Chemistry · Lesson

Choosing and comparing quantitative methods

Choose and compare quantitative methods using the same defined quantity, matrix and evidence requirements. Balance identity, selectivity, uncertainty and practical constraints rather than ranking methods by speed or tradition alone.

Back to the Chemistry map

By the end, you can

Compare two results with their stated independent uncertainties.
Solve an invented two-response amount model.
Assess a fixed weighing-uncertainty toy against a given relative target.
Filter supplied candidates by entity, matrix, range, uncertainty and time.

The intended result determines what evidence is needed

A classical method can be suitable when the chemistry provides a selective, sufficiently complete and observable relationship to the intended quantity. An instrumental method can supply other forms of information, but its name does not guarantee better selectivity or fitness. Compare the actual procedure and its evidence for the target and matrix.

Ask whether the method measures the intended entity or an operationally defined reacting fraction, whether sample preparation changes it, which interferents are accounted for, what interval and uncertainty are supported, and how references and controls are used. Available material, time and resources matter after the result requirements are clear.

Certificates and documented procedures can show the scope of references and approaches, including titrimetric, gravimetric and instrumental characterization. Their existence does not make every procedure interchangeable for every analyte or matrix.

Compare the same measurand with uncertainty on the same basis

Two invented independent results for the same specified analyte, matrix and basis are A=4.90 mg L⁻¹ with standard uncertainty 0.120 and B=5.15 mg L⁻¹ with standard uncertainty 0.160. Their difference B−A=0.250 mg L⁻¹ has standard uncertainty √(0.120²+0.160²)=0.200 mg L⁻¹.

A chosen illustrative compatibility rule uses |difference|≤2u(d). The bound is 0.400 mg L⁻¹, so these results meet that rule, with normalized difference 1.25. This neither proves absence of shared bias nor establishes one method more accurate. Correlated results need a covariance term.

Results for different chemical forms, extraction fractions or reporting bases cannot be made equivalent merely by using the same unit. Agreeing numbers also do not establish that both procedures measured the intended target quantity.

Two responses can determine what one response leaves unresolved

In an original exact abstract model, two amounts n_A and n_B contribute proton-transfer inventory q=n_A+2n_B and mass m=(100 g mol⁻¹)n_A+(150 g mol⁻¹)n_B. The coefficients are assigned teaching values for unnamed A and B, without a real compound or operating procedure.

Given q=0.00500 mol and m=0.400 g, substitute n_A=q−2n_B into the mass relation. Then m=(100 g mol⁻¹)q−(50 g mol⁻¹)n_B, giving n_B=0.00200 mol and n_A=0.00100 mol. Both reconstruct the supplied observations.

The numerical coefficient determinant is −50 when amounts are expressed in mol and mass in g; it carries the appropriate mass-per-amount dimension. A nonzero determinant allows inversion in this exact toy. One inventory alone cannot uniquely determine both amounts. Actual complementarity, identity, noise and conditioning still need evidence.

A fixed uncertainty has a quantity-dependent relative effect

For a separate invented two-weighing model, each reading has independent standard uncertainty 0.0500 mg, giving u(net)=√2×0.0500≈0.0707107 mg. At net masses 2.00 and 50.0 mg, relative standard uncertainties are approximately 3.53553% and 0.141421%.

If the sole supplied model criterion were relative standard uncertainty no greater than 1.00%, the algebraic minimum net mass would be u(net)/0.0100≈7.07107 mg. This is a model comparison, not a real recommended minimum specimen mass or a complete assessment of gravimetry.

Actual homogeneity, phase identity, recovery, matrix effects and reference contributions can add requirements. Increasing an inventory may reduce one relative uncertainty while leaving other constraints unresolved.

A fast precise method for the wrong entity does not answer the question

Choose an invented target: mass concentration of symbolic A in matrix M over 1–10 mg L⁻¹, with maximum standard uncertainty 0.100 mg L⁻¹ and maximum time 60 min. Four anonymous candidates are supplied as already assessed within their stated scopes; these are teaching data, not real-method rankings.

All supplied ranges cover the target interval. Candidate A measures A in M with u=0.150 mg L⁻¹ and time 10 min; it fails the uncertainty criterion. B measures A in M with u=0.0800 and time 40 min, meeting all supplied criteria. C has u=0.0200 and time 5 min but measures a different entity B. D has u=0.0400 and time 20 min but is validated only for another matrix N.

Only candidate B meets this specified set of requirements. The fastest or smallest-uncertainty number belongs to an unsuitable scope in this puzzle. Further requirements or incomplete validation evidence could change the decision; the table is not a universal preference for one technique.

Select a supplied candidate by scope as well as numerical performance

Candidate	Entity / matrix	Given standard uncertainty / mg L⁻¹	Given time / min	Assessment for this target
A	A / M	0.150	10	Fails uncertainty criterion
B	A / M	0.0800	40	Meets all supplied criteria
C	Different B / M	0.0200	5	Different chemical entity
D	A / different N	0.0400	20	Different validated matrix

Original anonymous-candidate puzzle. Target is symbolic A mass concentration in matrix M over 1–10 mg L⁻¹, maximum standard uncertainty 0.100 mg L⁻¹ and maximum time 60 min. Every supplied candidate range covers the target interval; only B meets all stated criteria. These are invented assessed scopes and values, not claims about actual methods.

Explain both the evidence and its remaining limits

Record the intended quantity, scope, performance criteria, references and reasons for the chosen procedure. When comparing methods, retain sample mapping, basis conversions, uncertainties and correlations, rather than only presenting a winner.

A justified method choice makes the result interpretable and reviewable. Classical amount relationships, separation behavior and instrument responses each contribute useful evidence when connected to the actual analytical question.

Worked examples

Use the uncertainty of the difference

Compare the stated independent results A and B.

Difference is 0.250 mg L⁻¹.
u(d)=0.200 mg L⁻¹.
Chosen 2u(d) bound is 0.400.
Normalized difference is 1.25.

Compatibility under this rule, without universal method superiority or bias exclusion.

Use both independent exact response equations

Use q=0.00500 mol and m=0.400 g with the assigned coefficients.

n_B=(0.500−0.400)/(50 g mol⁻¹)=0.00200 mol.
n_A=0.00500−2×0.00200=0.00100 mol.
Both reconstruct the given q and m.

Two determined amounts in the specified exact model; one response alone remains insufficient.

Normalize fixed uncertainty to net mass

Use u(net)≈0.0707107 mg.

At 2.00 mg, relative u≈3.53553%.
At 50.0 mg, relative u≈0.141421%.
For the supplied sole 1.00% model criterion, minimum is about 7.07107 mg.

A scoped arithmetic comparison, not a complete sampling or method recommendation.

Check each supplied requirement

Use the anonymous target/candidate puzzle.

A fails the supplied uncertainty criterion.
C measures a different entity.
D has a different validated matrix.
B meets the supplied entity, matrix, range, uncertainty and time criteria.

Only B qualifies under the stated information and requirements.

Try it yourself

Keep the same two-response coefficients but use q=0.00300 mol and m=0.250 g. Find n_A and n_B.
Show the answer and reasoning
Both 0.00100 mol.
n_B=(0.300−0.250)/50=0.00100 mol; substitute into q for n_A.
Use the same fixed two-weighing u(net) at net mass 10.0 mg. Find relative standard uncertainty.
Show the answer and reasoning
Approximately 0.707107%.
0.0707107/10.0×100 gives the relative value in this model.
If the candidate puzzle changes only its time cap to 30 min, does B still qualify?
Show the answer and reasoning
No.
Its supplied time is 40 min; changing a requirement changes the scoped decision.
Do two results in mg L⁻¹ necessarily measure the same chemical form and reporting basis?
Show the answer and reasoning
No.
Units alone do not establish entity, matrix, extraction fraction or basis equivalence.

What to carry forward

Choose for the intended quantity and matrix.
Compare results on the same basis with uncertainty.
Added responses need valid independent information.
A fixed uncertainty has a quantity-dependent relative effect.
Document the scope and evidence of the choice.

References

Teaching resources and reference material supporting this lesson’s explanations and examples.