EMPLOYEE RETICENCE: DEVELOPMENT AND VALIDATION OF ITS SCALE

Purpose of the study: The purpose of the study was twofold, first to find the antecedents and consequences of employee voice using the theory of planned behaviour, and second to develop, introduce and validate the construct of employee reticence (employee-silence attitude) and its scale. Methodology: Mixed-Method Research using interviews, and questionnaire was the methodology of the study. IBM SPSS 20.0 and SmartPLS 2.0 software were used for data analysis to test and confirm face validity, content validity, reliability, item reliability, convergent validity, discriminant validity, nomological validity. Main Findings: Employee reticence is a valid employee attitude. The employee silence behaviour is supported by the theory of planned behaviour. It was found that the Theory of planned behaviour applies to the situations of employee silence. The scale of employee reticence developed, as the result of this study, is valid and ready to be used by researchers, organizational behaviourists, psychologists, and HR experts. Applications of this study: The study is applicable for change intervention programs in organizations by organizational behaviourists. The researchers can use it to explore and develop organizational and team-level attitudes of silence. Its scale is a useful tool, largely for recruiters for ensuring PO fit. It is usable to researchers for further exploration and using it as in organizational research including but not limited to Employee silence, knowledge management, and stress. Psychologists can take the lead from the concept of ER to develop similar constructs for their field for analysing their patients of stress etc. from this perspective. Novelty/Originality of this study: The study is novel and original as the construct of employee reticence has been developed, introduced, and validated (employee-silence attitude). Further, the scale of the new construction has been constructed and validated, using standard methods.


INTRODUCTION
Employee silence is the intentional withholding of suggestion(s) intended for workplace improvement (Benevene, 2020;Morrison, 2014), it is a pervasive and internationally occurring phenomenon (Götz et al., 2019) across sexualities (Priola et al., 2014). The previous researchers have explored the dispositional and organizational level antecedents of employee silence (ES); however, they ignored a very vital aspect at the employee level that contributes towards the formation of silence behavior. This research has tried to capture this missing link of employee's attitudes towards silent behavior, which is a novelty and strong contribution to the existing body of knowledge.
One of the reasons why ES attitude has not been discussed in the previous literature is due to the lack of such a construct and unavailability of the scale to measure it. Thus, the main objective of the study is to define and to operationalize it. Secondly, its objective is to develop, validate and test its scale developed as the result of this study. Further, the rationale of addressing this gap derives from the theory of planned behaviour (TPB) (Ajzen, 1985) which postulates that attitude formulates intentions that contribute towards behavioral depictions.
The current research has identified and explored a very interesting variable, employee reticence, as an important antecedent of ES especially in the service sector, where competitive edge depends on employee's behavior.
Employee-attitude-towards-silence referred to as Employee Reticence (ER) is being proposed in this research as an attitude that pertains to the favorable evaluation of silent behavior at the workplace. According to the Oxford Dictionary, the word "reticence" originated in the mid-19th-century in Rome referred to as "remaining silent" and in English, it means "not revealing one's thoughts or feelings readily". Taking lead from this research we have used ER as favorable feelings and beliefs of an employee towards silent behavior in the organization. In literature silence behavior in an organization is referred to as ES (Knoll & van Dick, 2013) which is an omnipresent phenomenon in modern organizations (Lam & Xu, 2019) and dangerous for organizations (Abd El-Fattah Mohamed Aly et al., 2021). Its antecedents include factors at the individual level, group level, and organizational level (Knoll & Redman, 2016;Morrison, 2011;Tangirala & Ramanujam, 2012). However, the review of extant literature suggests that the individual level factor that is employee attitude towards silence has not to be investigated as an antecedent of ES. The significance of this research is embedded in introducing the concept of ER.
The significance of this research has multiple dimensions. First, the research makes a significant theoretical contribution by proposing a new variable titled ER. Secondly, the scale of ER has been developed, validated and tested using The main objective of the study is to define and operationalize employee reticence. Secondly, its objective is to develop, validate and test, through hypotheses testing, its scale developed as the result of this study. For the sake of testing the developed scale, there were three objectives of the study, which are: 1) To find out if employee reticence is positively related to employee intentions to remain silent, and 2) To find out if intentions to remain silent is positively related to employee silence, and 3) If intentions to remain silent mediates the relationship between employee reticence and employee silence.

Employee Reticence
Proposing the new job attitude warrants a brief discussion on attitudes. Extant literature exists on describing and understanding attitudes. The earlier researchers have described bi-polar evaluative-ness as an essential characteristic of attitude (Ajzen & Fishbein, 2005;Eagly & Chaiken, 1993;Fazio, 1990), however, a few researchers have defined it as the evaluative response itself (Kruglanski & Stroebe, 2005), while most contemporary researchers equate attitude with the hypothetical disposition (Eagly & Chaiken, 2005) and have concluded that attitudes can be inferred from evaluative responses towards the subject, object or behavior. In this research, the contemporary school of thought on attitudes has been adopted. That is evaluative responses towards job-related behaviors such as ES has been used to infer job-related attitudes such as ER. It is proposed in this research as a new job attitude. It has been defined as "Favorable feelings and beliefs of an employee towards silence behavior in organizations" based on the definition of attitudes by Eagly and Chaiken (2005).
In wholesome ER has been operationalised as favorable feelings and beliefs of an employee towards silence behavior of employees in the organization. An employee, who evaluates ES as positive behavior, will score high on ER scale and vice versa. Such employees not only prefer to remain silent but also advocate and encourage other colleagues to do so. Imagine a situation where a top-level manager inquiry about a process improvement initiative. Few employees will appreciate the initiative. Few others will not have anything to say about the initiative. There will still be employees who have suggestions for the improvement initiative but will prefer to remain silent. They will themselves remain silent and will also appreciate other employees who have similar attitudes (Byrne, 1961;Orpen, 1984) and consequently will remain silent. Meaning thereby that these employees evaluate ES behavior as positive and good. These employees will have high scores on ER scale. Therefore, they are likely to perform the same behavior in future interactions (Armitage & Christian, 2003).
ER and ES are different constructs. Employee reticence is an attitude while the construct of ES is a behavior that represents the intentional silence exhibited by employees. The relationship between ER and ES is proportional. However, this relationship is mediated by relevant intentions as proposed by TPB. It suggests that attitude contributes towards intentions which lead towards the exhibition of the behavior. Therefore, high ER would lead to intentions to remain silent which would lead to the manifestation of ES as given in Figure 1.  (Ajzen, 1985) The relationship between attitudes and intentions is also supported by empirical evidence. For example, significant relationships have been found between respective attitudes and turnover-intentions such as in the study of Valentine et al., (2011) a significant relationship was found between respective attitudes and turnover intentions. The TPB also supports this assertion. According to TPB one's positive attitude towards a behavior leads to the formation of one's intentions to perform the behavior. Therefore, getting support from the extant literature and TPB we hypothesize that ER would have a positive effect on intentions to remain silent. This argument leads to the formulation of the first hypothesis of the research as follows

Employee Intentions to Remain Silent and Employee Silence
Employee silence is an intentional behavior. It is not just a mere absence of voice. It does not occur when an employee has nothing to say or is mindless (Dyne et al., 2003). ES is exhibited when an employee intentionally withholds suggestion-for-improvement from a person perceived to be capable of implementing the improvement suggestion. It is withholding ideas/suggestions intentionally due to specific reasons (Dyne et al., 2003;Knoll & van Dick, 2013) which is also supported by the multiple intentional behavioral theories.

Employee Reticence
Intentions to Remain Silent Employee Silence The previous researchers have studied intentions as predictors of a wide array of behaviors and found a significant relationship between intentions and the exhibition of the behaviors (Howard et al., 2017). Based on the arguments, and further conceptualization the second hypothesis of this research is proposed as below:

H2: Intentions to Remain Silent is positively related with Employee Silence.
Supported by the extant literature and the TPB intentions have been found to mediate the relationship between attitudes and behavior (e.g. see (Ajzen, 1985;Chen et al., 2011;Shukla et al., 2013). Therefore, based on the TPB and empirical research the following hypothesis is proposed.

H3: Intentions to Remain Silent fully mediates the relationship between Employee Reticence and Employee Silence.
Before testing the hypotheses of the research, we developed the scale of ER. The methodology and procedure of scale development are explained below.

METHODOLOGY
In this research, we have developed the scale of ER by combining deductive and inductive approaches, as this is the best practise for the development of scale (Boateng et al., 2018). Moreover, following the contemporary school of attitudes and boundary conditions, the initial pool of items was generated.
In this research, we have used the qualitative method for a better understanding of ER and the generation of the initial items pool. The qualitative study was followed by quantitative analysis for item reduction, validity & reliability analysis, factor analysis, and testing the research model in light of the TPB.

Qualitative Study Phase
We conducted semi-structured, in-depth interviews with 41 employees and practitioners volunteering for the study from the service sector. This step was followed by the coding analysis of the interview transcripts. As a result, the items were generated in light of the contemporary school of attitudes. Though not the part of the objectives of the study, the structural dimensions of ER was also discovered.

Interview process
The interviews were conducted with the 67 employees of the service sector. It also helped in the generation of items. Sixty-seven employees voluntarily participated in the interviews, which were from telecom, hospital, education, oil & gas, and the courier sector. Demographic data of the interviews revealed 31.7% were the female participants. 68.2% of the participants were having 2-7 years of experience, while 17% of the employees were having 8-13 years of experience while the rest of the employees were having experience greater than 13.
Semi-structured and in-depth interviews were conducted during the interview phase. The interviews were conducted by experts who had in-depth knowledge of the phenomenon. Before the start of the interview, the participants were briefed about the purpose of the research. Two interviewers conducted each interview which was divided into groups of 5 each; the interview sessions lasted for approximately 1 hour. They remained very specific in explaining the meaning of attitudes / ER to have responses based on their cognitive, affective, and behavioral state of mind about ES. The interviews continued for two intervals. Each interval spanned over seven days. In the first interval, the interviews were summarized and relevant analyses were performed simultaneously. In the second interval, similar questions were asked. However, it was found that no new theme or ideas emerged. Therefore, it was concluded that increasing the sample size of the respondents would not affect the emergence of new ideas. Further, the three dimensions of ER were discovered in the first interval. Therefore, it was assumed not to take the interview phase in the third interval.

Coding process
Each interview interval was followed by the coding process. It was a two-stage process that involved (i) formal coding, and (ii) checking coding consistency. As recommended by Charmaz (2006) the interview content was grouped into firstorder code and second-order code. The coding process was performed independently by the two researchers; however, the same steps of the coding process were followed. It was cautiously done to check the coding consistency in the coding process.
During the open coding phase of the coding process, the distinct phenomena that emerged were conceptualized. As the result of the qualitative study three factors of ER were found, 1) cognitive employee reticence, 2) affective employee reticence, and 3) behavioral employee reticence. Some representative quotes from the interviews are given in Table 1. I do not like the colleagues who instead of focusing on their own departments, always find problems in other departments. I find it a total waste of time, whenever employees try to be ingenious for the problems of other departments. I am amused by the colleagues whose own department is a mess and yet find suggestions for other departments.

Cognitive
In organizations, disagreements should be avoided.
Disagreements should only be done, when you expect something good out of it. In my experience, disagreement is a waste of time and in the end, the boss wins in any case. Discussions are useless. In my organization, our new colleagues have to agree to all the disagreements, until the time they do disagree no more.
In organizations, it is better to stay quiet than to give suggestions.
In this job and in the previous jobs that I have done, it's better to hold suggestions and remain silent than staking your self-respect. I have always been told by my seniors to hold suggestions and keep a low profile for smooth functioning in the organization. My manager does whatever he has to do even when the majority of us is not in favor, so it's better to be quiet than to waste time in thinking and giving suggestions. Conflicts in an organization can be handled by remaining silent.
My advice to conflict management is to remain silent. There is no use of saying anything, it is likely to aggravate the conflict.
Remaining silent in organizations is beneficial.
What I have learned from my working experience is that, people who are silent in organizations are the one giving the most to the organization by supporting the current system. Above all, barking dogs seldom bite. People who are focused, get their work done, without making noise.
Making noise of what you are doing is useless, since what you are doing is what you are paid for, and making noise is useless.

Behavioral
The job assignment which may lead to a conflict is delayed by me to the last hour.
If I know there will be discussions and possible conflict, I prefer to delay the task until it becomes a necessity. I don't know, but for smooth functioning, I delay the work leading to conflicts as late as possible. I wait for others to raise concerns about a common problem rather than raising it by myself.
I am lucky that most of the times problems similar to mine are put by others. I wait and like it this way.
I often do the undone work of another employee without raising my voice.
I have lost my hair, as I am good at absorbing pressures, so much so that I do the task of others instead of asking them to do their task. My colleagues often choose me as their team member as I am reliable and mainly because they know I will go the extra mile to get the teamwork done without making noise.

Findings
Through semi-structured in-depth interviews, we found that employee reticence is three dimensional: affective, cognitive, and behavioral. Therefore, the ER concept is three-dimensional and encompasses affective, cognitive, and behavioral components. Since the dimensions were not part of the scope of the research, the qualitative analysis revealed three dimensions of ER and needs to be treated separately in future research. It has been left to the research to follow as part of future recommendations.

Item generation
The second phase of the qualitative study involved the item generation process. In this phase items developed were combined and an initial pool of 29 items to measure ER was generated. A few of the representative quotes from the qualitative interview phase are listed in Table 1. The generated pool of initial items also contained adapted items from existing scales, such from the scale of sportsmanship dimension of organizational citizenship behavior (1 item; Organ, 1988), Social Reticence Scale (two items; Jones & Russell, 1982).

Quantitative Study Phase
In order to empirically test the theoretical model, see Figure 1, we did two studies in the quantitative study phase. In the first study, the item reduction was performed using principal component analysis as given below.

Item reduction (principal components analysis)
The initial pool of 29 generated items was distributed among the 300 employees of the service sector. The response rate of the valid questionnaire was 72 %. The items were based on a Likert scale (five points) with 5 = strongly agree and 1 = strongly disagree. The demographics of the respondents revealed that 40.2% of them were female.
Five of the items were excluded in analyzing the responses through PCA. The exclusion was based on low correlations between them and their category mean. On the remaining of 24 items, the discriminant analysis was conducted by categorizing them into groups of high and low scores based on the total score of other items. Accordingly, 11 items were removed; with an insignificant difference between the high score and low score groups as recommended by DeVellis (2016). During this process, none of the items were deleted until the consensus was reached between the researchers and 3 experts. Other than these, discussions were also done on the remaining 13 items with the experts. The items which were confusing and found redundant were removed. Resultantly, the final questionnaire of ER had 11 items only. These 11 items were transformed in the form of a questionnaire for data collection and further analyses in study 2.

Study 2
Since the ES behavior is pervasive in organizations, the self-report survey-based research was selected as the methodological approach in the second phase. The self-report method was chosen mainly because of the implicit nature of ES. Lack of action in ER, intentions, and ES makes it difficult for the observers to detect behavior, therefore the selfreport method is a recommended method in such scenarios (Wang et al., 2020; Wang & Hsieh, 2013).

Sampling Design
The purposive sampling design was used. It served our specific purpose which was the selection of a respondent only if s/he answered the following questions positively 1) I have been working in the same organization from at least the last two years, 2) There are no major changes in the communication channels of my organization and 3) I have been with the same supervisor/line manager for at least last two years.
Data was collected in 3-time lags. Time lag research is a recommended way in research where intentions have been used as a predictor of a behavior (Fishbein & Ajzen, 2011). Further, all the items are self-reported which could lead to common method bias; therefore, we adopted a time-lagged research design to mitigate common method bias Podsakoff et al., (2003).
In the first lag data for ER was collected, after the gap of 2 weeks the data for a mediator, intentions-to-remain-silent was collected. Finally, in the third lag data of ES was collected. The respondents were briefed about the research before the start of the data collection process.

Data Collection Technique
700 questionnaires were distributed among the employees working in the service sector. 253 valid questionnaires were received back (RR 36.7%) which represents an adequate sample as determined through the G * Power application. It was used to run the power analysis as per the recommendations of Faul, Erfelder, Bucnhner, and Lang (2007). Keeping in view the research model, the sample size chosen for this research (n=253), was larger than the one proposed by G*Power 3.0, with the statistical sig. (α) level 5%, and the required level of power of 80% (Hair et al., 1998), and the effect size of 15 (Cohen, 1988).
Demographic analysis revealed that 18.5% of the respondents were female. The female respondents were less mainly because of the overall 17.71% female workforce of the country within the age range of 21-65.

Measures
The scale for ER developed as the result of study 1, was used. Accordingly, its validity and reliability analysis such as face validity, content validity, reliability, item reliability, convergent validity, discriminant validity, nomological validity, and analysis was performed. The scale of intentions to remain silent was adapted based on the recommendations by Ajzen (2006). One of the items from the scale was, "In my organization, it is likely that I will remain silent on the topic of concern in the two weeks".
The measures of ER and intentions to remain silent asked participants to rate each scale item using the Likert scale (five points) with 5 = strongly agree and 1 = strongly disagree.
The scale of ES was adapted from Vakola and Bouradas (2005). One of the sample items from the scale was, "How often do you express your disagreements to your managers concerning your company's issues? The respondents evaluated the statements/indicators given in Table 2. The content validation step was performed with 6 academicians in the area of human resource management and 4 managers. The face validation step was performed with 4 Ph.D. scholars and 3 consultants. As a result, the changes were incorporated, and 2 redundant items were deleted. The final scale of ER now contained only 9 items.  The quantitative analysis was done by bootstrapping directly in SmartPLS 2 with re samplings as recommended by Tenenhaus, Vinzi, Chatelin, and Lauro, (2005). The results of Cronbach's alpha and composite reliability are given in Table 3.  Table 3, showing that all items loaded on their respective construct with a value ranging from 0.65 to 0.90. The table also indicates that the items loaded highly on their constructs, in line with the recommendations by Howard (2016).
The convergent validity of each construct was evaluated by average variance extracted (AVE) as recommended by Hair et al., (2003). According to them the value of AVE should be at least 0.50 to ascertain convergent validity. The values of AVE in this research are shown in Table 3. It can be seen that the values of the AVE are greater than 0.50 for all the constructs, showing convergent validity of each construct. To complement the AVE findings, convergent validity was also tested by the values of cross-loadings of items on their latent constructs, which was found to be significant (p < 0.05). T-statistics results were also indicative of the same as shown in the loadings of outer model loadings in the output of PLS-graph. These values ranged from a low of 16 to a high value of 36. The constructs' convergent validity is confirmed from items' loadings and cross-loadings presented in Table 2. Further significant T-statistics for each individual item loading confirm the convergent validity of these indicators as representing distinct latent constructs. Therefore, the validation criteria for internal consistency and the convergent validity were found to be satisfactory, confirming our proposed construct of ER.
According to Fronell and Larcker (1981), the discriminant validity is confirmed if the square root of the AVE is higher than the correlation among all the latent constructs/variables. It is evident from Table 4, as the AVE square roots, show in diagonal, are greater than other elements in respective columns and rows.   Table 5 presents the hypotheses, outcomes, and conclusions as to the result of the analysis.

DISCUSSION
In this research, the employee reticence phenomenon was proposed, and its scale was developed and validated. ER has been defined as an attitude towards ES. During the qualitative phase, in-depth semi-structured interviews confirmed the presence of such attitudes in the employees. As the result of the qualitative phase of the study, three dimensions of ES that is cognitive employee reticence, affective ER, and behavioral ES were also discovered. In the quantitative phase, two studies were done. In the first study, the developed scale was finalized by performing PCA and expert opinion. In the second study using internal consistency, content validity, face validity, discriminant validity, nomological validity tests, and analysis of the research model were performed.
The PLS-SEM analysis of the study showed acceptance of the hypotheses, thereby suggesting a significant relationship between the ER and ES mediated by intentions to remain silent. In other words, the greater the ER the greater would be the employee's intentions to remain silent, consequently leading to the ES. This is in line with the theory of reasoned action (Fishbein & Ajzen, 2011), the TPB (Ajzen, 2006), functional attitude theory (Katz, 1964), and attitude-behavior consistency principle (Haddock & Maio, 2007). Thus, if an employee is high at ER the employee will intend to remain silent.
Finally, the intentions to remain silent (ITRS) was found to have a significant positive effect on ES. This is in line with the theory of reasoned action (Ajzen & Fishbein, 2005) and TPB (Ajzen, 1985) and consistent with findings of multiple research based on these theories. That is to say, if an employee has intentions to remain silent, the employee is likely to exhibit the behavior of silence. Therefore, in an organization, if an employee is asked for the input, the employee with intentions to remain silent will remain silent even if the employee has something important to contribute. Further, it was found that intentions to remain silent fully mediate the relationship between ER and ES. As shown in figure 2, the direct path from ER to ES is insignificant depicting full mediation of intentions. This is in line with the theory of reasoned action (Fishbein & Ajzen, 2011).

Figure 2: Structural Model Results
In this research, ER has been introduced, and its scale has been developed, validated, and tested. This is an addition to the silence literature by presenting a new attitudinal variable that is ER, which can be used in studies related to ES, respondent's bias, attitudinal and behavioral modification studies, etc. It can be used to develop and test new theories related to ES, personality, knowledge sharing, teamwork, stress, anxiety, employee health problems, harassment, and bullying.
The introduction of the intention to remain silent as a mediator between ER and ES will pave the way to find answers to why employees choose to remain silent despite being the witness and victim of bullying, harassment, and alike behaviors in organizations.
In organizations, successful management requires continuous feedback and input from lessons learned, however, ES impedes improvements and corrections (Dedahanov et al., 2016). It is a matter of prime importance to mitigate ES.
Attitudinal updating seems to be the easiest way of doing it. ER scale can be used for different types of research design including pre-test, post-test research design for attitudinal updating of employees. Further, change management programs, innovative and creative jobs require continuous knowledge sharing and feedback. The recruiters and managers can use ER scale for the initial screening of the candidates for such types of jobs.

CONCLUSION
In this research, the phenomenon of employee reticence has been proposed and its scale has been developed and tested. The scale development of ER builds upon and extends previous theoretical and empirical work on attitudes and behaviors by bringing in new perspectives in the existing body of knowledge on ES. Although previous research on ES has focused mainly on organizational factors, this research on employee attitude towards silence has furthered our understanding of why employees choose to remain silent. Conceivably, the recruiters, OB-interventionists, and researchers are greatly to benefit from this research because existing measures do not assess employee attitude towards silence. Keeping in view the abundance of evidence on the negative role of ES in cases such as the Challenger space shuttle disaster and Enron (Lalich et al., 2018), there is little doubt that organizations of all types could benefit from better understanding the antecedents of ES such as ER.

LIMITATION AND FUTURE RECOMMENDATIONS
Although this research has advanced our understanding of attitudes, it is just the beginning of building the foundations of ES attitude. This research has not only confirmed the presence of such an attitude but also developed and validated its scale.
The limitations of the scale of the ER can be removed by further delineating it with the dimensions of attitude that are cognitive, affective, and behavioral dimensions. Moreover, with the availability of ER scale, further research can be guided to predict its antecedents and consequences.
Based on the recommendation of TPB, other variables such as organizational silence, manager's attitude can also be included as IV to represent subjective norms, and communication opportunities can be included as IV to represent perceived behavioral control. Personality and gender may be used as moderators. In addition, organizational and teamlevel attitudes of silence can be explored. Further, the same research may be replicated in non-service sectors and projects, where ER and ES can have devastating effects on project success.