Obtaining P values for 2 non- parametric continuous variables

I am a bit of a novice of this so for give my ignorance.
Essentially i have two groups of data. One is for people with exon 19 mutations the other for those with exon 21 mutations.
Each group has different sample sizes with variables including overall survival etx. The overall survival data is a continuous variable.
My boss has asked my to calculate the means with confidence intervals of each group which I have been able to do.

She also wants p values to compare the OS between the 2 groups which has been a lot more difficult for me to calculate.
the first issue is the data is not normally distributed so i cant use t-test. I tried transforming but that did not work.
The next suggestion was to try to do a non-parametric test - Wilcoxan rank sum.
hoever the problem is I don't have a grouping variable; so i can't proceed with the test. I suppose the grouping would be the outcome would be Overall survival in exon 19 or exon 21 but this is not a separate variable in my data so i can't input it...

i'll attach an image of my problem

sorry for making any simple mistakes. thanks



Less is more. Stay pure. Stay poor.
Please define OS. What is the actual study question? Given the question, you shouldnt rule out survival analysis


TS Contributor
I supose that you do not mean overall survival (yes/no) but survival TIME?
In that case, in SPSS you put all 19+21 = 40 values for survival into 1 column.
In the next column, you indicate whether a case (the survival time) is from
group 1 or group 2. Since your total sample size is large enough, it is not necessary
that the data in each group are sampled from normally distributed populations.
So you can use t-test for independent groups. Preferably, you use the results
displayed in the "variances are not equal" row (this is the so-called Welch test).
If you do not feel comfortable with the t-test, you can use the Wilcoxon rank
sum / Mann-Whitney U-test.

With kind regards