Design and measurement
- •
Research design
- •
connects the question to data collection and sampling, comparison, evaluation and analysis. Quantittative, qualitative, and mixed-method designs answer different questions; examples include experiments, observational studies, surveys, case studies, and computational comparisons
- •
correlation alone does not establish causation. Computational comparisons shoudl specify a baseline, partitions, metrics, resources, repetitions, andcontrolled conditions
- •
- •
Variables and quality
- •
Variable -> observed or assigned value.
- •
Construct -> concept that may need several measures.
- •
Operational definition -> how a variable or construct is measured.
- •
Hypothesis -> testable relationship, not a result.
- •
Validity concerns the interpretation supported by a measure; reliability concerns its consistency. A consistent measure can still measure the wrong construct. Qualitative trustworthiness needs methods suited to the study.
- •