0
$\begingroup$

enter image description here

Here's the code for these graphs

par(mfrow=c(1,3)) plot(lm.all.income.factors, which=1) qqPlot(lm.all.income.factors$residuals, id=FALSE, main="QQ Plot") plot(lm.all.income.factors$residuals, main="Residuals vs Order")

Also I'm not quite sure how to fix the QQ Plot: it looks like the data is skewed. Perhaps a non-linear graph would be more appropriate?

Improve this question
$\endgroup$
1
  • $\begingroup$ You don't necessarily have to "fix" anything, however, the q-q plot as well as the other plots indicate that your distribution is more prone to extreme values that a normal distribution model would predict. If you want to try something different, try using a T distribution for the response variable. That will give you estimates that are less influenced by the extreme values, and also provide more sensible prediction limits for new observations. $\endgroup$ Commented Jul 15, 2023 at 21:00

1 Answer 1

Reset to default
3
$\begingroup$

I'm assuming that you've fit a linear regression model to independent observations. If so, your plots do not suggest any violations of the model assumptions. The "shotgun pattern" simply means that you have many fitted values in the middle range; such patterns will occur, for example, if you have a single predictor with lots of values in its middle range. As for the Q-Q plot, the deviations from the line look only minor. Given your apparently large sample size, these deviations shouldn't invalidate inference based on the linear regression model. In other words, based on the information you've provided, the model provides a reasonable description of the data.

Improve this answer
$\endgroup$

Start asking to get answers

Find the answer to your question by asking.

Ask question

Explore related questions

See similar questions with these tags.