Chapter 5- Producing Data



Ap Stats Chapter 5 - Producing Data

Observational Study - Observe, but don’t influence (subjects select their own group/treatment)

Experiment - Deliberately impose a treatment on individuals to observe their response

5.1 Designing Samples

Population - entire group that we want info about

Sample - the part of the group examined

Census - contacts every individual

Sampling - Studying a part to get info about the whole

Voluntary response sampling - people who care strongly will answer (Call-in polls, on-line polls, etc...)

biased by strong opinions

Convenience Sampling - get individuals who are easiest to reach; is biased (ex: interview at the mall)

Bias - systematically favors certain outcomes (unbiased if mean of sample = mean of pop)

Homework: p 333 1 – 8

Simple random sample (SRS) of size n: n individuals from the population chosen in a way that each

set of n individuals has an equal chance to be selected

Techniques: Draw names from a hat

Use computer programs

Use a table of random digits (back cover of textbook)

Using a Random Number Table to choose an SRS: (SRS survey term NOT experiment!!!)

Assign a label to every individual

1) Enter Table B at random

2) Pick numbers until you have obtained the amount of numbers you want

3) List the members of your sample

SRS’s

Probability sample - chosen by chance

Stratified random sample - divide into groups and get separate SRS for each group

Cluster sampling - divide into groups, randomly select a group and then use all individuals in the cluster

Are these SRS’s?

1) 6 picked out of a hat from 36 students

2) 6 students from same row in a classroom

3) every other name from an alpha list

Cautions about Sample Surveys: (doesn’t eliminate bias)

1. Undercoverage - some groups of the population are left out (homeless, those without phones, non-readers of the magazine)

2. Nonresponse - an individual cant be contacted or doesn’t cooperate

3. Response Bias - affected by behavior of interviewer and/or interviewee - sex, race, uncomfortable questions, faulty memories. Anything in the way data was collected that influences response.

4. Wording of questions – can’t be confusing or biased

Until you know the details of a poll, don’t trust it!!! Ex: exact questions used, sampling design, sample size, amount of non-response, data, etc….

Inferring from samples - larger samples = more accurate results to make guesses about the population.

***Samples vary, parameters are fixed!! s, r, mean

Homework: p 341 9, 10, 13, 14 p 347 15 - 20

................
................

In order to avoid copyright disputes, this page is only a partial summary.

Google Online Preview   Download