MGMC02H3 Lecture Notes - Lecture 6: Response Bias, Confidence Interval, Nonprobability Sampling
Document Summary
Sampling - accurate way of representing whole population without asking everybody. Asking everyone = cost time & money (even w/ best/most expensive effort) know population & list so that you need to ask everyone. # of ppl to be sampled = depends on range of opinions &behaviours & size of population. Resulting: cheaper, faster and almost as accurate (w/ greater knowledge of where/how results inaccurate) Sampling bias bias sampling = cases/ppl= not typical /representiative of population. Into real life some elements of population = more / less likely sampled. Eg: adv of knowing where or how samples are = may be biased or insufficient representing one or more pop &sub group. Use population parameters to estimate the probably margin of error. Utilizes some form of random selection through process / procedure that assures everyone in population has equal probabilities of being chosen.