At our school our assessment processes are undergoing a huge shake-up. This has involved the development of an assessment policy, followed by rearranging Powerschool Gradebook to fit our needs and lastly, developing a new report card which gives information on both achievement levels in various criteria, separated from learning behaviours, as well as the final overall level of achievement.
For the MYP program, this works perfectly: in mathematics we have four criteria, based on described achievement levels, and a mandated system for determining the overall achievement level.
In group 5 (Mathematics) of the diploma program, determining a student’s score works very differently. We have attempted to mirror the IB process as much as possible in our class assessment. The big problem is that the final achievement levels 1-7 are based on descriptions. Meanwhile the two largest components of the final assessment (totalling 80% of the grade) are based on the total points in a test, and then converted to the 1-7 scale with a grade boundary system.
So now, at our school, we are converting every score for a summative assessment into a 1-7 achievement level, and then at the end of the grading period, awarding students with a final achievement level using “best fit” judgement.
My questions about this process are as follows:
How can an indicator such as number of points accumulated on a test be converted into a descriptor? Two students could have exactly the same score, but their level of achievement may be described very differently. For example one student may be able to cope with all levels of problems, including being able to apply his knowlewdge to unfamiliar situations, but makes careless errors or does not communicate some of the working and loses points for those reasons. Meanwhile there could be another student who does the basics very well: communicates the thought process clearly and is accurate and careful in responses, but is unable to apply the concepts to unfamiliar situations.
Given that the boundary for a score of 7 may be 85%, is it right that a student with 84% and 78% are both awarded a 6, while a student who scores 85% is awarded a 7? My concern is that we are converting a continuous set of data to discrete scores. At least with the traditional letter grade system, this process only happened at the end of a grading period, not for each individual assessment. Plus that system made some allowances with such issues with the A- and B+ terminology.
In order to rectify these two problems I would like to propose to the IB that examinations and other work is scored differently. A criterion based examination score would be much more work than the current system, but at least it would give more detailed and equitable information about the students. Off the top of my head, I would envisage the criteria to be something like the following: Technical skills (can the student use the chain rule, are they accurate in their processes etc); Conceptual understanding (can the student make conclusions about the derivative when given the graph of the integral, etc.); Application skills (can the student work use the concepts learned in new and unfamiliar situations); and possibly Communication skills (do they produce clear graphs with a suitable scale, do they explain their thoughts through math working etc). There might be others too.
This might be a great deal more work, but I believe it is more in line with the IB philosophy of assessment, it is similar to the scoring process for the Internal Assessment component, and it gives a much more thorough and fairer picture of the student’s achievement level.