Choosing the Appropriate Test

1
Choosing the Appropriate Test

• Chapter 21

2
Choosing a Test

• Factors to use when choosing a test
• Relationships vs. Differences
• Type of Data
• Number of Variables
• Independence of the data

3
Z-tests

• One-Sample z-test
• ? is known
• We are testing whether or not a sample belongs to
  a particular population
• We are testing differences
• The IV is categorical the DV is measurement
• The data are independent

4
t-tests (three types)

• For all t-tests
• ? is unknown We are using s
• We are testing differences
• We have only 2 groups
• The IV is categorical the DV is measurement
• (1) One-sample t-test
• The data are independent
• We are testing whether a sample belongs to a
  particular population or not
• Like z-test, except that ? is unknown

5
t-tests

• (2) Independent Samples t-test
• The data are independent
• We are testing whether two samples differ on some
  variable
• (3) Repeated Measures t-test
• The data are non-independent We usually have
  multiple observations on an individual
• We are testing differences between groups on some
  variable

6
ANOVAs (two types)

• For all ANOVAs
• We are testing differences
• We have more than 2 groups
• The IV is categorical the DV is measurement
• (1) One-way ANOVA
• The data are independent
• We are testing differences between groups on a
  variable
• (2) Repeated Measures ANOVA
• The data are non-independent Multiple
  observations
• We are testing differences between conditions

7
Correlations (4 types)

• All correlations test relationships
• They differ on the type of data of X Y
• (1) Pearson correlation
• Both X and Y are continuous
• (2) Spearman correlation
• Both X and Y are ranked data
• (3) Point-biserial correlation
• X is dichotomous and Y is continuous (or vice
  versa)
• (4) Phi Coefficient
• Both X and Y are dichotomous

8
Regression

• We want to predict a criterion variable (Y) from
  a predictor variable (X)
• From what we have been talking about X and Y have
  been continuous data (at least interval scale)

9
Chi Square

• Both X and Y are categorical
• We are testing for differences between expected
  frequencies and observed frequencies
• Goodness of fit test
• We have only one classification variable
• Test of Independence
• We have 2 classification variables