Entering Data for Response Variables
main topics
    
 

Your data must be arranged in your worksheet in one of three ways: as binary data, as frequency data, or as event/trial data.

When the response is binary data the analysis uses a Bernoulli model. Binary data can preserve the order of data collection. Binary data and frequency data have more degrees of freedom for goodness-of-fit tests than event/trial data. When the response is event/trial data, the analysis uses a binomial model. With event/trial data, the patterns in the residuals versus fits plot can be meaningful.

The response entered as binary data or as frequency data:

Binary Data: one row for each observation

Frequency Data: two rows for each
combination of factor and covariate

C1

C2

C3

C4

 

C1

C2

C3

C4

 

Response

 

Factor

Covar

 

Response

Count

Factor

Covar

 

0

 

1

12

 

0

1

1

12

 

1

 

1

12

 

1

19

1

12

 

1

 

1

12

 

0

1

2

12

 

.

 

.

.

 

1

19

2

12

 

.

 

.

.

 

0

5

1

24

 

.

 

.

.

 

1

15

1

24

 

1

 

1

12

 

0

4

2

24

 

0

 

2

12

 

1

16

2

24

 

1

 

2

12

 

0

7

1

50

 

.

 

.

.

 

1

13

1

50

 

.

 

.

.

 

0

8

2

50

 

.

 

.

.

 

1

12

2

50

 

1

 

2

12

 

0

11

1

125

 

.

 

.

.

 

1

2

1

125

 

.

 

.

.

 

0

9

2

125

 

 

 

 

 

 

1

11

2

125

 

 

 

 

 

 

0

19

1

200

 

 

 

 

 

 

1

1

1

200

 

 

 

 

 

 

0

18

2

200

 

 

 

 

 

 

1

2

2

200

 

Summarized Data: one row for each combination of factor and covariate

C1

C2

C3

C4

 

Events

Trials

Factor

Covar

 

1

20

1

12

 

1

20

2

12

 

5

20

1

24

 

4

20

2

24

 

7

20

1

50

 

8

20

2

50

 

11

20

1

125

 

9

20

2

125

 

19

20

1

200

 

18

20

2

200