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.

NoteEstimation status

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.