4 ms·

I'm much more skeptical about the "use mouse movements as a proxy for eye gaze" part. For non-interactive elements like text blocks, cursor location has a NEGAT

I'm much more skeptical about the "use mouse movements as a proxy for eye gaze" part. For non-interactive elements like text blocks, cursor location has a NEGATIVE correlation with eye gaze (because users will often move the cursor out of the way of what they're reading). I would completely disregard any data that comes out of this. One way to potentially address this problem: dim out most of the page except a 200-300px diameter circle where the cursor is located. Force the user to use the cursor as a "spotlight" so they can read the resume. This allows you to see where they look & what catches/holds their attention. It's a wholly unnatural way to read, but should serve your particular purpose.

erickookoo 12y ago
You're correct about users not hovering over exactly where they are reading in order not to block their view with the cursor, but studies have shown that people do position their cursor very close to or just below where they are actually reading (there are, however, also other usage patterns such as the scroller who simply leaves his cursor in one place and scrolls down the page). We've allowed users to create links to their projects in order to make would normally be non-interactive text blocks on a resume into something clickable. There is plenty of literature that correlates eye gaze to cursor movement (e.g. http://scholar.google.com/citations?view_op=view_citation&hl=en&user=o8oJ6s4AAAAJ&citation_for_view=o8oJ6s4AAAAJ:HoB7MX3m0LUC http://scholar.google.com/citations?view_op=view_citation&hl...)
- mjhouse 12y ago
  Something I thought of coming home from a birthday dinner this afternoon: In order to get a better idea of what the viewer is focusing on, you might consider going to a vertical gradient instead of a pointer heat-map. Most people don't read to the bottom of the page before they start scrolling. Imagine two lines on the screen- the bottom line is the point at which the viewer will begin to scroll. The top line is the point that the viewer wants their reading area to be located. The space in between the lines is your "target" area. As the viewer scrolls down the page, you could average their scrolling speed, and increment the intensity for that area of the heat map only when they slow their scrolling speed drastically, or go back over an area. Just from paying attention to my own scrolling habits, I think the top line would probably be just above the horizontal center of the screen, and the bottom line would probably be about an inch down. You could do some testing to find the average line positions.

erickookoo 12y ago

You're correct about users not hovering over exactly where they are reading in order not to block their view with the cursor, but studies have shown that people do position their cursor very close to or just below where they are actually reading (there are, however, also other usage patterns such as the scroller who simply leaves his cursor in one place and scrolls down the page). We've allowed users to create links to their projects in order to make would normally be non-interactive text blocks on a resume into something clickable. There is plenty of literature that correlates eye gaze to cursor movement (e.g. http://scholar.google.com/citations?view_op=view_citation&hl=en&user=o8oJ6s4AAAAJ&citation_for_view=o8oJ6s4AAAAJ:HoB7MX3m0LUC http://scholar.google.com/citations?view_op=view_citation&hl...)

mjhouse 12y ago
Something I thought of coming home from a birthday dinner this afternoon: In order to get a better idea of what the viewer is focusing on, you might consider going to a vertical gradient instead of a pointer heat-map. Most people don't read to the bottom of the page before they start scrolling. Imagine two lines on the screen- the bottom line is the point at which the viewer will begin to scroll. The top line is the point that the viewer wants their reading area to be located. The space in between the lines is your "target" area. As the viewer scrolls down the page, you could average their scrolling speed, and increment the intensity for that area of the heat map only when they slow their scrolling speed drastically, or go back over an area. Just from paying attention to my own scrolling habits, I think the top line would probably be just above the horizontal center of the screen, and the bottom line would probably be about an inch down. You could do some testing to find the average line positions.