We use various statistical tests to comprehend the significant effect between the different stimuli under testing. Specifically talking about T test and ANOVA, the t test is used for comparing the mean between two samples. It is quite a simple test to administer and is applicable when we have to compare the mean between two samples. When we talk of more than two samples, we need to administer ANOVA.
ANOVA has great advantage over t test as it can test more than one treatment and it opens up the avenues for many further testing capabilities. ANOVA has its significance because it is otherwise impossible to compare the means of more than two samples. However, T tests are more simpler to administer and the question arises that, why not conduct t test for all possible interactions? The answer to this is that the more hypothesis tests you use, the greater is the chance of making type I error and consequently the power of the test diminishes. ANOVA that ways is a more powerful test and has more professional characteristics because it helps you to see the difference in the effectiveness of varied samples and also their longevity.
ANOVA rests on three assumptions:
- The population sample has to be normal
- The observations have to be independent in each sample
- There is homogeneity of variance
To conclude, it is imperative to use ANOVA when there are more than two conditions to compare. Though the T test is simple and appears to be less daunting but it is not worth taking the risk of the occurrence of Type I error. Sometimes, people opt for t test because of its simplicity to later realise that the experiment has turned out to be obsolete and a sheer waste of time and resources. Thus, never chose a test because of its simplicity, the outcome of the test should be relevant and useful. Else the entire research may turn out to be futile.