While far from finished, the project has jelled somewhat. All the pages except survey.php have been validated for XHTML 1.0 Strict, with survey.php validated for Transitional. I’ve prepared the downloadable zip file of the site source code, with some rudimentary installation instructions, copyleft notice, etc. There’s still no pigeonholing, but I’m now studying things like factor analysis and clustering. Considering that I didn’t know any PHP whatsoever before undertaking this project, there is hope. Oh, and I’ve learned that the “Likert scale” is the technical term for this all-too-familiar questionnaire format with responses ranging from “strongly disagree” to “strongly disagree.” The results pages (if we can call them that) in place so far compare one’s entries with those preceding it. This leaves high and dry the first few respondents. Since each is assigned a number, knowing that number allows one to revisit at a later time for comparison with subsequent respondents. For now, in addition to the “distance metric” from previous entries offered since the start, there is a page to compare any two profiles, and one to see how any particular survey item correlates with the others.
AEBHAM4D53ZT
Poor Richard has proposed re-wording some of the response items. I hesitate to do so, mainly because of concerns about statistical validity if profiles collected prior to the change were to be mashed up with later ones. But future plans definitely include enlargement of the questionnaire directly by visitors to the site. By all means, add proposed response items, future features and other proposed changes as comments to the present blog entry!
Leave a Reply