Over the last few weeks we have been exploring Firefox menu usage through data collected by the Test Pilot extension.  During our analysis, we encountered some irregularities and have since discovered that parts of the data set were affected by a few bugs.

The main issue was that context menus (the pop-up menus that appear after right clicking) were inaccurately tracked as mouse clicks on the menu bar.  Fortunately, we were able to identify these spurious events and have updated all the data sets on the Test Pilot website.  We are also in the process of revising our initial analysis and will continue to investigate Firefox menu usage using the adjusted data.

In fact, your comments helped lead us to these issues; we really want to encourage everyone to continue exploring the data and scrutinizing our analysis for themselves.  In addition, we are investigating different ways to ensure the most reliable Test Pilot data in the future.  Currently, we are considering pushing a trial version of each study to a small sub-sample, and doing some preliminary analysis before releasing the study to the full set of Test Pilot users.  Let us know if you have any suggestions related to this as well!