Help with Independent T-Test!

#1
Hi guys,

Really struggling with one part of analysis of my collected data - I'm sure it's probably something quite straightforward but just can't get my head around it!

To simplify (so as not to bore you all!), I have 2 questions, participants are coded based on whether they answered the questions correctly or incorrectly. I want to look and see if there are age differences between those who answered question 1 correctly, and those who answer question 2 correctly.

To try and answer this, I created two new variables (using if question 1 = 0 for example), one which contains only participants who answer question 1 correctly, the other containing only those who answer question 2 correctly. When I go to run an independent t-test to compare the two groups on age (I have a categorical variable which codes sample based on two age groups), I get a warning message which reads 'The independent samples table is not produced', 'a.t cannot be produced because the standard deviations of both groups are 0'.

Any help on this would be HUGELY appreciated! Head is starting to melt! Thanks for reading :)
 

gianmarco

TS Contributor
#2
May be you might want to try chi-square test on a 2x2 contingency table in which you cross-tabulate the right/wrong answer against age class. In this way, you will test for a dependence between gender and type of answer given.

Cheers
Gm
 

Karabiner

TS Contributor
#3
To simplify (so as not to bore you all!), I have 2 questions, participants are coded based on whether they answered the questions correctly or incorrectly. I want to look and see if there are age differences between those who answered question 1 correctly, and those who answer question 2 correctly.
If those groups overlap i.e. if there are cases which
answered both correctly, or both incorrectly, then you
cannot simply perform an independent samples test
with all cases. Moreover, as Gianmarco already stated,
t-test isn't appropriate if you have categorized age into
just 2 groups.

With kind regards

K.
 
#4
Thank you Karabiner and gianmarco for your suggestions. Im afraid I am still a little confused! :( conducting a chi-square test will tell me if there is an assoication between answering q1 correctly and age and answering q2 correctly an age - but what I want to discover is if there is an age difference between participants who answered q1 correctly and those who answered q2 correctly, im not sure if a chi square can tell me this?