19th August 2012 at 7:10 pm #2172smortensenMember
Hello all you clever statisticians! I am a physical therapist, completing a little research project about physicians use of physiotherapy.
Statistics are not my strong point, so I am struggling a little to see if there is a statistically significance in my data.
I wish to determine if younger physicians are more likely to refer patients immediately to physical therapy. However, my sample size is only 8. (Only 8 responded from a total questionnaire group of 40). The age groups and groups regarding when they refer are classified as nominal data, no? So theoretically, I could do a chi-square test, but since my sample size is so small, I thought a Fisher test was appropriate?
Anyway, here’s the data:
H0: There is no difference between younger physicians referral tendencies to physical therapy and older physicians referral tendencies to physical therapy
HA: There is a difference between younger physicians referral tendencies to physical therapy and older physicians referral tendencies to physical therapy
I made the following contingency table:
Now, this is where I become stuck. Do I need to use the df (degree of freedom) for anything? It is always 1 in 2×2 contingency tables (says so in my Polgar and Thomas book- Introduction to research in the health sciences), but what else do I do to determine the statistical significance? Do I use Excel or SPSS? Do I need a significance level, ie p=0.05?
I have made some calculations and have this result: The two-tailed P value equals 0.1071. The association between rows (groups) and columns (outcomes) is considered to be not statistically significant.
I thought the fact that 100% of the younger doctors referred immediately appeared stronger than perhaps it was. Have I done anything wrong?
I apologize for the extremely low level of this question.
Any advice would be greatly appreciated.
Many thanks in advance
- You must be logged in to reply to this topic.