← All articles

Mindset

Perfectionism Near the Due Date

October 5, 2026

The popular story says the page is late because you care too much. Piers Steel's 2007 meta-analysis in Psychological Bulletin looked at that story and did not find it. Across 691 correlations, the strong, consistent companions of procrastination were task aversiveness, delay before reward, self-efficacy, impulsiveness, and conscientiousness, including self-control, distractibility, organization, and achievement motivation. Neuroticism, rebelliousness, and sensation seeking were weak. Perfectionism, treated as one blob, was weaker still. In the table of that paper, 24 studies and 3,884 people produced an average correlation around .03. Self-oriented and other-oriented perfectionism came out lower than other irrational-belief measures. Only socially prescribed perfectionism, the belief that other people are holding you to their standard, showed a weak link, about r = .18, and Steel analyzed that figure together with other fear-of-failure measures. Haycock's earlier survey, which Steel cites, found about 7 percent of people naming perfectionism as a contributor to their delay.

Steel's own conclusion is easy to misquote. He writes that procrastinators are, if anything, less likely to be perfectionists, not more. A footnote immediately complicates it: even if perfectionists do not delay starting, they may still delay finishing while they chase a standard that does not quit. Starting and finishing are different clocks. The due date cares about finishing.

The 2017 split

Fuschia Sirois, Danielle Molnar, and Jameson Hirsch published a meta-analytic update in 2017 in the European Journal of Personality. They say Steel's review treated perfectionism as one dimension and found it unrelated to procrastination. Split into two higher-order dimensions, the picture changes direction depending on which dimension you mean. Perfectionistic concerns, the worry about mistakes, doubts, and other people's judgment, correlated positively with trait procrastination at r = .23 across 43 samples and about 10,000 people, with a 95 percent confidence interval from .19 to .27. Perfectionistic strivings, the high personal standards, correlated negatively at r = −.22 across 38 samples and 9,544 people, interval −.26 to −.18. Those averages stayed significant after the authors accounted for the overlap between the two dimensions. For concerns, but not for strivings, the size of the link depended on which perfectionism questionnaire was used. The respondent's sex and the procrastination scale did not change the result.

Read the size honestly. An r of .23 is a small-to-medium association in their words, not a destiny. Plenty of people with high concerns finish on time. Plenty of people with high standards do not freeze. The update's value is that "perfectionism" as a single excuse hides two opposite correlations. Standards that organize the work are not the same thing as doubt that keeps the file open.

This is not a diagnosis. Steel and Sirois were pooling mostly questionnaire studies. They did not sit with you at 11 p.m. If worry about mistakes is one part of a pattern that shows up in many situations, does not ease, and gets in the way of work or relationships, that is outside a deadline blog. The National Institute of Mental Health describes anxiety disorders that way: more than occasional worry, present across situations, able to worsen, and able to interfere with job performance, school, and relationships. NIMH says about a third of U.S. adolescents and adults experience an anxiety disorder at some point in their lives. That figure is about anxiety disorders, not about perfectionism, and it is not a test you can take from a paragraph. If that description fits, NIMH points to professional help and, for a crisis, to the 988 Suicide and Crisis Lifeline. Ordinary deadline nerves are not that description. Do not promote them into one, and do not wave off a real problem with a productivity trick.

Done is a description, not a feeling

The finishing problem in Steel's footnote needs a stop rule that does not live in your mood. The 2020 Scrum Guide's Definition of Done is a formal description of the Increment when it meets the quality measures required for the product. When a backlog item meets that description, an Increment exists. If it does not meet the description, it is not shown as finished and it goes back. The Guide says the description creates a shared understanding of what "complete" includes. Developers are the ones required to conform to it. A team that has no organizational standard has to write one appropriate to the product.

You can write a one-paragraph Definition of Done without claiming to run Scrum. Schwaber and Sutherland say partial Scrum is not Scrum. A personal stop rule is allowed to be smaller than a framework. Make it checkable by someone who is not inside your head. "Good enough" is not checkable. "The four required sections exist, the two numbers match the source file, the links resolve, and a named person has seen it" is checkable. Polish beyond that list is a new request. New requests are scope, and they need a date change or a cut. Otherwise the concerns dimension gets an unlimited shift, and the strivings dimension never gets to count the item finished.

Write the stop rule before the last day. On the last day, concerns are loud and the rule will be rewritten to match the anxiety. A rule written on day one is a promise you can keep. A rule written at 11 p.m. is a negotiation with a feeling.

A split you can use tonight

When you notice another pass starting, label it. Striving sounds like "this section fails the stop rule because the source number is wrong." Concern sounds like "someone might think the tone is off," with no item on the stop rule about tone. Fix the first. Park the second on a list that is explicitly after delivery, unless tone is in the stop rule because the audience requires it. If tone is required, it should have been in the rule on day one, with an example of acceptable and unacceptable, not as an infinite reread.

Here is an illustrative example, a fictional edit, not a client. The stop rule says: answer the three questions in the brief, cite the two source documents, and send by 4 p.m. At 3:40 the three answers exist and the citations resolve. The writer wants to rebuild the opening because the first sentence feels plain. That desire is not on the rule. Sending is. The plain sentence can be a note for a later version if the recipient asks. The 2017 correlations do not prove this writer will be late if they keep editing. They do say concern and delay travel together often enough to take the urge seriously, and that high standards are not the villain in the average result.

If the extra pass is really unfinished scope, do not hide it inside "quality." Add it through the decision on the scope page, or catch it at the next two-week lookback while the date can still move. Perfectionism is a bad name for a missing stop rule. It is also a bad name for a clinical problem. Tell those apart before you pick a remedy.

What these papers will not do for you

They will not tell you the right quality level for your job. A surgeon, a lawyer, and a newsletter have different Definitions of Done, and this site does not set them. They will not guarantee that shipping the stop-rule version will please the recipient. Sometimes the recipient wanted the thing you parked. That is information for the next stop rule, not proof you should never ship. Steel's footnote remains: finishing can lag even when starting does not. A stop rule is a counter to that lag. It is not therapy, and it is not a promise about grades, promotions, or income.

Which open file is still open because of a standard you wrote down, and which is open because of a worry that never made the list?

Papers and pages

← All articles · Try the deadline converter