This is a test version of Biostars. For the public version, visit https://www.biostars.org.
Help with graphing ANOVA results in R

Hi, I'm really new to using statistics and R and I need help with how to visualize my ANOVA results in a bar graph with error bars. I'm trying to follow this tutorial but I keep getting the error message: Error: 'names' attribute [20] must be the same length as the vector [1]

Here is my code:

data$group <- "NA" # make an empty column for the groups
data$group[1:4]<-"volant"
data$group[6:16]<-"terrestrial"
data$group[17:19]<-"semi-aquatic"
data$group[20] <- "aquatic"
cbind(data)

Group<-as.factor(data$group)
Species<-rownames(data)

x <- (data$functional_nonfunctional~data$habitat)
x<-data$functional_nonfunctional
Group<-data$habitat
names(x)<-rownames(data)
names(Group)<-rownames(data)

modelPHYLO<-aov.phylo(x~Group,tree)
summary(modelPHYLO)
var(data$functional_nonfunctional)
length(Group)
graph_summary <- ddply(data,Group, summarize, AVERAGE=mean(data$functional_nonfunctional), SE=sqrt(var(data$functional_nonfunctional)/length(functional_nonfunctional)))

I would SO appreciate any help! I'm feeling really stuck! Thank you

anova
• 9,349 views • link
updated by Ram Ram • written by sammysun • sammysun

It would help if you added a little of what your data looked like, even just the first few lines of your data.frame. Use:

head(data)

Also there are some inconsistencies in your code such as row 5 is not given a group (e.g. is it volant, terrestrial or other?) also you reassign the name Group and x to different objects which is confusing.

• 0 views • link
updated by Ram Ram • written by gtho123 gtho123

Oh wow I didn't even notice the row 5 thing thank you! The way that x is assigned to different things is confusing and doesn't make sense. The first x (x <- (data$functional_nonfunctional~data$habitat)) shouldn't be there now that I'm looking at it again.

Here is part of my table:

                   Olfactory_Bulb OlfactoryEp_Total Olfactory_resp functional_nonfunctional habitat 
Anoura_geofroyi                2.41              <NA>           <NA>                 9.985155       1
• 0 views • link
updated by Ram Ram • written by sammysun • sammysun

1 answer

I am unsure where the error occurred however if I understand you correctly you want to perform an ANOVA on your functional_nonfunctional column using the different habitats as factors. While I am unable to use the aov.phylo function as I don't have the phylogenetic information I think I can help you with graphing it.

I created some fake data as a demo to show you how to draw the graph I think you want.

data <- data.frame(Olfactory_Bulb = NA, OlfactoryEp_Total = NA, Olfactory_resp = NA,
                   functional_nonfunctional = c(rnorm(5, 10, 5), rnorm(11, 20, 5), rnorm(3, 30, 5),rnorm(1, 40, 5)),habitat = c(rep(1,5), rep(2,11), rep(3,3), rep(4,1)))

rownames(data) <- c("Anoura_aequatoris", "Anoura_cadenai", "Anoura_canishina", "Anoura_caudifera", "Anoura_cultrata", "Anoura_fistulata", "Anoura_geoffroyi", "Anoura_latidens", "Anoura_luismanueli", "Anoura_peruana", "Choeroniscus_godmani", "Choeroniscus_periosus", "Choeroniscus_minor", "Choeronycteris_mexicana", "Glossophaga_commissarisi", "Glossophaga_leachii", "Glossophaga_longirostris", "Glossophaga_morenoi", "Glossophaga_soricina", "Hylonycteris underwoodi")

This code then creates the graph:

library("plyr")
library("ggplot2")


graph_summary <- ddply(data, c("habitat"), summarize,
                       AVERAGE=mean(functional_nonfunctional),
                       SE=sqrt(var(functional_nonfunctional)/length(functional_nonfunctional)))

ggplot(data = graph_summary, aes(x = habitat, y = AVERAGE, colour = habitat))+
  geom_point()+
  geom_errorbar(aes(ymax=AVERAGE+SE, ymin=AVERAGE-SE))+
  theme(axis.text.x = element_text(angle = 90, hjust = 0, size=11),
       axis.title.x = element_text(size=14),
       axis.title.y = element_text(angle = 90, size=14))+
  scale_x_discrete("Habitat")+
  scale_y_continuous("Functional/Nonfunctional")

This gives this graph:

Note: habitat 4 has only one individual so no error bars.

• 0 views • link
updated by Ram Ram • written by gtho123 gtho123

Log in to answer this question.

More posts like this