STATS 7 Lecture Notes - Lecture 5: Response Bias, Simple Random Sample, Confidence Interval
Document Summary
Difficult to obtain (need a list of all members) Helps ensure data is representative: sample survey: subgroup of larger population. Special type of observatory study: sample vs. census. A sample is faster to acquire then a census. 2500 units gives 2% m. e: add or subtract to create 95% confidence interval, confidence interval: for 95% of properly conducted surveys the interval will contain actual proportion, interpreting. I am 95% confident that 53%-59%) of the population think religion is important : confidence interval (a%-b%) vs. confidence level (how confident you are, bias in samples, biased means data are consistently too high or low, types. Selection bias: selecting participants that don"t represent the population. Nonparticipation/nonresponse: a subset of population cannot be contacted or choose not to respond. )^2: everyone in the population has a chance to be in the sample, steps. Randomly select and assign using a program. Match numbers to units to population list: choosing sample type, stratified random sample.