Below is a snippet from the output of running a fp regression of the variable Glu on the variable T. 44 models are considered, but what are the terms in the final model? Is the final model Y= B0 + B1*sqrt(T) + B2*sqrt(T)*log(T) + e?
. fp <T> ,replace scale : regress Glu <T>, vce (cluster ID)
(fitting 44 models)
(....10%....20%....30%....40%....50%....60%....70% ....80%....90%....100%)
Fractional polynomial comparisons:
------------------------------------------------------------------------------
| Test Residual Deviance
T | df Deviance std. dev. diff. P Powers
-------------+----------------------------------------------------------------
omitted | 4 9282.744 26.307 238.420 0.000
linear | 3 9184.638 25.048 140.313 0.000 1
m = 1 | 2 9162.600 24.770 118.275 0.000 .5
m = 2 | 0 9044.325 23.346 0.000 -- -.5 -.5
------------------------------------------------------------------------------
Note: Test df is degrees of freedom, and P = P > F is sig. level for tests
comparing models vs. model with m = 2 based on deviance difference,
F(df, 985).
Linear regression Number of obs = 990
(Std. err. adjusted for 18 clusters in ID)
------------------------------------------------------------------------------
| Robust
Glu_T0 | Coefficient std. err. t P>|t| [95% conf. interval]
-------------+----------------------------------------------------------------
T_1 | 67.9893 10.23635 6.64 0.000 46.39249 89.58612
T_2 | 20.60366 2.990125 6.89 0.000 14.29505 26.91227
_cons | -40.80822 7.810368 -5.22 0.000 -57.28666 -24.32979
------------------------------------------------------------------------------------------------------------------------------------------
. fp <T> ,replace scale : regress Glu <T>, vce (cluster ID)
(fitting 44 models)
(....10%....20%....30%....40%....50%....60%....70% ....80%....90%....100%)
Fractional polynomial comparisons:
------------------------------------------------------------------------------
| Test Residual Deviance
T | df Deviance std. dev. diff. P Powers
-------------+----------------------------------------------------------------
omitted | 4 9282.744 26.307 238.420 0.000
linear | 3 9184.638 25.048 140.313 0.000 1
m = 1 | 2 9162.600 24.770 118.275 0.000 .5
m = 2 | 0 9044.325 23.346 0.000 -- -.5 -.5
------------------------------------------------------------------------------
Note: Test df is degrees of freedom, and P = P > F is sig. level for tests
comparing models vs. model with m = 2 based on deviance difference,
F(df, 985).
Linear regression Number of obs = 990
(Std. err. adjusted for 18 clusters in ID)
------------------------------------------------------------------------------
| Robust
Glu_T0 | Coefficient std. err. t P>|t| [95% conf. interval]
-------------+----------------------------------------------------------------
T_1 | 67.9893 10.23635 6.64 0.000 46.39249 89.58612
T_2 | 20.60366 2.990125 6.89 0.000 14.29505 26.91227
_cons | -40.80822 7.810368 -5.22 0.000 -57.28666 -24.32979
------------------------------------------------------------------------------------------------------------------------------------------
