Learning Science · Building reliable skills

Reliable routines and flexible expertise

Combine dependable routines with the ability to recognise exceptions, learn and adapt.

Use the examples to connect this lesson with the earlier topics. No specialist equipment is needed.

After reading, you can

Distinguish efficient routine execution from adaptive expertise.
Use changed conditions to test whether a familiar method remains appropriate.

A routine works inside conditions

A cook reliably doubles the ingredients in a recipe, then finds that a much larger pan changes cooking time. The routine of multiplying quantities was useful, but it did not preserve every physical condition. Expertise includes recognising when a dependable procedure still fits and when the situation requires a new analysis.

Routine expertise emphasises efficient performance of familiar tasks. Adaptive expertise includes using knowledge flexibly, monitoring uncertainty and learning when conditions change. These are complementary aims: reliable routines free attention for unusual demands, while understanding their basis helps prevent a routine from being applied blindly.

Teach what changes the decision

After a routine is established, vary a condition that matters. Ask the learner whether the method still applies, what needs adjustment and why. Compare the changed case with the familiar one. The purpose is not novelty for its own sake; the variation should make a boundary or deeper relationship visible.

Explicitly compare cases with the same structure but different surfaces, and cases with similar surfaces but different structures. Averages of equal-sized and unequal-sized groups look similar but require different weighting. The comparison makes the method-selection decision part of practice, rather than assuming that repeated execution will automatically teach it.

Adaptation includes seeking knowledge

An adaptive response may be to inspect a reference, run a small test or ask for expertise. Inventing a new procedure without checking it is not flexibility worth celebrating. State what is known, identify the uncertainty and choose evidence that could resolve it. The resulting revision should be tested against the task’s criteria.

Broad educational research treats adaptive expertise as continued learning and flexible use of organised knowledge. It does not promise that a general creativity exercise transfers to all domains. Flexible judgment depends on relevant knowledge and opportunities to use it. Evaluate a learner on routine reliability as well as justified responses to meaningful changes.

For a concrete authored check, take valid measurements 10, 12 and 14 plus the documented missing code −999. Treating all four as measurements gives −240.75; excluding the code leaves 36 ÷ 3 = 12. Report the excluded entry. This observed-data mean alone does not establish that missing observations would have matched those retained.

Authored dataset: (10 + 12 + 14)/3 = 12. Treating the code as data gives −240.75; exclusion alone does not remove possible missing-data bias.
Entry	Documented meaning	Use for this observed-data mean?
10	Valid measurement	Include
12	Valid measurement	Include
−999	Missing-value code	Exclude and report missing
14	Valid measurement	Include

Worked example

A spreadsheet formula meets a new dataset

What should change when a measurement file uses −999 as a missing-data code?

A learner has a routine for averaging a column of daily measurements. A new file contains blanks and a code of −999 for missing values. Applying the routine unchanged produces a meaningless average because the code is treated as a measurement. The competent first step is to inspect how missing data are represented.
The learner documents the code, excludes entries according to a justified rule, and reports how many observations remain. They test the revised calculation on a few hand-checked cases. Adaptation preserves the purpose of estimating the measured quantity; it changes the procedure when the old assumptions fail.

Inspect and document missing-data codes before averaging. Excluding a documented code repairs the calculation but does not establish that the remaining observations are unbiased.

Try it yourself

Think through each question before opening the answer.

Is abandoning a familiar method always evidence of flexibility?
Show the answer and reasoning
No.
The change should respond to relevant conditions and be justified. A dependable routine may remain the best method when its assumptions hold.
Why compare similar-looking cases with different structures?
Show the answer and reasoning
It teaches learners to attend to the relationship that determines method choice, rather than using surface familiarity as the only cue.
Equal-sized and unequal-sized groups can look alike while requiring different weighting.
What can an expert appropriately do when a new condition is not understood?
Show the answer and reasoning
Identify the uncertainty and seek evidence or help.
Acknowledging a knowledge boundary supports adaptation and prevents confident misuse of a routine.

What to carry forward

Build reliable routines and teach their conditions.
Adapt through relevant knowledge, comparison and checks.

References

Sources supporting the explanations and learning checks in this lesson.