Reliability Reflection

Discussion Post: Reflection on Questionnaire Reliability

Reliability Results

For my research, I calculated Cronbach’s Alpha to test the internal consistency of my Likert-scale items. The result was $\alpha$ = .948.

Interpretation

In social science research, a Cronbach’s Alpha above .70 is generally considered acceptable, and above .90 is considered excellent. My result of .948 indicates a very high level of internal consistency. This means the items in my scale are highly correlated and are likely measuring the same underlying construct (e.g., the perceived emotional intensity of metaphors).

Reflection Prompts

• Confidence in the Scale: I am very confident in the scale's reliability. The high Alpha coefficient suggests that if I were to administer this survey to a similar group again, I would likely get consistent results. However, a score this high (nearing .95) can sometimes suggest "redundancy," meaning some questions might be so similar that they aren't adding new information.

• Challenges in Writing Items: The biggest challenge was ensuring that the metaphors were categorized correctly. For example, distinguishing between an "internal" metaphor (ingestion) and a "physical" metaphor (external threat) required very precise wording to ensure respondents didn't get confused between the two categories.

• Improvements for the Next Draft: In the next version, I would look for "redundant" items. If two questions are asking almost the exact same thing, I might remove one to make the survey shorter and reduce participant fatigue without sacrificing the reliability of the data.