Reply With QuoteNo I haven't and haven't been told I should....but will....how should I do this?
In an experiment to test the hypothesis of group polarization there are two groups of 120 militarists and 80 pacifists. Both groups were asked to come to a group consensus on their position regarding war and decide on one of the four following options: 1- most dovish; 2- least dovish; 3-least hawkish; 4-most hawkish. Half of the militarists and half of the doves are provided with arguments supporting both view points before the group discussion. The other half are not provided with any arguments. The independent variables are group average which is the average of group members opinions before discussion and group decision which is the group decision after discussion.
I want to test the following hypotheses: (1) that doves should have reliably more pacifist decisions thatn their pre-discussion average opion i.e., group decision> group average for doves; (2) that hawks should have reliably more militarist decisions that their pre-discussion average opinion.
(3) that doves given arguments should have reliably less pacifist decisions than doves not given any arguments; and (4) that hawks given arguments should have reliably less warmongering decisions thatn hawks given no arguments.
I was going to use paired t-tests to test hypotheses 1 & 2 and then two independent t-tests to test 3 & 4 -one for each of the two dependent variables (i.e. group average and group decision). Does anyone know whether this would be ok or should i be using more sophisticated tests - because we are expecting doves to become more dovish and hawks to become more hawkish after discussion with themselves -how do i do pre-planned comparisons - any help would be greatly appreciated.....felix9999
No I haven't and haven't been told I should....but will....how should I do this?
Posting Permissions
|
|