0
$\begingroup$

I am trying to learn ttest and ANOVA for my project. I need to know that my subjects answers "effected" by the feedback they get from the first question.
What should I use to show "sth did effect or not sth"? ANOVA or ttest?

Thanks in advance.

My data is:

A person answers a question and receives a feedback as "good/bad" than he/she answers another question. Does the second answer effected by the feedback or not? That is what i want to learn.

$\endgroup$
4
  • $\begingroup$ Can you describe the data more clearly? ANOVA is a generalization of a two-sample t-test for more than two groups, so it is a matter of any of the two only if you want to compare the means of two or more groups (or means of a metric variable for two or more levels levels of a categorical variable). But a t-test can also be used with a single sample or a paired sample, so we need more info what it is you actually want to check and how your problem looks like. $\endgroup$
    – Momo
    Jun 19, 2014 at 20:28
  • $\begingroup$ If these are experimental data with randomly assigned conditions, I would be comfortable with using (causal) language like "X affects Y". But if these are data from surveys, you should really use more associative language. $\endgroup$
    – AdamO
    Jun 19, 2014 at 21:23
  • $\begingroup$ Thanks for the clarification. For the second answer, does the answer of the person get measured in a metric sense? E.g., is it response time or a freely assigned number of performance (say 0 to 100) or something like that, or are you recording if the person gets it correct/wrong the second time. And would you like to say something about each individual person or about all persons "on average"? $\endgroup$
    – Momo
    Jun 19, 2014 at 23:13
  • $\begingroup$ "sth"? Please expand your abbreviations. You don't have a 140-character limit. Please also be consistent with use of affect/effect $\endgroup$
    – Glen_b
    Jun 20, 2014 at 4:48

1 Answer 1

0
$\begingroup$

An ANOVA comparing 2 groups is exactly equal to a 2-sample pooled t-test with 2-tailed alternative. So it may not matter which you use.

Some advantages to using the t-test: you can do a 1-tailed test, you can do an approximation if you believe that the variances of the 2 groups differ. A confidence interval is more straight forward (automatic in some software).

Some advantages to using ANOVA: can compare more than 2 groups.

$\endgroup$

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service and acknowledge you have read our privacy policy.

Not the answer you're looking for? Browse other questions tagged or ask your own question.