Methods: how to test a quantum biology claim
Part 4 has met claims that are established, debated, and speculative. What separates them is not how exciting they are but whether they have passed a short list of tests, each designed so that a quantum explanation and a classical one predict different outcomes. This chapter collects the toolkit, the controls that keep it honest, and the statistics that decide whether a small effect is real. Then it puts each claim from Part 4 on trial.
On the plate: a tribunal for quantum-biology claims.
The toolkit
Isotope substitution: change mass (H→D) to test tunnelling, or nuclear spin (¹²C→¹³C, ¹²⁹Xe→¹³²Xe) to test spin chemistry, while keeping chemistry nearly fixed. Temperature dependence: tunnelling and thermally activated processes respond differently. Magnetic fields and radio frequencies: spin-dependent processes respond to weak static fields and to RF near the Larmor frequency; classical chemistry does not. Ultrafast and 2D spectroscopy: measure coherence lifetimes directly, and separate electronic from vibrational origins. Mutagenesis and structure: change specific residues and see whether the effect tracks the proposed mechanism. Each test has a known non-quantum explanation that must be excluded: impurities for isotope effects, heating for RF effects, vibrational coherence for spectroscopic beats. Established
Put a claim on trial
Choose a claim: . The plate scores it against nine criteria, from this book’s reading of the evidence in Chapters 163–175, and gives a verdict. Notice the pattern: claims that are established pass several independent tests with replication; speculative ones rest on one kind of evidence, often from one laboratory, without excluding the classical alternative. Neill Lambert and colleagues (2013), Jianshu Cao and colleagues (2020) and Adriana Marais and colleagues (2018) apply similar standards. Debated for the individual scores; they are judgements.
Controls, blinding, preregistration, power
Behavioural and physiological effects are noisy, and experimenters know which sample is which. Good practice: blind the person scoring outcomes to the condition; verify isotope purity by mass spectrometry and chromatography; randomize order; and preregister the hypothesis, sample size and analysis before collecting data, so that flexible analysis cannot manufacture significance. Power matters: to detect a medium effect (a standardized difference ) with 80% power at the 5% level, a two-group comparison needs about subjects per group. Studies far smaller than that produce unreliable “discoveries”. Finally, replication by an independent group is the single strongest test of all. Established
What you now have
The experimental toolkit (isotopes, temperature, fields and RF, spectroscopy, mutagenesis), the non-quantum explanation each must exclude, blinding, purity, preregistration and power, and a scored tribunal of Part 4’s claims.
· · ·
Three questions to carry with you.
And one to work out. Using per group, how many subjects per group detect ?
Sources: N. Lambert et al., Nat. Phys. 9, 10 (2013). J. Cao et al., Sci. Adv. 6, eaaz4888 (2020). A. Marais et al., J. R. Soc. Interface 15, 20180640 (2018). J. Cohen, Statistical Power Analysis for the Behavioral Sciences, 2nd ed. (Erlbaum, 1988). B. A. Nosek et al., Proc. Natl Acad. Sci. USA 115, 2600 (2018) (preregistration).