Table 4.

Inter-construct Pearson correlation matrix (N = 311).

Construct	M	SD	1	2	3	4
1. Intrinsic motivation	3.90	0.67	(0.73)
2. Extrinsic motivation	4.11	0.68	0.33***	(0.72)
3. Supervisory practices	3.58	0.66	0.48***	0.57***	(0.71)
4. Job satisfaction	3.46	0.71	0.53***	0.64***	0.61***	(0.71)

***p < 0.001 (two-tailed). All correlations are statistically significant. Cronbach’s α values: Intrinsic Motivation = 0.84; Extrinsic Motivation = 0.87; Supervisory Practices = 0.81; Job Satisfaction (3-item composite) = 0.83. Inter-construct correlations are Pearson r; all p < 0.001. Diagonal entries in parentheses report the square root of the Average Variance Extracted (√AVE) for each construct: IM (√AVE = 0.73); EM (√AVE = 0.72); Supervisory Practices (√AVE = 0.71); Job Satisfaction (√AVE = 0.71). Fornell–Larcker discriminant validity is supported when √AVE exceeds all off-diagonal inter-construct correlations in the same row/column: all four √AVE values satisfy this criterion, confirming discriminant validity for all construct pairs. Note that the Supervisory Practices–Job Satisfaction pair is the most proximal (√AVE = 0.71 vs. r = 0.61; difference = 0.10), as expected given shared MSQ-SF item origin; all other pairs show larger margins. 95% confidence intervals (Fisher’s Z transformation, N = 311, SE_Z = 0.057): IM–EM r = 0.33 [0.23, 0.43]; IM–Sup r = 0.48 [0.39, 0.56]; IM–JS r = 0.53 [0.45, 0.61]; EM–Sup r = 0.57 [0.49,0.64]; EM–JS r = 0.64 [0.57, 0.70]; Sup–JS r = 0.61 [0.54, 0.68].