Dear CSCI-220 and CSCI-316 Students,

This FYI email explains how I will score any
multiple-response items on this semester's exams.
Multiple-response items are fairly similar to
multiple-choice items that have exactly one
correct choice, but differ in that the number of
correct choices may be ZERO, ONE, or MORE THAN
ONE: Students are asked to write down ALL the
correct choices (if there are any). Students who
think none of the choices of a multiple-response
item is correct will be required to write a
specified phrase or word (e.g., "ALL FALSE" or
"NONE") that explicitly indicates this.

Not all of my exams include multiple-response
items, but some do. (On my exams, multiple-choice
items that have exactly one correct choice usually
offer exactly FIVE choices, but multiple-response
items that may have zero, one, or more than one
correct choice usually offer exactly FOUR or at
least SIX choices.)

Many instructors score multiple-response items
using the "all-or-none" method--see, e.g., the
last paragraph of:
https://naiku.my.site.com/s/article/scoring-multiple-response-items.
However, I use another, rather more lenient,
scoring method that's consistent with the attached
description of "Right Minus Wrong" scoring of
Multi-Select questions on Brightspace quizzes.
But, as mentioned above, in my multiple-response
items it's possible for none of the choices to be
correct, in which case the student is required to
write a specified word or phrase--e.g., "NONE" or
"ALL FALSE".

In more detail for those who are interested, this
scoring method calculates the score based on
(no. of RIGHT decisions) - (no. of WRONG decisions) 
where a "RIGHT decision" is writing down a correct
choice or NOT writing down an incorrect choice
(and "WRONG decision" means the opposite):
Students receive 0.25 point for each "RIGHT
decision" but lose 0.25 point for each "WRONG
decision". But a student who makes more wrong
decisions than right decisions will receive zero
rather than a negative score.

  Consider, for example, a multiple-response item
  that has two correct choices C1 & C2 and two
  incorrect choices C3 & C4. In this case:
  
  A student who writes just "C1" gets 
           0.25*3 - 0.25 = 0.5 pt.               
  for making 3 right decisions and 1 wrong
  decision. (The 3 right decisions are to write
  C1, to not write C3, and to not write C4; the
  wrong decision is to not write C2.) 
  
  A student who writes "C1, C2" gets 
           0.25*4 = 1 pt.              
  for making 4 right decisions and no wrong
  decision; this is the maximum possible score.

  A student who writes "C1, C4" gets 
           0.25*2 - 0.25*2 = 0  
  for making 2 right and 2 wrong decisions. (The 2
  right decisions are to write C1 and to not write
  C3; the 2 wrong decisions are to not write C2
  and to write C4.)
  
  A student who writes "C1, C3, C4" gets 0 because
  s/he has made more wrong decisions than right
  decisions (1 right decision [to write C1] vs. 3
  wrong decisions [to not write C2, to write C3, 
  and to write C4]).

  A student who writes "C1, C2, C3" gets 
           0.25*3 - 0.25 = 0.5 pt.
  for making 3 right decisions and 1 wrong
  decision. (The 3 right decisions are to write
  C1, to write C2, and to not write C4; the wrong
  decision is to write C3.)

When there are 4 choices (which has been quite
common in my multiple-response items over the past
15 years), this scoring method means a student
gets 1 point if s/he makes 4 right decisions
(i.e., his/her answer is completely correct), gets
0.5 point if s/he makes 1 wrong and 3 right
decisions, and gets 0 if s/he makes 2 or more
wrong decisions.

See you in class tomorrow.

==============================
T. Yung Kong, D.Phil.
Professor
Computer Science Department
Queens College, CUNY
Flushing, NY 11367, U.S.A.