Context-dependent vs state-dependent memory: the difference, and what it means for where you study

Context-dependent memory is better recall when your surroundings at the test match the surroundings where you learned. State-dependent memory is better recall when your internal state at the test, such as a drug state or a mood, matches your state when you learned. The difference is where the matching cue is: outside you or inside you. Divers who recall an underwater word list best underwater show context-dependent memory. People who recall material learned under alcohol better when they are under alcohol again show state-dependent memory. Both effects come from one rule, the encoding specificity principle.

The definitions are short. The exam questions are harder, because they mix in a third term, mood-congruent memory. Also, the most famous study of context-dependent memory failed a direct replication in 2021. Here is the distinction, two questions that sort the usual exam examples, and what the evidence means for where you study.

• Context-dependent memory uses external cues: the place, the sounds, the smells, or being underwater. State-dependent memory uses internal cues: a drug state, a mood or fear.
• Both are cases of the encoding specificity principle. A cue helps recall when it matches the memory trace.
• Mood-dependent memory is the mood form of state-dependent memory. Mood-congruent memory is a different effect: you recall material that matches the emotional tone of your current mood.
• Both effects are clearest in free recall. In the classic diver and alcohol studies, recognition tests showed little or no effect.
• The effects are real but usually modest. A 2001 meta-analysis found an average effect size of d = 0.28 for environmental context, and a 2021 replication of the diver study found no same-context advantage.

The difference in one table

A quick memory aid: context is the scene around you, and state is the condition inside you.

The principle behind both: encoding specificity

When you learn something, you store it together with the situation around it. Tulving and Thomson called this the encoding specificity principle in 1973. In the open textbook Noba, Kathleen McDermott and Henry Roediger summarise it: to the extent that a retrieval cue "matches or overlaps the memory trace of an experience", it "will be effective in evoking the memory".

Context-dependent and state-dependent memory are two applications of this one principle. In context-dependent memory, the cue is the environment. In state-dependent memory, the cue is your body or your mood. Students sometimes treat encoding specificity as a third, separate effect. It is not. Encoding specificity is the rule, and the two kinds of dependence are evidence for it.

The same chapter gives a limit. A retrieval cue "cannot be overloaded with too many memories". Watkins called this the cue overload principle in 1975. This helps to explain why a room can be a weak cue. If you learn hundreds of facts in one room, the room points to all of them at once.

Context-dependent memory: the diver study

In 1975, D. R. Godden and Alan Baddeley tested divers in two natural environments: on dry land and underwater. The divers learned lists of words in one environment. Then they recalled the words in the same environment or in the other one. The result, in the words of the paper's abstract: "Lists learnt underwater were best recalled underwater, and vice versa."

A second experiment checked a simple alternative. Maybe the divers did worse only because the move between land and water disrupted them. The authors concluded that this disruption "was unlikely to be responsible" for the result. The underwater sessions took place in cold open water near Oban, in Scotland, and each diver did all four combinations over four days, as Murre describes in his 2021 replication.

In 1980, Godden and Baddeley tested recognition. Divers learned lists of 36 words on land or underwater. Later they tried to pick out those words from a list of 72 words, in the same or in the other environment. This time, the authors observed "no trace of context dependency".

That contrast matters for exams. In free recall, such as an essay or a short written answer, you must find the answer yourself, and a cue helps you start the search. In a recognition test, such as a multiple-choice question, the answer is already on the page.

State-dependent memory: alcohol, drugs and mood

State-dependent memory is the internal version. A 2025 review by Liu and colleagues defines it as "better performance in memory retrieval when a person is in a similar state" to the state at learning. The review lists states such as "anxiety, fear, depression, sleep" and "states induced by alcohol or drugs".

The classic human study used alcohol. In 1969, Goodwin and colleagues gave male volunteers four memory tasks, either sober or under the effects of alcohol. Twenty-four hours later, the volunteers did the tasks again, in the same state or in a different one. In the recall tasks, the volunteers who were intoxicated in both sessions did better than the volunteers who were intoxicated only during learning. In the recognition task, the change of state had no significant effect.

Note what this comparison shows. It shows that a matching state helps recall. It does not show that alcohol helps memory.

Mood is the other common state. In a 1994 study by Eich, Macaulay and Ryan, people in a pleasant or an unpleasant mood thought of events from their own past in response to neutral words. Two or three days later, they tried to recall those events, in the same mood or in the other mood. Recall was better when the two moods matched. The researchers called this mood-dependent memory. They also proposed a "do-it-yourself" principle. The more you must use your own mind, not external aids, to produce the events and the cues, the more likely your memory is to be mood dependent.

The mood must also be real. In 2000, Eich and Macaulay asked students only to act happy or sad. They concluded that "only authentic affects have the power to produce mood-dependent effects".

Mood-dependent is not mood-congruent

This is a common exam trap. The two terms sound alike, but they compare different things.

• Mood-dependent memory compares two moods: your mood when you learned and your mood when you recall. A match helps recall, whatever the material is about.
• Mood-congruent memory compares your mood with the material. Holland and Kensinger describe it as the finding that information "congruent in valence to one's present mood" is more likely to be "perceived, attended to, stored, retrieved" than information that does not match.

Here is an example of each. When you feel sad and you remember more sad events than happy ones, that is mood-congruent memory: your mood matches the content. When you learn a neutral word list in a happy mood and recall it better the next time you feel happy, that is mood-dependent memory: your mood at recall matches your mood at learning.

So mood-dependent memory belongs under state-dependent memory. Mood-congruent memory needs no match between learning and the test. Holland and Kensinger also note that the congruence effect is often asymmetric: positive moods produce it more reliably than negative moods.

Where the line between context and state blurs

The two categories are clean in a textbook. In real experiments, they can mix.

Eich tested this in 1995. In three studies, he looked at how well memory transferred from one place to another. He concluded that the transfer "depends on how similar the environments feel rather than on how similar they look". When people learned and recalled in the same place but in different moods, their memory was worse. When they changed place but their moods matched, their memory was not worse. Eich concluded that data that look like place-dependent memory "may, at a deeper level, denote the presence of mood dependent memory".

Murre made a related point about the diver study itself. The original divers worked in cold open water. Murre suggested that their physiology possibly changed at depth, because of the higher pressure. In his words, that would be "evidence not so much for context-dependent memory but for state-dependent memory".

For an exam, call the diver study context-dependent memory. That is the standard answer. But the research shows that a change of place can change your state too.

Classifying exam scenarios

Two questions sort most examples:

1. 1.Is the matching cue outside you, such as a place, a sound or a smell? Then the example shows context-dependent memory. Is the cue inside you, such as a drug state or a mood? Then the example shows state-dependent memory.
2. 2.If the example is about mood, what matches? If your mood at recall matches your mood at learning, the example shows mood-dependent memory. If your mood matches the emotional content of the memory, the example shows mood-congruent memory.

The fourth and fifth rows are both state-dependent memory. Use the term mood-dependent memory for the fifth row when a question asks for the specific type.

How strong is the evidence?

The original diver result was large. From the statistics in the 1975 paper, Murre estimated a lower bound of d = 2.7 for the effect, which is "considered extremely large". For comparison, Murre notes that d = 0.8 is normally considered large.

But nobody published a direct replication for more than 40 years. Murre asked Baddeley in 2014, and Baddeley did not know of one. Murre's replication appeared in Royal Society Open Science in 2021. Sixteen divers heard word lists on land or underwater in a swimming pool. They recalled the words 4 minutes later, on land or underwater, and each diver did all four combinations. The result: "Contrary to the original study, we did not find that recall in the same context where the words had been learned was better than recall in the other context." Words learned on land were recalled better than words learned underwater, wherever the divers recalled them.

In January 2022, Murre published an addendum. A different statistical analysis, and the correction of one data entry, made the interaction between the learning place and the recall place statistically significant (p = 0.017). Murre still concluded that the errors "do not turn my study into a successful replication in the expected direction". In other words, the same-context advantage did not appear.

Murre lists several differences. His divers used a heated indoor pool about 180 cm deep, not cold open water. They did all four conditions in about 1.5 hours, not four days, which may have added proactive interference from the lists learned earlier. A TV crew also filmed the session. Murre concluded that more replications are necessary to establish whether the classic study deserves its place as the prime example of context-dependent memory.

The wider literature is steadier, but the effect is smaller. In a 2001 meta-analysis, Smith and Vela found that environmental context effects "were reliable". In a later paper, Smith and Handy describe them as "significant, but modest, with an average effect size of d = 0.28". Murre gives a similar figure, "around 0.25", for free recall.

State-dependent memory has limits too. According to Liu and colleagues, the effects are smaller than in free recall when a test gives strong support, "such as in recognition tests or heavily cued recall". The effects also tend to be weaker for highly overlearned material, and drug-induced effects often weaken over weeks.

There is one caveat about recognition. Murre notes that context-dependent recognition "was once thought to be non-existent". More recent studies found it under some conditions. So the safe statement is that recognition tests show a smaller effect, not that they never show one.

What this means for where you study

Should you study in the room where you will take the exam? The evidence gives three practical answers.

First, the room is a weak lever. The average effect is modest. Smith and Vela also found three things that reduce it: other useful cues during learning, other useful cues at the test, and mental reinstatement of the learning context at the test. This suggests that the room matters less when the test itself gives you other good cues, such as the wording of the questions.

Second, if the exam is in a different room, picture the place where you studied before you answer a recall question. That is mental reinstatement, one of the factors that reduced the context effect in Smith and Vela's meta-analysis.

Third, practise in more than one place. Smith and Handy taught people names for unfamiliar faces, with a video of a place in the background. In their first three experiments, each name had five retrieval practice trials. In one condition, the background was always the same. In the other, it was a new place on every trial. During practice, the constant background gave better scores. On the final test, with new backgrounds or none, the varied background gave better recall. In the first experiment, after two days, names practised in one constant context showed 47% forgetting. Names practised in varied contexts showed 3% forgetting.

That result has two limits. The advantage appeared only when people practised retrieval. In a fourth experiment with study trials alone, the two conditions did not differ. Also, Smith and Handy note that not all studies of varied contexts found an advantage. The study used videos and face–name pairs in a laboratory. The step to real study places is a reasonable interpretation, not a tested result.

Smith and Handy describe the aim as a memory that no longer needs its original context. A name first learned in one place depends on that place. After many encounters in many places, it becomes "a known fact". That is the move from an episode to semantic knowledge, which our article on how an episode turns into semantic knowledge describes. To get there, test yourself instead of rereading, and spread the sessions over time. Our guide to memorising psychology terminology shows how to combine both for an exam. Practise recall, not only multiple choice, because recall is where context and state effects showed most clearly.

Psychology World is built for this kind of practice. You save terms to your own lists and then drill a list in three ways: as flashcards, as multiple-choice questions once the list has four or more terms, or by typing the answer yourself. The Typing game is the recall practice. The app runs on iPhone, iPad and Mac, so you can do short sessions in different places, not always at the same desk. A daily word also brings one term back without a search. The four steps of the learning loop and what each study mode does are on the home page.