Regress dummy variable correctly

[Mat Sko]

Hello,

I have generated a dummy variable "cellar" with has 3 answers: 1 large basement, 2 small basement or 3 crawl space, using:

tab cellar, gen(c)

To use regression with the dependent variable logprice (=houseprice), should I use:

reg logprice i.cellar

or

reg logprice c2 c3

and c1 should be left out and is the base group.

Which one is correct?

Thanks in advance.

[David Benson]

They will be the same (although it is good practice to use i.cellar so I would use that one).

* I created some toy data
dataex id price cellar  // Data shared using dataex command. To install: ssc install dataex
clear
input byte id int price byte cellar
 1 316 2
 2 252 3
 3 334 3
 4 282 3
 5 304 2
 6 154 1
 7 218 2
 8 422 3
 9 478 3
10 251 1
11 193 2
12 457 1
13 354 3
14 189 3
15 116 3
16 469 1
17 372 1
18 313 2
19 208 1
20 119 1
21 307 3
22 272 2
23 457 3
24 417 3
25 350 1
end

gen ln_price = ln(price)
label define cellar_desc 1 "Large basement" 2 "Small basement" 3 "Crawl Space"
label values cellar cellar_desc
tabulate cellar, gen(c)  // creating the indicator variables

. reg ln_price i.cellar

      Source |       SS           df       MS      Number of obs   =        25
-------------+----------------------------------   F(2, 22)        =      0.33
       Model |  .115257474         2  .057628737   Prob > F        =    0.7219
    Residual |   3.8334791        22   .17424905   R-squared       =    0.0292
-------------+----------------------------------   Adj R-squared   =   -0.0591
       Total |  3.94873657        24   .16453069   Root MSE        =    .41743

------------------------------------------------------------------------------
    ln_price |      Coef.   Std. Err.      t    P>|t|     [95% Conf. Interval]
-------------+----------------------------------------------------------------
      cellar |
          2  |  -.0127374   .2254388    -0.06   0.955    -.4802688     .454794
          3  |   .1309974   .1939638     0.68   0.506    -.2712589    .5332537
             |
       _cons |   5.591398   .1475843    37.89   0.000     5.285326    5.897469
------------------------------------------------------------------------------


. reg ln_price c2 c3

      Source |       SS           df       MS      Number of obs   =        25
-------------+----------------------------------   F(2, 22)        =      0.33
       Model |  .115257474         2  .057628737   Prob > F        =    0.7219
    Residual |   3.8334791        22   .17424905   R-squared       =    0.0292
-------------+----------------------------------   Adj R-squared   =   -0.0591
       Total |  3.94873657        24   .16453069   Root MSE        =    .41743

------------------------------------------------------------------------------
    ln_price |      Coef.   Std. Err.      t    P>|t|     [95% Conf. Interval]
-------------+----------------------------------------------------------------
          c2 |  -.0127374   .2254388    -0.06   0.955    -.4802688     .454794
          c3 |   .1309974   .1939638     0.68   0.506    -.2712589    .5332537
       _cons |   5.591398   .1475843    37.89   0.000     5.285326    5.897469
------------------------------------------------------------------------------

[Mat Sko]

Thanks for the reply David.

Are they also the same in case of a multiple regression?

For instance:

reg loglprice i.yearquarter loglot unitsf floors i.cellar

[David Benson]

Yes (although test it for yourself and see--Stata won't care! :-)