Chapter 25: Paired Samples and Blocks

Chapter 25: Paired Samples and Blocks

AP Statistics

Paired/Matched Samples

• This situation arises when your groups are NOT independent.

• What are common situations where this occursCompare subjects with themselves before and

after treatmentCompare subjects with themselves on different

variables• NEED TO THINK!!!!!!!!!!

Logic

• If the groups are not Independent, you CANNOT use a 2-Sample Test for Means (2-Sample t-Test)

• If the groups are Paired/Matched, you are looking at the differences in the variables and that mean difference

• Since you are now only looking at one variable (difference in variables), you will perform a one-sample t-test.

Intervals and Tests

• When you groups are paired/matched, the intervals you create and the tests you perform are IDENTICAL (mostly) to the intervals you create and the tests you perform for ONE-SAMPLE T-INTERVAL and ONE-SAMPLE T-TEST

Assumptions/Conditions

Independence Assumption:Randomization Condition10% Condition

Normal Population Assumption:Need to check each group for normality. SHOW GRAPH.

Nearly Normal Condition

Paired Data Assumption Check reasonableness of this claim. BE CAREFUL

Things to remember• It is about the differences in the variables• The null hypothesis is typically written as:

• Once you start the “mechanics” of the problem, you will be looking at things like

• You are still dealing with a t-model• Don’t use a two-sample t-test for paired data.• Don’t use a paired –t method when the data are not

paired

0:0 H

dSEd ,

Calculator HelpFor a paired t test, use the same application as for

a one-sample t-test or one-sample t-interval.