Proactive vs retroactive interference: which memory gets in whose way

Proactive interference is when older learning makes it harder to recall something you learned more recently. Retroactive interference is the reverse: newer learning makes it harder to recall something you learned before. The prefix tells you which way the interference travels — "pro" forwards in time, "retro" backwards — and both are at their worst when the two sets of material resemble each other.

The two terms are easy to define and easy to mix up, because the direction word describes the interfering memory rather than the one that fails. This article gives you a single question that sorts any example, and the history that makes the pair stick.

• Proactive interference: old learning blocks new recall. Retroactive interference: new learning blocks old recall.
• The fastest test is to ask which memory you can no longer retrieve, not which one is interfering.
• Interference is a normal property of memory, not an injury or a disorder. It is not the same thing as amnesia.
• Proactive interference was the first of the two to be studied, then almost forgotten, until Underwood put it back at the centre of everyday forgetting in 1957.
• Similarity makes interference worse. Whether similarity is strictly required has been argued about since 1900 and is still open.

The difference in one table

Both terms name the same underlying event — one set of learning competing with another — seen from opposite ends of the timeline.

Both examples in that table come from OpenStax's Psychology 2e, which defines proactive interference as "when old information hinders the recall of newly learned information" and retroactive interference as "when information learned more recently hinders the recall of older information".

The one question that sorts any example

Most people learn both definitions and still hesitate under exam conditions, because they try to track the interfering memory. Track the failing one instead.

Ask: which memory can I no longer get to?

• If the newer memory is the one that will not come, the older one is reaching forwards to block it. That is proactive interference.
• If the older memory is the one that will not come, the newer one is reaching backwards to block it. That is retroactive interference.

This works because the prefix names the direction the interference travels, and interference always travels towards the memory that fails. Once you have identified the casualty, the prefix follows automatically.

Notice that two situations involving phone numbers and passwords land on opposite answers. The content is not what decides it; the direction of the failure is.

Where retroactive interference came from

The idea arrived with a German name. Müller and Pilzecker, working in Göttingen in 1900, called it "rückwirkende Hemmung" — retroactive inhibition. The older literature says "inhibition" throughout; "interference" is the modern word for the same effects.

The study that made the idea famous came in 1924. Jenkins and Dallenbach set out to show that time alone does not cause forgetting. Their two student subjects lived and slept in the laboratory, learned lists of nonsense syllables at various hours, and were tested after intervals of either sleep or ordinary waking life. Forgetting was dramatically less after sleep. Their conclusion is quoted to this day:

Two things are worth saying plainly about that study. It had two subjects, so treat it as a striking demonstration rather than a large sleep experiment. And the sleep finding was not actually new: as C. M. MacLeod records in his history of interference theory, Heine had shown the same advantage of sleep over waking a decade earlier, in Müller's own laboratory in 1914. The later paper simply became the better known one.

Jenkins and Dallenbach's evidence was one of the pieces that led McGeoch to argue in 1932 that decay theory should give way to interference theory — that we forget mainly because of what else we learn, not because of the passage of time as such. Time still matters, and interference is best described as a major cause of forgetting rather than the only one.

Proactive interference: found first, noticed last

The less intuitive of the two was, oddly, the first to be studied. In 1893 and 1894 John Bergström published experiments on what he called the "interference of associations". He timed people sorting a deck of cards into piles, then reshuffled the piles and had them sort again. They were slower the second time, and they kept making the old sorting response. MacLeod describes Bergström as apparently the first to report empirical studies of interference in memory. Today we would call that effect proactive interference.

Then it largely disappeared from view for decades. The term "proactive inhibition" was apparently first used by Maslow in 1934, and the concept did not really come to the fore until Underwood's 1957 paper.

What Underwood did to the forgetting curve

Underwood looked back at a long series of experiments in which people learned a list and were tested on it a day later, and noticed something nobody had made much of: recall was worse the more lists the person had already learned in earlier sessions. His summary was blunt — "I interpret this to mean that the greater the number of previous lists the greater the proactive interference" (p. 53).

The consequence was larger than the observation. Psychology's picture of how fast we forget came from Hermann Ebbinghaus, who in 1885 had run enormous numbers of nonsense-syllable lists on himself. If prior lists cause forgetting, then Ebbinghaus's famously steep curve was measuring his own accumulated interference as much as the passage of time. Underwood argued that earlier estimates of the rate of forgetting were dramatic overestimates — MacLeod puts it at perhaps three times the actual rate — and concluded that "such interference must result primarily from proactive interference" (p. 55).

That is the part most explanations of these two terms leave out. The reason proactive interference matters is not that it is a tidy mirror image of the retroactive kind; it is that it quietly rewrote one of the most reproduced findings in the study of memory.

Proactive interference in seconds, not days

The same effect turned up in short-term memory. Peterson and Peterson reported in 1959 that recall of a three-consonant string fell from almost perfect after three seconds to below ten per cent after eighteen seconds of counting backwards, and took this as evidence of rapid decay.

Keppel and Underwood looked again in 1962 at just the first few trials of that task, and found "a severe proactive effect produced by a single prior item" (p. 156). Recall on the very first trial barely dropped at all. The dramatic forgetting curve was largely interference building up across trials, invisible because the trials had been averaged together.

A year later, Wickens, Born and Allen showed that changing the kind of material across trials made most of that interference disappear — the effect they were the first to call "release from proactive interference". It is a striking result, though it was demonstrated in short-term memory tasks with word lists, and it is not a study technique that has been tested on learning terminology.

Why similarity matters, and where the evidence disagrees

The rule of thumb is that interference is worst between things that resemble each other. Underwood put it strongly: "similarity with other material and situational similarity are by far the most critical factors in forgetting" (p. 58).

The awkward part is that the people who discovered retroactive interference disagreed. Müller and Pilzecker claimed in 1900 that the effect did not depend on the similarity of the two sets of material at all. Ranschburg argued the intuitive opposite in 1905, and Robinson concluded in 1920 that the degree of retroactive inhibition is a function of the similarity between the interpolated activity and the original learning — while also noting that material which looked quite dissimilar could still interfere.

That older view has not gone away. Dewar, Cowan and Della Sala argued in 2007 that retroactive interference also arises from nonspecific mental effort after learning, not only from similar material: in their account even a demanding tone-detection task, which introduces no new meaningful material, reduces recall relative to a quiet rest.

So state it carefully. Similarity reliably makes interference worse, and that is the part you can rely on. Whether similarity is strictly required has been contested since 1900 and is still open.

One further caution, because it is often flattened into a rule: Heine's 1914 experiments found retroactive interference when the final test was recall, but not when it was recognition. That is a specific early result from one extensive series of experiments, not a general law about test formats.

Interference is not amnesia

Four terms get confused here, and two of the confusions are just prefixes that rhyme. Interference happens in ordinary, healthy memory. The amnesias are memory losses that follow brain injury or disease.

"Retroactive" and "retrograde" are not variants of one word. Retroactive interference is competition between two ordinary memories; retrograde amnesia is the loss of memories formed before an injury. The amnesia definitions above are OpenStax's.

The misinformation effect is a third thing again. It describes how exposure to inaccurate information after an event can make a person misremember the event itself. It resembles retroactive interference in that something later disturbs something earlier, but it concerns the distortion of a memory rather than the failure to retrieve one.

Keeping the terms apart

Psychology terminology comes in near-synonym pairs that differ by one prefix or one word: proactive and retroactive, anterograde and retrograde, classical and operant conditioning, negative reinforcement and punishment. Learning them one at a time, in isolation, is what leaves you confident about each one and uncertain about which is which.

In Psychology World you can save the four terms from the table above to one list and drill them together: as flashcards, as multiple-choice questions once the list has four or more words, or by typing the answer out, which is the mode that really tests you. What each study mode does explains the differences. Our guide to memorising psychology terminology covers the retrieval and spacing methods with the best evidence behind them.