Can Incompetence Be Defined as Raising Other People’s Cognitive Load?

A record of testing the claim that “incompetence in modern knowledge work means raising other people’s cognitive load; people who cannot self-regulate or self-correct raise others’ cognitive load and are therefore treated as incompetent; raw processing ability barely matters” against 58 academic sources and against industry practice. Full bibliographic details appear in the References section below (with DOIs/URLs). The internal corpora are at source/review/cognitive-load-competence/papers.md and source/review/cognitive-load-competence/industry.md (repository-internal, not published). For the general mechanics by which evaluation accumulates advantage, see the critique of meritocratic evaluation; for measuring human capability on the assumption of AI use, see ai-augmented-competency-measurement. This note records how far the empirical work reaches when the claim is put to it. The design that takes up what it fails to reach, by describing the field rather than measuring it, is From This Is Heavy to That Person Is Incompetent. Setup: three source-researcher agents (profile: scholarly), two industry-lead agents (persona: strategy-consulting / design-firm), one practice-auditor (lens: evidence-methodology), and two academic-critic agents (canon: design-education / design-industry). The key figures were verified by the orchestrator directly from publisher PDFs.

Everyone can name one

There is a colleague who processes quickly, and yet everything slows down once they are involved. There is another whose output volume is unremarkable, and yet whatever you hand them is reliably finished. As an account of that difference, the claim is attractive. It re-measures competence not by the performance of an individual head but by the cost pushed outward onto everyone else.

Being attractive is exactly why it should not be taken as it stands. What looks like one sentence joins three assertions of different character.

Splitting the claim in three

Following the decomposition produced by practice-auditor:

• Claim A (definition): incompetence is raising other people’s cognitive load.
• Claim B (pathway): people who cannot self-regulate or self-correct raise others’ cognitive load, and are therefore treated as incompetent.
• Claim C (exclusion): the level of raw processing ability barely matters.

Put to the test, the three behave in completely different ways.

A definition cannot be knocked down by observation

Claim A replaces the word “incompetence” with the phenomenon “raising other people’s cognitive load.” It is a definitional statement declaring how a word will be used, not an assertion whose truth is settled by observation. Being a definition, it offers no handle for refutation.

Claim B collapses into the same structure once Claim A’s definition is substituted in. “People who cannot self-regulate raise others’ load and are therefore treated as incompetent” becomes, on Claim A’s own reading, “are therefore treated as people who raise others’ load,” and the trailing clause adds nothing.

The minimal rewrite that makes Claim B testable is to strip out the evaluative label “incompetence” and assert only a relation among independently measurable variables.

A deficit in self-regulation and self-correction increases other people’s cognitive load (third-party ratings, or objective indicators such as response time) even when the level of processing ability is held constant.

In that form it becomes an empirical hypothesis about three variables, and data can say no to it.

Of the original claim, only Claim C is empirically testable as written.

Where the testable sentence leads

Claim C lands in a field that has just revised its answer.

In 1998 Schmidt and Hunter synthesized 85 years of selection research and estimated the operational validity of general mental ability (GMA) tests for job performance at r = .51, the highest of nineteen selection methods. That conclusion served as standard doctrine in personnel selection for more than twenty years. If the estimate holds, Claim C is plainly false.

In 2022 Sackett and colleagues argued that this line of meta-analysis had systematically overcorrected for range restriction, and recomputed the estimates. The paper’s own wording:

Taken together, the N-weighted mean reliability-corrected value across the GATB datasets, the Salgado et al.’s (2003) study, and the Bertua et al.’s (2005) study is .31. This is our estimate of cognitive ability test validity, in contrast with the .51 value presented by Schmidt and Hunter (1998).

The ordering changed as well. Structured employment interviews .42, job knowledge tests .40, empirically keyed biodata .38, work sample tests .33, and cognitive ability tests .31, in fifth place.

This is not the moment to hurry. A downward revision does not mean “does not matter.” A correlation of .31 is a non-trivial moderate association; the result is a change in relative standing, not a claim of zero. What Claim C asserts is that ability “barely matters,” which is an assertion about effect size, and assertions about effect size can be neither confirmed nor refuted without a measurement procedure.

Evidence pointing the same way comes from a different line. In organizational network studies, Casciaro and Lobo found that when people choose whom to work with, they favor being comfortable with someone over that person’s competence, and that once likability falls below a threshold, high competence can even carry a negative effect on the formation of task-related ties. What that study measures, however, is whether someone is chosen as a collaborator, not whether they are judged incompetent.

Claim C is neither supported nor refuted. The downward revision gives it a tailwind; it does not let it win.

Raters are not one bloc

The back half of Claim B, “are treated as incompetent,” reads as a prediction about a social fact. That can be tested. And the answer comes back in a shape the claim does not anticipate.

Rotundo and Sackett used a policy-capturing design in which 504 North American managers rated 34 hypothetical employees, measuring how the three dimensions of job performance (task, citizenship, counterproductive) are weighted into an overall rating. Raters fell into three clusters.

1. Raters who weight task performance most heavily
2. Raters who weight counterproductive performance most heavily
3. Raters who place equal and large weights on task and counterproductive performance

A one-cluster solution was not supported. “Raters all evaluate in the same way” is contradicted by the data.

More consequential is what failed to explain the clusters. From the paper:

the raters’ policies did not cluster according to the organizations they represented. Third, raters in the clusters did not differ significantly in terms of various demographic and background variables.

Claim B’s “treated as incompetent” holds for raters in clusters 2 and 3 and fails for cluster 1. Which kind of rater you draw cannot be predicted by choosing an organization or by looking at anyone’s attributes. The claim is one whose truth depends on the rater, with no means of identifying that rater in advance.

On average, both task and counterproductive performance carry more weight than citizenship. The claim extracts the counterproductive side alone and expands it into a definition of competence as a whole. A finding that each of three dimensions carries its own independent weight does not license a reduction to one.

What the search did not find

I looked for studies measuring the front half of Claim B directly: that people who cannot self-regulate or self-correct raise other people’s cognitive load.

There were none.

Twenty-one sources were confirmed across cognitive load theory, collaborative cognitive load theory, transactive memory, common ground, coordination cost, and interruption cost, and none places “a colleague’s capacity for self-regulation” as the independent variable with “my cognitive load” as the dependent variable. The nearest proxy is Kim and colleagues’ four-wave panel study showing that accumulated helping behavior leads to emotional exhaustion, but its independent variable is the frequency of helping, not the other person’s capacity for self-regulation.

The causal pathway at the heart of the claim is empirically untested.

Load arises in the relation, not in the person

There is a reason nothing turns up. Seen from cognitive load theory, load does not belong to an individual in the first place.

In Sweller’s formulation, extraneous load (load that does not contribute to learning the task) originates in instructional design, that is, in how information is presented. Intrinsic load is set by the interaction between the element interactivity of the task and the learner’s schemas.

The sharpest point is the expertise reversal effect demonstrated by Kalyuga and colleagues. The same material works as useful scaffolding for a novice and as redundant load for an expert. Hold the presentation constant, change the receiver, and the load reverses.

Whether a given utterance or artifact generates extraneous load is therefore settled not by the sender’s attributes but by the pairing of the sender’s presentation with the receiver’s prior knowledge. The claim attributes this relational quantity to the sender as an individual.

Does the extension to collaboration help that attribution? It does not. The collaborative cognitive load theory formulated by Kirschner and colleagues in 2018 introduces collective working memory together with its cost: transaction costs, the working memory resources consumed in exchanging information and building common ground among members. But transaction cost is treated as a function of group composition, the distribution of prior knowledge, and task complexity. It is not treated as a member’s trait. The 2011 experiment by Kirschner and colleagues likewise yields a conditional proposition (collaboration is more efficient than individual work when task complexity is high, and less efficient when it is low), not an argument locating load in a person.

A second point comes from inside cognitive load theory. The first principle Sweller and colleagues place in their account of human cognitive architecture is the borrowing and reorganising principle: taking information from others and reorganizing it into one’s own schemas. Asking others, consuming their attention, drawing out their explanations, is that principle in action. The claim converts use of the main channel of knowledge acquisition, as cognitive load theory itself describes it, into an indicator of incompetence.

Measurement of load is itself contested. The discriminant validity of instruments that separate the three load types remains in dispute. De Jong pointed out that germane load is inferred after the fact from the observation that learning occurred, a circularity. Skulmowski and Rey showed that two subjective load scales produce divergent results under identical learning conditions. Transferring a construct with that much internal dispute into the grounds for labeling a person’s ability is not supported by measurement theory.

When self-regulation is missing, who takes it on?

The theory that most strongly supports the claim sits inside the learning sciences.

Hadwin, Järvelä, and Miller distinguish three layers of regulation in learning: self-regulation, co-regulation, and socially shared regulation. Co-regulation names the process in which another person temporarily takes on part of the regulation.

When self-regulation does not operate inside an individual, the function does not vanish. It moves to someone else. The phenomenon the claim is reaching for can be described theoretically as that transfer.

Co-regulation, however, carries a built-in condition. The definition of scaffolding includes fading from the original paper by Wood, Bruner, and Ross onward. Puntambekar and Hübscher argued, from inside scaffolding research, that support without a plan for withdrawal should not be called scaffolding at all.

So far this favors the claim. Formulated as “a state that continually transfers regulation load to others without moving toward internalization lowers the learning productivity of the system,” it does not conflict with the learning sciences.

The problem is when that state can be diagnosed.

Whether withdrawal will happen cannot be read off a snapshot

In Alexander’s Model of Domain Learning, learners in the acclimation stage show fragmented knowledge and surface-level strategies. How long that stage lasts varies widely with domain and with support.

Meyer and Land’s threshold concept work says the same thing from another angle. Learners remain in a liminal state for extended periods, cycling through partial and unstable understanding and reverting to prior conceptions. That stumbling originates in what Perkins calls troublesome knowledge: knowledge that is counter-intuitive, alien, or tacit, which is a property of the concept.

A learner in a liminal state, by definition, demands a great deal of explanation and correction from those around them. The claim offers no reading of that demand other than as an indicator of “someone who cannot self-correct.” Separating concept-driven stumbling from person-driven stumbling requires independently observing how many other people stumble on the same concept, and the claim has no such procedure.

Nor do the learning sciences offer any guarantee that a snapshot in time predicts whether withdrawal will eventually occur.

A failure to self-correct may be a shortage of knowledge

When Claim C says processing ability does not matter, it tacitly assumes self-regulation and processing ability are separate variables. That assumption is doubtful.

Calibration, the accuracy of metacognitive monitoring at the core of self-regulation, depends on the amount of knowledge in the domain at hand. Dunlosky and Rawson showed experimentally that learners whose post-study self-evaluations are more overconfident go on to retain less. In Alexander’s framework, the very quality of strategy use changes with position inside the domain.

An observation that someone “cannot self-correct” may therefore be not the absence of self-regulation as an independent trait but a shortage of prior knowledge showing up as a calibration failure. The claim has no procedure for telling the two apart.

The consequence is ironic for the claim. By asserting “self-regulation over processing ability,” it recodes a remediable condition (missing prior knowledge) as a near-dispositional attribute (self-regulation). It works in the direction of hiding remediability.

Depending on others does not mean incompetence

The intuition that “someone who cannot work self-contained is low in ability” meets direct counter-evidence.

Nelson-Le Gall divided help-seeking in two: executive help-seeking, which asks for the answer itself, and instrumental help-seeking, which asks for limited assistance toward becoming able to solve the problem. In large college classes, Karabenick clustered help-seeking patterns and found that adaptive help-seekers score higher on learning outcomes and self-efficacy than both non-seekers and non-adaptive seekers.

What predicts outcomes is not how much someone leans on others but how they lean.

The opposite intuition, that people who actively seek corrective information perform better, also gets a cautious number in return. Anseel and colleagues meta-analyzed thirty years of feedback-seeking research and reported that the association between seeking behavior and performance is small. Any argument that places self-correction at the center of competence has to meet that result head on.

The Dunning-Kruger effect, often invoked as evidence that people cannot see their own incompetence, cannot be used as it stands either. Nuhfer and colleagues showed by Monte Carlo simulation that the quartile-mean graph typically used to display the effect is generated automatically by measurement error alone, as a regression artifact. Gignac and Zajenkowski pointed out that conventional tests confound the regression effect with the better-than-average effect, and reported that valid testing approaches failed to detect the effect in IQ data. A substantial part of the original causal interpretation has been refuted as a statistical artifact.

Practitioners are not watching the absolute level

Everything above comes from the academic side. Industry practice returns a more concrete set of corrections.

In the evaluation practice of strategy consulting, the claim holds conditional on hierarchy. At the analyst and associate level it broadly holds. The prototypical moment when a manager decides “I don’t want them staffed again” is when a junior’s numbers do not reconcile and the manager rebuilds the logic from scratch before it goes in front of the client, and the judgment forming in that manager’s head is closer to “working with this person eats my available hours” than to “this person is slow.”

Higher up it reverses. When an unexpected question lands in a client’s boardroom, whether you can rebuild the structure and answer on the spot is precisely speed of processing, and it decides whether the engagement continues or is lost. The same holds for precision in financial modeling and for depth of domain knowledge in heavily regulated industries. Claim C is too coarse unless it names the level it applies to.

The design studio returns a different correction. Juniors, by definition, raise other people’s load. Taken literally, the claim makes every junior incompetent, but that is not what managers on the ground are watching. They are watching how many times the same correction has now been given. Someone who drew the same correction ten times in month one, three times in month two, and none in month three becomes an investment. Someone still at ten after six months becomes a permanent internal cost.

What practice uses is not the absolute level of load but the rate at which it declines. The claim is written in absolute levels.

The studio adds a reversal by function. In production and design system maintenance, an artifact becomes the input to someone else’s work, so unclear work converts directly into downstream cost. In exploration and concept work, confusing people once is part of the job. A proposal that strikes everyone as clear and obviously right the moment they hear it is usually mediocre.

What happens the moment it becomes policy

Independently of whether the claim is true, what happens when it becomes a line item in an evaluation system can be predicted with some precision.

Load a construct you cannot measure into an evaluation, and it will be reduced to proxy indicators. The reductions are predictable too. Meeting hours, message counts, review round-trips, number of questions, number of spec changes. Ridgway laid out the dysfunctional consequences of performance measurement in 1956, and Campbell formulated in 1979 the law that the more a quantitative indicator is used for social decision-making, the more subject it is to corruption pressure. The moment rewards attach to an indicator, the indicator distorts behavior and loses its correlation with what it was meant to measure.

The people who end up rated highly are those who do not ask, do not check, and do not report problems.

This shows up as adverse selection. An artifact that is confidently wrong does not increase the reviewer’s immediate effort, so it tends to be rated well. Most accidents in front of a client come not from the junior who asks a lot of questions but from the junior who noticed something was wrong and submitted it silently.

And the criterion cannot be told apart from a preference for homogeneity. In 1977 Kanter named homosocial reproduction: managers under uncertainty select people like themselves in order to secure trust and smooth communication. The core of that explanation is the low cost of communicating. The claim’s criterion is functionally identical. Rivera documented how evaluators at elite professional service firms fold cultural similarity with a candidate into their assessment of ability under the heading of “fit.” Since fluency of communication is a function of similarity, low load becomes a proxy for homogeneity. The claim carries no procedure for making that distinction.

The hope that making the criterion explicit will reduce bias is also disappointed. Castilla and Benard reported the paradox that when an organization is presented as meritocratic, evaluators allocate rewards more unequally between men and women of equivalent performance. Displaying a criterion hands evaluators a certificate of objectivity.

An asymmetry of judgment remains. The party judging “other people’s cognitive load” is always the party on whom the load fell. Setting aside the coarseness of one’s own instructions and problematizing how the other person received them; attributing the other’s behavior to disposition rather than to time pressure or ambiguous direction; booking confusion caused by one’s own incomplete explanation as the other person’s failure to understand. The claim contains no mechanism for simultaneously measuring the quality of the judge’s instructions or the soundness of the work design.

What Norman rejected in The Design of Everyday Things was the convention of blaming people for accidents and errors. Carry the logic of the chapter titled “Human Error? No, Bad Design” over to organizations, and the comprehension cost others bear is set first by how the work is divided, how information is hidden behind interfaces, and how much information processing the design requires. Where module boundaries are poor, anyone who arrives raises everyone else’s load. The claim calls that structural quantity by the name of whoever touched it last.

Fixing the sign of friction is a further problem. De Dreu and Weingart’s meta-analysis found both task conflict and relationship conflict negatively related to team performance and satisfaction. De Wit and colleagues’ meta-analysis found task conflict positively related to performance under the condition that relationship conflict is low. The sign of friction is conditional, and the claim fixes it as negative unconditionally. What Edmondson showed is that teams with higher psychological safety engage in more learning behavior, including reporting errors. Pointing out an error reliably raises the cognitive load of the person being corrected.

Which way does AI push the claim?

Generative AI applies two opposing forces.

Start with what strengthens it. The irony of automation Bainbridge identified in 1983 is that automation removes the easy parts and leaves monitoring, verification, and exception handling to the human. Generative AI brings that structure into knowledge work. Generation cost collapses; verification cost does not. Someone who produces a great deal and does not verify their own output externalizes the verification cost onto others. Before AI, producing volume carried labor value in itself; now that the marginal value of the volume-producing portion is close to zero, verification capacity is what remains to be evaluated. In that sense the second step of the claim gets stronger after AI.

Now what weakens it. Once generative AI substitutes for speed of processing, observed output becomes a compound of individual processing ability and skill at using AI, and no measurement procedure separates the two. A claim that cannot be measured cannot be supported. Claim C becomes harder to test after AI, not easier.

One more change is the heaviest. In an AI environment, “other people’s cognitive load” can be implemented as proxy indicators computed automatically from logs. Review round-trips, size of correction diffs, recurrence rate of the same comment, acceptance rate of AI-generated artifacts. When an indicator is collected automatically, attached to rewards, and known to the people being measured, Campbell’s corruption pressure is maximized.

Rewriting the claim into a testable form

Taking the above together, the claim rewrites into three testable statements.

Rewrite 1 (causal pathway)

A deficit in self-regulation and self-correction increases other people’s cognitive load even when the level of processing ability, the clarity of instructions, the state of documentation, and workload are held constant.

The data required: an independent measure of processing ability; behavioral measures of self-regulation (time from comment to completed correction, recurrence count of the same class of comment); dependent measures of others’ load (ratings from multiple counterparts, response time); and control of confounders.

Rewrite 2 (time derivative)

The recurrence count of the same class of comment directed at a given person declines over time. A decline rate below a threshold, sustained over a defined period, is what justifies revisiting an assignment.

This is the axis industry actually uses, and by watching a rate of change rather than an absolute level it corresponds to observing fading.

Rewrite 3 (testing the attribution)

Holding receivers’ prior knowledge constant and presenting the same artifact to randomly assigned receivers, stable individual differences among senders reproduce, and those differences do not interact with receivers’ prior knowledge.

If that is shown, attributing load to individuals is justified. If it is not, load stays a relational quantity.

Disagreements that remain

Axis	Position A	Position B	Evidence needed to settle it
Where load belongs	A relational quantity arising from the sender’s presentation paired with the receiver’s prior knowledge (Kalyuga et al. 2003; Sweller 2010)	Group regulation failure is real and describable (Järvelä & Hadwin 2013; Kirschner et al. 2018)	Stability of load generation across senders, holding receivers’ prior knowledge constant
Does processing ability predict performance	It does; GMA is the strongest single predictor (Schmidt & Hunter 1998, r = .51)	It was substantially overestimated (Sackett et al. 2022, r = .31, fifth place)	Incremental validity studies under AI; a procedure separating processing ability from AI-use skill
Does friction raise or lower performance	Lowers it; both task and relationship conflict correlate negatively (De Dreu & Weingart 2003)	Raises it conditionally; positive when relationship conflict is low (de Wit et al. 2012)	Replication controlling relationship conflict level, hierarchy, and proximity of the outcome measure
Is depending on others incompetence	The manner of seeking help predicts outcomes (Nelson-Le Gall 1981; Karabenick 2003)	Permanent transfer lowers system productivity (the fading requirement in Puntambekar & Hübscher 2005)	Longitudinal observation of change in the amount and kind of support
Is self-regulation an independent trait	It can be described by a domain-general cyclical model (Zimmerman 2000; Pintrich 2004)	Calibration accuracy depends on domain knowledge (Dunlosky & Rawson 2012; Alexander 2003)	Correlation of self-regulation measures across domains within the same individuals
Can self-regulation be measured	Self-report instruments capture it (the operationalization of major SRL models)	Self-report and online measures diverge (Veenman et al. 2006)	Agreement rate between self-report and behavioral traces on the same task
The character of the “do not raise load” norm	A legitimate occupational norm of peer control	A jurisdictional claim about whose load gets counted (Abbott 1988)	Who operates the norm (peers or management), and the distribution of whose load is counted (upstream only, or both directions)

Conditions under which the claim is usable

Rather than rejecting it outright, the claim can function as a practical instrument when the following conditions hold.

• Roles and expectations are documented, and both parties can refer to the same document
• Load reports are collected from multiple counterparts in different relationships, rather than resting on a single evaluator’s judgment
• Load reporting runs in both directions: the load a reviewer imposes (vague comments, late-arriving direction changes, waiting time for review) enters the same ledger
• It is used as a relative indicator of change over time within one person, not as an instrument for attaching an absolute label of incompetence
• Processing ability is measured independently and explicitly separated as a control variable, so that only the effect of self-regulation is examined
• The record is returned to the people involved as material for redesigning interdependencies and module boundaries, rather than entering a personnel rating

When the last condition is not met, the claim stops being a definition of competence and becomes a procedure for calling a structural failure by an individual’s name.

Unverified items

• Empirical studies measuring directly how much another person’s regulation deficit raises a colleague’s cognitive load: none found within the search range [requires primary verification]
• A validated organizational instrument for measuring “load imposed on others” at the individual level: presumed not to exist, but an exhaustive check of I/O psychology instrument catalogs is needed [requires primary verification]
• Confirmation that “processing ability matters little” is not an artifact of range restriction in an already selected population [requires primary verification]
• Longitudinal data separating the direction of causation between regulation deficits and others’ load (before/after transfers, clarified instructions, workload adjustment) [citation to be confirmed]
• DOIs for Winne & Hadwin (1998) and Hadwin, Järvelä & Miller (2011) [citation to be confirmed]
• Publisher’s primary page for Hutchins (1995) (403; confirmed indirectly across bibliographic databases) [citation to be confirmed]
• Bibliographic record for Brooks (1975/1995) (no DOI; confirmed indirectly via publisher catalog) [citation to be confirmed]
• The specific validity coefficient in Barrick & Mount (1991) [citation to be confirmed]
• Any industry primary source that tracks and publishes review effort and rework cost at the individual level [to be confirmed]

References

Cognitive load theory

1. Sweller, J. (1988). Cognitive Load During Problem Solving: Effects on Learning. Cognitive Science, 12(2), 257–285. https://doi.org/10.1207/s15516709cog1202_4
2. Chandler, P., & Sweller, J. (1991). Cognitive Load Theory and the Format of Instruction. Cognition and Instruction, 8(4), 293–332. https://doi.org/10.1207/s1532690xci0804_2
3. Kalyuga, S., Ayres, P., Chandler, P., & Sweller, J. (2003). The Expertise Reversal Effect. Educational Psychologist, 38(1), 23–31. https://doi.org/10.1207/s15326985ep3801_4
4. de Jong, T. (2010). Cognitive load theory, educational research, and instructional design: some food for thought. Instructional Science, 38(2), 105–134. https://doi.org/10.1007/s11251-009-9110-0
5. Skulmowski, A., & Rey, G. D. (2020). Subjective cognitive load surveys lead to divergent results for interactive learning media. Human Behavior and Emerging Technologies, 2(2), 149–157. https://doi.org/10.1002/hbe2.184
6. Kirschner, F., Paas, F., & Kirschner, P. A. (2011). Task complexity as a driver for collaborative learning efficiency: The collective working-memory effect. Applied Cognitive Psychology, 25(4), 615–624. https://doi.org/10.1002/acp.1730
7. Kirschner, P. A., Sweller, J., Kirschner, F., & Zambrano R., J. (2018). From Cognitive Load Theory to Collaborative Cognitive Load Theory. International Journal of Computer-Supported Collaborative Learning, 13(2), 213–233. https://doi.org/10.1007/s11412-018-9277-y

Distributed cognition and transactive memory

8. Hutchins, E. (1995). Cognition in the Wild. MIT Press. ISBN 9780262581462 [citation to be confirmed: publisher page 403; matched across bibliographic databases]
9. Wegner, D. M. (1987). Transactive Memory: A Contemporary Analysis of the Group Mind. In B. Mullen & G. R. Goethals (eds.), Theories of Group Behavior (pp. 185–208). Springer. https://doi.org/10.1007/978-1-4612-4634-3_9
10. Ren, Y., & Argote, L. (2011). Transactive Memory Systems 1985–2010. Academy of Management Annals, 5(1), 189–229. https://doi.org/10.5465/19416520.2011.590300
11. Lewis, K., & Herndon, B. (2011). Transactive Memory Systems: Current Issues and Future Research Directions. Organization Science, 22(5), 1254–1265. https://doi.org/10.1287/orsc.1110.0647
12. Peltokorpi, V. (2008). Transactive Memory Systems. Review of General Psychology, 12(4), 378–394. https://doi.org/10.1037/1089-2680.12.4.378

Common ground and coordination cost

13. Clark, H. H., & Brennan, S. E. (1991). Grounding in Communication. In L. B. Resnick, J. M. Levine, & S. D. Teasley (eds.), Perspectives on Socially Shared Cognition (pp. 127–149). American Psychological Association. https://doi.org/10.1037/10096-006
14. Convertino, G., Mentis, H. M., Rosson, M. B., Carroll, J. M., Slavkovic, A., & Ganoe, C. H. (2008). Articulating common ground in cooperative work: content and process. CHI 2008. https://doi.org/10.1145/1357054.1357310
15. Brooks, F. P., Jr. (1975/1995). The Mythical Man-Month: Essays on Software Engineering (Anniversary Edition). Addison-Wesley. ISBN 978-0-201-83595-3 [citation to be confirmed: no DOI]
16. Cataldo, M., Herbsleb, J. D., & Carley, K. M. (2008). Socio-technical congruence. ESEM 2008. https://doi.org/10.1145/1414004.1414008
17. Gote, C., Mavrodiev, P., Schweitzer, F., & Scholtes, I. (2022). Big Data = Big Insights? Operationalising Brooks’ Law in a Massive GitHub Data Set. ICSE 2022. https://doi.org/10.1145/3510003.3510619

Cost of interruption

18. Mark, G., Gudith, D., & Klocke, U. (2008). The cost of interrupted work: more speed and stress. CHI 2008, 107–110. https://doi.org/10.1145/1357054.1357072
19. Mark, G., González, V. M., & Harris, J. (2005). No Task Left Behind? Examining the Nature of Fragmented Work. CHI 2005. https://doi.org/10.1145/1054972.1055017
20. Czerwinski, M., Horvitz, E., & Wilhite, S. (2004). A diary study of task switching and interruptions. CHI 2004. https://doi.org/10.1145/985692.985715

Self-regulation and co-regulation

21. Zimmerman, B. J. (2000). Attaining Self-Regulation: A Social Cognitive Perspective. In M. Boekaerts, P. R. Pintrich, & M. Zeidner (eds.), Handbook of Self-Regulation (pp. 13–39). Academic Press. https://doi.org/10.1016/B978-012109890-2/50031-7
22. Winne, P. H., & Hadwin, A. F. (1998). Studying as Self-Regulated Learning. In D. J. Hacker, J. Dunlosky, & A. C. Graesser (eds.), Metacognition in Educational Theory and Practice (pp. 277–304). Lawrence Erlbaum. [citation to be confirmed: DOI unverified]
23. Pintrich, P. R. (2004). A Conceptual Framework for Assessing Motivation and Self-Regulated Learning in College Students. Educational Psychology Review, 16(4), 385–407. https://doi.org/10.1007/s10648-004-0006-x
24. Järvelä, S., & Hadwin, A. F. (2013). New Frontiers: Regulating Learning in CSCL. Educational Psychologist, 48(1), 25–39. https://doi.org/10.1080/00461520.2012.748006
25. Panadero, E., & Järvelä, S. (2015). Socially Shared Regulation of Learning: A Review. European Psychologist, 20(3), 190–203. https://doi.org/10.1027/1016-9040/a000226
26. Hadwin, A. F., Järvelä, S., & Miller, M. (2011). Self-Regulated, Co-Regulated, and Socially Shared Regulation of Learning. In B. J. Zimmerman & D. H. Schunk (eds.), Handbook of Self-Regulation of Learning and Performance (pp. 65–84). Routledge. [citation to be confirmed: DOI unverified]
27. de Bruin, A. B. H., Roelle, J., Carpenter, S. K., & Baars, M. (2020). Synthesizing Cognitive Load and Self-regulation Theory. Educational Psychology Review, 32(4), 903–915. https://doi.org/10.1007/s10648-020-09576-4

Scaffolding and domain learning

28. Wood, D., Bruner, J. S., & Ross, G. (1976). The Role of Tutoring in Problem Solving. Journal of Child Psychology and Psychiatry, 17(2), 89–100. https://doi.org/10.1111/j.1469-7610.1976.tb00381.x
29. Puntambekar, S., & Hübscher, R. (2005). Tools for Scaffolding Students in a Complex Learning Environment. Educational Psychologist, 40(1), 1–12. https://doi.org/10.1207/s15326985ep4001_1
30. Alexander, P. A. (2003). The Development of Expertise: The Journey from Acclimation to Proficiency. Educational Researcher, 32(8), 10–14. https://doi.org/10.3102/0013189X032008010
31. Meyer, J. H. F., & Land, R. (2005). Threshold concepts and troublesome knowledge (2). Higher Education, 49(3), 373–388. https://doi.org/10.1007/s10734-004-6779-5
32. Lave, J., & Wenger, E. (1991). Situated Learning: Legitimate Peripheral Participation. Cambridge University Press. https://doi.org/10.1017/CBO9780511815355

Metacognition and calibration

33. Nelson, T. O., & Narens, L. (1990). Metamemory: A Theoretical Framework and New Findings. Psychology of Learning and Motivation, 26, 125–173. https://doi.org/10.1016/S0079-7421(08)60053-5
34. Dunlosky, J., & Rawson, K. A. (2012). Overconfidence Produces Underachievement. Learning and Instruction, 22(4), 271–280. https://doi.org/10.1016/j.learninstruc.2011.08.003
35. Veenman, M. V. J., Van Hout-Wolters, B. H. A. M., & Afflerbach, P. (2006). Metacognition and learning: conceptual and methodological considerations. Metacognition and Learning, 1(1), 3–14. https://doi.org/10.1007/s11409-006-6893-0
36. Kruger, J., & Dunning, D. (1999). Unskilled and Unaware of It. Journal of Personality and Social Psychology, 77(6), 1121–1134. https://doi.org/10.1037/0022-3514.77.6.1121
37. Nuhfer, E., Fleisher, S., Cogan, C., Wirth, K., & Gaze, E. (2017). How Random Noise and a Graphical Convention Subverted Behavioral Scientists’ Explanations of Self-Assessment Data. Numeracy, 10(1), Article 4. https://doi.org/10.5038/1936-4660.10.1.4
38. Gignac, G. E., & Zajenkowski, M. (2020). The Dunning-Kruger effect is (mostly) a statistical artefact. Intelligence, 80, 101449. https://doi.org/10.1016/j.intell.2020.101449

Feedback-seeking and help-seeking

39. Ashford, S. J., & Cummings, L. L. (1983). Feedback as an Individual Resource. Organizational Behavior and Human Performance, 32(3), 370–398. https://doi.org/10.1016/0030-5073(83)90156-3
40. Anseel, F., Lievens, F., & Levy, P. E. (2007). A Self-Motives Perspective on Feedback-Seeking Behavior. International Journal of Management Reviews, 9(3), 211–236. https://doi.org/10.1111/j.1468-2370.2007.00210.x
41. Anseel, F., Beatty, A. S., Shen, W., Lievens, F., & Sackett, P. R. (2015). How Are We Doing After 30 Years? Journal of Management, 41(1), 318–348. https://doi.org/10.1177/0149206313484521
42. Nelson-Le Gall, S. (1981). Help-Seeking: An Understudied Problem-Solving Skill in Children. Developmental Review, 1(3), 224–246. https://doi.org/10.1016/0273-2297(81)90019-8
43. Karabenick, S. A. (2003). Seeking Help in Large College Classes: A Person-Centered Approach. Contemporary Educational Psychology, 28(1), 37–58. https://doi.org/10.1016/S0361-476X(02)00012-7

Error management and psychological safety

44. Frese, M., & Keith, N. (2015). Action Errors, Error Management, and Learning in Organizations. Annual Review of Psychology, 66, 661–687. https://doi.org/10.1146/annurev-psych-010814-015205
45. Edmondson, A. C. (1999). Psychological Safety and Learning Behavior in Work Teams. Administrative Science Quarterly, 44(2), 350–383. https://doi.org/10.2307/2666999

Personnel selection and job performance

46. Schmidt, F. L., & Hunter, J. E. (1998). The Validity and Utility of Selection Methods in Personnel Psychology. Psychological Bulletin, 124(2), 262–274. https://doi.org/10.1037/0033-2909.124.2.262
47. Sackett, P. R., Zhang, C., Berry, C. M., & Lievens, F. (2022). Revisiting Meta-Analytic Estimates of Validity in Personnel Selection: Addressing Systematic Overcorrection for Restriction of Range. Journal of Applied Psychology, 107(11), 2040–2068. https://doi.org/10.1037/apl0000994
48. Sackett, P. R., Zhang, C., Berry, C. M., & Lievens, F. (2023). Revisiting the Design of Selection Systems in Light of New Findings Regarding the Validity of Widely Used Predictors. Industrial and Organizational Psychology, 16(3), 283–300. https://doi.org/10.1017/iop.2023.24
49. Rotundo, M., & Sackett, P. R. (2002). The Relative Importance of Task, Citizenship, and Counterproductive Performance to Global Ratings of Job Performance: A Policy-Capturing Approach. Journal of Applied Psychology, 87(1), 66–80. https://doi.org/10.1037/0021-9010.87.1.66
50. Barrick, M. R., & Mount, M. K. (1991). The Big Five Personality Dimensions and Job Performance: A Meta-Analysis. Personnel Psychology, 44(1), 1–26. https://doi.org/10.1111/j.1744-6570.1991.tb00688.x [citation to be confirmed: coefficient]
51. Curtis, B. (1981). Substantiating Programmer Variability. Proceedings of the IEEE, 69(7), 846. https://doi.org/10.1109/PROC.1981.12088

Knowledge-worker productivity and toxic workers

52. Drucker, P. F. (1999). Knowledge-Worker Productivity: The Biggest Challenge. California Management Review, 41(2), 79–94. https://doi.org/10.2307/41165987
53. Forsgren, N., Storey, M-A., Maddila, C., Zimmermann, T., Houck, B., & Butler, J. (2021). The SPACE of Developer Productivity. Communications of the ACM, 64(6), 46–53. https://doi.org/10.1145/3453928
54. Ko, A. J. (2019). Why We Should Not Measure Productivity. In C. Sadowski & T. Zimmermann (eds.), Rethinking Productivity in Software Engineering (ch. 3). Apress. https://doi.org/10.1007/978-1-4842-4221-6_3
55. Housman, M., & Minor, D. (2015). Toxic Workers. Harvard Business School Working Paper 16-057. https://dash.harvard.edu/handle/1/23481825 (unrefereed working paper)
56. Kim, A., Kim, Y., & Cho, Y. (2023). The consequences of collaborative overload: A long-term investigation of helping behavior. Journal of Business Research, 154, 113348. https://doi.org/10.1016/j.jbusres.2022.113348
57. Bergeron, D. M. (2007). The Potential Paradox of Organizational Citizenship Behavior. Academy of Management Review, 32(4), 1078–1095. https://doi.org/10.5465/amr.2007.26585791
58. Bolino, M. C., Hsiung, H-H., Harvey, J., & LePine, J. A. (2015). “Well, I’m Tired of Tryin’!” Organizational Citizenship Behavior and Citizenship Fatigue. Journal of Applied Psychology, 100(1), 56–74. https://doi.org/10.1037/a0037583

Competence, likability, and the social construction of evaluation

59. Casciaro, T., & Lobo, M. S. (2005). Competent Jerks, Lovable Fools, and the Formation of Social Networks. Harvard Business Review, 83(6), 92–99. https://pubmed.ncbi.nlm.nih.gov/15938441/
60. Casciaro, T., & Lobo, M. S. (2008). When Competence Is Irrelevant: The Role of Interpersonal Affect in Task-Related Ties. Administrative Science Quarterly, 53(4), 655–684. https://doi.org/10.2189/asqu.53.4.655
61. Kanter, R. M. (1977). Men and Women of the Corporation. Basic Books. homosocial reproduction. [citation to be confirmed: page]
62. Rivera, L. A. (2012). Hiring as Cultural Matching: The Case of Elite Professional Service Firms. American Sociological Review, 77(6), 999–1022. https://doi.org/10.1177/0003122412463213
63. Castilla, E. J., & Benard, S. (2010). The Paradox of Meritocracy in Organizations. Administrative Science Quarterly, 55(4), 543–576. https://doi.org/10.2189/asqu.2010.55.4.543
64. Abbott, A. (1988). The System of Professions: An Essay on the Division of Expert Labor. University of Chicago Press. jurisdiction. [citation to be confirmed: page]

Measurement dysfunction and friction

65. Ridgway, V. F. (1956). Dysfunctional Consequences of Performance Measurements. Administrative Science Quarterly, 1(2), 240–247. https://doi.org/10.2307/2390989
66. Campbell, D. T. (1979). Assessing the Impact of Planned Social Change. Evaluation and Program Planning, 2(1), 67–90. https://doi.org/10.1016/0149-7189(79)90048-X
67. De Dreu, C. K. W., & Weingart, L. R. (2003). Task versus Relationship Conflict, Team Performance, and Team Member Satisfaction: A Meta-Analysis. Journal of Applied Psychology, 88(4), 741–749. https://doi.org/10.1037/0021-9010.88.4.741
68. de Wit, F. R. C., Greer, L. L., & Jehn, K. A. (2012). The Paradox of Intragroup Conflict: A Meta-Analysis. Journal of Applied Psychology, 97(2), 360–390. https://doi.org/10.1037/a0024844
69. Norman, D. A. (2013). The Design of Everyday Things (Revised and Expanded Edition). Basic Books. Chapter 5, “Human Error? No, Bad Design.” [citation to be confirmed: page]
70. Bainbridge, L. (1983). Ironies of Automation. Automatica, 19(6), 775–779. https://doi.org/10.1016/0005-1098(83)90046-8