The History of Industrial Engineering: From Taylor to Tippett
From time studies to work sampling: different approaches, applications and limitations in industrial measurement.
Updated:
Taylor and Tippett: different questions about work
Work measurement history is not a linear replacement of stopwatches by a bias-free method. Time study, method study, and sampling still address different needs. Selection of workers, cycles, and conditions matters in all of them.
In The Principles of Scientific Management, published in 1911, Frederick W. Taylor presented his approach to systematic work study and organization. It is a historical source, not current validation of every proposal or justification for ignoring safety, autonomy, or employment relations.
The snap-reading contribution
L. H. C. Tippett published a 1935 article on snap-reading studies of machines and operatives in factory surveys: the original Journal of the Textile Institute Transactions reference. This dates the publication without inventing precise dates for every developmental stage.
The useful idea is observing states at selected instants rather than following every operation from start to finish. An appropriate sample can estimate the distribution of time. It does not automatically reconstruct each cycle's sequence or every event duration.
What statistics contributes
For a binary category, a binomial model requires independence and a common probability, or design and analysis that address their differences. Under that model, the state count has mean np and variance np(1−p). Industrial behavior itself is not therefore normally distributed.
Approximate planning n=z²p₀(1−p₀)/e² expresses a precision trade-off under assumptions. For example, p₀=0.35, z=1.96, and e=0.03 give 971.071…, rounding upward to 972 rather than 971 readings. This is a mathematical example, not a typical maintenance percentage. NIST documents the calculation.
Complementary methods
| Question | Required data |
|---|---|
| What fraction is observed waiting? | Representative readings and classification |
| How long does an operation take? | Start, finish, elements, and conditions |
| How many failures occur and how long do they last? | Complete events and exposure period |
| What is OEE? | Planned and run time, ideal cycle time, and counts |
Time study does not inherently imply ideal conditions, and sampling does not guarantee natural behavior. Reactivity and selection bias require investigation rather than a contest decided by method names.
Digital tools today
A PWA can support schedules, recording, and traceability. An algorithm does not make the observations independent, guarantee anonymity, or turn a human-activity indicator into OEE. Technology helps execute a protocol; it does not replace justification.
In WorkSamp, review categories, coverage, and reports, checking pending synchronization when working offline. Element-level times require a complementary time-study technique. Verify each tool's actual functions and commercial conditions against current documentation.
Practical lesson
Before selecting a technique, write down the quantity to estimate, the observational unit, and the decision the result will support. Retain definitions, versions, and limitations. These methods provide value through a measurable question and suitable evidence, not universal superiority attributed to a historical era.