Circle of Blue reported on October 23, 2026 that a new study suggests PFAS contamination in drinking water may be underestimated, in a piece by Alissa de Chassey. The summary of that report turns on a measurement point: "When PFAS concentrations in drinking water fall below laboratory reporting limits" the number that reaches a file depends on laboratory convention as much as on the water. Readers can review the original in Circle of Blue's report on the study.
Whatever a study concludes about how much contamination goes unseen, any trend built from older and newer results inherits a second question: what was the reporting limit on each sample date? A reporting limit is the lowest concentration a laboratory will report as a quantified value under its method, calibration, and quality controls. It is not the best detection capability an instrument can reach on a good day, and it is not a fixed property of the water.
How a Lower Limit Bends a Trend
When a laboratory adopts a more sensitive method, changes instruments, or moves testing to a different facility, the reporting limit can fall. Results that once read as non-detect now appear as low positive values. Plot both on one axis and the line climbs, even if the water has not changed. The rise is partly real detection and partly the ruler. A reviewer who cannot separate the two has no basis for calling the trend a change in the source.
Two coding errors often travel together. One treats non-detects as zeros, which pulls early averages down. The other treats the limit itself as a measured value, which pushes them up. Either convention can be defensible if it is written down and applied the same way across the record. Neither is defensible when the file is silent about which one was used.
What to Pull Before Comparing Years
Before any multi-year comparison, collect the reporting limit in force on each sample date, not the limit printed in the laboratory's current procedure manual. Record the method identifier and whether it changed during the period. Note how non-detects were coded in the dataset, whether as zero, as the limit, as half the limit, or as blank. Check the compound list as well. A newer panel can detect a compound the older panel never measured, which looks like a new contamination event but may simply be a longer menu.
Write this history into the trend file before the chart is drawn. The next reader, whether a board member, a consultant, or a regulator's reviewer, should not have to reconstruct it from lab correspondence that may no longer exist.
Where Home Testing Fits
Households that run their own tests face the same ruler problem. A private report with a reporting limit above a utility's can show a clean result for a compound that was present at a level the report could not resolve. A household comparing treatment options can check coverage for home systems and appliances to see what a device is certified to reduce, then test that claim against a result whose limit is printed on the report. A certification names the compounds a device was tested against. A test result names the floor below which it cannot say anything. Both need to be read together.
What the Coverage Does and Does Not Establish
The Circle of Blue summary frames the study around results that fall below reporting limits, which is a claim about how much a measurement can see, not a claim about any particular utility. Readers should not treat the headline as a revision to any compliance value. Compliance values come from the methods and dates in each system's own records. The study's authors may have accounted for all of this, but the summary does not say, and this piece does not assume it.