• Skip to primary navigation
  • Skip to main content
  • Skip to primary sidebar
The Analysis Factor

The Analysis Factor

Statistical Consulting, Resources, and Statistics Workshops for Researchers

  • our programs
    • Membership
    • Online Workshops
    • Free Webinars
    • Consulting Services
  • statistical resources
  • blog
  • about
    • Our Team
    • Our Core Values
    • Our Privacy Policy
    • Employment
    • Collaborate with Us
  • contact
  • login

polynomial regression

A Strategy for Converting a Continuous to a Categorical Predictor

by Jeff Meyer  Leave a Comment

At times it is necessary to convert a continuous predictor into a categorical predictor.  For example, income per household is shown below.Stage 2

This data is censored, all family income above $155,000 is stated as $155,000. A further explanation about censored and truncated data can be found here. It would be incorrect to use this variable as a continuous predictor due to its censoring.

[Read more…] about A Strategy for Converting a Continuous to a Categorical Predictor

Tagged With: Censored, continuous predictor, continuous variable, LOWESS, pairwise, polynomial regression, predictor variable, smoothing

Related Posts

  • A Useful Graph for Interpreting Interactions between Continuous Variables
  • The Impact of Removing the Constant from a Regression Model: The Categorical Case
  • What is Multicollinearity? A Visual Description
  • Recoding a Variable from a Survey Question to Use in a Statistical Model

R Is Not So Hard! A Tutorial, Part 4: Fitting a Quadratic Model

by guest contributer  15 Comments

 by David Lillis, Ph.D.Stage 2

In Part 3 we used the lm() command to perform least squares regressions. In Part 4 we will look at more advanced aspects of regression models and see what R has to offer.

One way of checking for non-linearity in your data is to fit a polynomial model and check whether the polynomial model fits the data better than a linear model. However, you may also wish to fit a quadratic or higher model because you have reason to believe that the relationship between the variables is inherently polynomial in nature. Let’s see how to fit a quadratic model in R.

We will use a data set of counts of a variable that is decreasing over time. Cut and paste the following data into your R workspace.

A <- structure(list(Time = c(0, 1, 2, 4, 6, 8, 9, 10, 11, 12, 13, 
14, 15, 16, 18, 19, 20, 21, 22, 24, 25, 26, 27, 28, 29, 30), 
Counts = c(126.6, 101.8, 71.6, 101.6, 68.1, 62.9, 45.5, 41.9, 
46.3, 34.1, 38.2, 41.7, 24.7, 41.5, 36.6, 19.6, 
22.8, 29.6, 23.5, 15.3, 13.4, 26.8, 9.8, 18.8, 25.9, 19.3)), .Names = c("Time", "Counts"),
row.names = c(1L, 2L, 3L, 5L, 7L, 9L, 10L, 11L, 12L, 13L, 14L, 15L, 16L, 17L, 19L, 20L, 21L, 22L, 23L, 25L, 26L, 27L, 28L, 29L, 30L, 31L),
class = "data.frame")

Let’s attach the entire dataset so that we can refer to all variables directly by name.

attach(A)
names(A)

First, let’s set up a linear model, though really we should plot first and only then perform the regression.

linear.model <-lm(Counts ~ Time)

We now obtain detailed information on our regression through the summary() command.

summary(linear.model)
Call:
lm(formula = Counts ~ Time)

Residuals:
   Min     1Q Median     3Q     Max 
-20.084 -9.875 -1.882   8.494 39.445 

Coefficients:
           Estimate Std. Error t value Pr(>|t|)   
(Intercept) 87.1550     6.0186 14.481 2.33e-13 ***
Time         -2.8247     0.3318 -8.513 1.03e-08 ***
---
Signif. codes: 0 ‘***’ 0.001 ‘**’ 0.01 ‘*’ 0.05 ‘.’ 0.1 ‘ ’ 1

Residual standard error: 15.16 on 24 degrees of freedom
Multiple R-squared: 0.7512,   Adjusted R-squared: 0.7408 
F-statistic: 72.47 on 1 and 24 DF, p-value: 1.033e-08

The model explains over 74% of the variance and has highly significant coefficients for the intercept and the independent variable and also a highly significant overall model p-value. However, let’s plot the counts over time and superpose our linear model.

plot(Time, Counts, pch=16, ylab = "Counts ", cex.lab = 1.3, col = "red" )

abline(lm(Counts ~ Time), col = "blue")

tn_image001

Here the syntax cex.lab = 1.3 produced axis labels of a nice size.

The model looks good, but we can see that the plot has curvature that is not explained well by a linear model. Now we fit a model that is quadratic in time. We create a variable called Time2 which is the square of the variable Time.

Time2 <- Time^2
quadratic.model <-lm(Counts ~ Time + Time2)

Note the syntax involved in fitting a linear model with two or more predictors. We include each predictor and put a plus sign between them.

 

summary(quadratic.model)
Call:
lm(formula = Counts ~ Time + Time2)
Residuals:
     Min       1Q   Median       3Q     Max 
-24.2649 -4.9206 -0.9519   5.5860 18.7728 

Coefficients:
             Estimate Std. Error t value Pr(>|t|)   
(Intercept) 110.10749   5.48026 20.092 4.38e-16 ***
Time         -7.42253   0.80583 -9.211 3.52e-09 ***
Time2         0.15061   0.02545   5.917 4.95e-06 ***
---
Signif. codes: 0 ‘***’ 0.001 ‘**’ 0.01 ‘*’ 0.05 ‘.’ 0.1 ‘ ’ 1

Residual standard error: 9.754 on 23 degrees of freedom
Multiple R-squared: 0.9014,   Adjusted R-squared: 0.8928
F-statistic: 105.1 on 2 and 23 DF, p-value: 2.701e-12

Our quadratic model is essentially a linear model in two variables, one of which is the square of the other. We see that however good the linear model was, a quadratic model performs even better, explaining an additional 15% of the variance. Now let’s plot the quadratic model by setting up a grid of time values running from 0 to 30 seconds in increments of 0.1s.

timevalues <- seq(0, 30, 0.1)
predictedcounts <- predict(quadratic.model,list(Time=timevalues, Time2=timevalues^2))
plot(Time, Counts, pch=16, xlab = "Time (s)", ylab = "Counts", cex.lab = 1.3, col = "blue")

Now we include the quadratic model to the plot using the lines() command.

lines(timevalues, predictedcounts, col = "darkgreen", lwd = 3)

tn_image002

The quadratic model appears to fit the data better than the linear model. We will look again at fitting curved models in our next blog post.

See our full R Tutorial Series and other blog posts regarding R programming.


About the Author:
David Lillis has taught R to many researchers and statisticians. His company, Sigma Statistics and Research Limited, provides both on-line instruction and face-to-face workshops on R, and coding services in R. David holds a doctorate in applied statistics.

Tagged With: linear regression, polynomial regression, R

Related Posts

  • What R Commander Can do in R Without Coding–More Than You Would Think
  • Linear Models in R: Improving Our Regression Model
  • Linear Models in R: Diagnosing Our Regression Model
  • Linear Models in R: Plotting Regression Lines

Regression Models:How do you know you need a polynomial?

by Karen Grace-Martin  4 Comments

A polynomial term–a quadratic (squared) or cubic (cubed) term turns a linear regression model into a curve.  But because it is X that is squared or cubed, not the Beta coefficient, it still qualifies as a linear model.  This makes it a nice, straightforward way to model curves without having to model complicated non-linear models.

But how do you know if you need one–when a linear model isn’t the best model? [Read more…] about Regression Models:How do you know you need a polynomial?

Tagged With: curvilinear relationship, linear regression, polynomial regression, quadratic terms

Related Posts

  • Interpreting Regression Coefficients
  • Linear Regression for an Outcome Variable with Boundaries
  • Using Marginal Means to Explain an Interaction to a Non-Statistical Audience
  • Understanding Interactions Between Categorical and Continuous Variables in Linear Regression

Primary Sidebar

This Month’s Statistically Speaking Live Training

  • Member Training: Moderated Mediation, Not Mediated Moderation

Upcoming Workshops

    No Events

Upcoming Free Webinars

TBA

Quick links

Our Programs Statistical Resources Blog/News About Contact Log in

Contact

Upcoming

Free Webinars Membership Trainings Workshops

Privacy Policy

Search

Copyright © 2008–2023 The Analysis Factor, LLC.
All rights reserved.

The Analysis Factor uses cookies to ensure that we give you the best experience of our website. If you continue we assume that you consent to receive cookies on all websites from The Analysis Factor.
Continue Privacy Policy
Privacy & Cookies Policy

Privacy Overview

This website uses cookies to improve your experience while you navigate through the website. Out of these, the cookies that are categorized as necessary are stored on your browser as they are essential for the working of basic functionalities of the website. We also use third-party cookies that help us analyze and understand how you use this website. These cookies will be stored in your browser only with your consent. You also have the option to opt-out of these cookies. But opting out of some of these cookies may affect your browsing experience.
Necessary
Always Enabled
Necessary cookies are absolutely essential for the website to function properly. This category only includes cookies that ensures basic functionalities and security features of the website. These cookies do not store any personal information.
Non-necessary
Any cookies that may not be particularly necessary for the website to function and is used specifically to collect user personal data via analytics, ads, other embedded contents are termed as non-necessary cookies. It is mandatory to procure user consent prior to running these cookies on your website.
SAVE & ACCEPT