Why peptide testing should be double-blind

Medicine learned a long time ago that knowing changes results. In a double-blind trial, neither the patient nor the researcher knows who received the treatment, because expectation and incentive shape outcomes even when everyone involved is honest. Product testing has the same problem, for slightly different reasons, and the same fix works.

To be clear at the outset: none of this assumes that laboratories or sellers are dishonest. Blinding removes pressure and opportunity, so that nobody has to be trusted on exactly the points where trust is hardest to check.

What double-blind means in product testing

It has two halves.

The laboratory is blind to the source. Samples arrive identified only by a code, with no seller, brand or batch information, and with product labels removed or covered.

The seller is blind to the laboratory. The seller doesn't choose which laboratory tests its product, and doesn't learn which one did until it no longer matters.

Each half closes a different door, and both are worth walking through.

Why the laboratory shouldn't know whose sample it is

Laboratories are businesses, and sellers are their clients. A laboratory that knows a sample came from a large, regular client is under pressure whether or not anyone applies it, and whether or not it gives in.

That matters because analytical work is full of small judgment calls. How to integrate a small shoulder next to the main peak. Whether an odd result deserves a re-run. What to make of a chromatogram that doesn't look like the last one. None of those calls is dishonest either way, and knowing whose sample is on the instrument can lean on them without anyone meaning it to.

Blinding takes the question off the table. It also protects the laboratory: one that could not have known whose sample it was cannot be accused of favoring anyone, and that is a defense worth having.

Why the seller shouldn't know which laboratory

Three reasons.

Laboratory shopping. Different laboratories use different methods, and methods can give different numbers for the same material. A seller who chooses the laboratory can choose the one whose method tends to flatter its product. Our piece on what a purity number does and doesn't tell you explains why the numbers differ.

The golden sample. A seller who knows exactly where a sample is going, and when, can send one prepared for the occasion rather than what actually ships.

Relationships. A seller that knows its laboratory will, over time, build a relationship with it. That's normal in business. It is also precisely what independent testing exists to rule out.

Who chooses the sample matters just as much

Blinding means little if the seller hand-picks the vial. The sample should come from the declared lot that is actually being sold, committed to before any result is known, and the record should be checked later against product bought anonymously at retail. That last step is the one that catches a sample that didn't represent the shelf.

Two laboratories, not one

A single laboratory can make an error and nobody would ever know. When two independent laboratories, neither told about the other, test the same lot, their agreement means something and their disagreement is information. Error isn't dishonesty. A second laboratory is how honest error gets caught.

What blinding doesn't do

It doesn't make fraud impossible. A determined enough party can try to work around any procedure. What blinding does is make casual gaming useless and deliberate gaming expensive and risky. It works best alongside a permanent public record and independent spot checks, so that anything that does slip through can still be found later.

Questions to ask about any test

• Did the seller choose the laboratory?
• Did the laboratory know whose product it was testing?
• Who chose the sample, and was it from the batch being sold?
• Did more than one laboratory test it?
• Are all the results published, including any that disagree?