# Which variables in the models below meet acceptable thresholds for accuracy, also, which model is more statically accurate

Which variables are most significant?

Weibull Model
Call:phreg(formula = Surv(time, v217y5_new) ~ v300e + v410e +
v406e + v309e + v320e, data = ERDDA, dist = "weibull", shape = 1)

Covariate          W.mean      Coef    Exp(Coef)  se(Coef)    Wald p
v300e               0.893     0.478     0.620        0.288    0.097
v410e              17.066    -0.023     0.977        0.011    0.029
v406e              48.821     0.006     1.006        0.005    0.205
v309e               3.687     0.231     1.259        0.117    0.048
v320e               2.250    -0.215     0.807        0.173    0.215

log(scale)                    8.464               0.448     0.000
Shape is fixed at  1

Events                    84
Total time at risk        399623
Max. log. likelihood      -783.03
LR test statistic         24.47
Degrees of freedom        5
Overall p-value           0.000176599

#################################################################
Call:
coxph(formula = Surv(time, v217y5_new) ~ v300e + v410e + v406e +
v309e + v320e, data = ERDDA, method = "efron")

n= 571, number of events= 84
(60 observations deleted due to missingness)

coef exp(coef)  se(coef)      z Pr(>|z|)
v300e -0.794299  0.451898  0.300964 -2.639  0.00831 **
v410e -0.024211  0.976079  0.010962 -2.209  0.02720 *
v406e  0.007030  1.007055  0.004839  1.453  0.14632
v309e  0.294238  1.342104  0.118667  2.480  0.01316 *
v320e -0.248514  0.779959  0.179505 -1.384  0.16622
---
Signif. codes:  0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1

exp(coef) exp(-coef) lower .95 upper .95
v300e    0.4519     2.2129    0.2505    0.8151
v410e    0.9761     1.0245    0.9553    0.9973
v406e    1.0071     0.9930    0.9975    1.0167
v309e    1.3421     0.7451    1.0636    1.6935
v320e    0.7800     1.2821    0.5486    1.1088

Concordance= 0.671  (se = 0.031 )
Rsquare= 0.051   (max possible= 0.797 )
Likelihood ratio test= 30.05  on 5 df,   p=0.00001
Wald test            = 29.66  on 5 df,   p=0.00002
Score (logrank) test = 31.48  on 5 df,   p=0.000008
#################################################################
Variable/label
v003e = cabinet
v300e = Does the cabinet represent the start of a new government
v410e = Cabinet Preference Range
v406e = Parliamentary Preference Range
v309e = Effective No. Parl Parties lower chamber
v320e = Number of Cabinet Parties

• (1) What do you mean by "most significant?" (2) Just dumping output at us is not a productive way to explain a problem. Consider providing a little context, a statement of your objectives, and focusing on a clearly-stated problem. There's more guidance about asking statistical questions available in our help center.
– whuber
Dec 11 '18 at 20:39
• Sorry to not be more specific. I am very new to the statistic models and don't understand the P-value or other things that show if something is statistically significant. So in these models, I can't tell which variables would meet acceptable thresholds for accuracy. I also cannot tell which of the models is more statically accurate. This particular problem is trying to determine what variables contribute to a change in the cabinet of a government. I hope this clarifies it more and thanks for taking the time to look at it. I'm really over my head here. Dec 11 '18 at 21:17
• What does your 'time' variable measure and in what units? What does your v217y5_new stand for and what are its possible values? How many models are you comparing? I edited your question by indenting all R output by 4 white spaces so that other people can actually read it - it was a mangled up mess. For the other variables in your model, what are their possible values? Remember that the more context you provide, the higher the chances of receiving a helpful answer. Dec 12 '18 at 3:52
• What were your variables measured on (e.g., countries?) and over what time period? Dec 12 '18 at 3:54
• There are many good books on survival analysis that you should study before getting this far. Dec 12 '18 at 13:30