which statistical method for my study?

#1
hi i am new to statistics and doing my dissertation study titled "effects of a pedometer based walking intervention on mental health and well-being". i am doing a 2 week data collection phase whereby i get participants to record there steps going about their normal routine so i can calculate average weekly steps taken for each individual participant.

After the 2 weeks are up i will split participants randomly into 2 groups (control and intervention). Control group will stick to whatever their average step count was for the next 4 weeks, intervention group will aim to increase upon average count by 15% for next 4 weeks. Each participant will complete a well-being questionnaire at the end of each week as well as pre study and post study. What i'm looking for (i think) is a causal relationship between steps taken and mental well-being score. I believe RCT such as mine are set up to be able to detect causation?

Anyway im confused as to what tests i should do, and im confused in 2 ways 1.) i know i should compare between the 2 groups but some people say that i should also do within group tests, what is the purpose of this?

2.) my between group test, would it be an independent t-test or some sort of repeated measure anova?

many thanks in advanced
 

Karabiner

TS Contributor
#2
You could use a "mixed" ANOVA with 1 repeated-measures factor (time of measurement: baseline versus post intervention), and 1 grouping factor (intervention group versus control group). The interaction between time of measurement and group memebershil will tell you whether the changes between baseline and follow-up were different between groups. Using the baseline scores and including within-subject factors uses each person as its own "control" and increases the power of your statistical analysis.

With kind regards

Karabiner