The first alternative is to manually create one dummy for each category of the variable.

Example
Variable   Educational attainment	Categories 1=Compulsory            2=Upper secondary 3=University	Dummy 1=Compulsory, 0=Other 1=Upper secondary, 0=Other 1=University, 0=Other

For example:

gen educ_comp=educ

gen educ_uppsec=educ

gen educ_uni=educ

gen educ_comp (1=1) (2=0) (3=0)

gen educ_upsec (1=0) (2=1) (3=0)

gen educ_uni (1=0) (2=0) (3=1)

In the regression analysis, all dummies for the specific variable should be included as x-variables, except one. The dummy that you exclude – and it is your own choice which one you exclude – will be the “reference category”. Each of the other dummies will be compared to the dummy that is excluded.