Case Study – Multiple Choice Exams

The Problems – SBG, Content Assessment, and Testing

The Multiple Choice exam is a classic traditional grading method that does not fit nicely into an Standards Based Grading model. In the Fall semester I worked out a system for incorporating the district mandated 55 Multiple Choice Question Common Assessments. These tests covered multiple decades chronologically over at least 6 sub standards. For example, common assessment 1 covered pre-colonial America to the end of the War of 1812. The 6 sub standards were (paraphrased):

  • 1.CO – compare the colonial regions
  • 1.CE – analyze the causes of the American Revolution
  • 1.P – determine the major events that makeup the period of the post-Revolution Era
  • 1.CX – explain the context of the early Federal Era
  • 1.CC – demonstrate the continuity and changes of the United States during this entire period
  • 1.E – analyze Primary and Secondary Source Evidence from this period

That is A LOT for a single test.

The data from this single MC test has limited value if the only number students and teacher ever see is the overall percentage score. To be useful beyond a ranking grade the data needs to be broken down and further analyzed. Standards Based Grading suggests that testing has the capability of becoming part of a grading story. To do so the assessment has to be combined with other data points to determine the trajectory of students understanding. Even with this mindset the percentage score is still too broad to be meaningful. Traditional gradebooks use the percentage score because it gives a false sense of accuracy (which I wrote about here in The Illusion of Accuracy). Instead what we need to do is break the test down into its component parts – the substandards.

My solution to using SBG with these required exams was to throw out the single percentage score, breakdown the test by content substandards, and give a score for each. To increase the feedback element students received an assessment report and were taught how to analyze their own data. This allowed them to see which areas they were strongest in, which areas they struggled in, and they took ownership of improvement. To pull the exam into the grading story students received a level 1-4 grade in each substandard which served as a data point in the Power Law Formula in the mastery level calculation. To promote the Growth Mentality students could then use their analysis and work through a credit recovery form which allowed them to earn up to full credit for each substandard.

This SBG testing system worked well. Students felt less pressure on the tests and most took advantage of the credit recovery system. Not only did this help their common assessment score but the forms were saved by the student and used as a study guide for the Final Exam. The growth mindset and learning were fantastic but the problem came with trying to mesh SBG into the mandatory final letter grades. One big complaint I have about traditional gradebooks is that they are too mathematica and obfuscate the student’s true learning.I found the SBG Mastery Level to Final Letter Grade calculation even more cumbersome and confusing. These were the steps for each student:

  1. Use the Power Law Function to find a 1-4 (to the hundredths place) for each substandard.
  2. Then average all of the substandards (to the hundredths place) to get a Standard average.
  3. Then average all of the Standards to get an Overall Class Average. This was only 80% of the overall Final Average.
  4. The Final Exam was required to be 20% of the Final Average which meant that the Final Exam score had to be translated from a 100 point score down to a 4 point scale (to the hundredths place).
  5. Then the 80% Overall Average and the 20% Final Exam were calculated… Again all on a 4 point scale.
  6. This was then compared to a chart that suggested the Final Average… all down to the hundreth of a point.

I told the students this number was the start of our grade conversation but there were only a few times when the students tried to have this conversation. Most just agreed with the number’s suggested letter grade. It felt like it ruined the whole SBG Concept. The Level Number was too obscure to provide a meaningful context to the student or parents. The grade book I was using was also not intuitive enough to understand without really being trained into understanding the power law function and the actual layout of the data.

Powered Up XP! A Potential Solution

Long story short (ok, not so short…) I am going to keep several pieces of the SBG Model but merge it into my older and far simpler XP Grading. I am going to call this Powered Up XP for now. I am keeping the Level 1 – 4 Scoring on the test for each substandard. Both the students and myself were able to see their areas of strength and needed growth quickly and clearly. I am also keeping the credit recovery possibility. The big change is converting the Percentage score for the substandard into an XP Score rather than a 1-4 score. In my Google Sheets Gradebook (you can see how I have modified from an Alice Keeler template and get a copy here) I created a new tab for the three mandatory exams we have remaining. In the screencap below you can see columns C- R in one of my gradebooks (A-B are the students’ names). 3.CO % is the percentage score for the 3.CO substandard and 3.CO LP is the Level “Grade” Points.

I have used a VLookup function to automate the conversion of the % score into the LP score. I have set a 10,000 LP goal for the semester. Each Common Assessment is worth 1800 LP meaning each substandard on the test is worth 300 LP. A Level 1 = 150LP, Level 2=200LP, Level 3=250LP, Level 4=300LP.

The D2 Cell is highlighted. The IF Statement is saying “IF” there is a number in C2, then use the VLOOKUP. the blank after the final comma tells the IF Formula to return a blank if there is no number. the 2nd C2 tells the vlookup to lookup the number in the cell C2. “$Q$2” tells the Vlookup where the top left cell in the chart starts and “R” tells the vlookup where the chart ends. Technically the chart goes down the column forever because there is no row ending number. the “2” tells the VLookup which column value to return. Because the value in C2=65 it falls in Level 2 and the 2nd column in the Q2:R chart is 200. The chart looks like its three wide but I did not include column P in the vlookup calculation.

In order to more easily enter the numbers from the student’s data reports I have hidden the LP Columns. This way I can simply TAB the columns.

This is what it looks like in the gradebook return. (There is a formula at work here too…)

So after all of this the student will get 6 numbers that clearly feed into one overall pot of LP. They can still return to the exam and grow their understanding through the credit recovery work and have it demonstrated in their gradebook. The benefit here is that it is not a complicated math formula to see their progress.

In my next few posts I intend on discussing a few other points. First is the new personal dashboards. Second, why I am calling this Powered Up XP grading but using the term LP or Level Points (maybe Life Points?). Third, I still need to take on the problem of Grinding – something I might be able to With the help of some of my EDU Twitter pals like Alicia Woody (@SuperEvansMath)