Step by Step: How do you run a T-Test and a 2 way ANOVA? Please help me.60 participants Quasi experimental they will have a pre-test on their knowledge of 50 words in Spanish, and a post test...

Step by Step: How do you run a T-Test and a 2 way ANOVA?

60 participants

Quasi experimental

they will have a pre-test on their knowledge of 50 words in Spanish, and a post test after the intervention

Should I just switch from quantitative to case study?

marilynn07 | High School Teacher | (Level 3) Associate Educator

Posted on

I can't help you on the "mathematical" parts of your question except that I have all of my formulas written out long-hand in my notebooks. I am including some links that I found useful when doing my research.

I am thinking that a case study would be better since your group is "small" by the standards that a T test or an ANOVA test would use.  I used case studies because my test groups were under 100 participants.  You will certainly want to check with your advisor regarding what you might be able to do or not do.  I used a "problem paper" rather than a thesis  for my final project.

neela | High School Teacher | (Level 3) Valedictorian

Posted on

ANOVA stands for ANalysis Of VAriance techniques used in testing the effect of a treatment . Usually we take the measurements before and after the treatment to study whether a treatment has made any effect.

You 60 students may be tested before training and your individual marks  obtained are recorded. After a period of training you may be given the test again and your individual marks  obtained are recorded.

Now it is a statistical test  where we calculate  a  static called t as below based on mean marks m1 obtained before the training and the mean marks m2 obtained after the training, and the mean sum of the squared deviations from the respective means:

The value of  t =  |(m1-m2)|/S[1/[n1 +1/n2], where  m1 and m2 are sample means and n1 and n2 are sample size ( in  our case n1=n2=60). And S = (sum of the squared deviation from the mean m1 of 1st sample + similar sum from the second sample) /(n1+2  -  2).

If the observed value ot the calculated  t from the experiment is very high compared to the theoretical  value of t , which is available in tables , we reject that there is any effect of treatment.

Hope this helps.