[R] Singularities in glm()

[Prof Brian Ripley]

Use alias() to help here.

It is likely that one of S T L only occurs (or does not occur) at level 6 
of C.

On Wed, 19 Apr 2006, voodooochild at gmx.de wrote:

> Hello,
>
> i have the following model,
>
> poi1<-glm(F~S+T+L+C,family=poisson,x=T)
>
> where F,S,T,L are metric and C is a factor variable with the levels "0",
> "1", "2", "3", "4", "5" and "6"
>
> if i do summary(poi1), i get the following
>
> Call:
> glm(formula = F ~ S + T + L + C, family = poisson, x = T)
>
> Deviance Residuals:
>
>     Min        1Q    Median        3Q       Max
>
> -2.44054  -0.80997  -0.04627   0.69402   2.90301
>
> Coefficients: (1 not defined because of singularities)
>              Estimate Std. Error z value Pr(>|z|)
>
> (Intercept)  -10.45553    1.16221  -8.996  < 2e-16 ***
>
> S              1.08024    0.13609   7.938 2.06e-15 ***
>
> T              1.63582    0.05170  31.643  < 2e-16 ***
>
> L              3.31684    0.49965   6.638 3.17e-11 ***
>
> C1             0.21256    0.16449   1.292   0.1963
>
> C2            -0.10895    0.06675  -1.632   0.1027
>
> C3             0.15159    0.06992   2.168   0.0302 *
>
> C4             0.50949    0.05870   8.680  < 2e-16 ***
>
> C5             0.11240    0.01686   6.667 2.61e-11 ***
>
> C6                  NA         NA      NA       NA
>
> ---
>
> Signif. codes:  0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
>
> (Dispersion parameter for poisson family taken to be 1)
>
>    Null deviance: 47365.29  on 495  degrees of freedom
>
> Residual deviance:   574.64  on 487  degrees of freedom
>
> AIC: 4091.8
>
> Number of Fisher Scoring iterations: 4
>
>
> my question now is why do i get those "singularities" and where do they
> come from, i guess because of the singularites i get the NA's, what can
> i do here to avoid them?
> I think i have done some wrong dummy coding or something like that? By
> the way, i have 7 levels and only got estimates for 6 levels, whats
> wrong here?
>
> best regards
> andreas
>
> ______________________________________________
> R-help at stat.math.ethz.ch mailing list
> https://stat.ethz.ch/mailman/listinfo/r-help
> PLEASE do read the posting guide! http://www.R-project.org/posting-guide.html
>

-- 
Brian D. Ripley,                  ripley at stats.ox.ac.uk
Professor of Applied Statistics,  http://www.stats.ox.ac.uk/~ripley/
University of Oxford,             Tel:  +44 1865 272861 (self)
1 South Parks Road,                     +44 1865 272866 (PA)
Oxford OX1 3TG, UK                Fax:  +44 1865 272595