A property owner rings me after receiving a single-sample meth test result of 0.3 µg/100cm² — below the 0.5 µg/100cm² Australian guideline. They believe their property is clear. But that single result, taken from one surface in one room, tells them almost nothing about the other 200 square metres of the property. This is the problem of statistical confidence in contamination assessment, and misunderstanding it costs Australian property owners millions of dollars every year.

What Confidence Intervals Mean in Plain Language

Imagine you want to know the average temperature in a city across a year. If you check the temperature once — say, on a Tuesday in July — you will get a number. But that single measurement tells you very little about the average annual temperature. It might be an unusually cold day, or an unusually warm one. You need many measurements across different days, months, and seasons before you can confidently state the average temperature within a meaningful range.

Contamination assessment works the same way. Each surface wipe sample measures the methamphetamine level at one specific point on one specific surface. That 100 cm² sample area represents a tiny fraction of the total surface area of a room, let alone an entire property. A three-bedroom house has roughly 400-600 m² of sampable surface area (walls, ceilings, floors). A single 100 cm² sample represents approximately 0.002% of that surface area.

A confidence interval expresses the range within which the true average contamination level is likely to fall, given the samples you have collected. A 95% confidence interval means that if you repeated the sampling exercise 100 times, the true mean would fall within your calculated range 95 times out of 100. The more samples you collect, the narrower that interval becomes — and the more confident you can be in your assessment.

Why a Single Sample Cannot Characterise a Whole Property

This is perhaps the most important concept in contamination assessment, and the one most consistently ignored by budget testing companies.

Methamphetamine contamination is not uniformly distributed across surfaces. Even in a property where someone smoked methamphetamine regularly, levels vary dramatically between rooms, between surfaces within the same room, and even between different positions on the same wall. Factors affecting this variation include proximity to where smoking occurred, air circulation patterns, surface material (porous vs non-porous), surface orientation (horizontal surfaces accumulate more than vertical), and whether the surface has been cleaned.

Consider a property where someone smoked methamphetamine primarily in the lounge room. The lounge room ceiling might show 3.2 µg/100cm². The lounge room wall near the smoking location might show 1.8 µg/100cm². The hallway outside might show 0.6 µg/100cm². A bedroom at the far end of the house might show 0.2 µg/100cm². The bathroom might show 0.1 µg/100cm².

If the single sample was taken from the bathroom, the property appears clean. If it was taken from the lounge room ceiling, it appears heavily contaminated. Both conclusions from a single sample would be misleading. Only by sampling multiple rooms and surfaces can you build an accurate picture of the contamination distribution and determine which areas require remediation.

Minimum Sample Numbers for Statistical Power

The question every property owner asks is: how many samples do I actually need? The answer depends on the property size, the purpose of the assessment, and the level of confidence required.

For a basic screening assessment of a typical three-bedroom house, I recommend a minimum of one sample per functional room: each bedroom, kitchen, lounge, bathroom, laundry, hallway, and garage. That gives you 8-10 samples. This provides a reasonable spatial overview but limited statistical power within individual rooms.

For a regulatory-grade assessment — the kind required for source determination, post-remediation verification, or legal proceedings — the sampling plan should aim for 95% confidence that at least 95% of the surface area has been characterised. This is the 95/95 standard referenced in AS/NZS guidance and consistent with NIOSH 9111 sampling protocols. For a three-bedroom house, this typically means 12-20 samples depending on room sizes and layout.

For larger commercial properties, apartment buildings, or properties with suspected manufacturing, sample numbers increase proportionally. I have conducted assessments requiring 50+ samples for large commercial premises or multi-unit residential buildings.

How Variance Affects Conclusions

Variance — the degree to which individual sample results differ from each other — has a profound effect on the reliability of your conclusions. Low variance (all samples returning similar values) narrows confidence intervals. High variance (samples returning widely different values) widens them.

In a property with use-only contamination, I typically see moderate variance. Living areas cluster around similar levels (say, 0.8-2.0 µg/100cm²) while non-living areas show lower levels (0.1-0.4 µg/100cm²). The variance is explicable and follows logical patterns.

In a property with manufacturing contamination, variance is often extreme. The cook area might show 85 µg/100cm² while a distant bedroom shows 0.8 µg/100cm² — a 100-fold difference. This high variance means you need more samples to characterise the property with the same level of confidence. Ironically, the most contaminated properties — where getting the answer right matters most — require the most extensive sampling.

If a assessor presents results with high between-sample variance but does not acknowledge or discuss the statistical implications, they may not understand the data they are presenting. This is a red flag.

Type I and Type II Errors in Contamination Assessment

Statistical decision-making involves two kinds of errors, and both have serious real-world consequences in contamination assessment.

Type I Error (False Positive)

A Type I error occurs when you conclude a property is contaminated when it actually is not, or when you conclude it is more contaminated than it actually is. In contamination assessment, this leads to unnecessary remediation, needless expense, failed property transactions, and unwarranted anxiety for occupants.

Type I errors can occur when a sample happens to be taken from a localised deposit (dust, residue from an object placed on the surface) that is not representative of general contamination levels. They can also occur from laboratory analytical uncertainty — all measurements have some degree of uncertainty, and results near the 0.5 µg/100cm² guideline may fluctuate above or below the threshold due to analytical variation alone.

Type II Error (False Negative)

A Type II error occurs when you conclude a property is clean when it is actually contaminated. This is the more dangerous error because it leaves occupants exposed to potentially harmful contamination and may result in contaminated properties being sold or rented without disclosure.

Type II errors are particularly likely with insufficient sample numbers. If contamination is concentrated in specific areas (as with manufacturing) and your limited samples happen to miss those areas, you will incorrectly conclude the property is clean. This is not an unlikely scenario — it is the predictable outcome of under-sampling.

The Cost of Under-Sampling vs Over-Sampling

Property owners understandably want to minimise testing costs. But there is a rational economic analysis that supports adequate sampling.

Each additional sample costs approximately $80-$150 in laboratory analysis (through independent NATA-accredited laboratories), plus the assessor’s time to collect it. For a typical assessment, the difference between a 4-sample budget test and a 12-sample professional assessment is perhaps $800-$1,500 in additional analytical costs.

Now consider the decision that rests on those results. If you are purchasing a property worth $600,000, spending an additional $1,000 on adequate sampling to avoid buying a contaminated property represents 0.17% of the purchase price. If that additional sampling reveals contamination that the budget test missed, you have potentially avoided a $50,000-$200,000 remediation liability.

Over-sampling — collecting more samples than statistically necessary — costs money but does not create risk. Under-sampling creates substantial risk. The rational choice is clear.

Regulatory Expectations for Sample Numbers

Australian state guidelines vary in their specificity about sampling requirements, but the general expectation is consistent: the sampling plan must be sufficient to characterise the property with adequate confidence.

The NIOSH 9111 method, which forms the basis for surface sampling methodology in Australian methamphetamine testing, specifies the wipe sampling technique but allows the assessor professional discretion in determining the number and location of samples. This places the responsibility on the assessor to design a sampling plan with adequate statistical power — and it means the assessor’s qualifications and experience directly affect the quality of the assessment.

For post-remediation verification (clearance testing), regulatory expectations are typically more prescriptive. Most state guidelines require sufficient samples to demonstrate that all previously contaminated areas now fall below the guideline value, with at least one sample from each room and additional samples from former hotspot areas.

How Professional Assessors Determine Sample Plans

An experienced forensic assessor designs the sampling plan before arriving at the property, then adjusts it based on site observations. The process involves several steps.

First, I review the property layout — floor plans, number of rooms, total area, and any available history (police reports, previous test results, tenancy information). This determines the baseline number of samples needed for spatial coverage.

Second, I assess the purpose of the testing. A pre-purchase screening requires fewer samples than a post-remediation verification or a source determination investigation. Legal or insurance matters require defensible sample numbers that will withstand scrutiny.

Third, I conduct a visual inspection upon arrival and adjust the plan based on observations. Evidence suggesting manufacturing (staining, modifications, chemical damage) increases sample numbers. Multiple suspected use areas may require additional coverage.

Finally, I consider the sampling strategy — whether representative sampling, hotspot sampling, or a combination is appropriate for the specific situation. Each strategy has different sample number implications.

Real Examples of Statistics Affecting Outcomes

Let me share two anonymised cases from my practice that illustrate why statistical considerations matter.

Case 1: The false clearance. A property owner had a single-sample test performed by a budget testing company. The result was 0.35 µg/100cm² — below the guideline. The property was sold as clean. The purchaser commissioned a comprehensive 14-sample assessment. Results ranged from 0.2 to 4.7 µg/100cm², with three rooms exceeding the guideline. The original single sample had been taken from the lowest-contamination room. The sale was disputed and the original owner faced remediation liability.

Case 2: The unnecessary remediation. A landlord received a two-sample test showing 0.7 and 1.2 µg/100cm². Both samples exceeded the guideline. A remediation company quoted $45,000. Before proceeding, the landlord sought a second opinion. My 12-sample assessment revealed that 10 of 12 samples were below 0.5 µg/100cm². The two elevated results were from localised deposits near where smoking had occurred — one on a shelf surface and one on a window frame. Targeted cleaning of those specific areas cost $3,500. The comprehensive sampling saved the landlord over $41,000.

In both cases, the outcome depended entirely on the number and placement of samples. Statistics is not an abstract academic exercise in contamination assessment — it directly determines financial outcomes.

If you need a contamination assessment that is statistically valid, forensically defensible, and interpreted by a qualified professional, contact Test Australia. We design every sampling plan to provide the statistical confidence your situation requires.