STRATA
Skip to study
XLVII

DISCOVERY

12 topics9 study records8 connections
An open optical instrument, prism and narrow orange beam.

Begin with a concrete example.

V.02 / Natural philosophy to physics

Follow an observation into a new model.

Take a sequence of recorded celestial positions and represent it in a simple diagram. Separate the observational record from assumptions about the bodies' arrangement. Compare how two geometric models account for the sequence and identify an additional observation that would discriminate between them.

V.03 / Serendipity and engineered discovery

Give an anomaly a structured record.

Suppose a sensor produces an unexpected spike during a routine run. Preserve the raw trace, instrument settings, environmental conditions, and neighboring observations before editing the dataset. Compare a repeat run with a changed cable and another with unchanged hardware; the sequence can distinguish an equipment artifact from a phenomenon worth further investigation.

V.04 / Hypothesis testing

Distinguish the threshold from the evidence.

For an illustrative quality-control problem, define an acceptable defect rate and a sampling rule. Compare a small and a large sample with the same observed percentage of defects. Their uncertainty differs, so the same percentage can support different decisions. Record the practical cost of each kind of error before choosing a threshold.

Source records

All studies

Research questions

Browse the atlas
  • V.01Alchemy to chemistry

    Which workshop techniques survived the change from alchemical recipes to quantitative chemistry?

    The movement from alchemical practices to modern chemistry is best investigated through instruments, procedures, substances, and changing explanations. Lavoisier's 1789 treatise supplies one inspectable historical document. It should be placed alongside earlier practices without treating a complex transition as a clean break. The practical question is how a more explicit accounting of materials and measurements changed what could be argued and reproduced.

  • V.02Natural philosophy to physics

    Which instruments and unresolved problems carried from natural philosophy into physics, and where did their methods change?

    Natural philosophy and physics overlap through changing questions about motion, matter, causation, and mathematical description. Galileo's printed observations offer a manageable historical case in which an instrument altered what could be seen and argued. A useful lineage follows a problem through particular texts and measurements, identifying where an old explanatory vocabulary was retained, revised, or replaced.

  • V.03Serendipity and engineered discovery

    What was entered in the notebook before the significance of an unexpected result became clear?

    Serendipity becomes scientifically productive when an unexpected observation is preserved, investigated, and connected to a question. The British Library's Fleming notebook record offers a material entry point into a familiar discovery history. The wider lesson is to examine the work following an anomaly: documentation, comparison, isolation of possible causes, and the development needed to make a result usable.

  • V.04Hypothesis testing

    Which outcomes would lead an investigator to reject the null hypothesis at a specified error rate?

    Hypothesis testing organizes decisions under uncertainty by specifying alternatives, a test statistic, and error behavior. Neyman and Pearson's work supplies a historical foundation for this design perspective. A useful analysis starts before the data are inspected: which difference matters, what procedure will be applied, and how often it could produce a mistaken decision under the assumed conditions.

  • V.05Information gain and experimental design

    Which affordable measurement would most reduce uncertainty about the decision at hand?

    Experimental design asks which observation is worth acquiring next. Lindley's information-based treatment supplies a historical anchor for comparing experiments by what they are expected to reveal. The practical decision also includes cost, feasibility, and the action the result could change. An experiment is valuable when its possible outcomes separate relevant explanations rather than merely producing another measurement.

  • V.06Bayesian updating

    How much does a new observation change the odds assigned to two competing explanations?

    Bayesian updating expresses how a prior distribution and a likelihood combine into a posterior distribution. Bayes's historical essay provides an early primary text in this lineage. In practical inquiry, the useful discipline is to state the competing possibilities and the probability of the observed evidence under each. This makes a change in belief inspectable instead of leaving it as an impression.

  • V.07Engineering iteration, laboratory notebooks, and provenance

    Which entries in a laboratory notebook let a second bench reproduce a revised procedure?

    Engineering iteration becomes research when changes, results, and decisions remain traceable. The Berkeley Protocol's treatment of preservation and provenance provides a useful discipline for records, while FAIR principles organize reusable research objects. A notebook should let a later reader reconstruct what changed, what was measured, and why the next version was chosen. Failed configurations can be as informative as the final working one.

  • V.08Replication, reproducibility, and preregistration

    What happens when an independent group collects new observations under a preregistered protocol?

    Reproducibility and replication become clearer when the intended repeat is named. Re-running the same analysis, collecting new observations under similar conditions, and testing a related mechanism answer different questions. The Open Science Collaboration's project provides a documented example of organized replication. Preregistration adds a time-stamped plan that helps distinguish intended analyses from decisions made after examining results.

  • V.09Causal inference and intervention

    Under what assumptions can an observational estimate predict the effect of an intervention?

    Causal inference asks what would change under an intervention, which requires more structure than observing an association. Pearl's account makes assumptions about causal relationships explicit through structural models. A diagram is useful when it identifies confounding, mediation, selection, or a missing variable that changes what can be inferred. The next step is to connect that structure to a concrete intervention or identification strategy.

  • V.10Contradiction diagnosis and model comparison

    Do the competing models predict different readings under the same experimental conditions?

    An apparent contradiction can arise from different definitions, samples, instruments, scales, or model assumptions. Diagnosing it requires placing claims in comparable form before deciding whether one must be rejected. Measurement uncertainty and Bayesian model comparison provide complementary tools: one checks whether the reported quantities are commensurable, while the other examines which model better accounts for specified observations.

  • V.11Negative results, retractions, and missing evidence

    What effect could the original experiment detect? Follow the data and subsequent corrections to see which conclusions changed.

    A negative result, an unavailable dataset, a correction, and a retraction represent different states of the record. COPE's retraction guidance provides an explicit editorial framework for the last of these. Research triage should preserve the actual notice and reason, then identify which claims are affected. Missing evidence remains a retrieval problem until there is a basis for a stronger conclusion.

  • V.12Recovery of abandoned, overlooked, or repurposed work

    Which change in materials, instruments or cost made an abandoned technique worth revisiting?

    Recovering overlooked or repurposed work begins with a bibliographic trail and a specific question about present usefulness. Author archives such as Cormack's bibliography help distinguish an original algorithm from later implementations and informal labels. An old method earns renewed attention through a documented fit to a current task; its age or obscurity alone does not establish either value or failure.

Related research

Bridge register

Research collections

Use these institutional collections to locate primary observations, methods and datasets.