SPSS Tips & Tricks #7 – Compute difference
Hi, this blog post will be a preparation for my blog post on Friday – about making a pairwise test. I have as you see a measurement of stress before and after a treatment period, for each person:
And on Friday I will test if there was an significant improvent of the stress. But first I have to check if the difference is normal distributed so I will calculate the difference between the 2 columns by choosing the command: Transform – Compute and then write aname of the new variable (“diff”) and the calculation (“click” in the variables or write manually):
Here is the new variable in the data file:
Also note what happens to the missing value on row 17, when calculation with missing value the new value will also be a system missing value (.):
Now I will study the new “diff”-variable and see if it’s normal distributed or not by using the command: Analyze – Descriptive Statistics – Explore:
From the button “Plots” I choose following choices:
And from the result I will have a look on 3 favorite output:
1. Boxplot: It doesn’t look quite normal because of the 2 outliers:
2. Histogram: It doesn’t look normal here either, because of the right “tail”.
3.Descriptives: If you compare mean and median they seems to differ –> not normal distributed. And i fyou look at the skewness value of 1.349 it’s more than double compare to the value of standard error (0.456) that also tell us that this is NOT normal distributed.
So I will keep this information to the blog on Friday how to get further in my Statistical journey.
If you want to learn more about how to choose the right method in your data, why not book some hours consulting with me, through email and Skype or why not book an online statistical training? Talk to your contact person at Crayon, if you are interested to talk about this.