Assessments - Depth and Frequency

A few unsorted observations that I've had on assessment, now that the first term using SBG is nearing an end:

• Assessing a standard 3, 4, or 5 or more times in a week is silly (especially because I report scores only weekly - those would all turn into one score).  Working with those standards that many times is great, but pick one to be an assessment, preferably the hardest.
• It's great (almost required) to assess each content standard before the test.  For me, this has meant more frequent small quizzes.
• It's tricky to deal with assessing a standard with a shallow-level question, in the absence of other data.  Getting it right doesn't really mean a 4, though getting it wrong could easily mean 0 or 1.
• Frequent assessments dealing with lots of standards are a pain to keep track of, and take longer to grade.  Sometimes, it's unavoidable.
• Sometimes, I feel like I need a "knew what the hell was going on in lab" standard, apart from any sort of content or skill.
• Broad and shallow assessments seem to give me the most pause when determining the scores to give.  Tests can be deep enough (and varied enough) that I can get a good picture of many standards at once.  Narrowly-focused quizzes, however, seem to be the best kind, because I can get the same kind of insight that I can on a test, just not over so many standards.  Lots of shallower questions (or a few questions smearing a variety of topics together) can tell me what you don't know anything about, but they're not great at telling me what you know at a 3.5 or 4 level.
• Even though the understanding of some overarching concepts is key in the overall prosecution of a problem, I tend towards mostly assessing them via "conceptual" questions, which can come as 'clicker' questions or ranking tasks, or which can come as the same types of questions on tests.  I'm not sure if that's good or not, but that's just how it seems that it has gone.
• Fewer kids than I've expected have come back to reassess.  It'll surely pick up towards the end of the term, but the "one standard per day" reassessment rule will keep it manageable for me.

Keeping Track

OK, so now we don't have a gradebook of points on assignments, but instead one of scores on standards, each of which can change multiple times during the term, and not even necessarily the same number of times for each student.  How do we keep track?


Method #1: just rewrite the old scores with new ones every time a new assessment is done

• Pros: relatively easy to do, compatible (mostly) with the FAWeb gradebook
• Cons: lots of entry into the gradebook software, FAWeb doesn't do the grading function correctly, but it could average the standard scores, doesn't track history of the scores, you have no record of a student's score on a particular assignment

Method #2: Have a different spreadsheet/record book for each kid, with a column or row for each standard

• You could potentially track scores by day or by assignment
• Pros: Lots of historical data
• Cons: Tracking by assignment certainly brings up lots of writing of assignment name on many sheets, you can't see the class at once, so many sheets!, not compatible with FAWeb (or only so by you updating each student each week or so in parallel with your other grade sheets

Method #3 (mine): Track student scores on a weekly tally sheet (maybe up to 3 per standard per student) - each class gets a grid with students in row, standards in columns, and scores are entered in the boxes, color-coded by assignment; come up with an average or something of those, by whatever algorithm or judgment you'd like to apply (I use judgment, looking at the different scores and their assignment types); enter those scores for each student for the week into Excel gradebook for that week; the gradebook uses the calculation method specified to come up with the current score for each student for each standard and then creates a report for each student tracking historical and current scores for each standard for each student, displaying those on graphs.

• Pros: Not as much work to implement as it sounds (really!); tracks history to a reasonable degree in the gradebook, while tracking history completely in the accumulated weekly tally sheets, if needed; historical graphs and reports for each student; class averages for each standard - overall and each week - so that I can adjust instruction to the current state of knowledge
• Cons: Takes maybe 30-45 minutes/week to enter grades, export spreadsheets [I'm planning to write an Excel macro to make that much quicker, actually], and upload them to wikiphys (so that families and students can access them) [I don't think that's so bad, so maybe it's a pro or neutral]; not compatible with FAWeb at all [I post links to their documents in the comments on FAWeb]; takes some work to set up the sheets, etc. at the beginning of the year/term, and a little to print the summary sheets each week

You see which I'm partial to, but observations and suggestions are welcome!


Here are sample documents for a class, with names redacted:

• Week 3 tally sheet
• Week 3 page from gradebook
• Summary page from gradebook (calculating the 'counting' scores for each student for each standard)
• One student's report page, showing current ('counting') scores and historical graphs

Endings

The big test on a unit isn't necessarily the last time that a standard is assessed - that's one of the awesome things about SBG - so students aren't off of the hook after the test is over.  That 3.5/4 that you got on acceleration goes away as soon as you show that your understanding's shaky now, and I love that.

That said, the content standards don't come up as often after their initial big unit tests.  In this way, the unit test is a kind of "final look" at content standards - nothing's ever really final, but some of those scores may not come up again, and those may stick.  It's at this point that kids would reassess individually, if they didn't quite get a concept or two on the test.

What about the skills standards?

• We're talking about algebra, 'no numbers 'til the end', units, graphical and algebraic models, making predictions, presentation, concept synthesis, etc. here
• These come up a lot - maybe they should be averaged over the term, or maybe the last, say, three scores should be averaged.  I don't quite know how to do that last one in my Excel gradebook (we'll look at that later!)
• They're so universal that I'm never 'closing the book' on them, so there's not so much of an end to these

How about the lab standards?

• We're talking about experimental design, lab writing, lab report format, error analysis, measurement here
• Perhaps there should be a final 'big' lab planned each term, so that I'm pretty sure that the last look at these standards is a good one
• I'd love to get some of these 'in the raw,' - that is, not in a group setting.  That's easy for writing and reports, but the measurement and especially the expt. design are usually found in group settings.  I need to get them isolated, but it's hard in a class of 17 to do that.  Perhaps on the last test of the term, somehow?

Beginnings

Here's my new mantra:

"Every great lesson begins with a great question...

...that the students generated."

If the kids come up with the questions, then they care more about figuring them out.  The art is steering them so that they're dealing with the desired concepts, and the hard part is giving up control of the lesson.

For example:

(I stole this video from the internets, of course!)

The Great Tricycle Race

I used this one this year.  My prompt: "What can you do with this?"  It took a few minutes, but some great questions came up:

• "How fast are they going?"
• "Where were they when the race started (assuming that they started at the same time)?"
• "When did each start (assuming that they started in the same place)?"
• "When/where are the unseen passes?"

Awesome questions - they took 35 minutes and took some data from the video, answered their own questions, and presented the answers to each other.  I was really pleased with the first trial for this new method.

In subsequent units, I'm going to tease the whole thing with a video, let the class develop the questions, and then come back to them right before the test and let them use their new knowledge to answer them.

For 2D kinematics, I'm thinking about this one:

Yikes

The State of Affairs

Here's how I'm doing it so far:

SBG Info Given to Students

Open Questions

Some open questions that I have about how I'll implement SBG:

• How many assessments per standard per week is too many? ...too few?
• How many assessments per standard per unit (up to and including the test) would be appropriate?
• Should I include a standard for timeliness of submission of work?
• What about conjunctive grading?  Should I require no grades below 2 (out of 4) for an A for the term, so that students don't blow off standards?  
• In general, is there a good way to assure that standards aren't over- or under-weighted in the final grade?

Let me know your thoughts!

Here we go!

I'm starting a place to post my reflections and experiences with Standards-Based Grading (SBG), and rehtorical (and non-rhetorical) questions about its implementation at Tatnall.  Comment early, comment often!