Multivariate Analysis Tuesday, October 30, 2018 3:25 PM . . Looking at - at least 3 variables Comparing DV with minimum of 2 iv's Example constructing a contingency table: . Identify ID or DV ID #1- gender (male, female) ID #1 - race ( black, white) . DV-attitudes to abortion (favrs, do not favor) . Determine which ones go where · IV's go in the column . DV's go in the row Constructing a contingency table: Step 1- label "table showing relationship between gender, race and attitudes towards abortion Step 2- pick any IV at random - "male, female" Step 3- Divide those by our second IV- either a black male or white male, black female or white female. (now we have our 4 columns) Step 4- Now we add our DV in the row- attributes are "favors abortion" or "do not favor abortion" Step 5- total our columns A.) Count how many black males are in favor (3) B.) Count how many black males who did not favor (0) C.) Count how many white males who favor (1) D.) Count how many white males (1) E.) how many black females who favor (1) F.) how many black females who do not favor (o) G.) how many white females who favor (1) H.) how many white females who do not favor (3) Step 6- percentage your table DOWN- (see photo you took in class) Step tt- Interpret our table- ACROSS - First comment- 100% of the black males favored abortion, compared to 50% of white males. Likewise, all the black females in the sample favored abortion compared to 25% of white females. Second comment- 50% of white males do not favor abortion, compared to tt5% of white females. 0% of black males and black females do not support abortion. Blacks, regardless of gender, are more supportive of abortion than white. Overall, males support abortion more frequently than females, particularly black males. White females are least supportive of abortion. · Go across and then one final statement explaining the whole thing The logic of multivariate analysis & social statistics Introduce a third variable An experiment, start with 2 v's and want to see what happens when we throw in a 3rd variable- is the relationship going to change . Elaboration paradigm
· 4 outcomes: (slides) 3rd variable is our control variable Partial: ex: black males, white males (all subgroups are partial relationships because we are not looking at entire relationship, only attributes) · Elaboration paradigm: does our 3rd variable come before or after our IV Intervening relationship: Ex: what is the relationship between education and spending? IV- education DV- spending Depending on your education level, will determine how much you spend Relationship doesn't occur directly, it occurs via income Education level determines your income then determines your spending Therefore it's an Intervening relationship ELABORATION PARADIGM ** Very important for final- need to know this table . Is relationship holds for one group, not the other- split- specification · Determine whether relationship