Dear Experts,
I have an event-related design with two conditions that were
defined post-hoc individually per subject: Condition A vs. Condition B.
The problem in that design is that we have an unequal number
of trials per condition depending on the subject. On average we have 50 trials
in condition A and 25 trials in condition B. In addition, there is a huge
variability between subjects, so there are some participants with 40-40 and some
with 10-70 trials, for instance.
The t-contrast between both conditions yields a much
stronger activation pattern for the contrast B > A than for A > B. Stated
differently, there is more activation in the condition with less trials.
So my question is: Is the unequal number of trials in the t
contrast problematic for the statistics? And, if yes, is there a way how I can
control for that?
Thank you so much for your answers in advance.
Best regards from sunny Zurich,
Roland
ETH
Zurich
Dr.
Roland H. Grabner
Research on
Learning and Instruction
Institute for
Behavioral Sciences
Universitaetsstrasse
6, CAB G 83.2
CH-8092
Zurich, Switzerland
+41 44
632 50 53 phone
+41 44
632 12 19 fax