Coding help for regression - dummy variables

[Paige LaPierre]

Hello!

Quick question, and very simple to answer.

I am wanting to look at whether or not participation in an early childhood education program will effect later school readiness. I will have two dummy variables: participation and non-participation. Since I want to the regression coefficient (or b) to be participation, how should I code my variables so that is what is being looked at in my regression?

Eg:
-Participation = 1, non-participation=0
OR
-participation=0, non-participation=1

Thank you for the clarification, I'm just all turned around and cant seem to figure it out.

Paige

[Carlo Lazzaro]

Paige:
you should create a two-level categorical precictor (say, 0=no participation; 1=participation); something along the line of:

. use "C:\Program Files\Stata16\ado\base\a\auto.dta"
(1978 Automobile Data)

. tab foreign

   Car type |      Freq.     Percent        Cum.
------------+-----------------------------------
   Domestic |         52       70.27       70.27
    Foreign |         22       29.73      100.00
------------+-----------------------------------
      Total |         74      100.00

. regress price i.foreign

      Source |       SS           df       MS      Number of obs   =        74
-------------+----------------------------------   F(1, 72)        =      0.17
       Model |  1507382.66         1  1507382.66   Prob > F        =    0.6802
    Residual |   633558013        72  8799416.85   R-squared       =    0.0024
-------------+----------------------------------   Adj R-squared   =   -0.0115
       Total |   635065396        73  8699525.97   Root MSE        =    2966.4

------------------------------------------------------------------------------
       price |      Coef.   Std. Err.      t    P>|t|     [95% Conf. Interval]
-------------+----------------------------------------------------------------
     foreign |
    Foreign  |   312.2587   754.4488     0.41   0.680    -1191.708    1816.225
       _cons |   6072.423    411.363    14.76   0.000     5252.386     6892.46
------------------------------------------------------------------------------

.

Please note the use of -fvvarlist- notation.

[Nick Cox]

The choice is substantive and hinges on whatever you want to explain or predict. Often the state of concern is a minority or negative condition, e..g. dying, becoming sick or unemployed, being a smoker. If your concern is with explaining participation that should be the indicator (*) variable.

(*) For an opinionated discussion of why this is a better term than "dummy", see https://www.stata-journal.com/articl...article=dm0099

[Joro Kolev]

If you are interested in the effect of participation on something, you code your dummy to be 1 if participation; 0 otherwise. If you are interested in the effect of non-participation, then you are coding your dummy to be 1 if non-participation; 0 otherwise.

This makes the interpretation of your results after that easier.