Are you wanting to make statistical data analysis easy to understand?
One stop shop for ways to avoid pitfalls and help things run that little bit simpler.
Specifically using SPSS to create a data analysis
- Label your subjects consistently
- Label your variables in a consistent and intuitive fashion that will work across different platforms
- Use the ‘freeze panes’ in Excel to easily translate your findings
- Keep an explanation of your variables
- Use one large file, hide the variables that you are not using
- Never name a file with the suffix ‘final’ and always remember to back up important data
- Look at your data results are they what you expected?
- Keep a log of your step by step analysis
- Use scripting in SPSS
- Check all analyses and double-check the results before publishing
It is really important to check all your analyses before publish your findings. It would not surprise me how often you could re-run an analysis and the outcome is different on each occasion. Even if you view the differences not to be serious in your results, they can still be a source of pain later on. In the worst case you may find you cut and pasted something into a file forgetting that the file had been sorted in a different order. leading to, a corrupted Excel or SPSS file.
If you’re curating data fields and just carrying out an analysis here are some good general guidelines to follow;
Remember this is by no means an exhaustive list, but a guide of a few of those things that could be helpful in reaching you main objective:
- Look at descriptive statistics first.
- Trim your data prior to analysis, making it easier to focus on analysis.
- Never perform analysis on the master copy of your data.
- Base your hypothesis in theory, not on a unformatted hunch (or on the data).
- Accept that you may not find “significance”.
- Check assumptions BEFORE you analyze your data.
- Carefully select your analysis to structure outcome.
- Try to remember that there is no such thing as “BAD RESULTS”.
- Use syntax to automate repetitive analysis.
- Form clear, specific hypothesis BEFORE analysis.
I hope this helps hake your findings more relevant to your working day and it makes that difficult task a little simpler!