* SOC593 [02/02/2001]
* STATA EXAMPLE OF GRAPHING INTERACTION EFFECTS BETWEEN TWO CATEGORICAL VARIABLES

. des

Contains data from stats.dta
  obs:            43                          Statistics students survey
 vars:             6                          19 Apr 1997 10:38
 size:           559 (99.5% of memory free)
-------------------------------------------------------------------------------
   1. gender    byte   %8.0g       genlbl     Gender
   2. age       byte   %8.0g                  Age
   3. party     byte   %8.0g       partylbl   Political party preference
   4. grade     byte   %8.0g       gradelbl   Expected grade in statistics
   5. msat      float  %9.0g                  Mathematics SAT score
   6. quiz      byte   %8.0g                  # concepts named from reading
-------------------------------------------------------------------------------
Sorted by:  party  

. codebook gender party grade


gender ----------------------------------------------------------------- Gender
                  type:  numeric (byte)
                 label:  genlbl

                 range:  [0,1]                        units:  1
         unique values:  2                    coded missing:  0 / 43

            tabulation:  Freq.   Numeric  Label
                            12         0  male
                            31         1  female


party ---------------------------------------------- Political party preference
                  type:  numeric (byte)
                 label:  partylbl

                 range:  [1,3]                        units:  1
         unique values:  3                    coded missing:  3 / 43

            tabulation:  Freq.   Numeric  Label
                            17         1  Democrat
                             7         2  Republic
                            16         3  Independ


grade -------------------------------------------- Expected grade in statistics
                  type:  numeric (byte)
                 label:  gradelbl

                 range:  [2,4]                        units:  1
         unique values:  3                    coded missing:  0 / 43

            tabulation:  Freq.   Numeric  Label
                             6         2  C
                            28         3  B
                             9         4  A

. * COLLAPSE DATA AND GET MEAN MSAT SCORES FOR EACH GENDER-GRADE GROUP
. sort gender grade

. collapse (mean) msat, by (gender grade)

. des

Contains data
  obs:             6                          Statistics students survey
 vars:             3                          
 size:            60 (99.5% of memory free)
-------------------------------------------------------------------------------
   1. gender    byte   %8.0g       genlbl     Gender
   2. grade     byte   %8.0g       gradelbl   Expected grade in statistics
   3. msat      float  %9.0g                  (mean) msat
-------------------------------------------------------------------------------
Sorted by:  gender  grade  
     Note:  dataset has changed since last saved

. list

       gender     grade       msat 
  1.     male         C   456.6667  
  2.     male         B   596.6667  
  3.     male         A   503.3333  
  4.   female         C   493.3333  
  5.   female         B    496.875  
  6.   female         A        590  

. * RESHAPE DATA SO THAT UNIT OF ANALYSIS IS GRADE, AND MSAT SCORES ARE TRANSFO
> RMED INTO TWO VARIABLES: ONE FOR MALE AND ONE FOR FEMALE.
. reshape groups gender 0-1

. reshape vars msat

. reshape cons grade

. reshape wide

. list

        grade      msat0      msat1 
  1.        C   456.6667   493.3333  
  2.        B   596.6667    496.875  
  3.        A   503.3333        590  

. * HURRAY!!
. * RENAME NEW MSAT VARIABLES
. rename msat0 mmsat

. label variable mmsat "male mean msat"

. rename msat1 fmsat

. label variable fmsat "female mean msat"

. * GRAPH IT!
. graph mmsat fmsat grade, connect (ll) xlabel ylabel