Thursday, April 9, 2015

Student-developed Driving Questions in Science PBL

One of my favorite aspects of project-based learning is that students identify the areas within a topic that are most fascinating to them, and then dive into that area. I'm very interested in this questioning process, and you can read about our question-storming around hydroelectricity and ecology at the beginning of this school year, and our more wide-open questioning as we started our science geniushour, or - as we're calling it - Science Innovation Time.

I'm still getting my sea legs, though... A few weeks ago, we launched a new PBL theme around paleontology, and we had some great successes as well as some clear areas to improve for future project launches. (I say "we" here because I was at SXSWedu for the critical week of the launch, and so that week was led by the awesome woman - Kiki - who will be moving into my position next year so that I can take charge of our BIG Lab.)

You could think of this as "What to do, and not do, supporting students in developing their own PBL driving questions."

Here follows the plan, the reality, and the reflections:

1.) Bloom's Taxonomy: Having introduced Bloom's Taxonomy in launching the Elwha Project, I re-accessed my students understanding of those question categories, and we practiced brainstorming questions that we already have about paleontology in the different categories. This was straightforward and easy... my kids understand Bloom's.

2.) A Launch Event: The fabulous Leonard Eisenberg of Evogeneao happens to be a cousin of one of my students, and he offered to come speak to my classes. He gave a fascinating presentation on the fossil record and how we understand relationships between organisms. For his presentation, I asked my students to only focus on taking notes of questions that pop into their heads during the presentation. Again, this was easy for my kiddos... they're full of questions!

In launching, I offered the following theme statement to my students, to guide their questioning and their projects:

We rely on tiny amounts of data to craft our theories about the history of life on Earth, from the earliest Precambrian life through the current Holocene era. These theories can sometimes change drastically when new evidence is discovered.
As I explained to my students, this is the ultimate goal of the project: to prove deep learning and understanding around that theme statement.

I crafted the theme statement from NGSS Disciplinary Core Ideas MS-LS4-1 and MS-LS4-2:
The collection of fossils and their placement in chronological order (e.g., through the location of the sedimentary layers in which they are found or through radioactive dating) is known as the fossil record. It documents the existence, diversity, extinction, and change of many life forms throughout the history of life on Earth. (MS-LS4-1)
Anatomical similarities and differences between various organisms living today and between them and organisms in the fossil record, enable the reconstruction of evolutionary history and the inference of lines of evolutionary descent. (MS-LS4-2)

3.) Whole-Class Practice Exploring Artifacts: Then I left for SXSWedu... Kiki's first activity was to present the kids as a whole class with two example artifacts: (1) an image of Tyranosaurus and Stegosaurus highlighting the fact that they lived farther apart than Tyranosaurus and us, and (2) a fascinating video about the discovery of a new anomalocarid. (You should watch it... it's really cool!)

A problem arose when Kiki asked the students to start brainstorming questions from the artifacts. You see, I had created a little template page with a sample question in each of the 6 Bloom's categories for each artifact, and then space for students to write their own question from each of the 6 categories. This really railroaded my students... they felt like they could only think of one question per category, and their many many other questions went uncaptured.

A future solution would be extremely simple... Instead, we should have had the kids brainstorm questions on Post-It notes... just all of their questions - without thought to Bloom's categories. Then the kids could have sorted their questions into categories afterwards. This would validate all of their questions, and if one or two of Bloom's categories have fewer questions, then so be it.

4.) Individual and Small-Group Deeper Artifact Exploration: Kiki and I together had collected some very exciting resources, including 3 "Burke Boxes" from our local natural history museum, and a few excellent children's books (that go much deeper than many children's books!) The idea here was to go back to brainstorming questions on Post-Its and then sorting, looking especially for those higher-level questions.

A problem arose because my students still really didn't have enough background knowledge to thoroughly investigate the boxes... Comparing 4 different hominid skulls is fascinating, but my kiddos simply didn't have the vocabulary to identify specific differences in the "zygomatic arches." When looking at the "Before The Dinosaurs" box, the trilobite fossils were amazing, but my kiddos simply didn't know how to compare them to other arthropods. Because Kiki is an incredible teacher and extremely knowledgeable paleontologist, she was able to more directly shepherd them through observations so that they could build truly deep questions.

A future solution could be to create some sort of guidance sheet to support students in comparing artifacts. Kiki found herself frequently asking students the same questions to guide their observations: What differences do you notice about these two fossils? What characteristics are recognizable? What might be start to infer about differences in how this or that organism lived?

An additional problem was that the hierarchical structure of Bloom's taxonomy led many of my students to interpret that the "create" level was the ultimate goal, and that "lower" level questions were unacceptable. This, of course, is not true! Especially in a field like paleontology, "analysis" level and other questions can lead to deep, rich inquiry.

A future solution could be to use a different scaffold for categorizing questions... perhaps one that is not hierarchical, but rather descriptive but with simple recall questions set aside - more in the googleable/nongoogleable framework.

A third problem (this just all-around wasn't perfect...) was that, to save space, I only posted sorting categories for the three "top" Bloom's levels. As I'm sure you can surmise from the previous problems, this just invalidated kids' "lower level" questions that were perfectly awesome for building bigger and deeper questions off of. Related to this was my students' voicing their beliefs that "create"-level questions were the ultimate goal, since those are the "top" of the pyramid...

A future solution, if I stick with Bloom's, would be to always give space for sorting all 6 levels, and build in explicit time for "growing" the lighter-weight questions into bigger, deeper. Better, though, I think is finding a Bloom's alternative, or at least de-hierarching the top levels so that "evaluate" questions are equally welcome as projects as "create" questions. I like the following graphic for that de-hierarchic-izing.

And I dig this little house graphic, but I don't love it...
Developing a questioning-support graphic may be a project!

5.) Sorting and Analyzing Questions: From their artifact-generated questions, my students selected their favorites to compile into a class database via a Google Form. It was during this database-entering that Kiki did lots of the support to move students from "understand" questions to deeper "analysis" and "evaluation" questions. Again, Kiki's domain expertise was absolutely critical in supporting this deepening... My own paleontology domain expertise would not have cut it!

The following image is literally a random, unfiltered, uncurated list of questions that made it into this database. I think their depth and breadth speaks for itself! You can see the whole database here.

6.) Final Project Proposals: From their own questions as well as from the whole-class database, my kiddos then defined their final project proposal. Here is the Google form I built for my students to enter their proposals, and I ran an Autocrat script to then return each student's proposal to them in the form of a Google doc shared with me and the student. This actually worked extremely effectively, and my students are now well underway in finding resources and scientific evidence to support answers to their questions.



1.) Bloom's is not the ultimate questioning support.
Bloom's is an effective starter to get kids thinking about different kinds of questions, but it can force artificial boundaries if it's held as the end-all, be-all. Questioning categorization needs to be presented flexibly, or at least with multiple different models, so that students' own thinking can also be flexible.

2.) Open-ended artifact exploration is insufficient without support.
We can present the most amazing artifacts in the world, but if the students don't have any ability to notice the amazing-ness, those artifact explorations won't elicit deeper questioning. This falls back to intensive training in observation, and teacher guidance in that observation. Not necessarily hand-holding - not "look at this piece right here" - but as Kiki suggested: some scaffolding to support looking for similarities and differences, and looking for small details.

3.) Launch events are quite effective!
Our visit from a domain expert launched some great thinking, and every other time I've had guest experts has also led to deeper discussions. A visitor with personal research, policy, field, whatever experience can bring so much to the table. Mr. Eisenberg's visit got my students asking big questions right off the bat, and they continued discussing his presentation and points more outside of class than other "typical" classroom activities.

4.) Deeper domain expertise from the teacher leads to deeper questioning from students.
I feel like I've known this in my own experience in my own areas of expertise (genetics, much of human anatomy), but it was fascinating to watch how much it made a difference watching a domain expert lead investigations in a field I'm weak in... Kiki's expertise in fossil details and paleozoology supported students in noticing key features and identifying keywords that I couldn't have supported.


Another nifty idea:
When Kiki and I presented this enactment and our reflections to our 4th-8th grade teacher colleagues, several additional questions arose further around how to support student questioning. They could essentially be summed up as "How do we wrangle student questions so that they (a) are answerable by the students and (b) call for an answer that will serve the learning objectives?"

A couple examples given were:
- (magnetism unit) - Why are some metals attracted to magnets and others aren't? not developmentally answerable by a 5th grader
- (human body unit) - Can I design a functional artificial heart? answering requires some anatomy learning and some really interesting materials investigations, but doesn't address human anatomy learning objectives beyond the circulatory system

A colleague and I came up with a potential solution while discussing the intersections between design thinking and PBL this weekend...

The ideation phase of design thinking and the process of developing potential driving questions for PBL have strong similarities. In the IDEO model, both are extremely divergent phases of a learning and development process. Sorting and specifying processes, then, are convergent phases. We already use many of the same techniques: identifying themes, grouping. What other ideation techniques could we steal to solve these problems?

What if we apply the matrix/axis sorting technique to our potential driving questions? A great technique in converging on wild and broad ideation outcomes is to sort ideas according to a two-axis framework, such as in the examples below:

What if we basically use those two concerns as a potential-driving-question sorting technique? With "highly answerable" and "highly meeting learning objectives" as two axes? This might then be followed up further sorting those top-right questions by two more axes: "most fascinating" and "most likely to be answered via a single google search" (in that case, you'd want bottom-right...)

What if we sorted our first giant brain-dump of questions into these matrices?

I'm definitely going to try this next time.


Infinite gratitude to Buck Institute for Education,, IDEO, NoTosh, and my wonderful colleagues in #dtk12chat for ALL my understanding of PBL and DT.

1 comment: