Presentation on theme: "AP Statistics C5 D2 HW: p.287 #25 – 30 Obj: to understand types of samples and possible errors Do Now: How do you think you collect data?"— Presentation transcript:
AP Statistics C5 D2 HW: p.287 #25 – 30 Obj: to understand types of samples and possible errors Do Now: How do you think you collect data?
Sampling Designs SRS (ensure that each individual has an equal chance of being selected for the sample AND that each subset has an equal chance of being the sample) Convenience sample Quota sampling -ensure that you have a certain # from each group of interest within the population -Ex: If you have a class that is 30% girls and 70% boys, you may want to choose a sample of size 10 that includes 3 girls and 7 boys.
Probability Samples SRS is one type – each element has an equal probability of being selected Stratified Random Sample - the population is divided into homogeneous groups (ex: urban, suburban, rural) called strata - get an SRS from each strata, then put all SRS together to form a sample - this ensures that all groups within a population are represented
Multistage Cluster Sample Ex: Suppose we want a sample of US households’ weekly spending on groceries. It would be a lot of work and cost a lot of money to take an SRS of households across the country. One day you might have to go to a house in Cleveland and the next day you have to go to New York, etc.
Instead you could take a multistage cluster sample: 1.Take an SRS of states in the US. 2.Take an SRS of towns within the states selected in stage 1. 3.Take an SRS of the streets in the towns selected in stage 2. 4.Take an SRS of the houses on the streets selected in stage 4.
This way, you end up interviewing 20 households on one block instead of 1 households on 20 blocks Multistage cluster sampling can be very efficient and cost effective while making sure that your sample is still randomly selected.
Systematic Random Sample - Survey every 50 th person who walks by.
Errors Sample frame error - when sample frame (list of possible subjects who could be selected in a sample) does not represent the population. Random sample error – chance variation (sample of students from this school just happens to contain only boys) Sampling method error – choosing the wrong method (convenience sampling)
Errors Response bias – wording of questions, order of answer choices, behavior of interviewer, dishonesty in responses Sample size is too small – larger samples give more accurate results