This chapter exhibits the findings of the study which are based on the analysis of the data collected. There are several categories in which the findings are discussed.
These categories include the regression approach and the steps, hypothesis testing results and the discussion of the findings. Each of these categories are given detailed explanation.
Demographic characteristics of the sample
Table 4.1 shows the demographic characteristics of the respondents participated in this study. The data collected for this study come from all the agencies within the public sector. The sample is rich in two level of positions which include subordinates and top-level management such as supervisor and managers. The respondents of this study consist of 185 males (62%) and 116 females (38%) with a mean age of 35. Around 50% of respondents have bachelor degree and 33% have some college qualifications while other qualifications are less than 15%.
Table 4.1.
Demographic Characteristics of Respondents.
Hypotheses Testing and Findings
Baron and Kenny (1986) devised four-step approach to test mediation with regression analysis. This study adopted and utilized it and several regression analyses were conducted. The significance of the coefficients was tested at each of these four steps. The figure 4.1 demonstrates the description to be followed.
Figure 4.1. Regression analysis approach
In figure 4.1, particularly the step 1-3, the aim is to establish the existence of zero-order relationships amid the variables. Supposing that one or more of these relationships among these variables are nonsignificant, the researcher should generally conclude that the mediation is unlikely or more likely, although such cannot be true in most instances (MacKinnon, Fairchild & Fritz, 2007). Supposing the result is otherwise which mean there are significant relationships from Steps 1 through 3 and progresses to Step 4. In Step 4, some form of mediation will be supported if the effect of M (path b) remains significant after controlling for X. In case that X is no longer significant when M is controlled, the finding supports full mediation. If X is still significant (i.e., both X and M both significantly predict Y), the finding supports partial mediation.
Besides, the internal consistency of the result measurement was assessed using a reliability test. In the study by Ueno and Sekaran (1992), coefficient alpha was used to assess the internal consistency between the scale of items. The Cronbach alpha cut-off value is .70 in which all values obtained were above the cut-cut-off value as indicated in table 4.3. This implies that these measurement items are very reliable to measure each construct. Construct validity assesses the degree to which a measurement represents and logically connects the observed phenomenon to the construct (Gates, McDaniel & Braunsberger, 2000). Table 4.2 exhibits the means, standard deviations and correlations among all the constructs while table 4.3 exhibits the hypothesis results.
The findings affirm that all the constructs are positively correlated (p< 0.01) and Cronbach’s Alpha reliability scores for all the constructs are above 0.7, indicating good reliability of all the measurements.
Table 4.2.
Descriptive Statistics, Reliability Coefficients, and Correlations among the Variables.
Table 4.3.
Hypotheses Testing Results
The mediation effect (indirect) of KS on the direct relationship between OC and OI and OCB and OI was assessed using the casual step method recommended by Baron and Kenny (1986) as indicated in table 4.2. The overall aim of the causal step method is to establish the conditions for mediation rather than a statistical test of the indirect effect (MacKinnon et al., 2007). In this study by Baron and Kenny (1986), four criteria for complete mediation was suggested. First, the independent and dependent variable must be significantly correlated. Second, the independent variable must be significantly related to the potential mediator. Third, the mediator and the dependent variable must be significantly correlated. Finally, the initially significant relationship between the independent and dependent variable must become insignificant once the role of the mediator is assumed in the process to determine complete mediation. If only steps one through three are met, then partial mediation is established.
In order to test the effect of the mediating variable, zero-order correlations between variables were calculated. Zero-order correlation assesses the relationship between two variables, while ignoring the influence of other variables in prediction.
First the zero-order correlation between OC with OI and OCB with OI was calculated (completely standardized with β = .24 and β = .22 respectively). Second, the zero-order correlation between OC with KS (potential mediator) and OCB with KS was calculated (completely standardized β = .49 and β = .36 respectively). Third, the zero-order correlation between KS and OI was calculated (completely standardized β = .22).
Findings suggest that all three zero-order correlations were significant at .05 significance level. Finally, the full model presented in Fig. 3.1 was tested. As presented in table. 4.3, the initially significant relationship between the direct relationship among the dimensions remained significant once the role of the mediator (knowledge sharing) is accounted for in the process. This finding provides support for the hypothesis
involving mediator by clearly indicating that the relationship between OC and OI, and the relationship between OCB and OI is partially mediated by KS.
Discussions
This study has empirically investigated the interplay between OC, OCB, KS and OI outcomes by empirically exploring theoretically derived models. The study found that KS partially mediated the relationship between OC and OCB with OI. First, the findings of this study show that both OC and OCB were strongly associated with KS and OI. This result between OC and KS implies that employees who has a firm sense of belonging to their organization, derive desire in sharing knowledge and thus, are motivated in helping others towards donating and collecting knowledge among colleagues. The empirical evidence suggests that higher OC would enhance the willingness amongst employees for to share their knowledge with other colleagues.
This is supported with the findings of McKenzie, Truc and Winkelen, (2001), which suggest that the motivation of employees to create and share their knowledge can be significantly influenced by OC. On the basis of this, OC can be viewed as a determinant of KS pursuit within an organization. Furthermore, the result of this study shows that OCB also has a significant positive influence on KS. This behavior motivates and encourages employees to extensively exhibit voluntary helping behaviours that are outside their job schedule which is an essential ingredient to KS. According to Davenport (1997), knowledge sharing is a voluntary act. This understanding of knowledge sharing implies that knowledge sharing amongst individuals is exchanged without pressure to do so and it is done unrestricted (Ipe, 2003). The workers’ extra discretionary behaviors go above minimum job requirement and produces opportunity for knowledge sharing and innovation. Social exchange theory also provides plausible theoretical lens to explain such finding. According to this theory, individual’s reactions with other individuals is regulated based on the “outcome of self-interest” analysis of
the costs and benefits of such an interaction in exchanging information. Social exchange is an important concept that explains the interaction among individuals (Kelley, Thibaut, Radloff & Mundy, 1962). In the study by Chadwick-Jones (1976), this theory is also commonly used for examining the knowledge-sharing behavior of individuals. Indeed, when employees become more committed to their organizations, they would become more willing to perform more in-role or even extra-role behaviors (e.g., knowledge sharing) in return for the organization. This could explain why this research found both OC and OCB are positively related to knowledge sharing.
Second, the results of this study show that KS is positively associated with OI.
Apparently organizational innovation initiatives tend to depend on the knowledge, skill, and experience possessed by employees. Knowledge sharing of the employees’ skills, knowledge and experience has been witnessed to affect organizational innovation. KS is changing traditional ideas about work styles and processes by providing new ideas, approaches, disciplines and cultures, thus constituting an OI (Darroch & McNaughton, 2002). It is apparent that the organization’s ability to transform and utilize the employees who possess the knowledge may determine the level of innovation in the organization, such as new problem-solving methods and new product for rapid reaction to the market demand (Goh, 2002; Tidd et al., 2005). Furthermore, organizations must begin to manage its employees’ knowledge effectively when employees are willing to share their knowledge within the organization. Constant sharing of knowledge has a huge potential in contributing to innovations in teams, units and the whole organization.
Innovative tasks result from employees sharing their tacit knowledge (skills or experience) with their colleagues or search for explicit knowledge (institutionalized approaches or practices) existing in the organizations. Both the explicit and tacit components of organizational knowledge sharing practices is crucial and play an essential role in innovation (Mascitelli, 2000; Xu, Houssin, Caillaud & Gardoni, 2010).
Drawing on this view, several studies show that knowledge sharing can be seen as valuable input for innovation (Chiang & Hung, 2010; Gachter et al., 2010). Thus, an organization that promotes KS practices within groups or organizations is likely to breed new ideas for developing new corporate opportunities, hence facilitating innovation activities (Lundvall & Nielsen, 2007; Heffner & Sharif, 2008).
Third, the findings of this study confirm that OC influences OI. Generally, employees with high level of commitment tend to be more committed to their organization by working harder than others. They are more willing and ready to make efforts to meet the organizational objectives and accomplish its goal. The OI is the reflection of its employees’ commitment and performance. With the enhancement of OC, it improves OI. Affective commitment which is related to individual value, beliefs and wishes helps improve employee’s attitude towards OI and promotes the OI. For example, in the study by Eisenberg et al. (1990) found that, the more closely employees’ emotion were connected with the organization, the more willing employees were to take up innovative activity. Furthermore, the findings also show that OCB positively influence OI. According to Smith, Organ and Near (1983), research on the antecedents of OCB such as job satisfaction, organization commitment and motivation are assumed to be the best predictors and must be given attention. OCB reduces the unconstructive interpersonal conflicts that could destroy the supportive teamwork climate which is conducive to innovation (Jehn, 1995). Correspondingly, several studies indicated that OCB helps create good collaborative relationships within the team and organization that may help to promote collaboration that leads to team creativity (Stewart, 1989; Yan & Sorenson, 2003).
Fourth, the results showed support for the mediating effect of knowledge sharing on the relationship of organizational commitment and organizational citizenship behavior with innovation respectively. That is, organizational commitment
and organizational citizenship behavior are beneficial for innovation via knowledge sharing. Indeed, since knowledge sharing is widely accepted as essential for organizational innovation (Chiang & Hung, 2010; Gachter et al., 2010), knowledge sharing could be an important reason why those committed employees and those who are willing to engage in organizational citizenship behaviors become beneficial for organizational innovation. In addition, as discussed earlier that knowledge sharing is a voluntary behavior with the attempt of benefiting the organization (Davenport, 1997), only those who are committed and are willing to conduct citizenship behaviors are likely to do so. Therefore, it is logical to find that knowledge sharing plays a partial mediating role for the relationship between organizational commitment and organizational citizenship behavior to organizational innovation, respectively.