7 Theory Choice and Scientific Change

How scientists compare theories

Theory choice is the process of comparing theories that could explain available evidence and deciding which is best supported or most useful for guiding further research. It is not a mechanical procedure that assigns each theory a single score. Scientists weigh several connected criteria, with evidence remaining central.

A theory is generally preferable when it accounts for reliable, varied observations and experiments better than its alternatives. Tests are especially informative when competing theories predict different outcomes; if both fit what is already known, a new test may help distinguish them. Measurements, experimental methods, and background assumptions may also need review when interpreting a test.

Evidence and explanatory virtues

Scientists assess several connected virtues. Accuracy concerns how closely predictions match observations, while scope concerns how many different phenomena a theory explains, including phenomena beyond those for which it was first proposed. Explanatory power and unification concern whether a theory explains why phenomena occur and connects observations that previously seemed unrelated.

Consistency asks whether a theory is internally coherent and fits with well-supported knowledge in relevant fields. Fruitfulness concerns whether it suggests new questions, experiments, or useful models. Together, these criteria help scientists compare how well a theory performs and whether it can guide further inquiry.

These criteria may reinforce one another, but they can also conflict. A more complex theory might explain a wider range of evidence, while a simpler theory might fit current observations but have less scope. Scientists judge which differences matter most in the particular research context; the criteria do not produce a fixed decision rule.

Predictions that distinguish theories

Scientists also consider whether a theory makes novel, risky predictions: predictions that were not simply built into its original formulation and that could have turned out differently. When such a prediction succeeds, it can provide stronger support than a theory’s ability to accommodate a result after it is already known.

No single test automatically settles a comparison. Scientists may need to review the measurements, experimental methods, and background assumptions associated with a result. A theory’s apparent success or failure must therefore be assessed in light of the broader evidence and the way the test was conducted.

An orbital anomaly as a comparison

Consider observations showing that a planet’s orbit differs slightly from the path predicted by a familiar model. One proposal adds a special adjustment that accounts for this planet alone; another proposes a general mechanism that explains the discrepancy and predicts related effects in other observations.

The second proposal could have greater scope and fruitfulness if it explains several findings and leads to successful new tests. It might also be simpler in an important sense because one shared explanation replaces several unrelated adjustments. But if its predictions fail, or the evidence for them is weak, its apparent elegance does not rescue it. The comparison depends on the combined record of evidence and explanatory performance, not simplicity alone.

What simplicity contributes

Simplicity can mean fewer assumptions, fewer kinds of entities, fewer adjustable parameters, or a more unified mathematical structure. These meanings do not always favor the same theory, so simplicity is best treated as a comparative virtue rather than a demand to choose the simplest imaginable explanation.

When two theories perform similarly, unnecessary complexity can count against one of them. But simplicity does not override evidence or establish that the simplest theory must be true.

Why scientific judgments remain revisable

Evidence can support a theory without proving it conclusively. Different theories may make the same predictions for the observations available so far, in which case further evidence or new methods may be needed to tell them apart. An apparent mismatch may also be traced to a measurement problem or to an auxiliary assumption used alongside the theory.

Scientists assess the broader body of evidence, test alternatives where possible, and revise theories when better-supported explanations emerge. Theory choice is therefore a reasoned but revisable judgment: the preferred theory is the strongest available option under the evidence and standards of inquiry at the time, not necessarily a perfect or final account.

Scientific change can occur as evidence accumulates, better tests become possible, or a theory with greater accuracy, scope, or explanatory coherence replaces an older one. Theories remain open to revision as these circumstances change.