What statistic/s should be run?

How many variables?

Categorical or continuous variables?

Number of levels for categorical?

Divided by IV and DV?

 Run by computer or by hand?

 Most likely analyses

  Chi square: All categorical variables

Correlation: All continuous variables

ANOVA: Both categorical and continuous variables.

 Use for two categorical variables

 Chi Square = Sum [(O – E)²/E]

 Example:

(30 – 35.45) / 35.45

(48 – 42.54) / 42.54

(20 – 14.54) / 14.54

(12 – 17.45) / 17.45

________________

Chi Square = 5.29

 data size;

input height weight;

datalines;

72 145

64 100

68 128

66 140

62 105

;

proc corr;

var height weight;

run;

Statements in bold are required just as shown.

 Statements in italics can be anything you want, but are limited to 8 characters, must begin with a letter, and can’t be the same as a SAS statement, e.g., you can’t have a variable called data or input.

 data exprment;

input treatmnt dvarible;

datalines;

1 12

1 18

1 14

1 10

2 20

2 16

2 14

2 18

;

proc anova;

classes treatmnt;

model dvarible = treatmnt;

means treatmnt;

run;

Copyright Paul E. Spector, All rights reserved. Last modified October 21, 1998.