WCPSS's "Next Big Thing"

O’Connor, Ken (2008).A Repair Kit for Grading.

My Comments in italics.


1. Don’t include student behaviors (effort, participation,
adherence to class rules, etc) in grades; include only
achievement.


A legitimate way to prevent teachers from grading for something other than achievement (behaviour, for example) is to allow a separate grade or rating scale for those behaviours we consider important. There’s no need to grade for behaviour in an academic grade if we have somewhere else to do it, but without that outlet (and 2 comments on a report card are not sufficient) people will still combine the two.


2. Don’t reduce marks on “work” submitted late;
provide support for the learner.

Claiming that penalties for late work distort the actual achievement level of the student represented by that mark is plausible and fair. The anecdotal evidence regarding the effects of penalties on motivation is nothing more than that, and cannot be reasonably treated as evidence of anything. The plural of anecdote is not data. Allowing students to renegotiate deadlines for turning in work is both excellent for teaching responsibility and problematic if poorly handled. Creating a separate record for these behaviours solves some of the problems, and requiring that all standards be met prior to moving to the next course solves the other (that students attempt to extend deadlines indefinitely).


3. Don’t give points for extra credit or use bonus points;
seek only evidence that more work has resulted in a
higher level of achievement.
Extra credit does indeed distort grades. This can be demonstrated by simple mathematics.


4. Don’t punish academic dishonesty with reduced grades;
apply other consequences and reassess to determine
actual level of achievement.

Punishment should be something other than a grade of 0, but also needs to be both a deterrent and something severe enough to convey the seriousness of the offense. A simple talking-to does not accomplish this for most of our students. The assessment also needs to be completed, without dishonesty, afterward in order for the student to demonstrate their actual achievement level on that standard. The sanctions imposed/recommended by MacDonald high are viable if and only if they apply to the given student (loss of extra-curricular for a student who doesn’t participate isn’t effective.). MacDonald is a private Catholic school, so loss of privilege and expulsion are much more realistic penalties to the students and parents than they might be to ours. (We have a limited 365-suspension/expulsion policy because we have a mandate to provide and education to all, not just to all that can pay the fee).


5. Don’t consider attendance in grade determination;
report absences separately.

O'Connor prescribes that attendance should not be considered in grades, yet WCPSS attendance policy is to fail students who miss more than a certain number of days without excuse. Excused absences do not imply that the student has actually made up the missed material; any more than unexcused absences imply that the student hasn’t mastered the material for which they were present or achieved missed objectives after their return. If we want to be honest with ourselves about our grading conversations, FF’s need to become INC’s until the objectives are met or not met. In addition, Gathercoal (2002) has a valid point: absent is absent, and we do no favours by excusing one form and not another.


6. Don’t include group scores in grades; use only
individual achievement evidence.


The argument against group assessment is valid. Cooperative learning products need to be used as formative assessment, if at all.


7. Don’t organize information in grading records by
assessment methods or simply summarize into a single
grade; organize and report evidence by standards/
learning goals.


Requires revision of our entire grading system, report cards and GPA included, to be worth implementing.


8. Don’t assign grades using inappropriate or unclear
performance standards; provide clear descriptions of
achievement expectations.


Marzano’s suggestions of criterion-referenced grades are no less arbitrary than using a simple % proficient, which is also more immediately communicable to parents and others. Criterion referenced grades need to be based, appropriately, on numeric marks with set levels of achievement needed to reach each one. In our case, these levels need to be predictive of performance on objective, outside examinations (CAT, EOC, Stanford 9’s, etc). Indicating both achievement and growth could be useful for everyone involved in using the grade in the future.


9. Don’t assign grades based on student’s achievement
compared to other students; compare each student’s
performance to preset standards.


Do we actually have people doing this? I’ve seen college profs grade on a Bell curve, but I’m unaware of anyone here who is… if so it certainly needs to be addressed. Achievement marks should, realistically, statistically distribute along something resembling a Bell curve in an on-level course simply because of the normal distribution of abilities. Skews towards high or low achievement are also natural if the students in the course are of skewed ability as well (i.e. a class load of true honors students being marked on the same scale as an on-level course. Or ICR graded in the same manner.)


10. Don’t rely on evidence from assessments that fail to
meet standards of quality; rely only on quality
assessments.


Right. This is not profound. The difficulty arises in determining what is actually a valid assessment prior to its use. This fix is fairly straight forward for anyone employing a semi-reflective practice. The footnoted Gardner text on time limits is apt yet we also need to convey, in some form or other, a student’s ability to read. Assessments may not need to rely solely on reading skill, nor should they ideally, but this then needs to open the door to oral or performance assessments that are still predictive of the student’s performance on an outside exam, at least until we allow for orally given EOC’s or do away with them altogether.


11. Don’t rely only on the mean; consider
other measures of central tendency and
use professional judgment.


O’Connor’s difficulty with the mean stems from a misunderstanding of the nature of statistics, and an unwillingness to simply drop the outliers. The mean is a valid, probably the most valid, measure of central tendency provided that the component marks are in a normal distribution. http://simon.cs.vt.edu/SoSci/Site/MMM/mmm.html . There may exist students for whom their mean mark does not reflect their level of understanding of a given objective, in which case something else must be done. O’Connor’s evidence for this topic is once again anecdotal at best.


12. Don’t include zeros in grade
determination when evidence is missing or
as punishment; use alternatives, such as
reassessing to determine real achievement
or use “I” for Incomplete or Insufficient
Evidence.


Giving INC’s in place of zeroes is possibly the best alternative for missing assignments I have seen proposed, provided we require the student to eventually demonstrate proficiency on the standard for which the assignment was missed. This would have to extend all the way to receiving an INC for any course in which the student did not demonstrate proficiency on a given standard (not that not demonstrating proficiency is not equal to demonstrating a lack of proficiency). The fact that they assign a numeric value to a grade that lacks a basis in reality is the only valid argument against the assigning of zeroes for missing marks. The other arguments presented are highly emotional and invalid without corroborating evidence. Adjusting to a 5 or 50 point scale brings the mean back toward centre, and towards the probable level of student achievement, but fails to account for missed assessments. For formative this may work out, but it would be an ineffective way of evaluating a student’s summative assessments because it still involves missing data.


13. Don’t use information from formative
assessments and practice to determine grades;
use only summative evidence.


Good, agreed. That’s what formative assessments are for.


14. Don’t summarize evidence accumulated
over time when learning is developmental and
will grow with time and repeated opportunities;
in those instances, emphasize more recent
achievement.

If the opportunity to use grade replacement (EOC score, or a section of a final summative assessment as applied to a topic which earned a lower mark earlier) were to be legitimately available, it would perhaps be the most accurate way of indicating exactly what a student achieved in a course. EOC data isn’t available rapidly enough for this to work (We’d need a goal by goal item analysis for each student to grade in this manner).

15. Don’t leave students out of the grading
process. Involve students; they can - and should
- play key roles in assessment and grading that
promote achievement.


Involving students in the grading process, at least as far as creating an understanding of that process so that they can monitor their own progress, could potentially provide motivation for those students (if they’re making progress). As with any other use of grades as an extrinsic motivator, it can work either way in the student’s mind. Students and parents, however, should not be left in the dark as to how their (child’s) grade has been determined.



O’Connor’s current work is a re-write of his two earlier works, none of which are based on empirical evidence obtained by O’Connor himself. A limited amount of his suggestions are based on someone else’s empirical evidence, and most are based on anecdotes at best. Some of them are still logically defensible if the initial premise is accepted as valid. The ideas presented are in many cases worth trying if the opportunity is to be fully and realistically given, rather than treated as a half-measure and abandoned after a brief trial period. This would require an overhaul of the grading system district-wide in order to be successful. However, this is once again likely to become something that is systematised to rest the burden solely on the teachers themselves, as our other policies seem to do.
___________________________________________________________________
Anonymous> I agree that it does help to establish bad habits if we don't somehow take into account attendance and timeliness of completing assignments. O'Connor proposes establishing a 2nd mark to convey this information, which I would accept at this level, perhaps not at post-secondary. From my personal experiences at university: I failed a course that I missed 6 times (5 absences were allowed). I averaged 98% on all my coursework and scored 100% on the final. Is it truly representative of my knowledge of my material that I received an F and had to retake that? (Yes, I know, I shouldn't have missed the 6th time... I do get that.) Students who miss that much coursework typically aren't going to pass in the first place. Firm deadlines for assignments are something I'm less likely to be lenient about; If an assignment is designed properly, it's intended as practice or assessment at a given point in time, and its usefullness has passed.