4.1.1 Missing Values Analysis
Knowing that the presence of missing values in the data set can lead to incorrect results and certain analysis cannot be performed, we did missing values analysis at first.
Missing Values Per Case
Table 1 shows the detail of missing values per case for all constructs. 271 cases are complete and free of any missing values. There are 13 cases with 1 missing value, 8 cases with 2 missing values, 11 cases with 3 missing values, 2 cases with 12 missing values, 3 cases with 17 missing values, and two with twenty missing values. All 39 cases with missing values were excluded from the data and we left with the final data size of 271.
Missing Values Analysis Result
Table 1: Missing Values Analysis Result
|
Frequency
|
Percent
|
Valid Percent
|
Cumulative Percent
|
0
|
271
|
87.41
|
87.41
|
87.41
|
1
|
13
|
4.19
|
4.19
|
91.60
|
2
|
8
|
2.58
|
2.58
|
94.18
|
3
|
11
|
3.55
|
3.55
|
97.73
|
12
|
2
|
0.65
|
0.65
|
98.38
|
17
|
3
|
0.97
|
0.97
|
99.35
|
20
|
2
|
0.65
|
0.65
|
100
|
Total
|
310
|
100
|
100
|
|
The Table 2 Frequency Percentage of Demographic Characteristics" presents a comprehensive overview of the distribution of demographic characteristics among the study participants. The data is organized in a tabular format, typically with columns representing different demographic categories and corresponding rows displaying the frequency, percentage, and potentially other relevant information for each category.
Each demographic characteristic, such as age groups, gender, educational qualifications, job positions, or any other relevant variables, is likely presented as a separate column. The "Frequency" column indicates the number of participants falling into each category, providing a count for each demographic group. The "Percentage" column represents the proportion of participants within each category relative to the total sample, expressed as a percentage.
For instance, if the table includes demographic variables like age groups, it might have rows for different age brackets (e.g., 20-30, 31-40, 41-50, etc.), with corresponding frequencies and percentages for each group. This presentation allows readers to quickly grasp the distribution of participants across various demographic characteristics.
Overall, the table serves as a concise and structured way to convey essential information about the composition of the study sample, aiding researchers and readers in understanding the demographic profile of the participants.
Table 2: Frequency Percentage of Demographic Characteristics
CFA: Transactional Leadership
We conducted CFA to test the three factor of transactional leadership [42]. The measurement model showed a good fit with all the values of fit indexes in acceptable range RMSEA=.08, GFI=.91, TLI=.86, and CFI=.98 [43] ( see Table 3).
Table 3: Fit Indices for Three Factor Model of Transactional Leadership
All the items were loaded well on their respective five factors of transformational leadership (see table 4). In addition, the value of item loading are in accordance to the threshold level (see Fig 2).
Table 4: Factor Loadings of 3 Factor Model of Transactional Leadership
Figure 2: Three Factor CFA Model of Transactional Leadership
4.2.4 Convergent and Discriminant Validities
The procedure proposed by [44] was used to establish the convergent and discriminant validities of the three factor model of transactional leadership. For convergent validity, the values of ρvc index for management by exception active (ρvc = 47%), management by exception passive (MBEP) (ρvc = 45%) and contingent reward (CR) (ρvc = 44%).
The shared variance between any of the two transactional dimensions was less than the variance extracted by respective dimensions. Discriminant validity was, therefore, established among three transactional leadership dimensions. Construct reliability was assessed by calculating Jöreskog’s rho [45], [46]. The values for Jöreskog’s rho for, MBEA, MBEP and CR were .78, .75, and .72 respectively (table 5).
Table 5: Average Variance Extracted and Shared Variance among Three factors of Transactional Leadership
|
MBEA
|
MBEP
|
CR
|
Construct Reliability
|
MBEA
|
0.47
|
|
|
0.78
|
MBEP
|
0.40
|
0.45
|
|
0.75
|
CR
|
0.26
|
0.01
|
0.44
|
0.72
|
Mediation Model Analysis
In model 1, transactional leadership is the independent variable (IV), task performance is the dependent variable (DV) and self-efficacy (MV), used as mediator. Path c is the direct link between IV and DV (see figure 3), a1 is the link between IV and MV while b1 is the link between MV and DV (see figure 3).
Fig 3. Mediation Model paths
Results mentioned in table 6 indicate that transactional leadership (IV) have positive and significant effects on employees’ task performance (DV) (path c) (Coeff= .25; p=00). Hence, H1 is supported. This is in line with the results reported [29] [30]. The results of [30] used only contingent reward factor of transactional leadership while we used both the factors (i.e. contingent rewards, and management by exception) mentioned by [22] in Multifactor factor leadership questionnaire (MLQ 5X-short). According to [30], research indicates that contingent rewards contribute to enhancing employees' performance. Additionally, the findings of a meta-analysis by [29] also reveal a positive correlation between transactional leadership and employees' performance, particularly in terms of task achievement. In summary, both studies emphasize the beneficial impact of contingent rewards and transactional leadership on enhancing employee performance, particularly in achieving tasks. Result of path a1 in table 6 shows that transactional leadership (IV) is positively and significantly related to employees’ self-efficacy (M1) (Coeff= .36; p= 00). This supports H1. We might get some interesting results of transactional effects on employees’ self -efficacy in our next models.
Self-efficacy (M1) and employees’ task performance (DV) are positively and significantly related (Coeff=.27; p= .00) (path b1). Hence, H2b is accepted. Transactional leadership style has been found to be associated with employees' task performance [29], as well as having positive effects on employees' efficacy beliefs [19]. Sufficient research supports the notion that employees' self-efficacy positively influences their task performance. We can observe from table 6 that, employees’ task performance is comparatively more responsive to employee’s self- efficacy (27%) or we can say that self-efficacy is the stronger predictor of employees’ task performance. This opens a window for an interesting discussion of comparative responsiveness of employees’ task performance to employees’ self- efficacy.
Table 6: Results of Model 1 For Direct Hypothesis
Path relationship
|
Mediation path
|
Coefficient
|
P
|
Results
|
Path c (IV---DV)
|
|
.250
|
.000
|
Accepted
|
Path a (IV---MV)
|
a1 (SE)
|
.360
|
.000
|
Accepted
|
Path b (MV---DV)
|
b1 (SE)
|
.270
|
.000
|
Accepted
|
Table 6 shows the results of Model 1 for mediating hypothesis. Results indicate that at 95% CIs, zero (0) does not exist for both self-efficacy (mediator). Hence, this proved that self-efficacy mediated the relationship between transactional leadership and employees’ task performance. Hence, our hypothesis H3 is accepted.