1

so we have to compare two categories , overweight to normal weighted persons

and wanna know if the health risks for overweight are higher than normal people. (which logically overweight will have more health risks)

minitab gives me this after doing the two-proportion test

"95% lower bound for difference: 0.0450190" im not sure how should I interpret it.

the thing is we only have number of subjects so i don't know what to compare the 95% to ?

HELP

emma
  • 11
  • Normally I'd expect it to be expressed as upper and lower bounds for a 95% confidence interval (or 90% or 99% or whatever). But there's also such a thing as a one-sided confidence interval, where you'd say that with 95% confidence the difference is more than $0.0450190$. I don't know if that's what's intended here, so this might be just a matter of language. – Michael Hardy Nov 15 '13 at 23:39
  • well the problem is the textbook doesn't give examples where we only use the lower bound, thats why I don't know how to interpret it. so you are saying that we are 95% confident the health risks for overweight person is 0.045 higher than normal persons? – emma Nov 15 '13 at 23:42
  • 1
    I think I'd want to see a verbatim statement of the problem before I could be sure I understand just what is being asked. – Michael Hardy Nov 15 '13 at 23:45
  • "Is the proportion of persons hospitalized higher among underweight, overweight or obese compared to people with normal BMI? " technically we wanna see if obesity is the reason for people being hospitalized. – emma Nov 15 '13 at 23:48
  • the only data we have in the number of subjects for all categories and if the number of those who were hospitalized. – emma Nov 15 '13 at 23:48
  • OK, it's making more sense now. How big are those numbers? It's possible that a chi-square test is being used. – Michael Hardy Nov 16 '13 at 01:34
  • the total is 1300, plus we are not allowed to use Chi-square test not just yet. – emma Nov 16 '13 at 04:55
  • actually the total is 113000, pretty big – emma Nov 16 '13 at 04:57

0 Answers0