R help - Sep 2012 - Interpretation of result in R

I am trying to do a quasipoisson regression to know if the frequency of
drinking of my subject is related to temperature. The problem is that I'm
not sure how to interpret my result.

1) Since my result is signifiant, can I tell that the frequency of drinking
of my subject increase linearly or exponentially?

2) When I want to quantify the increase, do I need to do an exponential
transformation of my coefficient before I can interpret the result?

See below the result of my regression.


Call:
glm(formula = Freqtot_drink ~ temper, family = quasipoisson, 
    data = data_t, na.action = na.omit)

Deviance Residuals: 
   Min      1Q  Median      3Q     Max  
-4.477  -1.663  -1.150  -0.196   8.613  

Coefficients:
            Estimate Std. Error t value Pr(>|t|)    
(Intercept) -7.88645    1.29215  -6.103 1.47e-08 ***
temper       0.33970    0.04525   7.508 1.44e-11 ***
---
Signif. codes:  0 ?***? 0.001 ?**? 0.01 ?*? 0.05 ?.? 0.1 ? ? 1 

(Dispersion parameter for quasipoisson family taken to be 7.589744)

    Null deviance: 1023.37  on 115  degrees of freedom
Residual deviance:  561.11  on 114  degrees of freedom
  (11 observations deleted due to missingness)
AIC: NA

Number of Fisher Scoring iterations: 6


Thank you for your help



--
View this message in context:
http://r.789695.n4.nabble.com/Interpretation-of-result-in-R-tp4643237.html
Sent from the R help mailing list archive at Nabble.com.

This is not actually a question about R but rather about understanding
generalized linear models.  If you are going to use such models you
need to learn what they are and what they mean.  The book "An
Introduction to Generalized Linear Models" by Annette Dobson and
Adrian Barnett (Chapman and Hall/CRC) might be a good place to
start.

     cheers,

         Rolf Turner

On 16/09/12 05:48, lpchre wrote:
> I am trying to do a quasipoisson regression to know if the frequency of
> drinking of my subject is related to temperature. The problem is that
I'm
> not sure how to interpret my result.
>
> 1) Since my result is signifiant, can I tell that the frequency of drinking
> of my subject increase linearly or exponentially?
>
> 2) When I want to quantify the increase, do I need to do an exponential
> transformation of my coefficient before I can interpret the result?
>
> See below the result of my regression.
>
>
> Call:
> glm(formula = Freqtot_drink ~ temper, family = quasipoisson,
>      data = data_t, na.action = na.omit)
>
> Deviance Residuals:
>     Min      1Q  Median      3Q     Max
> -4.477  -1.663  -1.150  -0.196   8.613
>
> Coefficients:
>              Estimate Std. Error t value Pr(>|t|)
> (Intercept) -7.88645    1.29215  -6.103 1.47e-08 ***
> temper       0.33970    0.04525   7.508 1.44e-11 ***
> ---
> Signif. codes:  0 ?***? 0.001 ?**? 0.01 ?*? 0.05 ?.? 0.1 ? ? 1
>
> (Dispersion parameter for quasipoisson family taken to be 7.589744)
>
>      Null deviance: 1023.37  on 115  degrees of freedom
> Residual deviance:  561.11  on 114  degrees of freedom
>    (11 observations deleted due to missingness)
> AIC: NA
>
> Number of Fisher Scoring iterations: 6
>