linear regression interpretation/categorial reference value

[Gianfranco Di Gennaro]

Dear all,
I have a question about the interpretation in STATA of the p values following the change of a reference category of a predictor.
I am studying the effect ("eff") of a treatment ("R") correcting for baseline ("bmpre", which can be 2,3,4 or 5), age ("AGE") and other covariates.

I inserted a baseline * age interaction factor into the model.
When I change the baseline reference category ("bmpre") from 2 to 5, the significance of the main effect of AGE changes.

How do you explain this?
Thank you all in advance.
PS: it's the first time I pubish a STATA output. I'm sorry if it won't be displayed in the best way.

. regress effpos ib1.R i.SEX c.AGE##i.bmpre i.dia

Source | SS df MS Number of obs = 114
-------------+---------------------------------- F(11, 102) = 14.46
Model | 106.194343 11 9.65403121 Prob > F = 0.0000
Residual | 68.0951304 102 .667599318 R-squared = 0.6093
-------------+---------------------------------- Adj R-squared = 0.5672
Total | 174.289474 113 1.54238472 Root MSE = .81707

------------------------------------------------------------------------------
effpos | Coef. Std. Err. t P>|t| [95% Conf. Interval]
-------------+----------------------------------------------------------------
0.R | .198729 .1944451 1.02 0.309 -.1869519 .5844099
1.SEX | -.1762637 .1651324 -1.07 0.288 -.5038029 .1512756
AGE | .0162547 .0436516 0.37 0.710 -.070328 .1028373
|
bmpre |
3 | .5395969 .4952407 1.09 0.278 -.4427107 1.521905
4 | .4252566 .9612166 0.44 0.659 -1.481312 2.331825
5 | 3.562841 .6331566 5.63 0.000 2.306978 4.818704
|
bmpre#c.AGE |
3 | -.0113402 .0458618 -0.25 0.805 -.1023069 .0796265
4 | .1434036 .149467 0.96 0.340 -.1530636 .4398708
5 | -.1795416 .0760778 -2.36 0.020 -.3304415 -.0286417
|
dia |
2 | -.2800586 .1732919 -1.62 0.109 -.6237822 .063665
3 | -.0587995 .2552767 -0.23 0.818 -.5651396 .4475407
|
_cons | .8820809 .4806346 1.84 0.069 -.0712556 1.835417
------------------------------------------------------------------------------

If I change the reference category of bmpre...

. regress effpos ib1.R i.SEX c.AGE##ib5.bmpre i.dia

Source | SS df MS Number of obs = 114
-------------+---------------------------------- F(11, 102) = 14.46
Model | 106.194343 11 9.65403121 Prob > F = 0.0000
Residual | 68.0951304 102 .667599318 R-squared = 0.6093
-------------+---------------------------------- Adj R-squared = 0.5672
Total | 174.289474 113 1.54238472 Root MSE = .81707

------------------------------------------------------------------------------
effpos | Coef. Std. Err. t P>|t| [95% Conf. Interval]
-------------+----------------------------------------------------------------
0.R | .198729 .1944451 1.02 0.309 -.1869519 .5844099
1.SEX | -.1762637 .1651324 -1.07 0.288 -.5038029 .1512756
AGE | -.1632869 .0629266 -2.59 0.011 -.2881015 -.0384724
|
bmpre |
2 | -3.562841 .6331566 -5.63 0.000 -4.818704 -2.306978
3 | -3.023244 .4799725 -6.30 0.000 -3.975268 -2.071221
4 | -3.137585 .9266609 -3.39 0.001 -4.975612 -1.299557
|
bmpre#c.AGE |
2 | .1795416 .0760778 2.36 0.020 .0286417 .3304415
3 | .1682014 .0645784 2.60 0.011 .0401105 .2962924
4 | .3229452 .1563563 2.07 0.041 .0128133 .6330771
|
dia |
2 | -.2800586 .1732919 -1.62 0.109 -.6237822 .063665
3 | -.0587995 .2552767 -0.23 0.818 -.5651396 .4475407
|
_cons | 4.444922 .4855814 9.15 0.000 3.481774 5.408071
------------------------------------------------------------------------------

[Carlo Lazzaro]

Gianfranco:
in your first code, the coefficient for -AGE- when -bmpre-==2 is reported;

in your first code, the coefficient for -AGE- when -bmpre-==5 is reported.

As what above is actually stating the obvious, a toy-example may be more helpful:

use "C:\Program Files\Stata17\ado\base\a\auto.dta"
. regress price i.foreign##i.rep78, allbase
note: 1.foreign#1b.rep78 identifies no observations in the sample.
note: 1.foreign#2.rep78 identifies no observations in the sample.
note: 1.foreign#5.rep78 omitted because of collinearity.

      Source |       SS           df       MS      Number of obs   =        69
-------------+----------------------------------   F(7, 61)        =      0.39
       Model |    24684607         7  3526372.43   Prob > F        =    0.9049
    Residual |   552112352        61  9051022.16   R-squared       =    0.0428
-------------+----------------------------------   Adj R-squared   =   -0.0670
       Total |   576796959        68  8482308.22   Root MSE        =    3008.5

-------------------------------------------------------------------------------
        price | Coefficient  Std. err.      t    P>|t|     [95% conf. interval]
--------------+----------------------------------------------------------------
      foreign |
    Domestic  |          0  (base)
     Foreign  |   2088.167   2351.846     0.89   0.378     -2614.64    6790.974
              |
        rep78 |
           1  |          0  (base)
           2  |   1403.125   2378.422     0.59   0.557    -3352.823    6159.073
           3  |   2042.574   2204.707     0.93   0.358    -2366.011    6451.159
           4  |   1317.056   2351.846     0.56   0.578    -3385.751    6019.863
           5  |       -360   3008.492    -0.12   0.905    -6375.851    5655.851
              |
foreign#rep78 |
  Domestic#1  |          0  (base)
  Domestic#2  |          0  (base)
  Domestic#3  |          0  (base)
  Domestic#4  |          0  (base)
  Domestic#5  |          0  (base)
   Foreign#1  |          0  (empty)
   Foreign#2  |          0  (empty)
   Foreign#3  |  -3866.574   2980.505    -1.30   0.199    -9826.462    2093.314
   Foreign#4  |  -1708.278   2746.365    -0.62   0.536    -7199.973    3783.418
   Foreign#5  |          0  (omitted)
              |
        _cons |     4564.5   2127.325     2.15   0.036      310.651    8818.349
-------------------------------------------------------------------------------

. regress price i.foreign##ib3.rep78, allbase
note: 1.foreign#1.rep78 identifies no observations in the sample.
note: 1.foreign#2.rep78 identifies no observations in the sample.

      Source |       SS           df       MS      Number of obs   =        69
-------------+----------------------------------   F(7, 61)        =      0.39
       Model |    24684607         7  3526372.43   Prob > F        =    0.9049
    Residual |   552112352        61  9051022.16   R-squared       =    0.0428
-------------+----------------------------------   Adj R-squared   =   -0.0670
       Total |   576796959        68  8482308.22   Root MSE        =    3008.5

-------------------------------------------------------------------------------
        price | Coefficient  Std. err.      t    P>|t|     [95% conf. interval]
--------------+----------------------------------------------------------------
      foreign |
    Domestic  |          0  (base)
     Foreign  |  -1778.407    1830.91    -0.97   0.335    -5439.538    1882.723
              |
        rep78 |
           1  |  -2042.574   2204.707    -0.93   0.358    -6451.159    2366.011
           2  |  -639.4491   1211.033    -0.53   0.599    -3061.059    1782.161
           3  |          0  (base)
           4  |  -725.5185   1157.969    -0.63   0.533    -3041.021    1589.984
           5  |  -2402.574   2204.707    -1.09   0.280    -6811.159    2006.011
              |
foreign#rep78 |
  Domestic#1  |          0  (base)
  Domestic#2  |          0  (base)
  Domestic#3  |          0  (base)
  Domestic#4  |          0  (base)
  Domestic#5  |          0  (base)
   Foreign#1  |          0  (empty)
   Foreign#2  |          0  (empty)
   Foreign#3  |          0  (base)
   Foreign#4  |   2158.296   2315.938     0.93   0.355    -2472.708      6789.3
   Foreign#5  |   3866.574   2980.505     1.30   0.199    -2093.314    9826.462
              |
        _cons |   6607.074   578.9845    11.41   0.000     5449.323    7764.825
-------------------------------------------------------------------------------

.

As an aside, whenever I'm not sure about what going on with my coefficients, I impose -allbaselevels- option.
-