How to explain correlations?

[Student Ravina]

Hi, I did a chi-square to test correlation between a categorial (nominal) independent variable and a dependent binary variable. There seems to be significance btween the 2 variables.
Now I want to explore the significance. So I've made dummy binary variables of the independent variable (ethnicity= creole: 1=yes, 0=no; indigenious: 1=yes, 0=no; etc.). What type of test should I do to explore how the variables are related to the dependent variable. The depentend variable is a type of coping strategy (1=yes and 0=no)

[Maarten Buis]

A test requires a hypothesis. If you want to explore how the variables are related to the dependent variable, then you want to look at descriptive statistics. In this case a simple cross-tabulation will do.

.  sysuse nlsw88, clear
(NLSW, 1988 extract)

. tab race collgrad , row nofreq

           |   College graduate
      Race | Not colle  College g |     Total
-----------+----------------------+----------
     White |     74.34      25.66 |    100.00
     Black |     82.33      17.67 |    100.00
     Other |     65.38      34.62 |    100.00
-----------+----------------------+----------
     Total |     76.31      23.69 |    100.00

So in this case we see that white are about 8 percentage points more likely to have a college degree than black, while other are 9 percentage points more likely to have a college degree than white.