Dream journals: what a decade of entries actually reveals

Most quantitative dream research uses short collections from many people. A smaller and more unusual literature analyses very long series from single individuals, and it produces findings that group studies cannot.

The Barb Sanders series

The best-studied case is a woman given the pseudonym Barb Sanders, who kept a dream journal for over twenty years and provided more than 3,000 dreams to researchers, along with extensive information about her waking life.

G. William Domhoff and colleagues coded the series using the Hall–Van de Castle system and compared it against her biography. The central findings are worth stating because they bear directly on what dreams are.

Consistency. The relative frequency of characters, the proportion of aggressive versus friendly interactions, the emotional tone — these remained remarkably stable across decades. A person's dream profile turns out to be a fairly durable individual characteristic.

Continuity. Who appeared, and how they were treated in dreams, tracked her waking relationships closely. Her ex-husband appeared frequently and the interactions were disproportionately negative. Her children appeared in ways that matched the actual state of those relationships. Someone reading the dream statistics blind could reconstruct much of her social world.

Change where life changed. The stability was not total. Where relationships altered materially, the dream statistics altered with them, on a lag.

Other long series

Several other individuals have provided decade-plus records, and the pattern replicates. A blind woman's series, an engineer's, a teenage girl's followed into adulthood — each shows the same combination of individual consistency and correspondence with waking life.

Domhoff has made many of these available in the DreamBank, a searchable database of tens of thousands of dream reports that has enabled other researchers to test hypotheses against real data rather than anecdote.

The database has also enabled word-frequency approaches: rather than hand-coding, researchers count how often particular words and categories appear relative to a baseline, which produces surprisingly informative profiles quickly.

Why the findings are deflationary in a useful way

The practice of dream journalling is usually sold on a promise of hidden insight — that the accumulated record will reveal something the dreamer did not know.

What the quantitative analyses find is that the record reveals what the dreamer's life was actually about, in proportion. That sounds like less. In practice it is more useful, because the proportions are frequently not what the person would have reported.

People are poor at estimating where their attention goes. A journal that shows one person appearing in a third of dreams, or that shows anxiety about a situation the dreamer considered settled, is providing information — not a coded message, but a measurement.

Keeping one that would be worth analysing

If the point is a record rather than an aesthetic exercise, a few things matter.

Record every morning, including blanks. The denominator matters. A journal containing only vivid dreams is a biased sample.

Write immediately, without moving. Dream memory decays within minutes and posture changes disrupt retrieval.

Note the emotion separately. It is the most stable and most informative element and it is frequently lost when only events are recorded.

Name the characters consistently, or use consistent initials. Otherwise counting is impossible later.

Do not interpret in the entry. Keep the report and any reflection separate, or the reflection contaminates future recall — a version of the eyewitness memory problem.

Note waking context briefly — anything significant that happened that day or week. Without it, continuity cannot be assessed.

The dream-lag effect, which a journal can show you

One finding a personal journal can actually demonstrate. Incorporation of waking material into dreams shows two peaks: one at one to two days after the event, and a second at around five to seven days.

The delayed peak was identified by Mark Blagrove and colleagues and appears to relate to memory consolidation processes operating over that timescale. It is one of the more specific and testable findings in dream research.

With a journal that records both dreams and daily events, an individual can look for it in their own data — which is a considerably more interesting exercise than looking up what a staircase means.

The caution

Journalling reliably increases dream recall, often substantially, within a fortnight. For most people this is the point.

For someone with frequent nightmares it may not be. Amplifying recall of distressing dreams is not obviously beneficial, and the relevant intervention there is imagery rehearsal therapy rather than more careful documentation.