate in this study. Thus, FSB was meaningful and helpful for their learning and reducing their performance uncertainty. Information was obtained from two distinct sources and in three parts. A portion of the data—including employees’ levels of goal achievement and control variables—was collected from the organization's database. Other data were posed through two online questionnaires at two different phases with an interval of 1 month. Using such online forms technically helped us detect participants’ IDs, which enabled us to link data collected from the two questionnaires and the organization's database. We had an agreement with the organization's management team to ensure that only the researchers had access to raw data, with no disclosure of participants' identities. The questionnaires’ introductions included this information. Because of the organization's previous cooperation in some other questionnaire-based studies unrelated to FSB, giving such an assurance to participants was helpful. Of the 600 participants, 317 answered all questions of the first questionnaire, 306 of whom answered the second questionnaire as well. Therefore, the response rate was 51%. We compared the sample's characteristics (e.g., mean age and experience) with the characteristics of the population (600 employees) and found no substantial differences. Seventeen percent of the individuals who filled in both questionnaires completely were women. Participants' mean age was 35 years (standard deviation [SD] = 6.6), their mean work experience in the organization was 72 months (SD = 45.7), and their mean overall work experience was 98 months (SD = 59.8).

We used several strategies to ensure data quality and reduce the probability of the common method bias (Podsakoff, et al., 2003). First, collecting data from two distinct sources and at two different phases might have helped reduce common method errors. Second, the questionnaire items of variables in each hypothesis were placed separately in different sections of the questionnaires and not close to each other. Third, assuring participants of the confidentiality of the procedure was considered by not presenting the company management team with identifiable data.

Although different control variables could be considered based on the literature, we were limited in the number of variables to consider due to the length of our questionnaire and the limited time the company permitted for the study. Thus, we included a few widely used control variables in the FSB literature. Age, organizational, and overall job experience (in months) and feedback-seeker's gender (male = 0, female = 1) were selected as previously used in other studies (Anseel, et al., 2015; Vandenberghe, et al., 2019). Organizational units and education (each with two dummy variables) were also selected, as in previous studies (Li and Qian, 2016; Wang, et al., 2021). Job seniority can also influence FSB due to its impact on the nature of the job and the risks FSB may cause, and has been statistically controlled in previous FSB studies (e.g., Qin, et al., 2021). The salary variable was chosen as a control variable to indicate job seniority.

We conducted principal axis factoring and direct oblimin rotation for each measure separately on the responses of 306 individuals who participated in both phases. Kaiser's eigenvalue criteria, parallel analysis, and Velicer's minimum average partial (MAP) test were considered and compared to determine the number of factors. For finalizing our factor solutions, the interpretation and soundness of factor structure in terms of the meaning of factor items and the theoretical backgrounds of the factors were essential.

The resulting factors were in line with expectations, except for IWP, where rather than two factors, only one factor was identified based on Velicer's MAP test. However, three factors were identified considering Kaiser's criteria and parallel analysis. The task performance factor obtained was almost similar to the factor in the original instrument described earlier. However, the contextual performance factor was divided into two meaningful factors: “updating knowledge and skills” and “facing challenges”. All three factors were interpretable and had acceptable reliabilities. Cronbach's α for task performance, updating knowledge and skills, and facing challenges were 0.76, 0.72 and 0.73, respectively. We preferred to keep the three-factor model because the one-factor model was complex for conceptual interpretation, given the varied nature of the items. In addition to interpretability and acceptable measurement reliability, the factors were more theoretically helpful for understanding the different aspects of performance in the study.

Factor analysis for feedback seeking frequency resulted in one factor, as expected, and the α was 0.77. Given the novelty of this approach, first, a pilot study with 80 participants was carried out for feedback seeking content, resulting in the revision of some items. Next, the revised questionnaire was adopted for the present study.

In exploratory factor analysis, items with weak factor loadings and strong cross-loadings were removed. Finally, the remaining three factors were identified by considering Kaiser's eigenvalue criteria, parallel analysis, and Velicer's MAP test. Six items were identified for diagnostic feedback seeking, three for normative feedback seeking, and two for assurance feedback seeking. Each factor's reliability coefficient and its corresponding items, taking into account all the data, are presented in Table 2. Because the two-item factor was interpretable with the conceptualization and its reliability was acceptable, we retained this factor.

We then utilized confirmatory factor analysis, as stated earlier. Based on Anderson and Gerbing's (1988) recommendations, the data were divided into two parts using a randomly sorted list of the whole sample. We managed the sample size of the second part to have an acceptable size for conducting confirmatory analysis (Boomsma, 1982). First, exploratory factor analysis was conducted on the first part (n = 106). Then, the factor structure identified in the first phase was used for confirmatory analysis in the second part (n = 200). The fitness measures for the feedback seeking content scale were χ² = 80.95, degree of freedom (df) = 32, χ²/df = 2.53, p < 0.01, goodness of fit index (GFI) = 0.93, adjusted goodness of fit index (AGFI) = 0.88, comparative fit index (CFI) = 0.95, root mean square error of approximation (RMSEA) = 0.088, and standardized root mean square residual (SRMR) = 0.06. In addition, the fitness statistics for the IWP scale were χ² = 62.65, df = 32, χ²/df= 1.96, p < 0.01, GFI = 0.94, AGFI = 0.90, CFI = 0.95, RMSEA = 0.069, and SRMR = 0.06. We followed Schermelleh-Engel, et al. (2003) recommendations for the final evaluation of our measurement models. Most fitness measures were in the acceptable ranges proposed by Schermelleh-Engel, et al., although the RMSEA of the feedback seeking content model was close to marginal (RMSEA =0.08).

Hierarchical regression was adopted to test the hypotheses. A separate analysis was conducted for each dependent variable. Control variables were considered in the first stages of variable entry, followed by each variable in different steps. The significance of R2 change was the criterion for the existence of an improvement in the regression model after each step. The variance inflation factor (VIF) criterion was used to evaluate multicollinearity. The VIF was not within the conservative multicollinearity range (higher than 3) for any of these variables (Hair, et al., 2016). Figure 1 illustrates the results. Regression results are presented in Tables 3 and 4. As shown in Table 3, the relationship between feedback seeking frequency and goal achievement is significant. Furthermore, as shown in Table 4, considering the three IWP dimensions, diagnostic feedback seeking was significantly related to the three aspects of IWP. No supports were identified for the relationships proposed between normative/assurance feedback seeking and the three dimensions of work performance.

Figure 1:

This study contributes to the FSB and performance management literature in several ways. First, this study extends the FSB literature and responds to calls for further investigation of how FSB may impact individual performance. More specifically, we conclude that the relationship between FSB and performance cannot be expected as a relationship between two simple phenomena or unidimensional variables. Different dimensions of both FSB and performance should be considered when investigating the relationship. Our results supported the existence of three types of feedback seeking content. This result is also theoretically significant, as future studies can examine the role of feedback seeking content and its consequences. This result is in line with previous papers which reported that the nature of the message transmitted by feedback seeking could impact FSB outcomes (Ashford, Blatt and VandeWalle, 2003; Ashford, Stobbeleir and Nujella, 2016).

This study also provides evidence that feedback seeking frequency is related to performance if achieving predetermined goals is regarded as performance. However, if IWP in terms of individual behaviors and actions associated with performance is examined, seeking diagnostic feedback as a type of feedback seeking content, rather than its frequency, is a significant determinant. Diagnostic feedback seeking is related to all three measured IWP dimensions, while normative and assurance feedback seeking are not associated with any of those dimensions. This issue has implications for the FSB literature, as feedback seeking, in general, may not improve an individual's performance. Our results can be discussed from different perspectives as follows.

Although we identified some statistically direct relationships, the relationships were not strong in the context of this study. Thus, our conclusions should not overestimate the magnitude of the relationships. We argue that the relationships are promising for future research, considering the lack of a conclusive result in previous studies regarding the direct relationship between FSB and performance (Anseel, et al., 2015). Indirect relationships may be stronger if some mediators or moderators are considered in these relationships. For example, as Ashford, et al. (2016) proposed that individual characteristics such as learning-goal orientation (Locke and Latham, 2002) and self-efficacy (Bandura, 1997) may also be used to explain the mechanism of the relationship. We argue that employees with learning-goal orientation may be more inclined to seek diagnostic feedback and learn more from the feedback they receive in terms of the details of required actions for improving their performance, compared to their coworkers. As the theoretical framework describes, normative/assurance feedback seeking can positively impact performance via self-efficacy. The leader–member exchange quality can moderate the relationship between FSB and IWP (Lam, et al., 2015). When there is a low-quality relationship between a feedback seeker and the feedback giver, there may be less chance of providing helpful feedback and less opportunity to use feedback in the future. Feedback usefulness can also be a mediator. Asking for more diagnostic feedback may encourage the feedback provider to take more time and provide more valuable and detailed feedback. More detailed feedback can help the seeker get the information necessary to improve performance (Ashford, Stobbeleir and Nujella, 2016). This result is in line with previous research reporting that self-focused negative feedback improves performance (Gong, et al., 2014), as diagnostic FSB focuses on self-focused negative feedback.

On the other hand, some previous research shows that seeking self-focused positive feedback performance is unrelated to performance (Ashford and Tsui, 1991; Gong, et al., 2014). Assurance FSB is focused on self-focused positive feedback. Thus, the result is similar to previous research. Kluger and DeNisi (1996) proposed that feedback intervention effectiveness increases as attention moves to task details and decreases as attention moves to self. Because diagnostic feedback seeking focuses on task details and assurance feedback seeking focuses on the self, the results of the present study are consistent with those of Kluger and DeNisi.

Since diagnostic feedback seeking focuses on performance analysis and improvement (Park, et al., 2007), it pays more attention to the details of actions and behaviors than the other two types of feedback seeking content. Because of the knowledge-based nature of the company, knowledge workers may rely more on accurate information regarding the details of their actions and behaviors (Carneiro, 2000). In addition, nonroutine tasks may require knowledge workers to learn and solve problems to accomplish their tasks. Another possible explanation is that diagnostic feedback seeking can be demonstrated after receiving information associated with normative/assurance feedback seeking. The initial information from normative/assurance feedback seeking may trigger further assessment by the individuals to examine the details of their performance. This possible explanation may be supported by the statistically significant correlation between normative/assurance feedback seeking and diagnostic feedback seeking in data identified in our analysis. Given that assurance feedback seeking may indirectly affect performance through the mediation of self-efficacy, the lack of a significant relationship between assurance feedback seeking and performance may be due to the characteristics of our sample. It is possible that knowledge workers who participated in our study might have possessed high self-confidence, making it unnecessary to use assurance feedback seeking to improve their self-efficacy.

Studying FSB in an organization in a collectivistic culture might have impacted our results. As Iran is among countries with high in-group collectivism (Dastmalchian, Javidan and Alam, 2001), being in harmony with others may be more crucial for individuals. Thus, we argue that seeking normative feedback in such contexts is likely to be used to gain information regarding the status of an employee's performance compared to his/her coworkers for the sake of harmony. However, seeking such normative feedback content in an individualistic culture may be more likely to identify performance gaps than high performers to improve individual performance. This result may also be in line with previous research reporting that feedback seeking process in collectivistic cultures could be different from feedback seeking in an individualistic one (Mao, 2022).

The knowledge-based nature of the company could also have affected our results. The organization units comprised structurally convenient teams for building relationships and constant communication. As a result, open conversations are more likely to occur in this context, which is beneficial for performance improvement. Therefore, a direct relationship between FSB and performance is more expectable in this context.

The feedback preference model proposed by Park, et al. (2007) was adopted to suggest three distinct types of FSB in terms of the content dimension. Accordingly, we defined these types of FSB, proposed required measurement items, and tested the measures in this study. Such modeling of feedback seeking content can yield more comprehensive results than previous studies. More studies are needed to strengthen the measurement items of these variables concerning the validity and reliability of the instrument proposed in this study.

Koopmans (2014) categorized questions into task performance and contextual performance. However, factor analysis findings revealed a different factor solution, according to which contextual performance was divided into two meaningful factors: “updating knowledge and skills” and “facing challenges”. These findings are similar to those of some previous studies. For instance, Viswesvaran, et al. (2005) have highlighted up-to-date job knowledge as an individual performance dimension. We further discuss that updating knowledge and skills can be distinct from facing challenges as dimensions of IWP. Knowledge workers may be active in updating their knowledge and skills during their activities, not for the sake of their current job but to be competitive in their future career path outside the organization.

First, individuals may have different types and levels of motives for each type of content when seeking feedback. Because of the practical limitations for including and measuring more variables in our study, we had no measure of feedback seeking motivation. However, considering the literature on feedback seeking motives (Ashford, Blatt and VandeWalle, 2003), motives for diagnostic feedback seeking may differ from motives for normative feedback seeking. This idea can also be investigated in future studies.

Second, the design of the study cannot support causality. Therefore, it is recommended that the hypotheses addressed in this study be tested using experimental methods to measure objective performance after FSB and assess the causality. In addition, future studies can use longitudinal designs to assess performance after measuring FSB. We had a practical limitation in this study to measure performance after a period of FSB.

Third, feedback seeking is one of the followership behaviors in performance management systems. Some other followership behaviors in performance management systems (e.g., feedback sharing between peers) can be studied in relation to performance. Fourth, the research was conducted in a knowledge-based company in the Middle East. Due to practical limitations, we had no measures for these contextual factors. Thus, the findings of this study may not be generalizable to other organizations in different contexts. Individuals in different contexts (different industries, task routineness, and culture) may view the performance concept differently and react differently to distinct feedback seeking content types. For example, the national context of this study is a country with a high power distance (Dastmalchian, Javidan and Alam, 2001). Less formal feedback seeking conversations in a lower power distance culture is more probable, so individuals can seek diagnostic feedback from their supervisors at lower costs.

Fifth, a self-reported performance measurement tool was used for IWP. This approach has been used in previous studies considering individuals’ awareness of the details of their behaviors and activities. However, there may be biases in this measurement. In our study, we measured goal achievement from the perspective of managers. Sixth, we decided to use only exploratory factor analysis for our measures because it was the first time feedback seeking content was measured using the normative, diagnostic, and assurance dimensions. We suggest using confirmatory approaches such as latent regression analysis in future studies. Finally, since the measurement instrument used for feedback seeking content has not yet been validated in the English language, a study is recommended for testing the validity of an English version of it.

Performance management systems aim to increase individual performance. Managers are also asked to help and support their employees to improve their performance. The results of this study can help organizations and managers in performance improvement. First, similar to other studies that take various types of individual performance into account, the present study urges managers to think about the specific types of employee performance they intend to improve because improving different types of performance requires different actions. Second, in cases analogous to the company under scrutiny, organizations looking to improve individual performance through FSB may find it helpful to steer FSB toward diagnostic content.

Organizations can use different practices to achieve this goal: for example, role modeling, coaching, educating and encouraging employees toward diagnostic feedback seeking, and arranging sessions suitable for diagnostic feedback seeking (e.g., meetings that have no image cost and where conversations remain private). These practices may require the training of employees for the diagnostic skills of knowing how to break down their activities and use information from performance appraisal to seek information regarding specific actions and behaviors that need improvement.

In addition, organizations can support employees to seek more diagnostic feedback because knowing the benefits of particular behavior does not necessarily motivate employees to demonstrate that behavior. Accordingly, managers are recommended to pay attention to their employees to realize when they are seeking diagnostic feedback. Thus, they put more time and concentration into their reactions because our evidence suggests that diagnostic feedback seeking can impact different aspects of IWP.

Practically speaking, this also has implications for performance management practices. Managers can increase the effectiveness of performance management when they are motivated and knowledgeable about different types of FSB and how they can react to them more effectively. Improving the quality of FSB and its consequences can improve conversations regarding performance required for the effectiveness of performance management (Pulakos and O’Leary, 2011). Based on this recommendation, training programs can be designed to help managers react effectively to different types of feedback seeking content and motivate them to provide relevant and helpful feedback that can benefit employees.

Third, our findings support a relationship between FSB and goal achievement. Therefore, it can be surmised that setting individual goals and matching managers’ and subordinates’ perceptions of expectations can lead to more useful feedback seeking processes. Providing training programs to both managers and employees about mutual goal setting at the beginning of the performance management cycle is recommended. In this context, FSB will be more related to their goals and how to achieve them. As FSB increases, goals and ways of attaining them will become more transparent for employees.