Reliability Reflection

Reliability Reflection

Hồ Thanh PhươngHUF04 -

Cronbach’s Alpha Result and Interpretation (click to see)

The Cronbach’s Alpha for the 6-item Usability scale is .880, which mean the questionnaire has good reliability ...

更多...

Cronbach’s Alpha Result and Interpretation (click to see)

The Cronbach’s Alpha for the 6-item Usability scale is .880, which mean the questionnaire has good reliability and good internal consistency. This means the six items (US1–US6) are reliably measuring the same underlying concept. The "Cronbach's Alpha if Item Deleted" table shows that removing any single item would lower the alpha (ranging from .850 to .873). This confirms that all items contribute positively to the scale's reliability.

How to improve the questionnaire: To improve the questionnaire based on these results, there are several options:

  • Review US4 (I can find the features I need in the AI tools without difficulty) as it has the lowest "Corrected Item-Total Correlation" (.612) and the highest "Alpha if Item Deleted" (.873). This suggests that the item should be re-examined its wording to ensure it is as clearly related to "Usability" as the other items.
  • Strengthen Item Correlations: The correlation between US1 and US4 is the weakest in the set at .413. The suggestion is improving the phrasing of these two questions to be more aligned to improve the Alpha score.
  • Scale Efficiency: Because the reliability is already quite high (.880), the questionnaire is good, but the survey can be shortened by removing the weakest performing item (US4) to reduce participant fatigue without damaging the reliability of the overall results.

Reliability Reflection

Trần Huỳnh Gia HânHUF04 -
I think your reliability interpretation is clear and well supported, especially your explanation of how each item contributes to the overall consistency of the scale. Your ...

更多...

I think your reliability interpretation is clear and well supported, especially your explanation of how each item contributes to the overall consistency of the scale. Your discussion of US4 is particularly insightful because it connects the statistics to practical questionnaire improvement rather than just reporting the numbers. However, I’m curious about your suggestion to remove US4 for efficiency. Since the reliability is already high and the item still contributes positively, do you think revising the wording might be more beneficial than removing it entirely? It would be interesting to consider how keeping the item could preserve content coverage while still improving clarity.