I have a particular preference for on-line survey collection. I like the fact that some systems have the ability to collect information electronically – without you leaving the house. I love the fact that your data gets routed to a remote location for further use. I event appreciate the systems which provide quick graphs and charts of your data. Marvelous! But as a statistician, I do not rely upon a survey collection system to make analytical decisions for me. I want to see the data as an exported file. I want to take that data and run my own frequencies and descriptive. It’s not that I don’t trust on-line survey… because I do. I just find that some systems lack the robustness of real statistical software. I do not consider Survey Monkey and these other survey tools as a statistical software package. It is a data collection tool that HAPPENS to have statistical capabilities. If I had a cluster analysis, survey monkey does not come to mind. I am not denouncing all data collection systems as useless to statistics. Rather, I am just pointed out that statisticians typically dump the data out of the system to get their hands dirty. It’s what we do!
-Moore to follow Amy