Barrier Factor Structure

The TABS Barriers scale includes 18 items and was developed through a concept-mapping process that identified four theoretical sub-constructs. Exploratory Factor Analysis (EFA) on the full TABS dataset included 124 responses, with 118 listwise valid responses for the factor analysis. An exploratory 3-group decomposition is also available for practitioner-oriented reporting.

See the Statistics Glossary for definitions of all psychometric terms used on this page, or the Instrument Validation page for the full results across all three constructs.

Level 1: Theory-Based Groupings (4 Groups)

The concept-mapping exercise (Appendix D) sorted the 18 barrier items into four sub-constructs based on thematic affinity and theoretical grounding in the adoption barriers literature.

Total: 2 + 6 + 4 + 6 = 18 items. The four groups have unequal sizes by design because real-world barrier categories differ in breadth.

Level 2: EFA-Derived Structure (2 Factors)

Horn’s Parallel Analysis compared actual eigenvalues against the 95th percentile of random-data eigenvalues and retained one factor. The single factor explains 26.2% of variance.

Level 3: Exploratory 3-Group Decomposition

Because F1 contains 14 of the 18 items, we explored whether it could be meaningfully sub-divided. Horn’s Parallel Analysis on F1 alone recommends retaining only 1 factor, so any split is not statistically mandated. However, a forced 2-factor extraction within F1 produces two interpretable, closely related sub-groups.

3-Group Reliability Summary

Item-Level Factor Loadings

Full 18-item loading matrix from EFA with Promax rotation (ML estimation, N=118). Primary loadings are bolded. Items are grouped by their dominant factor assignment.

Level 4: Multi-Group Stability and Cross-Validation

The 3-factor structure must hold across organizational subgroups (SMB vs Enterprise) and across random splits of the sample to support generalization. Multi-group CFA reports per-group fit; cross-validation reports Tucker congruence between independently estimated factor solutions.

Level 5: ESEM Sensitivity (Exploratory Structural Equation Modeling)

ESEM (Asparouhov & Muthen, 2009) sits between EFA and CFA: it estimates a target-rotated factor solution with all cross-loadings free. Items whose dominant loading shifts relative to the canonical 3-group assignment are candidates for reassignment in future revisions.

No items shifted dominant loadings under ESEM relative to the canonical 3-group structure.

Interpretation and Implications

Why 4 Theory Groups Become 2 Factors

The concept-mapping sub-constructs represent distinct theoretical traditions, but organizational leaders perceive barriers through a simpler lens: things within their control (internal organizational challenges) versus things imposed from outside (regulatory and compliance mandates).

The 14/4 Imbalance

F1 containing 14 items while F2 has only 4 is a legitimate asymmetry, not a flaw. Internal organizational barriers are inherently more diverse while external compliance constraints cluster tightly. The 3-group decomposition offers a more balanced practitioner view (9 / 5 / 4) for organizations seeking targeted intervention.

Practical Application

For academic reporting, use the statistically supported 2-factor structure. For practitioner dashboards and action planning, the 3-group decomposition provides more granular and actionable groupings.