What Piers Steel's 2007 meta-analysis reported, and what it did not

In 2007, Piers Steel, then at the University of Calgary, published a meta-analytic review of the procrastination literature in Psychological Bulletin. The paper pooled results from several hundred studies accumulated over roughly three decades and asked a narrow question: across all of this work, which variables show a consistent statistical relationship with self-reported delay?

The answer was shorter than the literature that preceded it. Four clusters showed the strongest and most stable associations: task aversiveness, low self-efficacy regarding the task, impulsiveness, and low conscientiousness — particularly its facets covering self-control, distractibility, organisation and achievement motivation. Steel also found that the length of delay before an outcome mattered: the further away a consequence sat, the weaker its recorded pull on present behaviour.

The perfectionism result

The finding that drew most comment was negative. Perfectionism, which the self-help literature of the period treated as a principal driver of delay, showed a correlation close to zero in Steel's synthesis, and in some analyses a slightly negative one. Neuroticism, rebelliousness and sensation seeking also performed weakly, despite each having a supporting literature of its own.

Steel's interpretation was that these variables had entered the field through clinical observation and small studies rather than through replicated measurement, and that pooling the evidence removed them. The perfectionism claim survives in popular writing largely undisturbed by this, which is itself a fact about how research reaches general audiences rather than a fact about perfectionism.

Temporal motivation theory

Alongside the synthesis, Steel proposed an integrative account he and Cornelius König had developed, describing motivation as rising with the expectancy and value of an outcome and falling with the delay before it arrives and the individual's sensitivity to that delay. The formulation draws on hyperbolic discounting work in behavioural economics, most closely associated with George Ainslie and reviewed by Leonard Green and Joel Myerson.

The theory is a modelling framework rather than a tested causal claim. It organises the correlates the meta-analysis found; it does not demonstrate that they operate the way the equation describes.

Sampling and measurement

The bulk of the pooled studies used undergraduate samples at a small number of institutions in North America and Europe. Undergraduates work under deadline structures, supervision arrangements and incentive schedules that are unlike most adult contexts, and results from that population do not automatically extend beyond it.

Outcomes were almost entirely self-reported. Where researchers have compared self-reported procrastination against observed delay — submission timestamps, task logs — the agreement has been moderate rather than close. Prevalence figures circulating from this period, including the frequently repeated estimate that around one adult in five reports chronic delay, rest on the same self-report base and should be read as descriptions of what people report about themselves.

Publication bias is a further constraint. Meta-analysis can estimate its extent but cannot recover studies that were never written up, and the period covered predates routine preregistration.

Standing

The 2007 review remains the most cited single reference in this literature and is the point of departure for most subsequent work. Its durable contribution is subtractive: it narrowed a crowded field of proposed causes to a short list with consistent statistical support, and marked several popular explanations as unsupported by the pooled evidence. Later authors, including Fuschia Sirois and Timothy Pychyl, have argued that the correlate list points toward emotion regulation rather than time management — an argument summarised separately in this catalogue.