How we classify questions and plan instruction: Bloom's Revised, SOLO, and Webb's Depth of Knowledge, with the neuroscience behind each, our five-tier difficulty model, and the reasoning for why we start every concept with a "Learn" question.
Each one answers a different question about a task. Great test prep uses all three.
Bloom's
What cognitive action is the student performing? From remembering a fact to creating something new. Our default for course questions.
SOLO
How well-integrated is the student's understanding? Moves from isolated bits to a connected schema. Best for tracking conceptual growth.
Webb's DOK
How deeply must the student think? From quick recall to extended, multi-step reasoning. Most aligned with how standardized tests classify items.
The neuroscience throughline. All three frameworks converge on one finding from cognitive neuroscience: the
depth of processing effect (Craik & Lockhart, 1972). Information encoded at deeper, more elaborative levels recruits
wider cortical networks (particularly the dorsolateral prefrontal cortex and parietal association areas) and leaves more
durable traces in long-term memory than shallow, rote repetition. Taxonomies are the practical scaffolding we use to engineer
that depth.
Bloom's Revised Taxonomy
Six verbs, from simple recall to creative synthesis. Anderson & Krathwohl's 2001 revision reframed Bloom's original 1956 nouns as cognitive actions, which is why it maps so cleanly onto question design.
Neuroscience: The Bloom's hierarchy loosely tracks cortical engagement. Lower tiers (Remember, Understand)
lean on hippocampal retrieval and temporal-lobe semantic networks. Upper tiers (Analyze, Evaluate, Create) recruit the
prefrontal cortex for working memory, inhibition, and executive function. Climbing the ladder literally means lighting up more brain
but only once lower tiers are fluent. Skipping to "Create" before "Remember" produces the cognitive equivalent of a building
without a foundation.
Biggs & Collis (1982). SOLO (Structure of the Observed Learning Outcome) looks at a student's response and asks how structurally complex it is. It's the clearest lens for seeing whether a student has isolated facts or a connected schema.
Neuroscience: SOLO is really a theory of schema construction. Neural binding, the process by which
distributed cortical representations are tied together via synchronous gamma-band activity, is what separates a multistructural
response (several facts, unintegrated) from a relational one (facts bound into a coherent whole). This is why asking a student
to "explain how X relates to Y" is a much stronger study move than re-reading: it forces the binding to happen.
Webb's Depth of Knowledge
Norman Webb (1997). DOK looks at the task itself, how much cognitive work is required, independent of difficulty. A "hard" recall question is still DOK 1. A "medium" multi-step problem is DOK 3. This distinction is critical for standardized test design.
Neuroscience: DOK correlates with sustained prefrontal activation. DOK 1 tasks resolve in under a second
via automatic retrieval. DOK 3–4 tasks require holding partial results in working memory while running additional operations
engaging the central executive (Baddeley's model) and extending dwell time in prefrontal networks. When a student says a test
"felt long," they're usually describing high-DOK items, not high-difficulty ones.
Compare & contrast
Click a row to highlight the equivalent levels across all three taxonomies. The same cognitive demand lines up, even though the vocabulary differs.
Cognitive complexity
Bloom's
SOLO
Webb's DOK
These alignments are approximate, real tasks cross cells. But the mapping is useful for planning: if you want a DOK 3 item, reach for Bloom's Analyze/Evaluate and a SOLO Relational response.
Our five-tier difficulty model
Difficulty ≠ cognitive complexity. A DOK-1 recall question can be easy or hard depending on how common the fact is. We use a separate five-tier ladder so teachers can sequence exposure intentionally.
Why five tiers, not three: adding a Learn tier below Easy and a Challenge tier above Hard gives us
two extra rungs that matter pedagogically. Learn items prime new content (see the priming section below). Challenge
items stress-test retrieval across concepts, which is what actually transfers to novel test items, the single strongest predictor
of real test-day performance (Roediger & Karpicke, 2006, on testing effect + transfer).
The Japanese priming philosophy
Why we lead with a "Learn" question before teaching the concept, not after.
In traditional Japanese language pedagogy, students meet a character or grammar pattern in context before the rule is
explained. They read an example, try a pattern-match, and only then does the teacher name the rule. This works because the
brain's pattern-recognition systems (primarily the basal ganglia and implicit-learning loops) begin firing the moment novel
stimuli appear, regardless of whether instruction has happened yet. By the time the teacher explains the rule, the
student's hippocampus is already asking the right question, which is exactly the state in which explicit instruction sticks
best.
That's what our Learn tier does. A Learn question is deliberately almost-too-easy, sits right before a lesson,
and asks the student to attempt the idea with no instruction. Success feels confidence-building. Failure does the heavier
cognitive work, it primes the student to notice the teaching moment that follows. Either outcome is a win for
encoding.
"Testing is not something we do after learning. It is the learning. A well-designed retrieval attempt, even one
the student fails, strengthens the trace more than an equal amount of study time."
Practical translation for our question bank: every unit should open with 1–2 Learn items, move through Easy → Medium to
build fluency, then finish with Hard and Challenge to test transfer. Course tests mix all five tiers. Diagnostics
deliberately skew harder because their job is to map the ceiling, not the floor.
learningwithsruthı