Where does control collide?

The mechanisms by which an organization gets its members to act toward its goals when their aims are only partly shared. Ouchi's 1979 A Conceptual Framework for the Design of Organizational Control Mechanisms distinguished controlling the way work is done from controlling its results, and later writers on project control, such as Kirsch's 1997 Portfolios of Control Modes and IS Project Management, group these into three formal and informal modes: behavior control (specifying and monitoring the steps), outcome control (setting targets and rewarding results), and clan control (relying on shared norms and socialization). This meaning is a social arrangement for steering people, not a feedback loop in a machine, a held-constant condition in an experiment, or a belief inside one person's head.[1][2]

Unit

One control mode applied to a task or project, as a mix of behavior, outcome, and clan control

Recovery means

Not applicable

Measured by

Assessed through surveys and case coding of which formal and informal controls a manager uses over a task

The regulation of a physical process so that a measured quantity tracks a chosen target, achieved by feeding the difference between the target and the measured value back into the input. Wiener's 1948 Cybernetics set out this feedback principle as a general theory of control and communication shared by machines and living systems, in which a controller continuously corrects for error. This meaning is a mechanism of measurement and correction acting on a signal, and it differs from a social scheme for directing people, a held-constant condition in a study, or a person's belief about their own influence, even though the management meaning later borrowed its vocabulary of feedback.[3]

Unit

One control loop holding a variable near its setpoint, scaled by the size of the residual error

Recovery means

The loop drives the error back toward zero after a disturbance, at a rate set by its gain and dynamics

Measured by

Characterized by the setpoint, the error signal, and the response of the output over time

In the design of experiments, control is the practice of holding other influences fixed or comparable so that an effect can be attributed to the factor under study. Fisher's 1935 The Design of Experiments formalized the comparison against a control condition and the removal of nuisance variation through randomization and blocking, which is why a study reports a control group and speaks of controlling for a variable. This meaning is a methodological safeguard against confounding, not a way of steering people, a feedback loop in a device, or a belief a person holds about their own power over events.[4]

Unit

One control condition or one variable held constant against which an effect is compared

Recovery means

Not applicable

Measured by

Built into a design through a control group, randomization, blocking, or statistical adjustment for covariates

A person's generalized belief about how far the outcomes they experience follow from their own actions rather than from luck, fate, or powerful others. Rotter's 1966 Generalized Expectancies for Internal Versus External Control of Reinforcement introduced this locus of control as a stable expectancy along an internal-to-external dimension, measured by a self-report scale. This meaning lives inside an individual as a perception, and it is distinct from an organization's control over its members, a controller acting on a machine, or a held-constant condition in an experiment.[5]

Unit

One person's placement on an internal-to-external control expectancy

Recovery means

Not applicable

Measured by

Scored by self-report scales such as Rotter's internal-external locus of control instrument