Transcript Why Sample?
Sampling Design But First a Sampling Experiment Each group of students should: 1. Pull 5 candies out of the bag 2. Weigh the candies 3. Write down the weight 4. Put the candies back in the bag!! 5. Pass the scale and bag to your neighbors 6. Silently multiply the weight of the 5 candies by 20. No Scale Candy Sample Type 3 Musketeers Carmels Tootsie Rolls Junior mints Peppermint Patties Weight (g) 60 8 7 18 17 Discussion Definitions • Sampling – procedure involving parts of the whole population • Sample – a subset of the pop. • Population – finite group of elements • Universe – infinite group of elements Why Sample? • Pragmatic reasons – Cheaper – Easier – Faster • Accurate and reliable results • Census? Sampling • Define the target population • A sampling frame – Mailing lists • Reverse directories • lists streets and the people that live on them • Sampling frame error when the entire population is not represented in the sampling frame • Sampling unit- Single Random Sampling Error vs. Nonsampling (Systematic) Error • Random Sampling Error – The difference between the sample results and the results of a census using the same methods • Systematic error – errors that are not due to chance fluctuations. Sampling frame error is a systematic error. Probability vs. Non-probability sampling • Nonprobability- the probability of any particular member of the population being chosen is unknown. • Therefore there are no appropriate statistical techniques for measuring random sampling error from a nonprobability sample. Thus making inference is inappropriate. Non-probability Sampling • Convenience Sampling – do you have a pulse? Non-probability Sampling • Judgment sampling – using your judgment to select the characteristics of interest Non-probability Sampling • Quota sampling – a min number of individuals with a certain characteristic. Non-probability Sampling • Snowball sampling – initial respondents selected with probability methods, and they refer others Probability Sampling • Simple random sampling – – everyone in pop has an equal probability of being selected Probability Sampling • Systematic Sampling– using every 50th name in a phone book after a random starting point is selected. • Sampling interval- in this case 50 • Periodicity- when the names are not ordered randomly Probability Sampling • Stratified sampling (increase homogeniety within strata, increase heterogeniety between strata) – Proportional vs. disproportional strata – Optimal allocation Probability Sampling • Cluster sampling – Area sample – Multistage area sampling Statistics • When sampling is not simple random sampling the statistics get much harder, ie more complex. • Observations need to be weighted based upon their probability of appearing in the sample. What is the appropriate sample design? But First a Sampling Experiment Each group of students should: 1. Pull 5 candies out of the bag 2. Weigh the candies 3. Write down the weight 4. Put the candies back in the bag!! 5. Pass the scale and bag to your neighbors 6. Silently multiply the weight of the 5 candies by 20. No Scale Candy Sample Type Nestle Crunch 3 Musketeers 3 Musketeers Mint Salted Nut Roll Twizzlers Starburst Tootsie Rolls Milk Duds Peppermint Patties Weight (g) 43.9 60.4 35.2 51 14 5 6.66 12 17 No Scale Candy Sample Type Twix Reese’s “Big Cup” Gum Milky Way Rolo Weight (g) 56.7 39 5.6 17 6 No Scale Candy Sample Type Crunch Heath Milk Duds 3 Muskateers Hot Tamales Weight (g) 43.9 14 12 60 14