goljan_fan
Forum Senior
Topics: 30
Posts: 96
|
can anybody explian how to chose following tests.They never stick to my mind.
a. Analysis of variance(ANOVA)
b. Chi-squared test
c. Fisher exact test
d. Multiple linear regression
e. Paired t-test
f. Pearson product-moment correlation formula
g. Pittman Welch permutation test
h. Simple linear regression
i. Spearman rank order formula
j. Student t-test
Edited by goljan_fan on 01/02/06 - 11:00 PM
___________________
GOD HELPS THOSE WHO HELP OTHERS.
|
| msyamp
Forum Fanatic
Topics: 60
Posts: 1,462
|
HAHAHA! I can never keep quite from seeing your caption down."god helps those..." so i would answer your question.
i will give the situation first and then the choice for each
you have three student groups (two or more) or some damn groups and you have their scores. even you can have single group measured twice or thrice to get different scores. and now you want to prove or disprove that the means of all the groups are equal or not equal. then use anova.
groups and their means being measured. anova eg. a treatment group's BP and placebo group bp and no treatment group bp is to mean for each groups are compared.
a 2x2 table requires chisquare tst.here not means of scores but the numbers are to be compared. for example no of persons with and without disease and with and without exposure are to be compare to then chisquare. now you can also do a relative risk here. but chi square is done to find if the difference is statistically significant. for eg. if the relative risk is high even then if the p value from chisquare is >.o5 then the RR is not statistically significant. that is it is due to sampling error.
___________________
If you think you can You can! If you think you cant you are right again!!
|
| sturge_weber
Forum Guru
Topics: 77
Posts: 1,042
|
well stats for me is also bad, thanks msyamp for clearing some doubts, is kaplan enough for stats
|
|
| |
| | | |
