Just curious and just getting my coffee fix today. What is the probability
our probability data are
representative? Haha Seriously now, many students believe if they are
going to guess an answer -
they should guess the same letter consistently throughout the examination.
This strategy should be
considered when considering guessing advantage. Using Marc's data
collection exercise for example,
consistent A's or B's guessed = loss of points, C's = break even - D's or
E's = gain of points. It is possible
ETS structures answers to use each letter the same number of times? If so,
then consistent letter guessing
is not an issue. If not, then the probability of one letter being used
more frequently in answers is
important. How many of us create exams with equal use of all answer
letters?
My apologies if this has been discussed earlier. Have a great day,
Diana
Diana J. Kyle, M.A.
Psychology Department
Fullerton, College
Office: 714-992-7166
We are what we repeatedly do. Excellence, then, is not an act, but a habit.-
Aristotle
The height of your accomplishments will equal the depth of your convictions.
--William F. Scolavino
----- Original Message -----
From: G. Marc Turner <[EMAIL PROTECTED]>
To: TIPS <[EMAIL PROTECTED]>
Sent: Monday, November 06, 2000 5:36 AM
Subject: Re: The madness continues: the guess mess
> Steven is claiming that given 5 questions, with 5 answer alternatives
each,
> that all students will get at least 1 item correct by random guessing in
> ALL situations. This 1 item correct is what is needed when guessing at 5
> items to "break even" with a 1/4 point deduction for wrong answers.
>
> The claim I'm making is that given the same situation, students will get 1
> item correct ON AVERAGE. This means that, although most students will get
> the statistical 1 item correct, others will get more than 1 item correct
> and some will get less than one item correct. I do not agree that EVERYONE
> will get at least 1 item correct, only some will get this score. My
> understanding of probability is that it works ON AVERAGE, not in isolated
> cases. So, most students will break even, or even benefit, from guessing,
> but some portion of students will actually lower their score by guessing
> because they fail to get the 1 item correct.
>
> Lucky for us, this is something we can test (no voting required)... Have
> your students take out a sheet of paper and randomly guess the answers to
5
> non-existent questions. Each question has a possible answer of a, b, c, d,
> or e. Since we want them to truly guess randomly, we will not actually
give
> them questions... we just want them to randomly give us the responses.
>
> Here is the key to use to grade the responses: C E D E D
>
> Now, grade the quizes... if ANY student fails to get at least 1 correct,
> the claim that it will NEVER hurt a person's score is false. At this point
> we know that Steven's claim does not have support from the data. If you
> continue scoring for everyone, you can figure out if my claim of getting 1
> correct, ON AVERAGE has support or not.
>
> Until we have actual data collected on this, I don't see much point in
> guessing as to the benefits or costs of guessing...
>
> - Marc
>
> G. Marc Turner, MEd
> Lecturer & Head of Computer Operations
> Department of Psychology
> Southwest Texas State University
> San Marcos, TX 78666
> phone: (512)245-2526
> email: [EMAIL PROTECTED]
>
