# What Statistical test would be most appropriate?

#### nickthorn

##### New Member
I have collected data on physical activity and well-being through two separate questionnaires. I want to check whether there is an association between the two variables. I have have used spearmans rank currently but I am not sure this is the right. My sample size is very small (n=7) making it very hard to distinguish what tests to run. Any help would be appreciated.

#### hlsmith

##### Less is more. Stay pure. Stay poor.
How are variables formatted? And data are matched right?

#### nickthorn

##### New Member
How are variables formatted? And data are matched right?
The data is not matched which is part of the problem, the first questionnaire gives the duration of moderate intensity exercise completed a week, whereas, the second questionnaire gives scores for well-being.
The variables are both ordinal, I want to see whether a higher score of 'duration of moderate intensity exercise completed per week' correlates with a higher score for well-being.

#### hlsmith

##### Less is more. Stay pure. Stay poor.
Matched her means you have an individuals value for both items. So Subject 1 has a value of 110 minutes and well-being of 6.

What scale is well-being on?

#### nickthorn

##### New Member
So the data is matched. Well-being is on the ordinal scale. The questionnaire was measured following a 6-point Likert scale ranging from 0 “never” to 5 “everyday”. Scores then summed to give total wellbeing that I want to compare with amount of physical activity completed.

#### hlsmith

##### Less is more. Stay pure. Stay poor.
But which is the true dependent variable? Does exercise make people healthier or do unhealth people opt not to exercise. Difficult to answer this with cross a sectional data.

Plot them via a scatter plot. You may have to jitter the likert questions and add some transparency to see them all. Then calculate Steadman correlation with confidence intervals.

Last edited: