How shall I do an anova? Without splitting the groups so I have 2 groups- this brings up an error. I can do a general linear model but how do I analyse this please? It brings up two rows? See below:
'...Table 2 shows that waist measurement reductions were observed in both conditions. In inches, the waist measurement reduction was greater for the control group than the treatment group, with a mean of -1.65 inches (SD= 2.27) and -0.71 inches (SD= 1.24) respectively by the end of the 12 week study. A higher percentage loss was also identified in the control with a mean of -4.36% (SD= 5.82), compared to -1.81% (SD= 3.2) for the treatment.'
I am now looking at the significance of the waist changes between the placebo and active groups:
An anova brings up the message:
Post hoc tests are not performed for Start Waist because there are fewer than three groups.
Post hoc tests are not performed for End Waist because there are fewer than three groups.
But I still get other parts of the test on output- is this ok not to have post hoc tests??
Heres my output for General Linear model. Can you tell me if this is significant please? I am looking at the significance of the waist changes between the placebo and active groups:
Effect Value F Hypothesis df Error df Sig. Partial Eta Squared
timeofMeasurement Pillai's Trace .986 1054.832a 1.000 15.000 .000 .986
Wilks' Lambda .014 1054.832a 1.000 15.000 .000 .986
Hotelling's Trace 70.322 1054.832a 1.000 15.000 .000 .986
Roy's Largest Root 70.322 1054.832a 1.000 15.000 .000 .986
timeofMeasurement * Batch Pillai's Trace .006 .083a 1.000 15.000 .777 .006
Wilks' Lambda .994 .083a 1.000 15.000 .777 .006
Hotelling's Trace .006 .083a 1.000 15.000 .777 .006
Roy's Largest Root .006 .083a 1.000 15.000 .777 .006
a. Exact statistic
b. Design: Intercept + Batch
Within Subjects Design: timeofMeasurement
How do I know if they are statistically different from each other?