# [R] Predict from glm

Luciano La Sala lucianolasala at yahoo.com.ar
Wed Mar 24 15:36:09 CET 2010

```Dear list members,

I fitted a glm model (See output below). My outcome is death, and weight
(continuous), ClutchSize (3-level factor), EggVolume (continuous), Sex
(obviously 2-level factor), and SiblingCompetence (2-level factor) are my
covariates.

I'd like to obtain the odds of death for a range of Weights, EggVolumes, and
different combinations of ClutchSize. I've tried using the "predict"
function but I just can't seem to get it right. Could anyone give me some
guidance? I am not including codes with predict function in my message
because I got it totally wrong (lots of errors) and I'd rather start from
scratch with some fresh guidance from the list.

Thank you very much. Luciano

Call:
glm(formula = Death ~ Weight + ClutchSize + EggVolume + Sex + SibComp,
family=binomial)

Deviance Residuals:
Min       1Q   Median       3Q      Max
-1.7205  -0.8476  -0.5362   0.9769   2.3132

Coefficients:
Estimate Std. Error z value Pr(>|z|)
(Intercept)           5.90543    2.60526   2.267   0.0234 *
Weight               -0.02521    0.01709  -1.475   0.1402
ClutchSizeTwo-eggs    1.61701    0.65699   2.461   0.0138 *
ClutchSizeThree-eggs  1.07775    0.66945   1.610   0.1074
EggVolume            -0.08200    0.03375  -2.430   0.0151 *
SexMale               0.77882    0.35374   2.202   0.0277 *
SibCompPresente       0.72333    0.40077   1.805   0.0711 .
---
Signif. codes:  0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1

(Dispersion parameter for binomial family taken to be 1)

Null deviance: 238.12  on 182  degrees of freedom
Residual deviance: 200.09  on 176  degrees of freedom
(84 observations deleted due to missingness)
AIC: 214.09

Number of Fisher Scoring iterations: 4

```