Re: (Trying to catch students cheating) Cronbach alpha & semantic differentia

[Björn Türoque]

I thought I should offer my advice, but I myself have never taught a college
course, I am currently a grad student and TA quite a bit. Hopefully next
year I'll get to teach some 100 level survey classes, as prep I had to take
a workshop where they had us read some books on the subject of teaching.

For anyone who has to teach college students I reccomend Ken Bain's What the
best college teachers do:
http://www.amazon.com/What-Best-College-Teachers-Do/dp/0674013255/ref=sr_1_1/104-9502469-8172759?ie=UTF8&s=books&qid=1181163549&sr=8-1

And Peter Fileens' The Joy of Teaching
http://www.amazon.com/Joy-Teaching-Practical-College-Instructors/dp/0807856037/ref=sr_1_3/104-9502469-8172759?ie=UTF8&s=books&qid=1181163549&sr=8-3

Both excelent reading to prepare first time teachers, and good for those who
have been doing it for a while.If I ever get to teach an advanced class
using SPSS and surveys I'll definitely use the trick I wrote about.

Don

On 6/6/07, Don Asay <[hidden email]> wrote:

>
> Shoaib,
>
> You can't get SPSS to detect if students are cheating, and not filling out
> surveys correctly, but with some deception on your part and a few clever
> stats tricks you can make them think you can find the surveys they cheated
> on and hopefully those who did cheat within the class to confess.
>
> The procedure you would have to follow is a bit complicated but should
> work for the most part, first divide the class in to 2 halves and have every
> student create a sheet with the results of maybe 100 to 200 coin tosses.
> Have one half make fake results, and have the other half actually do it with
> a coin.
>
> Upon examining the results you should be able to pick out the random
> tosses done by a coin and the non-random human created results, because
> people do not often have long strings of 5, 6, 7 or 8 heads or tails in a
> row. Theese long runs of one particular type are likely to happen if you
> have a large enough sample of flips the papers done with actual coins will
> most likely have theese runs.
>
> An internet article on theese runs is here:
> http://everything2.com/index.pl?node_id=774249
>
> Once you have shown the class that you are reliably able to pick out human
> data from real random data, state that you know some students cheated, and
> that you will give them the opportunity to redo their work, if they confess.
>
>
> Ask everyone to take out a piece of paper and write down their name and if
> they cheated or not, and the numbers of the surveys they submitted, (so
> these ones can be removed from the data) have the students fold the paper in
> half and pass it up. With any luck, and some good acting skills, you'll
> probably catch a good proportion of your cheaters.
>
> I saw the coin trick successfully executed in one of my stats classes, and
> it does work about 90% of the time, just enough for you to decieve the
> students. Additionally, I have TAed a class where the professor caught 2
> people cheating on papers, and allowed students who confessed to cheating,
> through the folded paper method to redo their work. Interestingly 4
> students in the class confessed to cheating on their papers. Cheating is
> rampant among undergrads these days, so a little deception on your part
> won't be too horrible.
>
> I am sorry if my ramblings are not as coherent as they could be, I'm
> shooting from the hip here, and trying to get this out quickly. Let me and
> the list know how this works in your class, some other professors might
> start using this method if it is successful.
>
> Don
>
>
>
> On 5/29/07, shoaib ul-haq <[hidden email]> wrote:
> >
> > Hello All,
> >
> > I am a novice user of SPSS and I want to ask the
> > following questions.
> >
> > 1. How can I code a question asked on semantic
> > differential scale? For example, the question is
> >
> > The customer care representative was...
> > Very Much  Somewhat Neither Somewhat Very Much
> >   (1)       (2)     (3)      (4)       (5)
> >
> > Helpful                                 Unhelpful
> > Friendly                                Unfriendly
> > Polite                                  rude
> >
> > How should I enter the output of question like the
> > above in the data editor.
> >
> > 2. I suspect that some of my students have filled the
> > questionnaires by themselves instead of going to other
> > people. How can I detect it in SPSS. Will calculating
> > Cronbach alpha help?
> >
> > 3. If value of Cronbach alpha for likert scale item
> > questions is 0.90, can I say that the data is
> > tempered, since it is too repetitious?
> >
> > Thanks a lot for your help.
> >
> > Regards
> > Shoaib
> >
> >
> >
> >
> >
> >
> >
> >
> >
> >
> >
> > ____________________________________________________________________________________Got
> > a little couch potato?
> > Check out fun summer activities for kids.
> > http://search.yahoo.com/search?fr=oni_on_mail&p=summer+activities+for+kids&cs=bz
> >
> >
>
>