CODEWITHSUNDEEP
rplotlminteraction
rowbust
rowbust

Reputation: 461

Interplot labels when interacting a factor

So I need to plot the interaction of a factor variable using interplot in R. I have been able to figure everything out except an important part of it: how to change the label of the factor that gets plotted. Here's a replicable example showing the issue:

set.seed(507)
df <- data.frame(
  outcome = sample(1:7, 1000, replace = T),
  scale = sample(1:7, 1000, replace = T),
  dummy = sample(0:2, 1000, replace = T))

# factor the dummy
df$dummyf <- factor(df$dummy)

# linear model
lm.out <- lm(outcome ~ scale * dummyf, data = df)

# interplot
library(interplot)
interplot(lm.out, "dummyf", "scale", plot = T, hist = F, ci = 0.95)

Once I plot the interaction here's what I get:

enter image description here Now, I need to be able to change the dummyf1 and dummyf2 labels in the facets to read essentially LABEL1 and LABEL2. Here's a possible solution I tried but that isn't getting me what I need:

# possible solution?
levels(df$dummyf)[levels(df$dummyf) == 1] <- "LABEL1"
levels(df$dummyf)[levels(df$dummyf) == 2] <- "LABEL2"

# linear model
lm.out.1 <- lm(outcome ~ scale * dummyf, data = df)

# interplot
library(interplot)
interplot(lm.out, "dummyf", "scale", plot = T, hist = F, ci = 0.95)

I also tried to modify the facets of ggplot2 since interplot uses ggplot2 but haven't been able to get it to work either. Any suggestions? Thanks in advance!

Upvotes: 1

Views: 798

Answers (1)

Roman Luštrik
Roman Luštrik

Reputation: 70653

You will have to modify the underlying code. It just appends integers to the variable name, not levels.

A workaround would look like this:

library(interplot)
set.seed(507)
df <- data.frame(
  outcome = sample(1:7, 1000, replace = T),
  scale = sample(1:7, 1000, replace = T),
  dummy = sample(0:2, 1000, replace = T))

# factor the dummy
df$LABEL <- factor(df$dummy)
# df$LABEL <- df$dummyf
lm.out.1 <- lm(outcome ~ scale * LABEL, data = df)
interplot(lm.out.1, "LABEL", "scale", plot = T, hist = F, ci = 0.95)

Upvotes: 1

Related Questions