2 groups - one who have taken a drug, one who hasn't. both doing tests - want to see what effect the drug has on the results of the test. what statistical test do i use? a t test? related or unrelated? i dont know! ahh!
2006-11-30 03:24:13 · 8 answers · asked by Helen 4 in Social Science ➔ Psychology
no people, its different.... i need to either use a t test or a anova.. but is it one or two tailed?? im not sure???
2006-11-30 03:29:06 · update #1
one group is cocaine users one group are not. want to see what the effect of the cocaine is on their memory.
2006-11-30 03:30:27 · update #2
Depends on what your hypothesis is.
First of all, if you're comparing 2 groups, you want to use a t-test. ANOVA is for multiple groups; a 2-group ANOVA on one variable is basically just a t-test. If you're looking at normal people, cocaine users, and marijuana users, for example, and looking to see whether the groups perform the same or differently on a memory test, you do an ANOVA to see if there's an overall effect of drug use on memory. Then you can use additional tests you probably haven't learned yet in order to figure out which groups have better or worse memory scores. So anyway, for the study you asked about, you'd use a t-test.
Then, you want to know whether the groups you're comparing are independent. Since these are totally separate groups of people, one group taking cocaine and one that's not, it's an independent-samples t-test. (If you were pairing members of each sample by age or something, it would be a different type of t-test.)
Then you asked about 1- or 2-tailed. If you have no idea whether cocaine will improve or damage memory, and your question is "What does cocaine do to memory?" then you're doing a 2-tailed test. If your question is "Does cocaine improve memory?" (in other words, your hypothesis is that it improves memory and you don't care if it actually makes it worse), then it's a 1-tailed test, because you only care about 1 tail of the curve. If the question is "Does cocaine make memory worse?" and you don't care if it improves it, then you also do a 1-tailed t-test but the tail will be on the other side of the curve.
2006-11-30 04:48:31 · answer #1 · answered by Lee A 2 · 2⤊ 0⤋
Correlation and regression analysis are heavily correct, and probable the two would want for use. regardless of the undeniable fact that, if the question replaced into approximately how age reasons weight substitute (or any opposite direction around), then that could many times point out using regression. because of the fact you're in basic terms being requested if the "substitute at the same time" that's probable a correlation question.
2016-10-13 10:37:40 · answer #2 · answered by rosen 4 · 0⤊ 0⤋
You want a simple comparison between "undrugged" test takers and "drugged" test takers? Who scored better? Who finished quicker? Did the quicker do better or worse and did that depend on being "drugged" or "nondrugged"? Define what you are looking for a bit more, then the statistical tool/method will become more clear.
2006-11-30 03:28:52 · answer #3 · answered by Anonymous · 0⤊ 2⤋
More people involved the better the result 50% take the drug the other half take the placebo. Whatever you do don't let on who's take the dummy pill. If 50% die hey presto! You got great results, drug works.
2006-11-30 03:28:38 · answer #4 · answered by Anonymous · 0⤊ 3⤋
Why don't you just use a standard IQ test? You can get a free one on the internet. Give both groups the same amount of time to complete the test.
2006-11-30 03:26:50 · answer #5 · answered by ekinevel 4 · 0⤊ 5⤋
Chi squared test to examine degree of similarity?
2006-11-30 03:28:41 · answer #6 · answered by Anonymous · 1⤊ 3⤋
just use a normal hypothesis test
2006-11-30 03:26:55 · answer #7 · answered by Gucci_xoxo 2 · 0⤊ 3⤋
compare and contrast the results
2006-11-30 03:57:19 · answer #8 · answered by G-Unit 3 · 0⤊ 3⤋