HED 373 Lecture Notes - Lecture 10: Stratified Sampling, Simple Random Sample, Cluster Sampling
3 Apr 2017
School
Department
Course
Professor
Document Summary
Researchers draw a sample from a population & make an inference to the population with that generalizable sample. Unbiased sampling: simple random sample- every member of a population has an equal chance of being included in sample. Failure to identify all members of a population. Are you looking for small differences in a population? (need large sample) Error associated with sampling: population of inference(cid:0)target population(cid:0)sampling frame (coverage error)(cid:0)sample (sampling error)(cid:0)respondents (non-response error) Valid instrument (or measure): measures what it"s supposed to measure, relative to testing purpose, a matter of degree (how [not whether] valid), some traits are hard to measure (ex. Judgmental validity: content validity: based on coverage of subject matter, important components with right depth. Appropriate skills covered: face validity: on superficial inspection, appears to measure what it purports to measure. Concurrent validity (current status): validate by comparing measure to a criterion gold standard , criterion variable measured at same time. Validity coefficients: typically range between 0. 3 & 0. 5.
