anova with categorical variables

[Theofanis Nikolopoulos]

egion	reg_dummy	group	b var
MAINLAND	0	yes	0,23
MAINLAND	0	no	0,15
MAINLAND	0	yes	0,2
MAINLAND	0	no	0,05
MAINLAND	0	no	0,06
IONIAN SEA	1	no	0,07
N. AIGEAN	2	no	0,02
C. AIGEAN	3	no	0,02
C. AIGEAN	3	yes	0,19
C. AIGEAN	3	no	0,18
C. AIGEAN	3	no	0,23
C. AIGEAN	3	yes	0,29
S.E. AIGEAN	4	no	0,36
S.E. AIGEAN	4	no	0,09
CRETE	5	no	0,15
CRETE	5	no	0,25
CRETE	5	no	0,11
CRETE	5	yes	0,1

I need to make an anova test in order to control for differences and significance of var b in relation to reg_dummy and group. The variables reg_dummy is categorical and has 6 values ( 0-5). The Group variable also is categorical with values of yes/no. Finally I need to make an interaction effect ( or new variable ) of region and group. What is the anova command in order to run the above? What other statistical tests should I run in orden to further strengthen the robustness of the test?

[Carlo Lazzaro]

Theofanis:
I'd go:

regress  var_b i.Group##i.Region, vce(cluster Region)

Please note that clustering the standard errors on -Region- makes sense if you have 30 Regions at least.