1  Too Many Theories

NoteIn a nutshell
  • Theory proliferation, reviewing culture, and writer anxiety form a feedback loop that makes papers theoretically exhaustive and argumentatively weak
  • In Psychological Science alone, 359 distinct theories appeared across a decade of articles — most cited only once
  • Preregistration was designed to help, but only does the work if the theoretical thinking was done first
  • Each individual decision to add more coverage is understandable; the cumulative result is manuscripts that argue nothing
  • The writer cannot fix the incentive structure, but can choose not to be driven by it

When you sit down to draft a manuscript, the cursor blinks against a massive, invisible weight: the accumulated literature of your field. Because psychology has generated frameworks faster than it has retired them, the rational response to that blank page is defensive writing. A paper that should make one clear argument instead surveys a landscape of loosely related theories, hedges every claim, and leaves the reader wondering what the point was.

The cause is structural, not individual.

1.1 The theory overload problem

Psychology — and cognitive psychology in particular — is a theory-rich field. For almost any phenomenon you study, there are multiple accounts, frameworks, and models that are at least partially relevant. Event cognition alone has produced a remarkable number of competing and overlapping theories: Event Segmentation Theory, Situation Models, Narrative Comprehension models, and more.

When a PhD student sits down to write, they face a genuine dilemma. They know this landscape well. They also know that reviewers are drawn from that same landscape — and that reviewers notice when their preferred theory is not discussed. The rational response is to mention everything. The result is a manuscript that mentions everything and argues nothing.

The scale of the problem is visible in the record. McPhetres et al. (2021) examined 2,225 articles published in Psychological Science between 2009 and 2019 and identified 359 distinct theories mentioned in text. The majority were cited only once. Theories accumulate in the literature faster than they are tested, refined, or retired — and the introduction section is where that accumulation is most visible.

1.2 Where it comes from

The problem has three interlocking sources.

Theory proliferation. Psychology has accumulated theories faster than it has retired them. Many theories overlap substantially but use different terminology, which makes them look more distinct than they are.

Reviewing culture. Reviewers sometimes treat citation of their preferred framework as a condition of acceptance. This is rarely stated explicitly, but early-career researchers learn quickly that omitting a theory can sink a paper—so they throw in an extra paragraph about Theory X just in case Reviewer 2 happens to be its author.

Writer response. Faced with these pressures, writers rationally expand their theoretical framing to pre-empt criticism. This makes each paper slightly worse, which makes the overall problem slightly worse.

These three forces form a feedback loop. No single actor is to blame, and no single fix works in isolation.

1.3 Why preregistration is not enough

In recent years, the field has leaned on preregistration to solve this exact problem of shifting, post-hoc theoretical framing. And for good reason: it creates a public record of your theoretical commitment before you see the data, protecting against the temptation to adjust the framing retrospectively to fit the results — what Kerr (1998) called HARKing, Hypothesizing After Results are Known. But preregistration only does this work if the registered hypothesis names the theory and the mechanism, not just the expected effect. A preregistration that states “we expect memory to be better at event boundaries” records a prediction without a reason. It does not specify which theory generates the prediction, what mechanism connects theory to outcome, or what result would count against it. Oberauer and Lewandowsky (2019) make this point directly: preregistering a hypothesis that is only loosely connected to a theory does not increase the credibility of that hypothesis, because the inferential link from theory to prediction was weak to begin with. Preregistration records your thinking. It cannot replace it.

The feedback loop is a structural problem, and you cannot fix the incentive structure of the field alone. But you can choose not to be driven by it. The rest of this guide is about taking back control of your manuscript — starting with how you filter the overwhelming noise of the literature before you even write a single word.