Conducted in Python, this analysis measured the internal consistency reliability of a 20-item, four-construct questionnaire using Cronbach's alpha (α). The instrument measures perceptions across four technology-and-investigation constructs: Artificial Intelligence (AI), Blockchain Technology (BCT), Big Data Analytics (BDA), and Perceived Effectiveness of Forensic Fraud Investigation (FFI).
This script provides a construct-level reliability analysis for a 20-item Likert-type instrument (5 items per construct, N = 30 respondents). Cronbach's alpha is computed for each subscale to assess the degree to which items within each construct measure the same underlying dimension. Results are interpreted against standard psychometric thresholds, and a comparative bar chart is generated to visualise construct reliabilities at a glance.
- AI (Artificial Intelligence): α = 0.908 — Excellent
- BCT (Blockchain Technology): α = 0.931 — Excellent
- BDA (Big Data Analytics): α = 0.945 — Excellent
- FFI (Forensic Fraud Investigation): α = 0.925 — Excellent
- Sample Size: 30 respondents
- Total Items: 20 (4 constructs × 5 items)
- Cronbach's Alpha computed per construct using the standard variance-based formula
- Construct-Level Reliability Table with interpretation labels
- Automated Interpretation using conventional psychometric thresholds
- Comparative Bar Chart with Acceptable (0.70) and Good (0.80) reference lines
| Method | Purpose |
|---|---|
| Cronbach's Alpha (α) | Internal consistency of items within each construct |
| Item Variance Summation | Numerator component of the alpha formula (Σσ²ᵢ) |
| Total Score Variance | Denominator component of the alpha formula (σ²ₜ) |
| Descriptive Reliability Table | Side-by-side comparison of constructs |
| Bar Visualisation | Graphical comparison against threshold benchmarks |
Where k = number of items, σ²ᵢ = variance of item i, and σ²ₜ = variance of total scores.
| Coefficient | Range | Interpretation |
|---|---|---|
| Cronbach's α | ≥ 0.90 | Excellent |
| 0.80 – 0.89 | Good | |
| 0.70 – 0.79 | Acceptable | |
| 0.60 – 0.69 | Questionable | |
| < 0.60 | Poor |
- All four constructs exceeded the 0.90 threshold, indicating excellent internal consistency.
- BDA recorded the highest reliability (α = 0.945), followed by BCT (α = 0.931), FFI (α = 0.925), and AI (α = 0.908).
- No construct fell below 0.90, suggesting robust item homogeneity within each subscale.
| Rank | Construct | α | Interpretation |
|---|---|---|---|
| 1 | Big Data Analytics (BDA) | 0.945 | Excellent |
| 2 | Blockchain Technology (BCT) | 0.931 | Excellent |
| 3 | Forensic Fraud Investigation (FFI) | 0.925 | Excellent |
| 4 | Artificial Intelligence (AI) | 0.908 | Excellent |
The uniformly high alpha values suggest the instrument is highly reliable for measuring the four targeted constructs in this sample.
© 2026 Rioborue Alexander Oghenerume. All Rights Reserved. This repository is for viewing purposes only. No part of this work may be copied, reused, modified, reproduced, or redistributed without prior written permission. Unauthorised use will be pursued legally.
BY ACCESSING THIS REPOSITORY, YOU AGREE TO THESE TERMS!