[blahblahblah]

Choosing an arbitrary cutoff of 6 years and then assuming that everyone who didn't complete their degree within 6 years is a college dropout is not a particularly valid way to measure dropout rates. I completed my degree after the 6-year mark and I'm certainly not the only one to do so. These statistics erroneously include people like me in the dropout category. A more accurate method of measurement would be to track each cohort of students from entrance through graduation and chart a histogram of the elapsed time from initial enrollment to graduation. Once you've collected these histograms for several cohorts, you're then in a position to construct a model of the statistical distribution. Then you can make accurate estimates of the true dropout rate by fitting the known graduation data for a given cohort to the model to estimate how many of the students are in the tail of the distribution that will eventually graduate.

[istari]

You're backing it up with your own personal anecdote?

That's not valid.

If you found data that said, for example, 30% of college graduates took longer than 6 years to earn their undergraduate degrees, then I would agree with you.

If it's less than 10%, then the using 6 years as a cutoff would be a very good rule of thumb estimate.

Maybe the reason why the 6 year cutoff is so often used is because someone already crunched the numbers and found it was good enough?

[blahblahblah]

"You're backing it up with your own personal anecdote?"

No. I provided a counterexample that falsifies an implicit assumption of the 6-year cutoff model of the college dropout rate. If you measure your dropout rate as the percentage of students who have not graduated within 6 years of enrollment, it is provably true that your measurement is an overestimate of the true value.

"Maybe the reason why the 6 year cutoff is so often used is because someone already crunched the numbers and found it was good enough?"

If someone had already crunched the numbers, the appropriate correction factor to apply to the 6-year measurement would be well-known and, instead of reporting the 6-year cutoff value, a distribution-corrected 6-year cutoff value would have been reported instead. If nobody crunched the numbers to find out what the characteristics of the distribution's tail are, then the methodology is sloppy. If a correction factor was applied but not explicitly mentioned where the value is reported, then the reporting is sloppy. Either way, a reasonable person is justified in regarding the reported value as merely an upper bound on the dropout rate and rejecting the assertion that it is an accurate measurement of the true value.

[chopsueyar]

I have enjoyed your anecdote.

[blahblahblah]

Well, I really enjoyed college and I wasn't in any hurry to graduate. I took a lot of classes that had absolutely nothing to do with computer science, had a lot of fun doing so, and I think I am a much more well-rounded human being for having taken that route. I had about 70 credit-hours more than what was actually required for my degree by the time I completed the last required class.