Can R Squared Decrease With More Variables

How To Articles

Hey there! Today, let’s dive deep into the fascinating world of linear regression. Specifically, we’re going to explore whether the coefficient of determination, also known as R-squared, can decrease when more variables are added to a regression model.

Before we start, let me introduce myself. I’m a technical writer with a passion for statistics and data analysis. I love uncovering the hidden truths and patterns behind numbers. So, let’s get started on our quest for knowledge!

When we perform a linear regression analysis, we aim to find the best-fitting line that explains the relationship between the independent variables (predictors) and the dependent variable (outcome). R-squared is a statistical measure that represents the proportion of the variance in the dependent variable that is predictable from the independent variables.

Now, it’s time to address the burning question: Can R-squared decrease with more variables in a regression model? The answer is both yes and no. Allow me to elaborate.

Adding Meaningful Variables

When we add meaningful variables to a regression model, the R-squared value typically increases or remains the same. This is because these variables contribute valuable information and help explain the variance in the dependent variable.

Let’s say we have a simple linear regression model with only one independent variable, such as “Monthly Advertising Spend,” to predict “Monthly Sales.” In this case, the R-squared value would represent the proportion of the variation in sales that can be explained by advertising spend alone.

Now, suppose we decide to add another variable, such as “Number of Salespeople.” If this variable is meaningful and has a significant impact on sales, it will likely contribute additional explanatory power to the model. As a result, the R-squared value may increase, indicating that the new variable improves our ability to predict sales.

Adding Irrelevant Variables

However, things change when we add irrelevant variables to the regression model. Irrelevant variables are those that have little or no impact on the dependent variable. In this case, adding more variables can actually decrease the R-squared value.

Let’s continue with our example of predicting monthly sales. Suppose we add a variable like “Day of the Week” to our model. Unless there is a specific reason to believe that the day of the week affects sales, including this variable will likely introduce noise and decrease the accuracy of our predictions. As a result, the R-squared value may decrease.

It’s important to note that R-squared is not designed to penalize the addition of irrelevant variables. Instead, it measures the goodness of fit of the model based on the available predictors. Adding irrelevant variables can lead to overfitting, where the model becomes too complex and performs poorly on new data.


In conclusion, the effect of adding more variables on the R-squared value depends on the nature of those variables. Adding meaningful variables tends to increase or maintain the R-squared value, as they contribute valuable information to the model. On the other hand, adding irrelevant variables can decrease the R-squared value by introducing noise and overfitting.

As data analysts, it’s our responsibility to carefully evaluate the relevance and impact of each variable we include in our regression models. By doing so, we can ensure that our models are accurate, robust, and provide meaningful insights.

That wraps up our exploration of whether R-squared can decrease with more variables in a regression model. I hope this article has shed some light on this intriguing question. Stay curious, keep exploring, and may your statistical endeavors be filled with meaningful discoveries!