Choose your reading route
eduKateSG · Why Mathematics? · Horwitz Equations, HorRat and Concentration-Dependent Precision
Start with the question, audit the worked example, or jump to the assumptions and practice. Every chapter stays open for uninterrupted reading.
Understand the question
Define the quantity and follow the mechanism.Route 2 · Chapters 4–7
Reproduce the example
Audit the numbers one step at a time.Route 3 · Chapters 8–13
Check assumptions
See where design and calibration matter.Route 4 · Chapters 14–18
Explore applications
Connect the method to a concrete investigation.Route 5 · Chapters 19–22
Study and verify
Practise, inspect failure cases and read further.
Full chapter index · Practice and checking · Mathematics Learning Hub
Analytical methods often become less precise when they measure very small concentrations. The Horwitz equation models that broad empirical relationship, and the HorRat compares observed relative precision with the equation's predicted value. It gives students a striking lesson: the same percentage variation is not equally surprising at every concentration.
Mathematics is important here because powers and ratios create a scale-aware benchmark. HorRat is a rough diagnostic for analytical-method precision, not a universal law, a measure of trueness or permission to skip method-specific requirements.
Full chapter index
Chapters 1–3 · Understand the question
Chapters 4–7 · Reproduce the example
Chapters 8–13 · Check assumptions
Chapters 14–18 · Explore applications
Chapters 19–22 · Study and verify
An interlaboratory relative standard deviation of 6% might look large at a major-component concentration and modest at a trace concentration. A useful comparison therefore needs to account for concentration level.
The Horwitz relationship was developed from interlaboratory analytical data. AOAC guidance expresses the predicted reproducibility relative standard deviation as PRSD(R) = 2C^-0.15, where C is the concentration as a mass fraction and the result is expressed as a percentage.
CHAPTER 2 OF 22 · Understand the question
2. Understand mass fraction before calculating
A concentration of 0.1% is not entered as 0.1. As a mass fraction it is 0.001, or 10^-3. This conversion is the most important checkpoint in the calculation.
Percent means per hundred. Divide a percentage by 100 to obtain the fraction. Keep the exponent visible so an order-of-magnitude error is easy to catch.
Let C = 0.001 = 10^-3. Then C^-0.15 = (10^-3)^-0.15 = 10^0.45. The value of 10^0.45 is about 2.818.
Therefore PRSD(R) = 2 × 2.818 = 5.636%. The equation predicts a reproducibility relative standard deviation of roughly 5.64% for this concentration under the guidance's empirical model.
| Step | Calculation | Result |
|---|---|---|
| Mass fraction | 0.1 / 100 | 0.001 |
| Power | (10⁻³)⁻⁰·¹⁵ | 10⁰·⁴⁵ ≈ 2.818 |
| PRSD(R) | 2 × 2.818 | 5.636% |
Suppose an interlaboratory study observes RSD(R) = 7.0%. HorRat(R) = observed RSD(R) divided by PRSD(R) = 7.0/5.636 = about 1.242.
AOAC's published guidance describes an original interlaboratory acceptability range of 0.5 to 2.0 for HorRat(R), while stressing that the statistic is rough and must be interpreted with care. This illustrative value lies inside that range.
| Observed RSD(R) | Predicted RSD(R) | HorRat(R) |
|---|---|---|
| 7.0% | 5.636% | 1.242 |
A HorRat near one means the observed relative reproducibility is close to the empirical prediction. A high value means observed variation is larger than predicted. A very low value may seem excellent but can also prompt checks for hidden averaging, prior knowledge, rounding or an unusually restricted study.
The ratio is dimensionless because percentage is divided by percentage. It compares precision with a concentration-dependent reference.
CHAPTER 6 OF 22 · Reproduce the example
6. Repeatability and reproducibility versions
AOAC distinguishes HorRat(R), using observed reproducibility RSD, from HorRat(r), using observed repeatability RSD. The denominator remains the predicted reproducibility value. The guidance gives different indicative ranges for the two contexts.
Do not label an within-laboratory study as interlaboratory reproducibility. The study design owns the symbol.
At C = 10^-6, PRSD(R) = 2 × (10^-6)^-0.15 = 2 × 10^0.9. Since 10^0.9 is about 7.943, the prediction is about 15.887%.
The predicted relative variation is larger at the lower concentration. This does not mean low-concentration work is careless; it reflects the empirical challenge of trace measurement.
Take base-10 logarithms: log10(PRSD/2) = -0.15 log10(C). This is a straight line in log-log coordinates with slope -0.15. A multiplicative change in concentration becomes an additive movement on the log axis.
Students can use this form to check exponents and understand power laws. A factor of ten decrease in C multiplies the prediction by 10^0.15, about 1.413.
C must be a dimensionless mass fraction for the stated equation. Milligrams per kilogram, micrograms per gram and percentages must be converted consistently. Matrix composition and analyte definition still matter.
A number copied without its unit or basis cannot be repaired by a formula. Record the conversion beside the calculation.
RSD describes dispersion relative to a mean. A method can have an attractive HorRat yet remain biased against a reference. Recovery, selectivity, calibration and other validation evidence answer different questions.
Do not translate precision into accuracy or safety. A method decision needs the full intended-use requirements.
AOAC guidance warns that the relationship can overpredict variability at concentration extremes, including above about 0.1 mass fraction and below about 10^-8. Method-specific performance requirements can supersede a generic expectation.
This is a valuable lesson about models: even a celebrated equation has a domain of useful application.
An observed reproducibility RSD should come from a suitable interlaboratory design with independent laboratories, comparable material and a clear procedure. Too few laboratories produce an unstable estimate. Selective exclusions can make precision look better than it is.
The value also depends on material level, homogeneity, stability and how replicates are handled.
For a high HorRat, examine uncontrolled time, temperature, mass, volume, concentration, extraction, transfer and instrumental conditions. For a low value, check prior knowledge, undisclosed averaging, rounding and study restriction.
Use the ratio to ask questions. Do not use it as a substitute for reading the data.
Create a column of mass fractions from 10^-1 to 10^-8. Compute PRSD(R) for each and make a log-log plot. Verify that each tenfold concentration decrease raises the prediction by about 1.413.
Then choose several observed RSD values and compute HorRat. Label every row with concentration basis and study type.
Validate that C is positive and dimensionless. Compute on a log scale when numerical ranges are extreme: log(PRSD) = log(2) – 0.15log(C). Return both predicted RSD and HorRat with units and design labels.
Add tests for 0.001 and 10^-6 using the worked values. A program should reject C = 0 and warn about the guidance's extreme ranges.
The Youden plot article shows labelled laboratory patterns for two runs. The mountain plot guide shows a folded distribution of differences. HorRat instead compares an observed relative precision with a concentration-based empirical prediction.
Each tool preserves a different piece of the evidence.
Students can practise percentages, scientific notation, indices, logarithms and ratios in one coherent problem. Ask them to explain why 0.1% becomes 0.001 and why the final HorRat has no unit.
Parents can encourage estimation before calculator use. If concentration decreases, the predicted RSD should increase; a result moving the other way signals an exponent or conversion error.
Chemistry, food science, environmental analysis, pharmaceuticals and laboratory quality use concentration-dependent precision reasoning. Mathematics helps teams set expectations and investigate methods. It does not replace professional standards or guarantee a career outcome.
Continue through the Mathematics Learning Hub and the modelling discussion in How Mathematics Works.
Calculate PRSD(R) for C = 10^-2 and 10^-4. Confirm that the two-decade decrease multiplies the prediction by 10^0.30. If observed RSD(R) is 4% at C = 0.01, compute HorRat. Then repeat after incorrectly entering 1 instead of 0.01 and explain the conversion error.
Finally, compare two methods at different concentrations without using HorRat, notice the unfairness, and then use the ratio.
### Is HorRat a test of accuracy? No. It is a ratio about relative precision compared with an empirical prediction.
### Why is C dimensionless? The power-law equation requires a consistent concentration fraction. Units must be converted before applying the exponent.
### Is a HorRat below 0.5 always excellent? No. It can prompt checks for hidden averaging, rounding or a restricted study, and the governing method criteria still apply.
### Can I use the equation at every concentration? AOAC guidance identifies limitations at extremes. Use method-specific requirements and current professional judgement.
Use AOAC INTERNATIONAL's Guidelines for Standard Method Performance Requirements, especially its definitions and calculations for PRSD and HorRat, ranges and stated limitations. The worked concentrations and observations here are original teaching examples.
For broader mathematical context, compare interlaboratory variation in Youden plots and fixed tolerances in coverage probability.
HorRat is a compact power-law lesson with practical reach. Convert concentration correctly, predict relative precision, compare observed variation and respect the model's limits. The mathematics is most useful when it produces a better technical question rather than an automatic verdict.
Extended practice laboratory
Practice case 1: Convert 1 percent to fraction
Setup: Use the article’s Horwitz-and-HorRat example and change one declared input for case 1. Task: Recalculate from the raw values, preserve units and state which design condition changed.
Check: Check arithmetic against the invariant described in the chapter, then separate the numerical finding from the practical claim. Rebuild the relevant counts before interpreting the coefficient, preserve the stated design, and write one sentence separating the numerical result from the claim it can support. Then vary one defensible assumption, calculate again, and explain whether the conclusion changes because of new data, a new reference model, or a new decision purpose. Finish by naming the unit, tolerance, and evidence needed for transfer. Record one graph or table that would expose failure hidden by the summary.
Practice case 2: Convert 0.1 percent to fraction
Setup: Use the article’s Horwitz-and-HorRat example and change one declared input for case 2. Task: Recalculate from the raw values, preserve units and state which design condition changed.
Check: Check arithmetic against the invariant described in the chapter, then separate the numerical finding from the practical claim. Rebuild the relevant counts before interpreting the coefficient, preserve the stated design, and write one sentence separating the numerical result from the claim it can support. Then vary one defensible assumption, calculate again, and explain whether the conclusion changes because of new data, a new reference model, or a new decision purpose. Finish by naming the unit, tolerance, and evidence needed for transfer. Record one graph or table that would expose failure hidden by the summary.
Practice case 3: Decrease concentration tenfold
Setup: Use the article’s Horwitz-and-HorRat example and change one declared input for case 3. Task: Recalculate from the raw values, preserve units and state which design condition changed.
Check: Check arithmetic against the invariant described in the chapter, then separate the numerical finding from the practical claim. Rebuild the relevant counts before interpreting the coefficient, preserve the stated design, and write one sentence separating the numerical result from the claim it can support. Then vary one defensible assumption, calculate again, and explain whether the conclusion changes because of new data, a new reference model, or a new decision purpose. Finish by naming the unit, tolerance, and evidence needed for transfer. Record one graph or table that would expose failure hidden by the summary.
Practice case 4: Increase concentration tenfold
Setup: Use the article’s Horwitz-and-HorRat example and change one declared input for case 4. Task: Recalculate from the raw values, preserve units and state which design condition changed.
Check: Check arithmetic against the invariant described in the chapter, then separate the numerical finding from the practical claim. Rebuild the relevant counts before interpreting the coefficient, preserve the stated design, and write one sentence separating the numerical result from the claim it can support. Then vary one defensible assumption, calculate again, and explain whether the conclusion changes because of new data, a new reference model, or a new decision purpose. Finish by naming the unit, tolerance, and evidence needed for transfer. Record one graph or table that would expose failure hidden by the summary.
Practice case 5: Use scientific notation
Setup: Use the article’s Horwitz-and-HorRat example and change one declared input for case 5. Task: Recalculate from the raw values, preserve units and state which design condition changed.
Check: Check arithmetic against the invariant described in the chapter, then separate the numerical finding from the practical claim. Rebuild the relevant counts before interpreting the coefficient, preserve the stated design, and write one sentence separating the numerical result from the claim it can support. Then vary one defensible assumption, calculate again, and explain whether the conclusion changes because of new data, a new reference model, or a new decision purpose. Finish by naming the unit, tolerance, and evidence needed for transfer. Record one graph or table that would expose failure hidden by the summary.
Practice case 6: Check the negative exponent
Setup: Use the article’s Horwitz-and-HorRat example and change one declared input for case 6. Task: Recalculate from the raw values, preserve units and state which design condition changed.
Check: Check arithmetic against the invariant described in the chapter, then separate the numerical finding from the practical claim. Rebuild the relevant counts before interpreting the coefficient, preserve the stated design, and write one sentence separating the numerical result from the claim it can support. Then vary one defensible assumption, calculate again, and explain whether the conclusion changes because of new data, a new reference model, or a new decision purpose. Finish by naming the unit, tolerance, and evidence needed for transfer. Record one graph or table that would expose failure hidden by the summary.
Practice case 7: Calculate at one part per million
Setup: Use the article’s Horwitz-and-HorRat example and change one declared input for case 7. Task: Recalculate from the raw values, preserve units and state which design condition changed.
Check: Check arithmetic against the invariant described in the chapter, then separate the numerical finding from the practical claim. Rebuild the relevant counts before interpreting the coefficient, preserve the stated design, and write one sentence separating the numerical result from the claim it can support. Then vary one defensible assumption, calculate again, and explain whether the conclusion changes because of new data, a new reference model, or a new decision purpose. Finish by naming the unit, tolerance, and evidence needed for transfer. Record one graph or table that would expose failure hidden by the summary.
Practice case 8: Compare observed RSD
Setup: Use the article’s Horwitz-and-HorRat example and change one declared input for case 8. Task: Recalculate from the raw values, preserve units and state which design condition changed.
Check: Check arithmetic against the invariant described in the chapter, then separate the numerical finding from the practical claim. Rebuild the relevant counts before interpreting the coefficient, preserve the stated design, and write one sentence separating the numerical result from the claim it can support. Then vary one defensible assumption, calculate again, and explain whether the conclusion changes because of new data, a new reference model, or a new decision purpose. Finish by naming the unit, tolerance, and evidence needed for transfer. Record one graph or table that would expose failure hidden by the summary.
Practice case 9: Double observed RSD
Setup: Use the article’s Horwitz-and-HorRat example and change one declared input for case 9. Task: Recalculate from the raw values, preserve units and state which design condition changed.
Check: Check arithmetic against the invariant described in the chapter, then separate the numerical finding from the practical claim. Rebuild the relevant counts before interpreting the coefficient, preserve the stated design, and write one sentence separating the numerical result from the claim it can support. Then vary one defensible assumption, calculate again, and explain whether the conclusion changes because of new data, a new reference model, or a new decision purpose. Finish by naming the unit, tolerance, and evidence needed for transfer. Record one graph or table that would expose failure hidden by the summary.
Practice case 10: Halve observed RSD
Setup: Use the article’s Horwitz-and-HorRat example and change one declared input for case 10. Task: Recalculate from the raw values, preserve units and state which design condition changed.
Check: Check arithmetic against the invariant described in the chapter, then separate the numerical finding from the practical claim. Rebuild the relevant counts before interpreting the coefficient, preserve the stated design, and write one sentence separating the numerical result from the claim it can support. Then vary one defensible assumption, calculate again, and explain whether the conclusion changes because of new data, a new reference model, or a new decision purpose. Finish by naming the unit, tolerance, and evidence needed for transfer. Record one graph or table that would expose failure hidden by the summary.
Practice case 11: Distinguish HorRat R and r
Setup: Use the article’s Horwitz-and-HorRat example and change one declared input for case 11. Task: Recalculate from the raw values, preserve units and state which design condition changed.
Check: Check arithmetic against the invariant described in the chapter, then separate the numerical finding from the practical claim. Rebuild the relevant counts before interpreting the coefficient, preserve the stated design, and write one sentence separating the numerical result from the claim it can support. Then vary one defensible assumption, calculate again, and explain whether the conclusion changes because of new data, a new reference model, or a new decision purpose. Finish by naming the unit, tolerance, and evidence needed for transfer. Record one graph or table that would expose failure hidden by the summary.
Practice case 12: Test a low ratio
Setup: Use the article’s Horwitz-and-HorRat example and change one declared input for case 12. Task: Recalculate from the raw values, preserve units and state which design condition changed.
Check: Check arithmetic against the invariant described in the chapter, then separate the numerical finding from the practical claim. Rebuild the relevant counts before interpreting the coefficient, preserve the stated design, and write one sentence separating the numerical result from the claim it can support. Then vary one defensible assumption, calculate again, and explain whether the conclusion changes because of new data, a new reference model, or a new decision purpose. Finish by naming the unit, tolerance, and evidence needed for transfer. Record one graph or table that would expose failure hidden by the summary.
Practice case 13: Test a high ratio
Setup: Use the article’s Horwitz-and-HorRat example and change one declared input for case 13. Task: Recalculate from the raw values, preserve units and state which design condition changed.
Check: Check arithmetic against the invariant described in the chapter, then separate the numerical finding from the practical claim. Rebuild the relevant counts before interpreting the coefficient, preserve the stated design, and write one sentence separating the numerical result from the claim it can support. Then vary one defensible assumption, calculate again, and explain whether the conclusion changes because of new data, a new reference model, or a new decision purpose. Finish by naming the unit, tolerance, and evidence needed for transfer. Record one graph or table that would expose failure hidden by the summary.
Practice case 14: Inspect a concentration extreme
Setup: Use the article’s Horwitz-and-HorRat example and change one declared input for case 14. Task: Recalculate from the raw values, preserve units and state which design condition changed.
Check: Check arithmetic against the invariant described in the chapter, then separate the numerical finding from the practical claim. Rebuild the relevant counts before interpreting the coefficient, preserve the stated design, and write one sentence separating the numerical result from the claim it can support. Then vary one defensible assumption, calculate again, and explain whether the conclusion changes because of new data, a new reference model, or a new decision purpose. Finish by naming the unit, tolerance, and evidence needed for transfer. Record one graph or table that would expose failure hidden by the summary.
Practice case 15: Plot on log axes
Setup: Use the article’s Horwitz-and-HorRat example and change one declared input for case 15. Task: Recalculate from the raw values, preserve units and state which design condition changed.
Check: Check arithmetic against the invariant described in the chapter, then separate the numerical finding from the practical claim. Rebuild the relevant counts before interpreting the coefficient, preserve the stated design, and write one sentence separating the numerical result from the claim it can support. Then vary one defensible assumption, calculate again, and explain whether the conclusion changes because of new data, a new reference model, or a new decision purpose. Finish by naming the unit, tolerance, and evidence needed for transfer. Record one graph or table that would expose failure hidden by the summary.
Practice case 16: Recover the slope
Setup: Use the article’s Horwitz-and-HorRat example and change one declared input for case 16. Task: Recalculate from the raw values, preserve units and state which design condition changed.
Check: Check arithmetic against the invariant described in the chapter, then separate the numerical finding from the practical claim. Rebuild the relevant counts before interpreting the coefficient, preserve the stated design, and write one sentence separating the numerical result from the claim it can support. Then vary one defensible assumption, calculate again, and explain whether the conclusion changes because of new data, a new reference model, or a new decision purpose. Finish by naming the unit, tolerance, and evidence needed for transfer. Record one graph or table that would expose failure hidden by the summary.
Practice case 17: Check dimensionlessness
Setup: Use the article’s Horwitz-and-HorRat example and change one declared input for case 17. Task: Recalculate from the raw values, preserve units and state which design condition changed.
Check: Check arithmetic against the invariant described in the chapter, then separate the numerical finding from the practical claim. Rebuild the relevant counts before interpreting the coefficient, preserve the stated design, and write one sentence separating the numerical result from the claim it can support. Then vary one defensible assumption, calculate again, and explain whether the conclusion changes because of new data, a new reference model, or a new decision purpose. Finish by naming the unit, tolerance, and evidence needed for transfer. Record one graph or table that would expose failure hidden by the summary.
Practice case 18: Reject zero concentration
Setup: Use the article’s Horwitz-and-HorRat example and change one declared input for case 18. Task: Recalculate from the raw values, preserve units and state which design condition changed.
Check: Check arithmetic against the invariant described in the chapter, then separate the numerical finding from the practical claim. Rebuild the relevant counts before interpreting the coefficient, preserve the stated design, and write one sentence separating the numerical result from the claim it can support. Then vary one defensible assumption, calculate again, and explain whether the conclusion changes because of new data, a new reference model, or a new decision purpose. Finish by naming the unit, tolerance, and evidence needed for transfer. Record one graph or table that would expose failure hidden by the summary.
Practice case 19: Compare unlike concentrations
Setup: Use the article’s Horwitz-and-HorRat example and change one declared input for case 19. Task: Recalculate from the raw values, preserve units and state which design condition changed.
Check: Check arithmetic against the invariant described in the chapter, then separate the numerical finding from the practical claim. Rebuild the relevant counts before interpreting the coefficient, preserve the stated design, and write one sentence separating the numerical result from the claim it can support. Then vary one defensible assumption, calculate again, and explain whether the conclusion changes because of new data, a new reference model, or a new decision purpose. Finish by naming the unit, tolerance, and evidence needed for transfer. Record one graph or table that would expose failure hidden by the summary.
Practice case 20: Audit an RSD source
Setup: Use the article’s Horwitz-and-HorRat example and change one declared input for case 20. Task: Recalculate from the raw values, preserve units and state which design condition changed.
Check: Check arithmetic against the invariant described in the chapter, then separate the numerical finding from the practical claim. Rebuild the relevant counts before interpreting the coefficient, preserve the stated design, and write one sentence separating the numerical result from the claim it can support. Then vary one defensible assumption, calculate again, and explain whether the conclusion changes because of new data, a new reference model, or a new decision purpose. Finish by naming the unit, tolerance, and evidence needed for transfer. Record one graph or table that would expose failure hidden by the summary.
Practice case 21: Inspect hidden averaging
Setup: Use the article’s Horwitz-and-HorRat example and change one declared input for case 21. Task: Recalculate from the raw values, preserve units and state which design condition changed.
Check: Check arithmetic against the invariant described in the chapter, then separate the numerical finding from the practical claim. Rebuild the relevant counts before interpreting the coefficient, preserve the stated design, and write one sentence separating the numerical result from the claim it can support. Then vary one defensible assumption, calculate again, and explain whether the conclusion changes because of new data, a new reference model, or a new decision purpose. Finish by naming the unit, tolerance, and evidence needed for transfer. Record one graph or table that would expose failure hidden by the summary.
Practice case 22: Check method-specific criteria
Setup: Use the article’s Horwitz-and-HorRat example and change one declared input for case 22. Task: Recalculate from the raw values, preserve units and state which design condition changed.
Check: Check arithmetic against the invariant described in the chapter, then separate the numerical finding from the practical claim. Rebuild the relevant counts before interpreting the coefficient, preserve the stated design, and write one sentence separating the numerical result from the claim it can support. Then vary one defensible assumption, calculate again, and explain whether the conclusion changes because of new data, a new reference model, or a new decision purpose. Finish by naming the unit, tolerance, and evidence needed for transfer. Record one graph or table that would expose failure hidden by the summary.
Practice case 23: State model limits
Setup: Use the article’s Horwitz-and-HorRat example and change one declared input for case 23. Task: Recalculate from the raw values, preserve units and state which design condition changed.
Check: Check arithmetic against the invariant described in the chapter, then separate the numerical finding from the practical claim. Rebuild the relevant counts before interpreting the coefficient, preserve the stated design, and write one sentence separating the numerical result from the claim it can support. Then vary one defensible assumption, calculate again, and explain whether the conclusion changes because of new data, a new reference model, or a new decision purpose. Finish by naming the unit, tolerance, and evidence needed for transfer. Record one graph or table that would expose failure hidden by the summary.
Practice case 24: Write a precision-only conclusion
Setup: Use the article’s Horwitz-and-HorRat example and change one declared input for case 24. Task: Recalculate from the raw values, preserve units and state which design condition changed.
Check: Check arithmetic against the invariant described in the chapter, then separate the numerical finding from the practical claim. Rebuild the relevant counts before interpreting the coefficient, preserve the stated design, and write one sentence separating the numerical result from the claim it can support. Then vary one defensible assumption, calculate again, and explain whether the conclusion changes because of new data, a new reference model, or a new decision purpose. Finish by naming the unit, tolerance, and evidence needed for transfer. Record one graph or table that would expose failure hidden by the summary.