Negative Transfer

What it is

Thorndike and Woodworth published their work on the influence of improvement in one mental function upon the efficiency of other functions in Psychological Review in 1901, in three parts. What they established was that transfer is narrow and specific, which demolished the formal-discipline theory of education that preceded them and founded the study of transfer of training as a subject.

Negative transfer sits inside that tradition alongside proactive and retroactive interference. The systematic treatment of it as a function of stimulus and response similarity came later, in the interference literature, rather than from Thorndike and Woodworth themselves, which is why this publication records medium confidence on the originator credit and points to the tradition as much as to the founding paper.

The mechanism is not mysterious. Where a new task presents stimuli that resemble an old one but demands a conflicting response, the practised response arrives first, because practice is exactly what made it fast. The better the old skill, the stronger the interference.

In effect

The construct is standard in instructional design and in human factors, and it underpins interface design conventions and a good deal of training doctrine. The everyday cases are familiar: a driver moving between left-hand and right-hand traffic, a typist moving between keyboard layouts, a clinician moving between two drug charts laid out differently.

The cleanest field case in this library comes from Hasard Lee, who uses the term correctly at page 95 of The Art of Clear Thinking and cites nobody for it. In the 2017 F-35 scaling problem, pilots arriving from the F-16, F-15, A-10 and F-22 brought habits that were a detriment in the new aircraft, and unlearning those habits took more time than teaching a less experienced pilot from scratch. The syllabus answer was a custom course by prior airframe, smaller classes, and grouping students by the background they came from.

That is an unusually clean observation, because the population is large, the prior skills are documented and the outcome is measured in whether people qualify. It is still a practitioner's report rather than a measurement, and this publication cites the learning literature for the effect and Lee only for the case.

What it does not say

It does not say that experience is a liability in general. Transfer is more often positive than negative, and the same specificity that produces interference is what makes a well-matched prior skill useful. The prediction is about conflicting responses to similar cues, not about experience as such.

It does not say that the F-35 case is evidence for the effect. Nothing in that account was measured against a comparison group, no figure for the extra time is given, and a reader who met the idea only there would have no way of knowing a century of research existed behind it.

It does not say that unlearning is a matter of effort. The whole point of the construct is that the old response is automatic, which is why the standard remedies are structural: separate the cues, change the interface, or teach the new response in conditions that do not resemble the old one.

---