In 2005, when we started this, we simply wanted to identify an interesting set of assessments and try to relate a few of our findings to something that has a research anchor. Consequently, the main thing we were looking for was the use of graphic elements in these presentations because Richard Mayer has presented some interesting research on the use of graphics, the percent of “screen” area used for text or graphics, the proximity of graphics to text and whether supporting graphic elements were simply decorative or really contributed to learning.

So to answer your question about PPT selection, we simply searched for PPT extensions through Google and downloaded 75 presentations in 2005 and another 100 presentations in 2007. There was no method used related to which of the available files we chose.

From a research perspective, could you let us know the criteria you used to find and select the PowerPoint files you analyzed? Did you employ a random sampling methodology so we would be able to extend the findings from your sample to all PowerPoint presentations? Are you assuming that the PowerPoint files that are available on the web are a true sample of the total universe of presentations? (For example, many presentations reside only behind corporate firewalls because of sensitive content, and/or are often greater than 50MB in size because of graphics and multimedia, either of which would preclude them from being on the web.)

It’s quite possible that the files available on the web are indeed representative of typical presentations. But I think it’s important to elaborate on the research design and assumptions at the start to make them clear. This is something Tufte did not do in his “study” on PowerPoint, which was limited to “Typical PowerPoint Presentations that Edward Tufte Could Find on the Web”, ignoring the many types of presentations he was not able to access.