Question
Guidance on statistics
My fellow is conducting research to identify the best bone age predictor for chronological age. Currently used method is G&P. His hypothesis is that other tests (three methods he is going to compare, I have forgotten actual names) are superior to G&P in predicting age. He is taking school children (8 - 15yrs both M/F). What statistical method he should use? And on what basis he should calculate his sample size from previous studies? (mean/variance of: a] chronological age OR b] G*P age)
All Answers (3)
-
according to my understanding, one group will be administered with the standard procedure in use (like G & P in ur case) and the other group will have the new procedure or method that the investigator is interested to test if they are superior to the standard ones. then to find the difference between two groups, t-test for independent two samples can be used.if there are more than two groups, then ANOVA can be used. However, these tests are used to compare means... -
Yup, I agree with Salima. Since you are comparing three methods to the standard one (the G&P) the you have to use the ANOVA test (given that the outcome chronological age i guess in this case is continuous). Make sure to do the ad-hoc test to compare the difference between each two groups; in this case between each method and the standard one.
As for sample size, you need to look at similar data and take out the outcome that they calculated; in this instance it would be the mean chronological age for each method. -
What is the objective
is to determine which best bone age for prediction
or compare different technique to predict
if the second you must apply the different tecnique to all group
again sample size and test used depends on your objective
