« Coil Lamp | Home | World's most expensive watermelon: a snip at $3,740 »

June 21, 2011

Bubble-filling technique can identify test-takers

Sdcvgbnm

Who knew?

From Christopher Shea's June 16, 2011 Wall Street Journal story:

People who fill out bubble forms like those ubiquitous fill-in-the-circle tests use distinctive pencil strokes that can be used to identify them, researchers report. They programmed a computer to take stock of 804 potentially tell-tale aspects of people's pencil strokes on such forms. These include the mark's center of mass, the variance of pencil strokes from the bubble's radius, and the depth of shading, as well as more mathematically advanced measures.

The computer analyzed 92 student surveys, checking a dozen marks from each respondent. Then the researchers scrutinized eight marks from a randomly picked person. The computer identified its man or woman 51% of the time. The correct answer was among the computer's top three choices 75% of the time and was 92.4% of the time among the top 10.

The method could be used to catch students who hire proxies to take their SATs and teachers who change answers on their students' high-stakes tests. But employers, the researchers said, could also use it to monitor the voting habits of their employees, since some jurisdictions, in the interest of transparency, release scans of voters' bubble forms, without attaching their names.

The paper "Bubble Trouble: Off-Line De-Anonymization of Bubble Forms," by Joseph A. Calandrino, William Clarkson and Edward W. Felten, will be presented at the USENIX Security Symposium in August.

Can't wait that long?

No problema, I've got you covered.

Read the paper in its entirety here.

Fair warning: there goes the day.

[via Freedom to Tinker]

June 21, 2011 at 05:01 PM | Permalink


TrackBack

TrackBack URL for this entry:
https://www.typepad.com/services/trackback/6a00d8341c5dea53ef0154331921ad970c

Listed below are links to weblogs that reference Bubble-filling technique can identify test-takers:

Comments

I've read the 14 page paper. Alas, I can find no reference to those tests that are "signed" by the takers. I'm fairly certain that a good 20% come with a printed name - whether it is really the taker's name - or the name of the prof's pet - is a question that the study wholly fails to address.

Posted by: 6.02*10^23 | Jun 21, 2011 7:59:27 PM

The comments to this entry are closed.