I am having trouble with using SPSS as I'm relatively new to this. I have two groups, control group and treatment group. Do I enter the data for each in separate sheets or do I create a variable for ''groups''. I want to find if there is a significant difference between the two groups and and significant difference in pretest vs. post test in the treatment group.
Use one sheet - it will be less complicated. I'm assuming you have pre- and post-measurements. So use the second option, but even in that case you can enter the data in two ways.
1) The wide format:
Make separate variables (columns) for the outcome measurements (e.g. outcome_baseline, outcome_followup) also enter a group variable for the groups. It is a good idea to have a case identifier variable (id) identifying the subjects.
2) The long format:
Make one column for the outcome measurements with baseline and follow-up values on top of each other, include a new variable time indicating whether the value came from baseline (time=0) or follow-up (time=1). Also include the group variable and an id-variable identifying the study subjects (note that each case will have two rows in long form data - baseline and follow-up, and the two rows need to have the same id-value).
The wide format is used e.g. for repeated measures ANOVA. Also, t-tests utilize this format. If you are going to use the mixed model or GEE for analysis, use the long format.
You can switch between these two forms by using the data/restructure.
Regards,
Timo
This data format-array issue is ubiquitous in data science, and the format of the data array (also known as rows vs columns) differs among the statistical platform you are using (SPSS, SAS, R, STATA, MINITAB, etc.). The good news is that manipulation of the data array into either the "Stacked Variable Values" format (each row of data is an observation of a unique subject and each column is an observation of that subject variable at a different time (initial, secondary, tertiary, ...). The subject had two observations pre and post? The subjects data will have two rows, one with the pre value, one with the post value "stacked" on tob of each other. The stacked format is required in SPSS in order to run all parametric and non-parametric hypothesis tests for comparing the distributions of 2 or more groups (pre,post, etc., ....) (t-tests, Mann-Whitney U, General Linear Model one-way ANOVA, Kruskal-Wallis non-parametric ANOVA, ...).
On the other hand (and it is quite a BIG other hand) is that the parallel format (where each row is a single observation of a single subject and each COLUMN is an observation at a unique time (Pre column, Post column, additional times, ...). The stacked format is required in a single data array in order to run all of the "paired samples" (also known as 'two related samples') tests, both parametric (paired samples t-test) and non-parametric (Wilcoxon Signed Rank, etc.).
I hope this helps. Your question is a fundamental and essential question that must always be addressed and resolved at the first step of data analysis:
1. Inspect the data!
You should use one spreadsheet and create a variable that distinguishes your two groups. You will need that variable for any analysis (e.g., t-Test) that compares your two groups.
Dear Mariyam,
I think you already found the answer. But yes. When you start using SPSS, this is a common problem I guess. But if you have entered data into a spread sheet, it is not difficult to prepare data for the tests. Deciding if you want to have all the data related to pre test and post test in one column or you need to create two column can be done by yourself. Go to the relevant test and check how you need to add data for the testing purpose: if you have to add as variable 1 and 2 or you have to have group them. This way you can reorganize if it is necessary.
Good luck!
- Badges
- Science topic
- Similar topics
- Mathematical Sciences
- Statistics
- Statistical Software
- SPSS