Table 2

ANCOVA models, and solution for final model

Initial model

Augmented Model

Final Model

Solution in final model


Effect

Level

df

F

p <

df

F

p <

df

F

p <

Parameter Estimate

Std Error

Confidence Limits


Intercept

1.65

2.98

-4.22

7.51


Treatment assignment (tx)

SS

1

2.79

0.0974

1

3.24

0.0722

1

3.84

0.0510

-8.18

4.17

-16.39

.04


Race

Non-white

1

1.02

0.3165

1

0.79

0.3754

1

1.80

0.1827

-2.03

1.51

-5.03

0.97


Sex

Male

1

0.06

0.8136

1

0.13

0.7177

1

0.00

0.9977

0.00

0.94

-1.86

1.86


Phase I weight change

1

0.55

0.4644

1

2.92

0.0903

1

2.62

0.1102

0.19

0.12

-0.04

0.42


Randomization weight

1

0.67

0.4111

1

0.09

0.7670

1

0.05

0.8248

-0.01

0.03

-0.06

0.05


Phase I weight change*tx

SS

1

3.03

0.0829

1

4.45

0.0359

-0.32

0.15

-0.61

-0.02


Randomization weight*tx

SS

1

2.59

0.1088

1

2.46

0.1176

0.06

0.04

-0.02

0.14


Education

1. High School or less

3

3.71

0.011

4.44

1.34

1.80

7.09


2. Some College

1.32

0.86

-0.39

3.02


3. College degree

0.69

0.92

-1.12

2.50


5. Part college (reference)

0.00

.

.

.


Phase II %attendance

1. < 50%

4

3.62

0.006

2.30

1.33

-0.35

4.95


2. 50-63%

4.35

1.13

2.13

6.56


3. 75%

3.11

1.11

0.93

5.30


4. 87.5%

2.53

1.07

0.43

4.63


5. 100% (reference)

0.00

.

.

.


Elder et al. BMC Complementary and Alternative Medicine 2012 12:19   doi:10.1186/1472-6882-12-19

Open Data