For the first time in 2006, Canadian households had the option of responding to the Census via the Internet. Almost one in five households chose to report their data with this new collection mode. A study was undertaken to check for the presence of an Internet mode effect in the data. Different means were used to assess the presence of a mode effect: non-response rates to each question by response mode were looked at, as well as answers distributions. Among other means, the propensity score method was used to compare answer distributions. This method allowed us to take into account some of the differences in characteristics of Internet reporters and Paper reporters and hence make them more comparable. The findings of the 2006 Census Internet Mode Effect Study are given in this paper.