Jacklyn Giampa Research Methods 241 Lecture 5 2 2 16 Factorial Designs Each IV is called a Factor You could have many factors but we are sticking to two factors rare to see more than three Since we have 2 IV s you can manipulate either one as an independent group manipulation or repeated measures every subject experiences all levels manipulation In factorial designs the independent group manipulations are called between subject manipulations 3 possible designs involving 2 IVs 1 2 way between subject both IVs are manipulated between different sets of subjects 2 2 way repeated measures both IV are manipulated within the same set of subjects 3 2 way mixed one IV is manipulated between subjects and the other within the same repeated measures set of subjects repeated Naming Factorial Designs way x Designs 1st blank number of IVs 2nd and 3rd show the number of levels of each IV you need as many blanks here as you have IVs 4th blank has the name of the design for designs with 2 IVs 2 way There are three possible designs that could fill in this blank between subjects repeated measures or mixed 2 way 3x4 mixed design translates as Two independent variables three levels of one and four levels of the other mixed one IV is between subject IV and the other is repeated measures IV 3 way 2x2x3 between subject design translates as Three IVs two with 2 levels and one with three levels each manipulated between different sets of subjects Rote 1 3 Another three separate subjects 10 12 Imagery 4 6 as you want different subjects for each level 13 15 Loci 7 9 16 18 Memory study example 2 way 2x3 Between Subjects each do one level of each of the two variables 18 subjects Then make a table for 2 way 2x3 Repeated Measures and a 2 way for Mixed Design 2 way 2x3 Repeated Measure measuring the same set of subjects total of 3 subjects Imagery 1 3 1 3 Loci 1 3 all subjects have now been exposed to all levels 1 3 all subjects have been exposed to both IVs 2 way 2x3 Mixed Design concrete versus abstract will be the repeated measure and the mnemonic variables will be between subjects total number of 9 subjects Imagery 4 6 4 6 Loci 7 9 7 9 Concrete Abstract Concrete Abstract Concrete Abstract Rote 1 3 1 3 Rote 1 3 1 3 2 way factorial examples asking and answering the 3 questions Why do a study with more than one IV Why not do two separate studies Interaction the result of two or more variables combining to produce a result different from that produced by either variable done you see much more information by studying the two the effect of one IV will differ depending on the level of the other IV There are 3 questions that you can ask and answer with any 2 way factorial design 1 Is there a main effect of IVa 2 Is there a main effect of IVb 3 Is there an interaction between both Three null hypotheses three alternative hypotheses three statistical statements one for each question the 3rd question is by far the most interesting and important Example 1 Jury simulation participants read a passage describing a vehicular homicide due to drunk driving and then make a decision about how long a sentence the defendant deserves In this example the defendant was a man The IV was the sex of the participant and the hypothesis was that women would give more lenient sentences than men Results IV sex of the participant DV sentence length Mean for men 7 Mean for women 13 So the results showed the opposite of our hypotheses Women gave a harsher sentence than the men BUT WITH A FACTORIAL DESIGN 2 Ivs a Sex of participants male female b Sex of defendant male female DV jail sentence length Sex of participant Sex of defendant Male Female 7 13 10 yrs 13 7 10 yrs Mean for sex of participant 10 yr 10yr 1 2 Is there a main effect of sex of the defendant no Is there a main effect of sex on the participant yes 3 Is there an interaction of sex of participant and sex of defendant On the graph if there are no parallel lines then YES there is an interaction If parallel then NO interaction Male participants gave a more lenient sentence to john However female participants gave a more lenient sentence to Jane In other words participants gave a more lenient sentence to a defendant of the same sex Example 2 Hypothesis students of differing academic ability will rate the relevance of course materials differently depending on the style of the course being taught IVa academic success high vs low GPA IVb style of course lecture independent study DV relevance rating 1 10 1 not relevant at all 10 is extremely relevant 2 way 2x2 between subjects design High low Mean style of course Lecture IS Mean ac Success 2 4 8 6 5 5 3 7 1 2 3 Is there a main effect on academic success no Is there a main effect of style of course yes Independent study had a higher mean rating than the lecture Is there an interaction of academic success and style of course graph yes there is an interaction as they cross The effective style of course was different for low vs high GPA Style of course made little difference in rating for low GPA students but high GPA students gave much higher ratings when the course was taught as an independent study style