trouble with hypotheses & t-test results interpretation

alic62

New member
Joined
Oct 9, 2009
Messages
2
Hello :) I am having some trouble with my stats homework -could you please please help me a bit... :(

This is the assignment:

A company that offers a motivational course claims that “healthy self-esteem is an
important factor in achieving your potential. Successful people are able to take measured
risks and avoid wallowing in self-pity. They are ready to see the best in all people and are
willing to trust others with responsibility”. They say that their course improves participants’
self esteem.
In order to assess the effectiveness of their course, 14 people were randomly chosen and
given an attitude assessment before and after taking the course. The assessment
consisted of them taking a short quiz which asked them to give a numerical score to
statements like “I am comfortable accepting a compliment”. The scoring system is shown
overleaf:
3 = always
2 = more than half the time
1 = Occasionally
0 = Never

The results of the attitude assessment before and after the motivational course:

Before After
24 25
10 14
17 16
36 38
12 10
18 17
28 31
31 38
11 14
39 39
29 31
16 13
26 30
37 39

(b) Set up an appropriate t-test for this investigation. You should clearly specify the null
and alternative hypotheses for your test and explain the reason for your choice of
alternative hypothesis. Perform this test using SPSS. Ok, I can perform the t-test but I cannot specify the hypotheses... how come we get multiple hypotheses and how do we figure which ones they are?

(d) State clearly any assumptions required for the application of this t-test and, by
presenting a relevant plot, comment on the validity of the test. No idea about this one again...

Could you please give me some insight into questions b & d.

Thank you soooooooooooooo much :D
 
The null hypothesis is that there is no difference between "Before" and "After". A logical alternative is that "After" is greater than "Before". The assumptions required for a valid t-test is that both samples come from a normal distribution with the same variance in this case.
 
Thanks so much :D

Now I don't mean to sound silly :oops: .. but what would be a relevant plot... a boxplot perhaps?
 
Top