Getting character Z values instead of numerical z values

General
1

I am trying to produce a GLM using glmulti, though when I produce z values for the model, it says that they are characters rather than numerical. This also makes the p value calculation incorrect. The code for producing the importance, z values and p values is as follows:

mmi1<-as.data.frame(coef(res1))
mmi1<-data.frame(Estimate=mmi1$Est, SE=sqrt(mmi1$Uncond), Importance=mmi1$Importance,row.names(mmi1))
mmi1$z<-mmi1$Estimate/mmi1$SE
mmi1$p<-2*pnorm(abs(mmi1$z),lower.tail = FALSE)
names(mmi1)<-c("Estimate", "Std. Error", "Importance", "z value", "Pr(>|z|)")
mmi1$ci.lb<-mmi1[[1]]-qnorm(.975)*mmi1[[2]]
mmi1$ci.ub<-mmi1[[1]]+qnorm(.975)*mmi1[[2]]
mmi1<-mmi1[order(mmi1$Importance, decreasing = TRUE), c(1,2,4:7,3)]
summary(mmi1)

I believe the issue is with the 3rd line of code but I can't figure out what the issue is.
Any help would be appreciated, thanks!

2

Please provide a minimal REPRoducible EXample .
See e.g. whats a reproducible example reprex

It is not clear to me what you are trying to do:
did you already create a glmulti model with results in res1 and are you now deriving z-values for a next step? In that case also show how you created the model.
If you use one of the standard datasets of R (e.g. iris) you don't have to show you own dataset (that could be confidential).

3 
res1<-glmulti(Diadema~Poma+EV+Agraham+Rock, data = X10m, 
              crit = "aicc", confsetsize = 100, fitfunction = "glm", level = 1)

print(res1)
plot(res1)

top1<-weightable(res1)
top1<-top1[top1$aicc<=(top1$aicc)+2,]
top1
summary(res1@objects[[1]])

plot(res1, type="s")

eval(metafor:::.glmulti)
coef(res1)

mmi1<-as.data.frame(coef(res1))
mmi1<-data.frame(Estimate=mmi1$Est, SE=sqrt(mmi1$Uncond), Importance=mmi1$Importance,row.names(mmi1))
mmi1$z<-mmi1$Estimate/mmi1$SE
mmi1$p<-2*pnorm(abs(mmi1$z),lower.tail = FALSE)
names(mmi1)<-c("Estimate", "Std. Error", "Importance", "z value", "Pr(>|z|)")
mmi1$ci.lb<-mmi1[[1]]-qnorm(.975)*mmi1[[2]]
mmi1$ci.ub<-mmi1[[1]]+qnorm(.975)*mmi1[[2]]
mmi1<-mmi1[order(mmi1$Importance, decreasing = TRUE), c(1,2,4:7,3)]

Here is the rest thanks, it's not confidential so it's ok

4

X10m is not a dataset that I could google on the internet.
Can you provide a small subset so that we can redo your analysis?
See the link in my previous answer for methods to do that.

5

Apologies, try this

res1<-glmulti(Sepal.Length~Sepal.Width+Petal.Length+Petal.Width, data = iris, 
              crit = "aicc", confsetsize = 100, fitfunction = "glm", level = 1)

print(res1)
plot(res1)

top1<-weightable(res1)
top1<-top1[top1$aicc<=(top1$aicc)+2,]
top1
summary(res1@objects[[1]])

plot(res1, type="s")

eval(metafor:::.glmulti)
coef(res1)

mmi1<-as.data.frame(coef(res1))
mmi1<-data.frame(Estimate=mmi1$Est, SE=sqrt(mmi1$Uncond), Importance=mmi1$Importance,row.names(mmi1))
mmi1$z<-mmi1$Estimate/mmi1$SE
mmi1$p<-2*pnorm(abs(mmi1$z),lower.tail = FALSE)
names(mmi1)<-c("Estimate", "Std. Error", "Importance", "z value", "Pr(>|z|)")
mmi1$ci.lb<-mmi1[[1]]-qnorm(.975)*mmi1[[2]]
mmi1$ci.ub<-mmi1[[1]]+qnorm(.975)*mmi1[[2]]
mmi1<-mmi1[order(mmi1$Importance, decreasing = TRUE), c(1,2,4:7,3)]

summary(mmi1)

Thanks again for the help

6

Hello @tomjohnson10 ,
I think the problem was that in your names(mmi1) <- .... statement you forget to name the variable/intercept term. Your calculations look alright.

library(glmulti)
#> Warning: package 'glmulti' was built under R version 4.1.1
#> Loading required package: rJava
#> Loading required package: leaps

res1<-glmulti(Sepal.Length~Sepal.Width+Petal.Length+Petal.Width, data = iris, 
              crit = "aicc", confsetsize = 100, fitfunction = "glm", level = 1)
#> Initialization...
#> TASK: Exhaustive screening of candidate set.
#> Fitting...
#> Completed.

# statements not relevant for this question (??) omitted

mmi1<-as.data.frame(coef(res1))
mmi1<-data.frame(Estimate=mmi1$Est, SE=sqrt(mmi1$Uncond), Importance=mmi1$Importance,row.names(mmi1))
mmi1$z<-mmi1$Estimate/mmi1$SE
mmi1$p<-2*pnorm(abs(mmi1$z),lower.tail = FALSE)
sapply(mmi1,class) # inserted HO
#>        Estimate              SE      Importance row.names.mmi1.               z 
#>       "numeric"       "numeric"       "numeric"     "character"       "numeric" 
#>               p 
#>       "numeric"
names(mmi1)<-c("Estimate", "Std. Error", "Importance", "variable"," z value", "Pr(>|z|)") # changed HO
mmi1$ci.lb<-mmi1[[1]]-qnorm(.975)*mmi1[[2]]
mmi1$ci.ub<-mmi1[[1]]+qnorm(.975)*mmi1[[2]]
mmi1<-mmi1[order(mmi1$Importance, decreasing = TRUE), c(1,2,4:8,3)] #changed HO
mmi1
#>     Estimate Std. Error     variable   z value     Pr(>|z|)      ci.lb
#> 3  1.8561143 0.25084071  (Intercept)  7.399574 1.366224e-13  1.3644756
#> 4  0.7090615 0.05677313 Petal.Length 12.489386 8.530993e-36  0.5977882
#> 2  0.6508207 0.06665394  Sepal.Width  9.764174 1.604132e-22  0.5201814
#> 1 -0.5563173 0.12767546  Petal.Width -4.357277 1.316908e-05 -0.8065566
#>        ci.ub Importance
#> 3  2.3477531  1.0000000
#> 4  0.8203348  1.0000000
#> 2  0.7814601  1.0000000
#> 1 -0.3060780  0.9997029

summary(mmi1)
#>     Estimate         Std. Error        variable             z value     
#>  Min.   :-0.5563   Min.   :0.05677   Length:4           Min.   :-4.357  
#>  1st Qu.: 0.3490   1st Qu.:0.06418   Class :character   1st Qu.: 4.460  
#>  Median : 0.6799   Median :0.09716   Mode  :character   Median : 8.582  
#>  Mean   : 0.6649   Mean   :0.12549                      Mean   : 6.324  
#>  3rd Qu.: 0.9958   3rd Qu.:0.15847                      3rd Qu.:10.445  
#>  Max.   : 1.8561   Max.   :0.25084                      Max.   :12.489  
#>     Pr(>|z|)             ci.lb             ci.ub           Importance    
#>  Min.   :0.000e+00   Min.   :-0.8066   Min.   :-0.3061   Min.   :0.9997  
#>  1st Qu.:0.000e+00   1st Qu.: 0.1885   1st Qu.: 0.5096   1st Qu.:0.9999  
#>  Median :0.000e+00   Median : 0.5590   Median : 0.8009   Median :1.0000  
#>  Mean   :3.292e-06   Mean   : 0.4190   Mean   : 0.9109   Mean   :0.9999  
#>  3rd Qu.:3.292e-06   3rd Qu.: 0.7895   3rd Qu.: 1.2022   3rd Qu.:1.0000  
#>  Max.   :1.317e-05   Max.   : 1.3645   Max.   : 2.3478   Max.   :1.0000
Created on 2021-08-28 by the reprex package (v2.0.0)
7

Thankyou! I appreciate the help

8

This topic was automatically closed 7 days after the last reply. New replies are no longer allowed.

If you have a query related to it or one of the replies, start a new topic and refer back with a link.