Planning and Execution: Two Times Can Tell Different Stories

Someone who takes longer to think before starting a task may also take longer to complete it. Does that show that thinking makes execution slower? Not necessarily. The comparison may combine differences in people’s general working speeds. Within a particular person-and-task encounter, extra preparation might instead be associated with shorter execution. The findings can point in opposite directions without contradicting each other, because they answer different questions.

Zhaojun Li, Paul De Boeck and Jian Li examined this distinction in a 2020 study of planning and execution times. This was not a planning-fallacy experiment asking people to predict how long future work would take. The researchers recorded time spent in two stages of a box-pushing game. They then separated overall relations between people from the remaining relationship associated with particular solutions. The purpose was to find information that one aggregate correlation could conceal. [1]

A puzzle with an irreversible beginning

The assessment contained ten tasks developed from Sokoban. Players moved a character to push boxes onto target locations. They could not pull boxes or push two or more together. Each task was redesigned to have only one correct first move; other initial moves produced an impasse. Participants were told their moves could not be undone and were advised to plan before beginning. There was no time limit. [1]

This arrangement makes the opening move more than a convenient starting point. Moving the nearest box without considering later access can make completion impossible. Looking ahead is part of the task. Nevertheless, the recorded planning time was not a detailed record of every element of a mental plan. It was the interval from task onset to the first move. Execution time ran from that move to successful completion.

A player can continue thinking after the first move. Under these definitions, that thought remains inside execution time. The two intervals therefore do not perfectly isolate all mental planning from all motor behaviour. Their operational boundary matters: one measures initial preparation, while the other measures progress from the first action to completion. Naming the stages does not make their psychological contents entirely separate.

Why successful solutions defined the main sample

The study began with 266 students from a Chinese university. The main analysis focused on 119 who completed all ten tasks successfully. After an incorrect move led to an impasse, there was no successful completion time for execution. The researchers also reported supplementary analyses using groups who completed smaller common sets of tasks, with the principal findings remaining consistent. [1]

That check is useful, but does not turn failed trials into completed ones. The main relationship is built from observations with available execution times. It cannot automatically describe every error, abandoned attempt or unfinished job. Nor does a student sample playing ten puzzles establish a fixed performance rule for workers in every occupation. The controlled task gives repeated comparisons of a particular kind, not an exhaustive sample of daily problem solving.

What a slower person tells us

The researchers analysed log-transformed times, rather than treating all raw seconds as equivalent increments in the statistical relationship. They used two linear mixed-model approaches and a factor-model analysis. Across these approaches, the aim was to distinguish a person’s general speed and a task’s time demands from the remaining relationship between the two intervals in a particular solution. [1]

At the person level, general planning-time and execution-time tendencies were positively related. People who tended to take longer to plan also tended to take longer to execute than others. The authors considered this consistent with a general mental-speed account. Consistency with that account is not direct observation of a separate speed mechanism. More importantly, the result does not answer what would happen if a particular person were made to spend additional time planning. That is a different comparison.

There was also a positive relationship across tasks. Some puzzles required longer preparation and longer execution. The paper investigated route length in connection with this pattern: a longer solution path can place demands on both stages. Such covariation does not establish that planning harms performance. Differences between tasks can make both measured times rise together. Pooling people and puzzles into one relationship risks losing the contribution of each source.

The relationship after accounting for those differences

When person and task differences were included in the models, the average conditional relationship was negative. More initial planning than expected was associated with less execution time. The second approach specifically used residual planning time: the difference between the observed time and the time predicted from the person’s and task’s general tendencies. Residual here does not mean wasted. It names variation remaining after those modeled influences. [1]

The association was not equally supported in every puzzle or every analytic detail. Variation across tasks had clearer support, while variation across people was supported differently by the two mixed-model approaches. In the factor analysis, not every individual task’s residual relationship was significantly negative. These results limit a simple universal slogan. Extra preparation may be more helpful for some problem structures, and an average negative dependency does not promise shorter execution for everyone on every task.

Both levels can coexist. Relatively slow planners can also be relatively slow executors overall, while additional preparation within a particular modeled encounter is associated with faster execution. One concerns differences in general tendencies; the other concerns the remaining relationship for an individual observation. Calling the results contradictory would confuse the units of comparison rather than expose a flaw in the finding.

Shorter execution is not automatically shorter total work

There is another boundary to preserve. If preparation grows while execution shrinks, the total time need not decline. That depends on how much execution time is saved relative to the extra preparation. The conditional association does not establish that every additional second spent waiting saves more than a second later. Nor does it justify unlimited delay before acting. Time allocated to one stage must not be mistaken for a guaranteed reduction in the whole task.

Longer planning periods were not randomly assigned. Participants generated their own times as they solved the puzzles, and the researchers modelled those observations. Statistical adjustment for person and task differences is not experimental control of every possible cause. The authors interpret the pattern as compatible with efficient planning, but a rule that compulsory waiting will certainly improve performance would exceed the design. The quality of a plan is also not identical to the length of time before the first move.

The evidence-based conclusion is that planning and execution do not have one relationship that can be stated independently of the comparison. General speed between people, route demands between puzzles and preparation within a modeled solution contain different information. This study offers no universal recipe for how long to think. It instead shows why positive and negative associations can belong to the same data. Before deciding whether thought helps action, a sound first step is to identify which people, tasks and levels the claimed relationship actually compares.

Sources

[1] Li, Z., De Boeck, P., and Li, J. (2020). “Does planning help for execution? The complex relationship between planning and execution.” PLOS ONE, 15(8), e0237568. Open access, Creative Commons Attribution; used with independent exposition and interpretation. https://journals.plos.org/plosone/article?id=10.1371/journal.pone.0237568