Hello,
I was hoping that someone could give me some advice on an analysis I'm
trying to complete.
I have scanned 15 individuals completing a learning task where individuals
have to pick between 2 options (A and B), one of which, lets say option B,
they should learn is not only more lucrative but also the more risky choice.
Anyway, I have analysed the data and someone has suggested an extra analysis
but I'm having problems figuring out the best approach to analysing it.
Basically, the contrast of interest is change over time in the differential
contrast between risky choices and safe choices.
so (B-A)*time
Now obviously I could break risky and save choices into say 4 blocks
then take 4 contrasts [1 0 0 0 -1...] [0 1 0 0 0 -1...] [0 0 1 0 0 0 -1 ...]
and [0 0 0 1 0 0 0 -1] to a second-level 4x1 anova.
However, as learning is involved there are far fewer trials in B later on
(at least in those who learned to avoid the risky choice).
If i took A*time and B*time (where *time signifies a polynomial expansion)
and did a differential contrast either within the individual subject or
using a paired sample T-test at the second level [1 -1] {this would show me
areas where there was an interaction between time and choice(so basically,
areas that show linear decrements as they choose A(for example) but linear
increments in as they choose B, correct?
I'm not sure that is exactly the same result as what i'd get from my 4x1
anova. am i mistaken? Is it possible to impliment such an analysis without
using the anova approach, am i right about what a differential contrast
between the polynomial expansions is showing?
Sorry if it is a stupid question, any advice would be great.
Cheers
Owen
|