Theoritical Framework
Structure
The framework retains TOE’s three dimensions but replaces the additive combination rule with a congruence rule on the Technology–Environment pair, with Organization entering as a moderator whose effect is expected to be non-monotonic.
- T — digital transformation awareness within the firm
- E — smart-system infrastructure available to the firm
- O — organizational size
- Outcome — change in DT competency from pre- to post-training
Hypothesis
| H | Statement |
|---|---|
| H1 | A linear additive TOE specification fails to predict DT training effectiveness. |
| H2 | A Technology × Environment interaction adds explanatory power beyond linear main effects. |
| H3 | The T×E response surface exhibits a ridge along the congruence line and an asymmetric penalty for incongruence. |
| H4 | Organizational size moderates training effectiveness in an inverted-U pattern, with an interior optimum. |
| H4a | At the small-firm end, effectiveness is attenuated by resource constraints (liability of smallness). |
| H4b | At the large-firm end, effectiveness is attenuated by a ceiling effect. |
| H5 | Training causally increases infrastructure and awareness. |
| H6 | A single self-report outcome measure captures substantially less of the training effect than an objective outcome measure. |
| H7 | More than one configuration of conditions is sufficient for high training effectiveness (equifinality). |
| H8 | A compensation path exists in which strong infrastructure substitutes for weak awareness. |
Surface predictions
Under the Edwards (2002) parameterization, the P-E Fit reading of TOE makes directional predictions about the surface statistics rather than about individual regression coefficients:
| Statistic | Meaning | Prediction |
|---|---|---|
| a1 | Slope along the congruence line | Not the primary claim; may be near zero |
| a2 | Curvature along the congruence line | Non-zero, indicating an optimum exists rather than a monotonic gain |
| a3 | Asymmetry between the two directions of misfit | Non-zero if one failure mode is worse than the other |
| a4 | Curvature along the incongruence line | Negative — the signature prediction. Misfit in either direction imposes a penalty |
a4 is the hypothesis that distinguishes this framework from a conventional interaction model. An interaction term alone shows that T and E are non-additive; a negative a4 shows that the non-additivity takes the specific form P-E Fit predicts.
All hypotheses were evaluated against Draft v1 and are being re-evaluated under the pending revision, which changes the outcome measure. Estimates are therefore not reproduced here — see Products.