Analysing data

Table of Contents

1. Analysing data
2. Contents:
3. The database:
4. Data cleaning
5. Correcting the database
6. Analysis plan
7. Describing the sample
8. Averages
9. T-test:
10. Doing the T-test yourself
11. The results of the T-test
12. Redoing the T-test
13. In Summary

T-test:

 

We have described our sample, now we want to know if the intervention worked. Now we are going to do some statistical tests.

A T-test is used when we have ratio data and we want to compare the mean in one group compared to another.

For our study we want to see if the well-being in the intervention group improved compared to the control group. Now at a glance we can see that when we did our averages to describe the data, we also could look at the average of the wellbeing at baseline and the well being at the end of the study.

At baseline Group 0 had a wellbeing of 29 (on a 0-100 scale) and Group 1 had a wellbeing of 30 (on a 0-100 scale). They don't look different but to double check we can do a T test.

You can do this using the web program in the book.

Here we will do a t-test using Excel.

The formula is:

=ttest(first set of data to compare, second set of data to compare, number of tails, type of test)

First set of data is : the cell reference for the baseline well being of group 0

Second set of data is: the cell reference for the baseline well being of group 1

Number of tails is 2

Type of t-test is : 1 if the test is for groups who are matched or are the same person measured on two occasions.

2 if we are comparing two sample (group 1 and group 0) which have similar standard deviations

3. if we are comparing two samples (group 1 and group 0) which don't have similar standard deviations)

So for our test we put:

=ttest(h2:h18,h19:h31,2,2)

This should give us the result of 0.84

This means the p=0.84

P is the p value that tells us if there is a significant difference.

1

P=0.84 is a significant difference a) True b) False No it is not significant. To reach significance we need a p value of 0.05 or 0.01 or lower. Redo the course on understanding data if this is not clear No it is not significant. To reach significance we need a p value of 0.05 or 0.01 or lower Redo the course on understanding data if this is not clear. Your answer has been saved.