Experimental Open Problems: Choose a Discriminating Measurement
An experimental research question asks which observation would distinguish competing physical explanations or establish a missing capability. The decisive object is a measurable contrast with a credible uncertainty and control strategy. A more elaborate apparatus is useful when it improves that inference, not merely when it produces a larger signal or system.
Helpful background. Why benchmarking is hard explains how the chosen observable and comparison determine a claim.
Choose the missing evidence
Section titled “Choose the missing evidence”| Experimental question | Topic owner | Essential control |
|---|---|---|
| Does more protection reduce logical failure? | Logical milestones | Matched physical task and acceptance |
| Is a candidate phase distinguished from alternatives? | Spin liquids, non-Abelian statistics | Competing phase and disorder explanations |
| Is a weak signal new dynamics or environmental noise? | Collapse tests, precision AMO | Independent nuisance calibration and null configurations |
| Does a link deliver useful entanglement? | Entanglement distribution | Heralding, background and state-quality validation |
| Can a proposed gravitational interaction generate entanglement? | Gravity-mediated entanglement | Electromagnetic and environmental alternatives |
Established: experiments support bounded claims with specified preparation and measurement. Active: exclusion of alternative mechanisms and reliable extrapolation to new regimes. Conjectural: a unique interpretation inferred from a signal that several models predict.
Design the likelihood before collecting the decisive data
Section titled “Design the likelihood before collecting the decisive data”Write the signal distribution , where describes the effect and the nuisance parameters. Identify which calibration constrains each nuisance and whether those calibrations are independent of the signal fit. An apparent detection can disappear when a previously fixed nuisance is allowed to vary.
The certification review by Eisert and colleagues explains why a witness certifies a particular object under particular assumptions. Logical-memory experiments illustrate the importance of a matched reference and duration. Certification review, memory experiment.
Entry project: a sensitivity and identifiability study
Section titled “Entry project: a sensitivity and identifiability study”Choose one published small experiment and reconstruct its signal model and dominant background. Generate synthetic data under the null and under an alternative; fit both while varying nuisance priors or calibration uncertainty. Deliver expected confidence intervals, false-positive behavior and the measurement setting that best separates the hypotheses. A laptop suffices before laboratory access is required.
The baseline is the original analysis with its stated assumptions. Failure tests include drift, correlated noise, miscalibrated readout and an alternative background with similar shape. If the effect is not identifiable at feasible resources, revise the observable or stop that experiment design. Research begins with a demonstrably better control, new discriminating observable or independently calibrated experiment, not a favorable synthetic significance alone.
Novelty and resource checks
Section titled “Novelty and resource checks”Before proposing a new experiment, identify the apparatus, data access, sample preparation, acquisition time and calibration expertise actually available. Check whether a recent result has already supplied the intended contrast. Reviewed 11 October 2026; selected examples are 2020 and 2025 sources, not a comprehensive current experimental-status list. Topic assessments own the changing evidence.