Wait, what? Sometimes giving a student a precise goal can make learning harder.
In ordinary problem solving, clarity about the goal is usually valuable. But for a novice facing a complex unfamiliar problem, the demand to keep comparing “where I am” with “where I must end up” can consume attention that might otherwise be used to notice the structure of the problem.
Quick Answer
Owned learner job: temporarily replace a tightly specified end goal with a broader instruction so the novice can work forward from known information and learn useful problem moves.
A classic Mathematics example changes “Find angle ABC” into “Find as many angles as you can.” The point is not vagueness. It is to reduce a particular kind of search while the learner is still building the schema needed for efficient problem solving.
Why a clear goal can still create a poor learning operation
A conventional unfamiliar problem often encourages means–ends search: compare the current state with the target, identify a difference, choose a subgoal, search for an operation that reduces the difference, and repeat. That may solve the problem, but for a novice it can require many interacting elements to be held and coordinated at once.
A goal-free version changes the operation. Instead of repeatedly asking “How do I get to the target?”, the learner asks “What can I validly derive from what I know now?” This can direct attention toward relationships and legal moves rather than toward an exhausting backward search.
What this learner state can look like
- The student knows several relevant rules but cannot coordinate them toward the target.
- They repeatedly restart because they cannot see which subgoal to choose.
- They can follow a worked route but conventional problems trigger long, unproductive search.
- When asked what can be derived immediately, they suddenly produce valid steps.
Do not infer “poor problem solving ability” from these observations. Missing prerequisite knowledge, weak retrieval, representation difficulty, language load, or a badly designed problem can look similar.
A discrimination test
Give two structurally similar problems. Keep one conventional. Reframe the other with a non-specific goal such as “Calculate as many relevant quantities as you can” or “Write every justified inference you can make.” If the learner suddenly generates valid forward moves, the bottleneck may involve goal-directed search rather than absence of all relevant knowledge.
The Goal-Free Protocol
1. Use it for the right problem. The task must have a manageable set of instructionally useful moves. If “find anything” creates dozens of irrelevant possibilities, the format has failed.
2. Make the broad goal bounded. “Find as many relevant values as you can” is better than “do whatever you want.”
3. Ask for justification. Every forward move should be attached to a rule, relationship or piece of evidence.
4. Compare routes. After solving, ask which derived steps turned out to be useful and why.
5. Fade back to ordinary goals. Goal-free problems are a learning format, not a permanent replacement for conventional problems. Return to specific-goal tasks and test whether the learner now sees the structure more efficiently.
How do we know?
Goal-free problems are a long-standing instructional effect within cognitive load research. Reviews describe how they can reduce the extraneous load of means–ends search for novices and support schema construction. A major review of cognitive architecture and instructional design explains the mechanism and its place alongside worked examples and guidance fading. Read the review.
The boundary matters. A 2025 study with 150 seventh-grade students found lower cognitive load with goal-free presentation but no significant main effect on retention or transfer in that experiment. That is a useful warning against turning a robust theoretical idea into a universal promise. See the study.
When not to use it
- When the learner already has an efficient schema and needs realistic conventional practice.
- When the problem admits too many irrelevant moves.
- When the real weak link is missing prerequisite knowledge.
- When removing the goal changes the construct you actually need to practise.
Parent and tutor guide
If a learner is staring at a target and repeatedly saying “I don’t know how to start,” do not immediately give the first step. Try changing the question: “Forget the final answer for a moment. What can you work out from the information you already have?”
If valid moves appear, let them accumulate. Then reconnect those moves to the original goal. This keeps more of the reasoning with the learner than simply supplying the route.
Transfer and return test
After several goal-free items, restore the specific goal on a new problem. Can the learner now choose useful moves without a long search? If not, the operation may have reduced immediate load without yet building a transferable schema. That return result matters.
MindOS direction
Observe: novice has relevant rules but gets trapped in target-driven search → Discriminate: compare specific-goal and bounded goal-free versions → Operate: derive valid forward moves → Fade: restore the conventional goal → Transfer: new problem → Return: can the learner now coordinate the route independently?
Neighbouring owners include Working Memory Load, Worked Example State, and Strategy Selection. This page owns the narrower operation of temporarily reducing goal specificity to change the novice’s problem-search behaviour.
eduKate Sengkang 