---------------------------------------------------------------------------------------------------------------- log: C:\Documents and Settings\jlee\My Documents\EC671\duration.log log type: text opened on: 20 Nov 2004, 02:17:34 . . insheet using duration.txt, clear (2 vars, 62 obs) . * declare t as survival time data . stset t failure event: (assumed to fail at time=t) obs. time interval: (0, t] exit on or before: failure ------------------------------------------------------------------------------ 62 total obs. 0 exclusions ------------------------------------------------------------------------------ 62 obs. remaining, representing 62 failures in single record/single failure data 2646 total analysis time at risk, at risk from t = 0 earliest observed entry t = 0 last observed exit t = 216 . . . * graph the Kaplan-Meier survivor function . sts graph failure _d: 1 (meaning all fail) analysis time _t: t . . * graph the Nelson-Aalen cumulative hazard function . sts graph, na failure _d: 1 (meaning all fail) analysis time _t: t . . * Cox's proportional Hazard model for T with covariates . stset t failure event: (assumed to fail at time=t) obs. time interval: (0, t] exit on or before: failure ------------------------------------------------------------------------------ 62 total obs. 0 exclusions ------------------------------------------------------------------------------ 62 obs. remaining, representing 62 failures in single record/single failure data 2646 total analysis time at risk, at risk from t = 0 earliest observed entry t = 0 last observed exit t = 216 . stcox prod, nohr failure _d: 1 (meaning all fail) analysis time _t: t Iteration 0: log likelihood = -197.3651 Iteration 1: log likelihood = -193.27501 Iteration 2: log likelihood = -193.27217 Refining estimates: Iteration 0: log likelihood = -193.27217 Cox regression -- Breslow method for ties No. of subjects = 62 Number of obs = 62 No. of failures = 62 Time at risk = 2646 LR chi2(1) = 8.19 Log likelihood = -193.27217 Prob > chi2 = 0.0042 ------------------------------------------------------------------------------ _t | Coef. Std. Err. z P>|z| [95% Conf. Interval] -------------+---------------------------------------------------------------- prod | 9.073035 3.225299 2.81 0.005 2.751565 15.39451 ------------------------------------------------------------------------------ . . * log-linear parametric hazard model . stset t failure event: (assumed to fail at time=t) obs. time interval: (0, t] exit on or before: failure ------------------------------------------------------------------------------ 62 total obs. 0 exclusions ------------------------------------------------------------------------------ 62 obs. remaining, representing 62 failures in single record/single failure data 2646 total analysis time at risk, at risk from t = 0 earliest observed entry t = 0 last observed exit t = 216 . streg prod, distribution(weibull) time failure _d: 1 (meaning all fail) analysis time _t: t Fitting constant-only model: Iteration 0: log likelihood = -102.25125 Iteration 1: log likelihood = -101.92653 Iteration 2: log likelihood = -101.92639 Iteration 3: log likelihood = -101.92639 Fitting full model: Iteration 0: log likelihood = -101.92639 Iteration 1: log likelihood = -97.474014 Iteration 2: log likelihood = -97.285696 Iteration 3: log likelihood = -97.285419 Iteration 4: log likelihood = -97.285419 Weibull regression -- accelerated failure-time form No. of subjects = 62 Number of obs = 62 No. of failures = 62 Time at risk = 2646 LR chi2(1) = 9.28 Log likelihood = -97.285419 Prob > chi2 = 0.0023 ------------------------------------------------------------------------------ _t | Coef. Std. Err. z P>|z| [95% Conf. Interval] -------------+---------------------------------------------------------------- prod | -9.332198 2.937555 -3.18 0.001 -15.0897 -3.574695 _cons | 3.779774 .1367012 27.65 0.000 3.511845 4.047704 -------------+---------------------------------------------------------------- /ln_p | .0078269 .1005017 0.08 0.938 -.1891528 .2048066 -------------+---------------------------------------------------------------- p | 1.007858 .1012914 .82766 1.227288 1/p | .9922036 .0997181 .8148049 1.208226 ------------------------------------------------------------------------------ . . * the regression above is accelerated failure-time form. Following regression is log relative-hazard form . streg prod, distribution(weibull) failure _d: 1 (meaning all fail) analysis time _t: t Fitting constant-only model: Iteration 0: log likelihood = -102.25125 Iteration 1: log likelihood = -101.92653 Iteration 2: log likelihood = -101.92639 Iteration 3: log likelihood = -101.92639 Fitting full model: Iteration 0: log likelihood = -101.92639 Iteration 1: log likelihood = -97.474014 Iteration 2: log likelihood = -97.285696 Iteration 3: log likelihood = -97.285419 Iteration 4: log likelihood = -97.285419 Weibull regression -- log relative-hazard form No. of subjects = 62 Number of obs = 62 No. of failures = 62 Time at risk = 2646 LR chi2(1) = 9.28 Log likelihood = -97.285419 Prob > chi2 = 0.0023 ------------------------------------------------------------------------------ _t | Haz. Ratio Std. Err. z P>|z| [95% Conf. Interval] -------------+---------------------------------------------------------------- prod | 12155.38 37699.24 3.03 0.002 27.84919 5305474 -------------+---------------------------------------------------------------- /ln_p | .0078269 .1005017 0.08 0.938 -.1891528 .2048066 -------------+---------------------------------------------------------------- p | 1.007858 .1012914 .82766 1.227288 1/p | .9922036 .0997181 .8148049 1.208226 ------------------------------------------------------------------------------ . . * to specify that coefficients rather than exponentiated coefficients are to be displayed . streg prod, distribution(weibull) nohr failure _d: 1 (meaning all fail) analysis time _t: t Fitting constant-only model: Iteration 0: log likelihood = -102.25125 Iteration 1: log likelihood = -101.92653 Iteration 2: log likelihood = -101.92639 Iteration 3: log likelihood = -101.92639 Fitting full model: Iteration 0: log likelihood = -101.92639 Iteration 1: log likelihood = -97.474014 Iteration 2: log likelihood = -97.285696 Iteration 3: log likelihood = -97.285419 Iteration 4: log likelihood = -97.285419 Weibull regression -- log relative-hazard form No. of subjects = 62 Number of obs = 62 No. of failures = 62 Time at risk = 2646 LR chi2(1) = 9.28 Log likelihood = -97.285419 Prob > chi2 = 0.0023 ------------------------------------------------------------------------------ _t | Coef. Std. Err. z P>|z| [95% Conf. Interval] -------------+---------------------------------------------------------------- prod | 9.405527 3.101446 3.03 0.002 3.326804 15.48425 _cons | -3.809474 .4449054 -8.56 0.000 -4.681473 -2.937476 -------------+---------------------------------------------------------------- /ln_p | .0078269 .1005017 0.08 0.938 -.1891528 .2048066 -------------+---------------------------------------------------------------- p | 1.007858 .1012914 .82766 1.227288 1/p | .9922036 .0997181 .8148049 1.208226 ------------------------------------------------------------------------------ . . * exponential distribution . streg prod, distribution(exponential) time failure _d: 1 (meaning all fail) analysis time _t: t Iteration 0: log likelihood = -102.25125 Iteration 1: log likelihood = -97.525189 Iteration 2: log likelihood = -97.288888 Iteration 3: log likelihood = -97.288441 Iteration 4: log likelihood = -97.288441 Exponential regression -- accelerated failure-time form No. of subjects = 62 Number of obs = 62 No. of failures = 62 Time at risk = 2646 LR chi2(1) = 9.93 Log likelihood = -97.288441 Prob > chi2 = 0.0016 ------------------------------------------------------------------------------ _t | Coef. Std. Err. z P>|z| [95% Conf. Interval] -------------+---------------------------------------------------------------- prod | -9.333816 2.9599 -3.15 0.002 -15.13511 -3.532519 _cons | 3.776512 .1311242 28.80 0.000 3.519513 4.033511 ------------------------------------------------------------------------------ . . * log-normal distribution, the option time is disabled, so this is a log relative-hazard form . streg prod, distribution(logn) failure _d: 1 (meaning all fail) analysis time _t: t Fitting constant-only model: Iteration 0: log likelihood = -132.94873 Iteration 1: log likelihood = -113.59021 Iteration 2: log likelihood = -103.52111 Iteration 3: log likelihood = -103.47558 Iteration 4: log likelihood = -103.47554 Iteration 5: log likelihood = -103.47554 Fitting full model: Iteration 0: log likelihood = -103.47554 Iteration 1: log likelihood = -100.69154 Iteration 2: log likelihood = -99.929197 Iteration 3: log likelihood = -99.928761 Iteration 4: log likelihood = -99.928761 Log-normal regression -- accelerated failure-time form No. of subjects = 62 Number of obs = 62 No. of failures = 62 Time at risk = 2646 LR chi2(1) = 7.09 Log likelihood = -99.928761 Prob > chi2 = 0.0077 ------------------------------------------------------------------------------ _t | Coef. Std. Err. z P>|z| [95% Conf. Interval] -------------+---------------------------------------------------------------- prod | -9.180774 3.348935 -2.74 0.006 -15.74457 -2.616981 _cons | 3.205657 .1583704 20.24 0.000 2.895256 3.516057 -------------+---------------------------------------------------------------- /ln_sig | .1928157 .0898027 2.15 0.032 .0168057 .3688256 -------------+---------------------------------------------------------------- sigma | 1.212659 .1089 1.016948 1.446035 ------------------------------------------------------------------------------ . . * log-logistic distribution, the option time is disabled, so this is a log relative-hazard form . streg prod, distribution(logl) failure _d: 1 (meaning all fail) analysis time _t: t Fitting constant-only model: Iteration 0: log likelihood = -116.44451 Iteration 1: log likelihood = -108.53143 Iteration 2: log likelihood = -104.80926 Iteration 3: log likelihood = -104.78231 Iteration 4: log likelihood = -104.78229 Iteration 5: log likelihood = -104.78229 Fitting full model: Iteration 0: log likelihood = -104.78229 Iteration 1: log likelihood = -101.72506 Iteration 2: log likelihood = -101.34104 Iteration 3: log likelihood = -101.34034 Iteration 4: log likelihood = -101.34034 Log-logistic regression -- accelerated failure-time form No. of subjects = 62 Number of obs = 62 No. of failures = 62 Time at risk = 2646 LR chi2(1) = 6.88 Log likelihood = -101.34034 Prob > chi2 = 0.0087 ------------------------------------------------------------------------------ _t | Coef. Std. Err. z P>|z| [95% Conf. Interval] -------------+---------------------------------------------------------------- prod | -9.543597 3.587275 -2.66 0.008 -16.57453 -2.512668 _cons | 3.292277 .163065 20.19 0.000 2.972676 3.611879 -------------+---------------------------------------------------------------- /ln_gam | -.3446479 .1042584 -3.31 0.001 -.5489906 -.1403052 -------------+---------------------------------------------------------------- gamma | .7084698 .0738639 .5775325 .8690929 ------------------------------------------------------------------------------ . . * Same results are yielded as those of limdep . * Note: Limdep makes no explicit distinction between accelerated failure-time form and log relative-hazard f > orm . * the results produced by Dr Lee's limdep codes amounts stata command without option "nohr" . end of do-file