Figure 7

Regression trees quantifying variable importance in the determination of the time on market. For each node, the number on top is the average time on market (in days), and the percentage below is the sample size in that node. Nodes are colored in darker shades of blue for longer time on market. From each node, the edge to the left is a TRUE to the condition, while the edge to the right is a FALSE. (A) The relative number of clicks is by far the most important variable because it determines the first splits. It is followed by geographical and price variables. Higher number of clicks (edges to the left) predicts shorter time on market. (B) The first split distinguishes between ads that received no contacts (left branch), and ads that received at least one (right branch). In the right branch, contacts are the most important variable. The figure has been generated using R package rpart.plot