Ch 6 Sallis: Questionnaires Flashcards
In the Sallis et al. (2021) textbook there is a diagram showing a theoretical plane and an empirical plane. How does the operational definition of a construct relate to the theoretical definition?
It refines the theoretical definition into something measureable.
What is content validity?
he extent to which the measurement captures the entire theoretical construct. ex: if you’re measuring mathematical ability, a test with only algebra questions would have low content validity because it doesn’t cover geometry, calculus, or other areas of math.
What is construct validity based on? (2)
it is expressed and evaluated as two subdimensions: covergent validity and discriminant validity
What is construct validity?
is about whether a test or measurement truly measures the concept or idea (the “construct”) it is supposed to measure.
For example: if a test claims to measure intelligence, it should actually assess skills like reasoning, problem-solving, and understanding—not just memory or trivia knowledge.
It’s like asking: Does this test really capture the thing we’re trying to measure?
is of particular importance when measuring several constructs and the relationships between them
the degree to which measures are related to a specific construct and not related to other constructs.
What is Face Validity?
Face validity is about whether something looks like it measures what it’s supposed to measure, based on first impressions.
For example:
- If a questionnaire is supposed to measure stress, do the questions seem clearly related to stress (like asking about feeling overwhelmed)?
It’s a quick, common-sense check—often done by asking experts or a small group of people if the test seems right. It’s not as detailed as other types of validity but helps ensure the questions make sense.
What is statistical conclusion validity?
Statistical conclusion validity is about whether the results of a study are backed up by proper statistical analysis. It checks if the conclusions you draw from the data are reliable and accurate.
In short, it ensures your data and methods are solid enough to trust the conclusions.
What is reliability?
Reliability is the extent to which a measurement produces consistent results when repeated. All measurements are subject to random error. A measurement with low random error has high reliability. If we measure customer satisfaction with a questionnaire, there will be random factors that influence how respondents answer. If we immediately measure it again with the same respondents, assuming that nothing substantive has happened to change it, the results will be slightly different. In general, if respondents have well-developed opinions about what we are measur- ing, repeated measures will be quite similar. If they do not have developed opinions, they effectively guess the answers, and randomness increases. Reliability, in this sense, is not just how good the measurement instrument is (the questionnaire), but also a function of the context and respondents.
Which are the four types of validity?
- content
- construct
- face
- statistical conclusion
What is a parametric statistical method?
Parametric statistical methods (like advanced statistical tests) require data measured at the interval or ratio level because these levels provide precise and meaningful numerical information.
What is likert scales?
Ordinal scales but with opinions.
What is semantic differential scales?
Semantic differential scales are used to measure how people perceive things, like words, concepts, or brands. The process involves identifying key traits (attributes) and then creating a scale with two opposite ends (anchors) for each trait.
What is scale value measurements?
Scale value measurement turns opinions into numbers so they can be studied and understood.
What are two types of distinctive scales?
comparative and non-comparative
What is comparative scales?
Comparative scales ask people to compare different options instead of judging them on their own.
When should you not use parametric methods?
With ordinal or nominal data because they don’t meet the necessary requirements (like equal intervals or an absolute zero).
