What Counts as Evidence in a Field Without Stable Objects?
Psychology’s evidentiary standards are often discussed as if the field were studying stable, well-defined objects. The language of variables, effects, mechanisms, and outcomes implies a level of ontological consistency that psychological phenomena rarely possess. Unlike the objects of the physical sciences, psychological objects are not fixed entities. They are dynamic, context-sensitive, self-interpreting, and often transformed by the very act of being studied. This instability poses a fundamental problem for evidentiary reasoning, one that methodological refinement alone cannot resolve.
Evidence presupposes an object of reference. To count something as evidence is to assume that it bears on a relatively stable phenomenon whose properties can be inferred, compared, and accumulated across observations. In psychology, however, many of the phenomena of interest do not maintain this stability. Beliefs change as they are articulated. Emotions shift as they are labeled. Identities reorganize in response to social feedback. Psychological constructs are not merely observed; they are enacted.
This does not make psychological inquiry impossible, but it does complicate its epistemology. When the object under study is mutable, evidence cannot function in the same way it does in fields where the object remains unchanged by observation. Repeated measurement does not simply reveal more about the same phenomenon; it may participate in its alteration. The meaning of a response today may not be equivalent to the meaning of the same response tomorrow, even if the measure remains constant.
By the time I began studying psychology in the mid-1980s, this tension was already present but rarely named. The field operated as though psychological objects could be stabilized through operational definition and statistical aggregation. Constructs were treated as if they possessed an underlying essence that measures imperfectly captured. Evidence was evaluated primarily in terms of consistency, convergence, and effect size, with less attention to whether the phenomenon itself remained invariant across contexts.
The difficulty becomes especially apparent when psychological evidence is treated cumulatively. Meta-analyses aggregate findings across studies under the assumption that they are indexing the same underlying construct. Yet small differences in context, framing, population, or historical moment can fundamentally alter what a construct means. What counts as anxiety in one cultural or institutional setting may not be functionally equivalent in another. Aggregation can obscure these shifts while creating the appearance of robustness.
Case material again exposes the fragility of evidentiary assumptions. Individual cases often demonstrate that psychological phenomena reorganize around life events, relational dynamics, or shifts in self-understanding. The same person may exhibit different patterns of behavior or report different experiences under identical measurement conditions, not because of error, but because the phenomenon itself has changed. Treating such variability as noise risks mistaking transformation for instability.
The problem is not limited to subjective experience. Even behavioral and physiological measures are embedded in interpretive frameworks. A behavior observed in one context may carry a different meaning in another. A physiological response may signal threat, excitement, or engagement depending on situational interpretation. Evidence does not speak for itself. It acquires significance through theoretical framing.
This places a particular burden on theory. In a field without stable objects, theory must do more than generate hypotheses. It must specify what kind of stability is being claimed, under what conditions, and for how long. A theory that treats psychological constructs as static entities invites evidentiary confusion. A theory that treats them as processes unfolding over time can better accommodate variability without collapsing into relativism.
The evidentiary hierarchy often invoked in psychology implicitly assumes object stability. Randomized trials, large samples, and replication are valued because they are designed to reduce variability and isolate effects. When variability reflects genuine transformation rather than error, these strategies can misfire. Evidence that looks weak or inconsistent under one epistemic lens may be entirely coherent under another.
This has implications for how psychology evaluates disagreement. Conflicting findings are often attributed to methodological differences or sampling error. Less often are they treated as indicators that the phenomenon itself may be context-dependent or historically contingent. When evidence fails to converge, the field tends to tighten methods rather than reconsider its assumptions about object stability.
The discipline’s discomfort with unstable objects also explains its preference for constructs that can be more readily stabilized. Traits, capacities, and dispositions are attractive because they promise continuity across situations. Yet even these constructs are increasingly understood as probabilistic tendencies rather than fixed properties. The more psychology attends to development, culture, and context, the harder it becomes to sustain the fiction of stable psychological objects.
Recognizing this does not entail abandoning evidentiary standards. It entails refining them. Evidence in psychology must be evaluated not only in terms of reliability and validity, but in terms of interpretive fit. Does the evidence illuminate the process by which the phenomenon unfolds? Does it specify the conditions under which the phenomenon takes a particular form? Does it acknowledge the role of meaning, history, and context in shaping what is observed?
A mature evidentiary framework for psychology would tolerate a plurality of evidence types. Quantitative patterns, qualitative narratives, longitudinal trajectories, and theoretically informed case analyses each reveal different aspects of unstable phenomena. No single form of evidence can claim epistemic supremacy across all domains. Hierarchies of evidence must be aligned with the nature of the object, not imposed uniformly.
This alignment also requires intellectual humility. Psychological knowledge will often be provisional, context-bound, and subject to revision. Evidence will sometimes point in different directions depending on the questions asked and the lenses applied. Treating this as a failure misunderstands the field’s subject matter. Instability is not an obstacle to psychological knowledge; it is one of its defining features.
What counts as evidence in psychology cannot be determined solely by method. It must be determined by an ongoing negotiation between theory, phenomenon, and context. Evidence earns its status not by conforming to abstract standards, but by illuminating how psychological life actually unfolds. In a field without stable objects, evidence must be flexible enough to move with what it seeks to understand.
Letter to the Reader
If you have ever felt uncertain about what kind of evidence should “count” in psychology, that uncertainty is not a weakness. When I was training in the mid-1980s, evidentiary standards were already being presented as settled, even as the phenomena themselves resisted stabilization.
Learning to work in this field means learning to live with that tension. Ask not only whether evidence is strong by conventional criteria, but whether it is appropriate to the kind of phenomenon being studied. Psychology advances not by pretending its objects are stable, but by developing forms of evidence that can move with them.