Wednesday, February 9, 2011

Grammarly as a Scam & Plagiarism Detector

Earlier I posted wondering if Grammarly could be a viable plagiarism and scam detection alternative to my old favorite Copyscape, I decided to bite the bullet and try it out. I signed up for the seven-day free trial, partially because curiosity was getting my cat, and partially because I do get a little bit sick of the signal to noise ratio with Copyscape's free tool.

The first thing I noted when I started my free trial was that my suspicions from last week were right - you can't just enter a URL into the system and have it bring back a bunch of comparisons. You actually have to plug the text into their editor, which means one more step in terms of copying and pasting. Not a big deal, just something I noticed. My method for the test was to use a random story from CNN, which I knew would have all kinds of similarities across the Internet, to see what each system brought back.

Like I said, it's all about intent. Copyscape is trying to find every single instance of that same copy shows up online, to see who's trying to scam you and steal your copy. Grammarly, on the other hand, is trying to identify the single source it thinks is most likely the original. My sample was from CNN and neither tool seemed to quite get there in terms of finding my real source, and I have to say each result had its flaws.

Copyscape showed me every single retweet and legit use of that copy, while Grammarly just showed me a page that it hoped was the original. Neither one is quite giving me what I want when I'm looking for duplicate content and spammers, and all things considered I'm going to stick with the free tool.

No comments:

Post a Comment