Measuring contract cheating
"If you can't measure it, you can't improve it"* is the inspiration behind this blog post. In this post, I discuss why a way to measure contract cheating is necessary and propose a measurement metric.
The motivation behind this (and future) posts is to journal the process of building this cheating measurement tool, collecting feedback and getting some help along the way. So, if anyone has any thoughts or is interested in helping please feel free to comment 
.
Okay, so, the question is...
Why do we need to measure cheating?
Over the years, we've all seen interventions in the area of contract cheating increase. And interventions come in many forms: technological (software), political (bans) and pedagogical (less writing assignments/raising awareness). While all such news is great, there is a larger question: how do we know the interventions are working? I feel this is a difficult, yet crucial question to ask (and answer!).
A measurement tool is as necessary as the interventions themselves. Why? Because we will eventually need the measurement tool to gauge the efficacy of the detection/prevention tools. How else can we tell if any government policy/technology is really hurting the businesses of essay mills?
The next question then becomes...
How do we measure contract cheating?
Self-reporting seems like a sub-par method to measure contract cheating interventions in my opinion. Since that approach is a bit biased (un-verifiable), my tiny brain proposes the following way: we measure the popularity of contract cheating websites and essay mills. I mean if cheating is decreasing, contract cheating websites will be less popular and vice versa right?
Since we obviously don't (and never will) have the actual data of students cheating, I think the popularity of contract cheating websites is the ideal proxy/stand-in to measure the cheating market.
The most straight-forward (and reliable) data we can get on a website's popularity is its traffic/analytics data. But then there are hundreds and thousands of essay mill and contract cheating websites.
The next question then becomes..
How do we monitor all odem websites?
Fortunately, other people have run into the same problem and they do it as such: they create an index. For example, there are 2,400 companies listed on the stock exchange but the DJIA (Dow Jones Industrial Average) only pools the data of the 30 largest companies and monitors their prices over-time. This 'average' then becomes a proxy for the entire stock market and the economy (by extension). Much like how how our website traffic data will be the proxy for the entire cheating economy .
The next question then becomes...
What do we call this cheating measurement tool?
I'm going to go out on a limb and call it the 'Contract Cheating Index (CCI)'. But if you have any better names, please feel free to suggest. Anyway, I feel we have something to build upon now.
Which begs the question...
Where do we start?
The plan of action is:
- Analyze the traffic of a sub-set of contract cheating websites over-time
- Pick the top 30
- Create an 'index' which shows an upward or downward movement (much like the DJIA)
- Automate the process
- Display it
In the next post I shall do task 1 and task 2 and get a sense of the data. Just a heads-up our traffic data will come from Alexa (not the speaker, the website), so if anyone can find the time to collaborate with me on this that would be fun. Maybe @kona , with your statistics experience?
For now, this journey has to stop here. I hope you enjoyed reading, as much as I did writing. What is getting me excited is: in the next post I'll actually have some numbers to play with and data to share! Ain't that fun!
*I think the quote is attributed to Peter Drucker.
