? Part 1.  Descriptive statistics
      SAMPLE  ; all $
      DSTAT   ; rhs = * $
      CREATE  ; loginc=log(income)$
      CALC    ; xbar=xbr(loginc);sdv=sdv(loginc)$
      CREATE  ; normal=rnn(xbar,sdv)$
      KERNEL  ; rhs=loginc,normal$

      NAMELIST ; demogrfc = age, female, married $
      NAMELIST ; years = year1984,year1985,year1986,year1987,year1988,year1991 $

Part 2.  Basic linear regression
      REGRESS ; lhs = income ; rhs = one, demogrfc, years $
      CALC    ; rsq0 = rsqrd $
      REGRESS ; lhs = income ; rhs = one, demogrfc, years, educ $
      CALC    ; rsq1 = rsqrd $
      CALC    ; list ; fstat = ((rsq1 - rsq0)/1) / ((1-rsq1)/(n-kreg)) $

Part 3. Least absolute deviations
      SAMPLE  ; 1 - 100 $  (use the first 100 observations)
      REGRESS ; lhs = income ; rhs = one,demogrfc $
      REGRESS ; lhs = income ; rhs = one,demogrfc ; alg=LAD ; nbt=25 $
      REGRESS ; lhs = income ; rhs = one,demogrfc ; alg=LAD ; nbt=25 $
      CALC    ; ran (12345) $ (use a 5 digit odd number)
      REGRESS ; lhs = income ; rhs = one,demogrfc ; alg=LAD ; nbt=25 $
      CALC    ; ran (12345) $ (use a 5 digit odd number)=
      REGRESS ; lhs = income ; rhs = one,demogrfc ; alg=LAD ; nbt=25 $

Part 4.  A cluster corrected variance matrix
      SAMPLE  ;all $
      REGRESS ; lhs = income ; rhs = one,demogrfc,years; test:demogrfc$
      REGRESS ; lhs = income ; rhs = one,demogrfc,years ; cluster=id $

Part 5. Wald test of a hypothesis
      SAMPLE  ; all $
      REGRESS ; lhs = income ; rhs = one,demogrfc,years ;cluster=id
              ; test: years = 0 $
      REGRESS ; lhs = income ; rhs = one,demogrfc,years ; cluster=id$
      MATRIX  ; by=b(5:10) ; vy=varb(5:10,5:10) $
      MATRIX  ; list ; wld = by'<vy>by $
      REGRESS ; lhs = income ; rhs = one,demogrfc,years; cluster=id 
              ; test:demogrfc$

Part 6. Partial effects and simulation
      REGRESS ; Lhs = loginc ; rhs = one,age, educ, female, age*educ,
            age^2, age*female, educ*female $
      PARTIAL ; effects: age | educ = 12,14,16,18,20 $
      PARTIAL ; effects: educ & age = 25(3)64 ; plot(ci) $
      PARTIAL ; effects: female | educ = 12,16,20 & age = 25(5)64 ; plot $
      SIMULATE; scenario: female | educ = 12,16,20 & age = 25(5)64 ; plot $

Part 7. Normality tests
      REGRESS ; Lhs = loginc ; Rhs = one,demogrfc,years ; Res=e ; quietly $
      CREATE  ; v2 = e^2 ; v3 = v2*e ; v4 = v3*e $
      CALC    ; sv = sqr (xbr (v2)) ;  m3 = xbr(v3) ; m4 = xbr (v4) $
      CALC    ; List ; Chisq = N *( ( m3/sv^3)^2 / 6  +  ((m4/sv^4) - 3)^2 /24 )  $
      CALC    ; List ; kst(e)$
      KERNEL  ; Rhs = e ; normal $

Part 8. Panel Data Regressions
      SAMPLE ; all $
      SETPANEL ; group = id ; pds = ti $
      REGRESS  ; Lhs = income ; Rhs = one,age,educ,female,female*age 
               ; panel ; Fixed effects $
      REGRESS  ; Lhs = income ; Rhs = one,age,educ,female,female*age 
               ; panel ; Random effects $

Part 9. More panel data analysis
      NAMELIST ; x=one,age,educ,hsat$
      CREATE   ; ageb  = group mean(age, pds=ti)$
      CREATE   ; educb = group mean(educ,pds=ti)$
      CREATE   ; hsatb = group mean(hsat,pds=ti)$
      REGR     ; lhs = income ; rhs = age,educ,hsat ; Panel ; fixed effects $
      REGR     ; lhs = income ; rhs = age,educ,hsat,ageb,educb,hsatb $

Part 10.  A random parameter model
	REGRESS  ; Lhs = income ; Rhs = one,age,educ,hsat,working
               ; Panel ; RPM=female ; Pts=10 ; Halton 
               ; Fcn = educ(n) ; Parameters $
      CREATE   ; b_educ = beta_i(_stratum)$
      CREATE   ; t=prd(id)$
      KERNEL   ; if[t=1] ; rhs=b_educ ; group=female$