Case Study: Likert Scale Transformation & Analysis in a Public Health Dissertation

---

1. Research Context

In likert scale transformation case study a Public Health PhD dissertation explored healthcare workers’ perceptions of operational protocols via a 10-item Likert scale (1 = Strongly Disagree … 5 = Strongly Agree). The goal was to rigorously reduce response bias, aggregate conceptually related items, and derive a robust categorical measure for subsequent inferential testing .

---

2. Analytical Workflow for likert scale transformation

Level 1: Scale Reduction (5-Point → 3-Point)

To mitigate individual differences in response style (e.g., tendency to avoid extremes), each 5-point response was collapsed into three categories:

• Agree (4–5) → 3
• Not Sure (3) → 2
• Disagree (1–2) → 1

This transformation ensured comparability across respondents with similar backgrounds but varying psychological biases .

Level 2: Combining Related Items

Conceptually linked items (e.g., two questions on “staff training”) were combined by computing each respondent’s arithmetic mean of their Level-1 scores. For triplets of related items, cut-offs were set based on possible mean ranges:

• Mean > 7/3 → 3 (Agree)
• Mean ∈ [5/3, 7/3] → 2 (Not Sure)
• Mean < 5/3 → 1 (Disagree)

Similarly, for pairs of items, the boundary of 2 separated “Agree” from other responses .

Level 3: Deriving Categorical Outcomes

Finally, across all combined item groups, a respondent’s overall mean (from Level 2) was calculated. Only those with a mean of 3 (i.e., consistent “Agree” across key constructs) were classified as “High Perception/Knowledge”; all others fell into “Low/Moderate” categories. This yielded a single categorical variable suitable for chi-square or logistic regression analyses .

---

3. Application & Hypothesis Testing in likert scale transformation

With the derived categorical measure, the dissertation tested whether the proportion of “High Perception” respondents differed by hospital type (public vs. private) using a chi-square test of independence:

• H₀: High-perception rates are independent of hospital type.
• H₁: High-perception rates differ by hospital type.

By anchoring each analytical step in clear transformation rules and combining strategies, the study ensured transparent, reproducible, and psychometrically sound use of Likert data in inferential contexts.

---

4. Key Methodological Insights on likert scale transformation

1. Bias Mitigation via Scale Collapsing: Reducing scale granularity can neutralize extreme-response tendencies without discarding meaningful variation.
2. Conceptual Aggregation: Grouping and averaging related items preserves construct validity while simplifying the dataset.
3. Data-Driven Cut-Points: Selecting cut-off boundaries based on possible mean values guarantees that each category reflects genuine consensus among related items.
4. Reproducibility: Documenting three distinct transformation levels (raw → reduced → combined → deduced) enhances auditability and peer review.

---

This methodology-centered case study exemplifies best practices for transforming and analyzing Likert-scale data in a Public Health dissertation, enabling robust categorical outcomes for downstream inferential testing.

---

Want to explore more PhD-level case studies? Check out our Comprehensive Case Studies on PhD Statistical Analysis guide page.