Why PhD Students Stall: The Decision Load Behind Dissertation Paralysis

The standard advice for dissertation paralysis is to sit down and write. Just start. Break it into smaller tasks. Use a timer.

This advice is not wrong, exactly. But it addresses the wrong problem.

Cognitive load research suggests the stall is rarely about motivation or time. It is about decision load — the number of unresolved choices competing for working memory.

The Decision Load Framework

The Decision Load Index measures the number of unresolved decisions an individual is actively holding at any given time. Cognitive load theory (Sweller, 1988) and decision fatigue research (Baumeister et al., 1998) motivate the hypothesis that this relates to burnout symptoms independently of hours worked. Neither body of work tested that comparison, and ego depletion itself has failed large replication attempts (Hagger et al., 2016).

The hypothesis: the people most at risk of burnout are not always the people working the most hours. They may be the people holding the most unresolved decisions.

Among PhD students experiencing the highest stall rates, one pattern is worth investigating: the attempt to hold an unusually large number of open choices in working memory simultaneously.

The Dissertation Is Not One Decision

When someone sits down to write and cannot start, the internal experience is often described as being stuck. What we suspect is actually happening: the person is not stuck because they do not know what to do. They are stuck because too many things need to be decided before they can do it.

A single dissertation chapter generates decisions across multiple layers simultaneously:

Layer	Example decisions
Scope	Does this section belong here or in a later chapter? How much literature do I need to cover? Is this a footnote or a paragraph?
Argument	Do I claim X or hedge to "X may be the case"? Does this evidence support my thesis or complicate it? What do I do with the finding that contradicts my framework?
Advisor dynamics	Will my advisor accept this framing or push back? Should I flag this gap now or address it in the draft? How much do I deviate from their suggestions?
Field positioning	Am I citing the right people? Is this contribution original enough? Would a reviewer consider this a minor or major finding?
Process	Do I outline first or draft? Which citation manager format? Should I write the intro or the methods first?

None of these decisions can be made in isolation. Resolving one often opens several more. The person sitting at the blank document is not failing to act. They are trying to navigate a decision tree that has no clear root node.

Why ABD Syndrome Makes Sense Through This Lens

"All But Dissertation" describes a real phenomenon: students who complete every structured requirement of a PhD but cannot finish the dissertation itself. The coursework, exams, and qualifying papers all have clear external constraints. The dissertation has almost none.

External constraints reduce decision load. A seminar paper has a deadline, a word count, a format, and a professor who defines the scope. The number of open choices is bounded.

The dissertation, by design, removes most of these boundaries. The contribution must be original. The scope must be self-defined. The timeline is flexible. The feedback cycle with the advisor is irregular. Every one of these freedoms is also a source of unresolved decisions.

This is not a claim that ABD syndrome is purely a decision-load problem. There are funding constraints, mental health factors, life circumstances. But the hypothesis — that high decision load predicts stall, and that the dissertation stage is structurally optimized to maximize decision load — is worth examining. We have no dataset that tests it.

Why Productivity Systems Often Make This Worse

The instinct when stalled is to add structure: a new writing system, a daily word count, a project management tool, a Pomodoro timer.

We suspect that adding new productivity systems during high-stall periods often does not reduce burnout or speed progress. The likely reason is decision overhead: every new system introduces its own set of choices about how to use it, how to classify tasks, and when to follow its rules versus override them.

The system designed to reduce load generates its own. The person ends up managing the system instead of writing.

What Might Help

Three practices are worth trying in high-complexity knowledge work, including academic writing. They are suggestions, not measured findings:

Pre-committing scope decisions before each session

Not deciding what to write during the writing session. Deciding the day before — or at the start of a dedicated 10-minute planning block — exactly which section, which argument, which scope boundary applies to today. The idea is to arrive at a writing session with a narrowed set of valid choices rather than a blank document.

Externalizing the unresolved

Keeping a separate log specifically for unresolved decisions, distinct from a to-do list. "I still need to decide whether to cite Smith or Jones here." "I am not sure if this belongs in Chapter 2 or Chapter 4." Writing these down — not resolving them, just logging them — may reduce the cognitive cost of holding them, because items you have given yourself permission to defer need less re-checking.

Separating discovery writing from commitment writing

Some writers distinguish between sessions where the goal was to generate text without committing to it (low-stakes drafting, voice memos, freewriting) and sessions where the goal was to refine and commit. The distinction can reduce the decision pressure on each session. Discovery sessions do not require scope decisions. Commitment sessions do, but by then the material already exists.

A Note on What This Does Not Mean

This is a hypothesis, not a study. We have no retained dataset of PhD students or stall rates, and we cannot establish causation. People who stall less may have other advantages — better advisor relationships, clearer research questions, more external structure from funding deadlines.

The claim is narrower: decision load may relate to stall rate, and the structural features of dissertation work are well-designed to maximize decision load. If interventions that reduce load help in general knowledge work, they may be worth trying in academic writing contexts specifically.

We do not know yet. That is the honest answer.

This is a research field note, not a clinical or academic finding. It offers hypotheses, not measured results. This content is educational and does not constitute career or psychological advice. Results vary.

Corrections, 28 September 2026: An earlier version presented several patterns as findings from “our data” and cited a “Decision Load Index: 90-Day Knowledge Worker Study, N=733”; no such dataset of PhD students or stall rates exists (733 was a sign-up count from our retired consumer assessment), so we removed the citation and reworded those passages as hypotheses and suggestions. The ~50% figure is a commonly cited overall doctoral attrition estimate, not a dissertation-stage rate, and we corrected the title and journal of the Sverdlik et al. (2018) review. We removed a claim that cognitive load and ego-depletion research show decision load predicts burnout better than hours worked; neither tested that.