In statistics, **added variable plots** are individual plots that display the relationship between a response variable and one predictor variable in a multiple linear regression model, while controlling for the presence of other predictor variables in the model.

**Note:** Sometimes these plots are also called “partial regression plots.”

These type of plots allow us to observe the relationship between each individual predictor variable and the response variable in a model while holding other predictor variables constant.

To create added variable plots in R, we can use the **avPlots()** function from the **car** package:

#loadcarpackage library(car) #fit multiple linear regression model model <- lm(y ~ x1 + x2 + ..., data = df) #create added variable plots avPlots(model)

The following example shows how to use this syntax in practice.

**Example: Added Variable Plots in R**

Suppose we fit the following multiple linear regression model in R, using data from the **mtcars** dataset:

#fit multiple linear regression model model <- lm(mpg ~ disp + hp + drat, data = mtcars) #view summary of model summary(model) Call: lm(formula = mpg ~ disp + hp + drat, data = mtcars) Residuals: Min 1Q Median 3Q Max -5.1225 -1.8454 -0.4456 1.1342 6.4958 Coefficients: Estimate Std. Error t value Pr(>|t|) (Intercept) 19.344293 6.370882 3.036 0.00513 ** disp -0.019232 0.009371 -2.052 0.04960 * hp -0.031229 0.013345 -2.340 0.02663 * drat 2.714975 1.487366 1.825 0.07863 . --- Signif. codes: 0 ‘***’ 0.001 ‘**’ 0.01 ‘*’ 0.05 ‘.’ 0.1 ‘ ’ 1 Residual standard error: 3.008 on 28 degrees of freedom Multiple R-squared: 0.775, Adjusted R-squared: 0.7509 F-statistic: 32.15 on 3 and 28 DF, p-value: 3.28e-09

To visualize the relationship between the response variable “mpg” and each individual predictor variable in the model, we can produce added variable plots using the **avPlots()** function:

#load car package library(car) #produce added variable plots avPlots(model)

Here is how to interpret each plot:

- The x-axis displays a single predictor variable and the y-axis displays the response variable.
- The blue line shows the association between the predictor variable and the response variable,
*while holding the value of all other predictor variables constant*. - The points that are labelled in each plot represent the two observations with the largest residuals and the two observations with the largest partial leverage.

Note that the angle of the line in each plot matches the sign of the coefficient from the estimated regression equation.

For example, here are the estimated coefficients for each predictor variable from the model:

**disp:**-0.019232**hp:**-0.031229**drat:**2.714975

Notice that the angle of the line is positive in the added variable plot for *drat* while negative for both *disp* and *hp*, which matches the signs of their estimated coefficients:

These plots allow us to conveniently visualize the relationship between each individual predictor variable and the response variable.

**Additional Resources**

How to Perform Simple Linear Regression in R

How to Perform Multiple Linear Regression in R

How to Perform Logistic Regression in R

Zach, thank you. Your posts are beneficial and enlightening, particularly for a novice R user. Your blog posts are the first port of call when I come across any new function.