Need urget help for one sample z test of proportions

#1
Hi All,

I need to confirm whether I am doing it correctly.

Survey question is:
Have you ever witnessed violent episodes/incidents in the last week?
1. NO
2. Yes.

58% witnessed/experienced violent incidents and others 38% didn't with 4% missing responses.

I just need to test whether this 58% is statistically significant. I am using one sample z test of proportions and using hypothesized value of 0.38 (38% i.e those who didn't) from the same sample. I am approaching it correctly. I know I can use 0.5 as hypothesized value but I am not concerned whether this 58% is statistically significant from 50%. Another choice would be use literature as a hypothesized value from some previous research.

I need to know whether statistically I am doing it right since I am using dependent proportion.

Immediate reply would be helpful and thanks for any help.

Best Regards
 

hlsmith

Less is more. Stay pure. Stay poor.
#2
I personally would throw out the 4% number if you were confident that they were missing at random. Then recalulate your percentages. Everything seems pretty straightforward. If you post your counts or sample size we can verify your results.
 
#3
Sample size is 350. Just assume that they are 61 and 39. Can I use hypothesized value of 0.39 to test whether this 61% is significant proportion?

And Can you please let me know what would be the case if there are three categories yes, no and somewhat with 40% 30% 30% respectively. Its command for stata?
 
#6
Any one? I just want to put it like this..... If I have three categories (yes, no and don't know) for a single variable (say Effective_policies_in_hospital) with 50% 25% and 25% respectively... I am only interested to test whether this 50% is significant... Can I run prtest from stata by coding these three categories 1 0 0 and testing it against hypothesized value of 0.25....

prtest Effective_policies_in_hospital==0.25