How to add an interaction term to a model

Description

Sometimes, a simple linear model isn’t sufficient for our data, and we need more complex terms or transformed variables in the model to make adequate predictions. How do we include these complex and transformed terms in a regression model?

Related tasks:

• How to add a polynomial term to a model
• How to add a transformed term to a model

Solution, in R

View this solution alone.

We’re going to use the ToothGrowth dataset in R as example data. It contains observations of tooth growth for guinea pigs who received various doses of various supplements. You would use your own data instead.

1
df <- ToothGrowth

Let’s model tooth length (len) based on the product of two predictors, the supplement given (supp) and its dosage (dose). We simply use the ordinary multiplication operator in R, written *, to express the product of these two factors when creating the model, as shown below.

Note that supp is a categorical variable with two values, so the model will include a binary variable for whether the supplement was equal to “VC.”

1
2
3
# Build the model
model <- lm(len ~ supp*dose, data = df)
summary(model)

1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
Call:
lm(formula = len ~ supp * dose, data = df)

Residuals:
    Min      1Q  Median      3Q     Max 
-8.2264 -2.8462  0.0504  2.2893  7.9386 

Coefficients:
            Estimate Std. Error t value Pr(>|t|)    
(Intercept)   11.550      1.581   7.304 1.09e-09 ***
suppVC        -8.255      2.236  -3.691 0.000507 ***
dose           7.811      1.195   6.534 2.03e-08 ***
suppVC:dose    3.904      1.691   2.309 0.024631 *  
---
Signif. codes:  0 ‘***’ 0.001 ‘**’ 0.01 ‘*’ 0.05 ‘.’ 0.1 ‘ ’ 1

Residual standard error: 4.083 on 56 degrees of freedom
Multiple R-squared:  0.7296,	Adjusted R-squared:  0.7151 
F-statistic: 50.36 on 3 and 56 DF,  p-value: 6.521e-16

Now we have a model of the form $\hat L = 11.55 - 8.255s + 7.811d + 3.904sd$, where $L$ stands for tooth length, $s$ for whether the VC supplement was given, and $d$ for the dose given.

Content last modified on 24 July 2023.

See a problem? Tell us or edit the source.

Topics that include this task

• Bentley University MA252

Opportunities

This website does not yet contain a solution for this task in any of the following software packages.

• Python
• Excel
• Julia

If you can contribute a solution using any of these pieces of software, see our Contributing page for how to help extend this website.