Analyzing data from a dependent sample experiment with 2 conditions

alfred

New Member
#1
Hello there,
I have the following dataset resulting from an experiment where I have two conditions:


Each row is a subject. Each subject was exposed to both condition1 and condition2 (depentent-sample design) and the numbers in the respective columns are the subjects' scores for each condition. The order of the exposure to condion 1 first and then 2 (or vice versa) was randomized (to avoid a training effect) and its reported in the column "order".

Now, Condition1 and Condition2 are normal (K-S test and Shapiro-Wilk test both >.05) therefore I ran a "Paired-Sample T-Test" in SPSS. The result is significant, therefore there is a significant difference between condition 1 and 2.

However... I would like to check if the Order influenced in some way the scores of the subjects in condition 1 or 2. How can I test for this in general and, specifically in SPSS? (e.g. How can I check if there is a significant difference between condition 1 and 2, taking into account also the order column? Did the order of presentation play a role in the resulting scores or not?).

Thank you in advance for your answer
 

hlsmith

Less is more. Stay pure. Stay poor.
#2
Will subjects do better on the test the second time regardless of the condition?

Is there no baseline prior to either condition?

Were you going to test effect of gender?


FYI, ALF is an acronym for Artificial Life Form.
 

alfred

New Member
#3
Thanks for the reply.
Subjects have been exposed to both condition 1 and 2, with the order randomized.
the t-test is significant, therefore subjects under one condition scored more than under the other.
Now my question is, how can I check with a test in SPSS whether the order actually influenced the results? e.g. how can they know if participants scored more in condition 2 when condition two was the 2nd condition presented, and not the first? I have an idea of couse, I could just check the means and see if the order influenced the results. However I wonder: which statistic test I should use in SPSS? (Anova? Can I use the two conditions as dependent variables and the order as Grouping condition?)
 

Karabiner

TS Contributor
#4
Maybe you are looking for a repeated measures analysis
of variance with condition as within-subject factor and
order als between-subject factor. The interaction between
the factors would be of interest.

Or you analyse both condition variables seperately with
independent samples t-tests where "order" is the grouping
variable.

With kind regards

K.