Multiple Factor Experiments

We tested whether the purpose of a loan affects loan amount funded and found that it does. However, we also know that it's unlikely that loans are funded based only on their intended purpose. It's more likely that the company is looking at a holistic picture of an applicant before they decide to issue a loan.

We can examine more than one explanatory factor in a multiple factor experiment. Like our experiments on ToothGrowth from Chapter 1, an experimenter can try and control two (or more!) different factors and see how they affect the outcome. We're using open data, so we can't quite control the factors here (they're submitted as someone fills out their loan application), but let's look at how a few other factors affect loan amount funded.

Diese Übung ist Teil des Kurses

Experimental Design in R

Kurs anzeigen

Anleitung zur Übung

Use aov() to build a linear model and ANOVA in one step, examining how purpose_recode and employment length (emp_length) affect the funded amount. Save as an object purpose_emp_aov and print the result out.
The printed purpose_emp_aov does not show p-values, which we might be interested in. Display those by calling summary() on the aov object.

Interaktive Übung

Vervollständige den Beispielcode, um diese Übung erfolgreich abzuschließen.

# Use aov() to build purpose_emp_aov
___ <- aov(funded_amnt ~ ___ + ___, ___ = ___)

# Print purpose_emp_aov to the console
purpose_emp_aov

# Call summary() to see the p-values
summary(___)

Code bearbeiten und ausführen