Assessing the Day-Residue and Dream-Lag Effects Using the Identification of Multiple Correspondences Between Dream Reports and Waking Life Diaries

Abstract

Several studies have found a high incorporation of waking life events into dreams that occur during the following night (day-residue effect), then a decrease in incorporation into dreams over the next 2 to 4 nights, followed by a resurgence of incorporation into dreams 5 to 7 days after events (dream-lag effect). These studies involve dream diary and daily diary keeping across a 1 to 2 week period, after which participants or judges give a single rating to the degree of correspondence between each dream report and each diary record. In the current study, participants (3 males, 11 females; mean age = 50.62 years) rated separately the intensity of as many correspondences as they could identify between each dream report and each diary record. From these multiple ratings, summary variables, including total number and total intensity of correspondences, were computed for periods between the daily diary and occurrence of the dream of 1 to 10 days. The dream-lag effect was not found. The day-residue effect was found for a group (n = 7) defined as having identified a below median total number of correspondences across the study. It appears that individuals who identify large numbers of correspondences dilute the day-residue effect. Suggestions are made for personality characteristics of such individuals, who display what may be akin to a Barnum effect in their response to the comparison of dream reports to daily diary records.

Full text

Assessing the Day-Residue and Dream-Lag Effects Using
the Identification of Multiple Correspondences Between
Dream Reports and Waking Life Diaries
Josephine A. Henley-Einion and Mark T. Blagrove
Swansea University
Several studies have found a high incorporation of waking life events into
dreams that occur during the following night (day-residue effect), then a decrease in
incorporation into dreams over the next 2 to 4 nights, followed by a resurgence of
incorporation into dreams 5 to 7 days after events (dream-lag effect). These studies
involve dream diary and daily diary keeping acrossa1to2week period, after which
participants or judges give a single rating to the degree of correspondence between
each dream report and each diary record. In the current study, participants (3 males,
11 females; mean age ⫽50.62 years) rated separately the intensity of as many
correspondences as they could identify between each dream report and each diary
record. From these multiple ratings, summary variables, including total number and
total intensity of correspondences, were computed for periods between the daily diary
and occurrence of the dream of 1 to 10 days. The dream-lag effect was not found. The
day-residue effect was found for a group (n⫽7) defined as having identified a below
median total number of correspondences across the study. It appears that individuals
who identify large numbers of correspondences dilute the day-residue effect.
Suggestions are made for personality characteristics of such individuals, who display
what may be akin to a Barnum effect in their response to the comparison of dream
reports to daily diary records.
Keywords: dream, dream-lag, day-residue, overinclusion, Barnum effect
Several studies have addressed the relationship between dream content and
the events and occurrences of waking life, as a function of time between the waking
life events and the occurrence of the dream (Blagrove, Fouquet, Henley-Einion,
Pace-Schott, Davies, Neuschaffer, & Turnbull (2011);Blagrove, Henley-Einion,
Barnett, Edwards, & Seage, 2011;Nielsen, Kuiken, Alain, Stenstrom, & Powell,
2004;Nielsen & Powell, 1989,1992). In general the finding has been of a U-shaped
function, such that waking life events tend to be referred to in the dreams of the
following night, and in dreams occurring 5–7 days after waking life events, but less
Josephine A. Henley-Einion and Mark T. Blagrove, Sleep Laboratory, Department of Psychology,
Swansea University, Swansea, United Kingdom.
Correspondence concerning this article should be addressed to Mark T. Blagrove, Sleep
Laboratory, Department of Psychology, Swansea University, SA2 8PP, Swansea, United Kingdom.
E-mail: m.t.blagrove@swansea.ac.uk
This document is copyrighted by the American Psychological Association or one of its allied publishers.
This article is intended solely for the personal use of the individual user and is not to be disseminated broadly.
71
Dreaming © 2014 American Psychological Association
2014, Vol. 24, No. 2, 71–88 1053-0797/14/$12.00 DOI: 10.1037/a0036329

so for dreams occurring 2–4 days, or more than 7 days, after the waking life events.
The immediate incorporation of events from waking life into dreams has been
termed the day-residue effect, and the 5–7 day delayed incorporation into
dreams has been termed the dream-lag effect. The dream-lag effect has been
claimed to indicate a memory processing function for sleep (Nielsen &
Stenstrom, 2005), and, in particular, Rapid Eye Movement (REM) sleep
(Blagrove, Fouquet, et al., 2011).
In these studies the method has involved participants recording reports of their
dreams, and completing a daily diary, so as to covera1to2week period, depending
on the details of the design. That initial data collection method has been followed
here, but the current study differs from previous ones in the manner of the rating
of similarities between dream reports and daily diary records, in that here
participants can identify multiple correspondences between each dream report and
each daily diary record. So as to explain the current rating method, a review of
previous rating methods follows.
In Blagrove, Henley-Einion et al. (2011) participants were instructed “to
match each journal entry with each dream report” using a 6-point scale, where 0 ⫽
no matches at all between dreams and events and 5 ⫽multiple strong matches
between dreams and events.InBlagrove, Fouquet et al. (2011) participants rated
“how much correspondence there is between the dream and each diary day” using
a scale from 0 ⫽none to 4 ⫽extremely high. Although participants in these studies
provided ratings that, when analyzed statistically, showed the day-residue and
dream-lag effects, it is unclear how participants were making these global judg-
ments of one score to represent the overall correspondence between a diary record
and a dream report, given that participants could be identifying multiple corre-
spondences between a dream report and a diary record.
These two studies allowed for participants or judges to take into account that
more than one event from a day, as recorded in a diary, could be matched to various
parts of the dream report, even though only one global summary score was
recorded. However, in contrast, some previous work on the dream-lag allowed the
dream report to be compared with only one wake life event. For example, in
Nielsen et al. (2004) participants were asked to consider “correspondences between
their selected dream and events occurring on a target day prior to the dream...
Participants were asked to recall events (‘everything that you did on this day’) from
one of seven randomly determined days prior to their dream . . . Participants were
also asked to write down the event from the target day that seemed most closely
related to their dream. Finally, they were asked to rate the apparent correspon-
dence between their chosen dream and this target event (‘What is the extent of
correspondence between your event and any part of your dream?’).”
Similarly, in the first experiment in Nielsen and Powell (1989), the daily diary
record was of “brief” descriptions of events, which would be “one or two sentences
in length.” However, in their second experiment, the waking life experience that
home dreams were compared with was the experience of having slept in the sleep
lab, with the dream reports being compared with “any aspect of the lab procedures
or physical environs” by two judges who had been experimenters in the laboratory
study, and who thus knew the details of the procedures used.
A further method was used in Nielsen and Powell (1992). On each evening for
14 days, participants recorded their emotionally meaningful daytime events. “One
This document is copyrighted by the American Psychological Association or one of its allied publishers.
This article is intended solely for the personal use of the individual user and is not to be disseminated broadly.
72 Henley-Einion and Blagrove

judge then selected a negative daytime event from each participant’s diary without
any reference to the dream reports and according to the following criteria: (a) the
event seemed relatively important, (b) the description of the event was relatively
specific, and (c) the event occurred relatively early in the 14-day recording period.”
Two judges then made ratings for the incorporation of the event into dream reports
recorded after the event.
The above studies did find day-residue and dream-lag effects, except in Nielsen
and Powell (1992), where limited support was found by an experienced judge, and
no support by an inexperienced judge. However, the literature has not addressed
whether some of the above methods are more optimal for the identification of the
day-residue or dream-lag effects. The literature has also not addressed how, if the
consideration of more than one event per diary day is allowed, participants or
judges combine heuristic judgments of the relative weights of several correspon-
dences if more than one is present in a diary record/dream report pairing.
A further question also arises, of whether participants or judges are influenced
by the length of a dream report, or the amount of text in the dream report not
related to any waking life events in a diary record, in deciding on a correspondence
score. This question is highly pertinent given that, according to Saredi, Baylor,
Meier, and Strauch (1997), just 17.3% of dream text refers to waking life concerns,
and, in Edwards, Ruby, Malinowski, Bennett, and Blagrove (2013), waking life
memory sources were found for only 14.4% of dream report text. The impact of
such unrelated dream report text on the ratings by participants in the Blagrove,
Henley-Einion, et al. (2011) and Blagrove, Fouquet, et al. (2011) studies is indeed
unknown, and may be a source of uncontrolled variance for the ratings.
The reasoning for not wishing participants to take into account text in a dream
report that is unrelated to the diary record is as follows: Suppose that a participant
saw their friend in the day and the friend appeared briefly in the dream; there might
be a score x given to this level of correspondence. If the participant saw their friend
in the day and the friend appeared in the dream and said many of the things that
they also talked of in the day, a higher score of x ⫹y might then be given. But, if,
as part of these dream reports there was then a long scene unrelated to the day or
the friend, we propose that that extra text should not lead to a decrease in either
of these scores, just as it would not decrease the scores if that portion of dream text
had been forgotten on waking. Indeed, in Nikles, Brecht, Klinger, and Bursell
(1998) a method of calculating incorporation of current concerns into dream
reports used a scoring method that did not take into account any parts of the dream
that were unrelated to the waking concerns. This was done by using a 0–5 scale that
varied from 5 ⫽the dream report makes direct reference to the concern by describing
or mentioning the concern directly...thedream must have the same meaning as the
concern and uses some of the same language as the concern . . . and 0 ⫽no
relationship at all. Participants and judges performed this rating exercise, but dream
reports were found to refer only very minimally to waking life concerns, even using
a method that only requires the concern to be referred to somewhere in the dream
report. It can indeed be argued that these minimal incorporation findings give a
further reason for not wishing to dilute correspondence scores by taking into
account dream text that is unrelated to the daily diary record.
The current study thus has a first aim of using a method where participants
identify for each dream report and each diary record all instances of a correspon-
This document is copyrighted by the American Psychological Association or one of its allied publishers.
This article is intended solely for the personal use of the individual user and is not to be disseminated broadly.
Comparing Dream Reports and Daily Diaries 73

dence between part of the dream report and part of the diary record, and rate the
strength of each correspondence separately. This method has the advantage of
ensuring that, in producing a score, participants do not have regard to text in a
dream report, or the amount of text in a dream report, that is unrelated to the diary
record. We note that Marquardt, Bonato, and Hoffmann (1996) also allowed
participants to identify multiple waking life sources of dream reports. However, in
that study the waking life sources had not been recorded in a daily diary, but were
instead elicited by association to the dream content.
The second aim is to assess and compare different methods of deriving a single
summary score, for the comparison of a dream report with a daily diary record,
from these multiple correspondences. The variables to be computed that summa-
rize a set of correspondences are: (a) the sum of correspondence intensities; (b) the
number of correspondences identified, irrespective of their intensities (following
Marquardt, Bonato, and Hoffmann, 1996); (c) the percentage of text in the dream
report for which waking life correspondence(s) are identified; and (d) the sum of
the products of correspondence intensities and length of dream text referred to by
the correspondence, divided by the total dream length.
These summary variables are computed for each diary day to dream period,
resulting in single summary scores for each diary day to dream period, as in the
previous studies. As only similarities in text are marked by participants, instances
of dream texts and diary record texts that are not related do not affect these
variables.
The third aim of the study is to investigate whether these new methods of
deriving variables that quantify the correspondence between dream reports and
daily diary records would identify a dream-lag with a greater effect size than that
found in Blagrove, Henley-Einion, et al. (2011) and Blagrove, Fouquet, et al.
(2011). The final, fourth aim of the study is to investigate whether individual
differences in the number of correspondences identified by participants should be
taken account of in the statistical analyses. The multiple correspondences approach
allows these aims to be specified and to be addressed.
METHOD
Participants
Fourteen participants (11 female, three male; mean age ⫽50.62 years, SD ⫽
11.16, minimum ⫽31 years, maximum ⫽65 years; one participant did not provide
an age) were recruited to the study and completed the daily diaries and the dream
diaries for 14 days and provided scores for the matching of all their daily diary
records with all their dream reports. Recruitment was elicited by e-mail advertise-
ment through the University, and posters and flyers in the University and in cafes
and bars in the South Wales area. In addition, articles about the study eliciting
recruitment were printed in two United Kingdom national publications: the Mensa
magazine and the Writers Magazine. These magazines were chosen as subscribers
were considered potentially to be more interested in the research than are the
general public. A Web site, a post on a knitting social network, and a Facebook
group were also created to publicize the study. Recruitment numbers through these
This document is copyrighted by the American Psychological Association or one of its allied publishers.
This article is intended solely for the personal use of the individual user and is not to be disseminated broadly.
74 Henley-Einion and Blagrove

sources were: Mensa, n⫽7; writers, n⫽5; Web sites, n⫽2. The study was carried
out in the participants’ own homes; 13 participants were based in the United
Kingdom and one in the Republic of Ireland.
Procedure
Ethics approval for the study was obtained from the Research Ethics Com-
mittee of the Department of Psychology, Swansea University. The study was
conducted in accordance with the APA and British Psychological Society ethical
standards for the treatment of human participants. Respondents to the advertise-
ments were screened so as to meet criteria for being frequent dream recallers
(defined as recalling dreams on “most nights”), having English as their first
language, and not having excessive alcohol intake (i.e., no higher than the United
Kingdom Government recommended consumption of 21 units per week for males
and 14 for females). Heavy drinkers were excluded due to the possibility of alcohol
affecting memory of dreams and also diminishing the amount of REM sleep.
Several participants did mention drinking alcohol in their diaries, but none to
excess.
Participants gave written informed consent to take part and were then given a
14-day daily diary and a 14-day dream diary to complete. After completion of the
diaries the participants returned them to the experimenter so that scoring docu-
ments could be compiled and sent to the participants. For the scoring, participants
were required to compare each daily diary record with each dream report, so as to
identify any correspondences between each diary record and each dream report,
and to give an intensity rating for each correspondence.
Materials
Daily Diary
Participants were asked to complete a diary record during the evening of each
day. They were given instructions to describe their day in a narrative description of
what occurred during each hour of the day. The hours were listed along the left
hand side of the page from 07:00 to 24:00 and then “Later than midnight.” They
were given the instruction:
Please describe the events of today as fully as you can. Your report should contain a
description of what you did during each hour of the day, where you were, the people you
were with, conversations, and any feelings/emotions you had. Continue your diary report on
the other side and on additional sheets if necessary.
Participants were instructed to record any emotions they felt in two separate
columns to the right of the narrative, one column for the name of the emotion and
one column for an intensity rating. Participants rated the intensity of each emotion
on a 1–3 scale (1 ⫽low;2⫽medium;3⫽high).
This document is copyrighted by the American Psychological Association or one of its allied publishers.
This article is intended solely for the personal use of the individual user and is not to be disseminated broadly.
Comparing Dream Reports and Daily Diaries 75

Dream Diary
Participants were asked to write down a report of any dreams they had in the
night. The instructions were:
Please describe the dream exactly and as fully as you remember it. Your report should
contain, whenever possible: a description of the setting of the dream, whether it was familiar
to you or not; a description of the people, their age, sex, and relationship to you; and any
animals that appeared in the dream. If possible, describe your feelings during the dream and
whether it was pleasant or unpleasant. Be sure to tell exactly what happened during the
dream to you and the other characters. Continue your report on the other side and on
additional sheets if necessary.
Again, names of any emotions were recorded in a separate column on the right
of the narrative, and participants gave an intensity rating of 1–3 (1 ⫽low;2⫽
medium;3⫽high) for each emotion.
Participants were offered the diaries in electronic or hard copy format. A
Microsoft Word document containing a single instruction page for the daily diary,
14 copies of the daily diary template, a single instruction page for the dream diary
and 14 copies of the dream diary template was produced. An example of a
completed diary record and dream report was given so that participants could see
the type of information that could be reported: This was the experimenter’s own
diary and dream report, produced for this study. Eleven participants completed the
diary electronically by typing in text and three participants requested hardcopy
diaries they could complete by hand.
Materials and Method for Recording Ratings of the Correspondences Between
Diary Records and Dream Reports
After all the daily diaries and dream reports were completed by each
participant they were returned to the experimenters either by e-mail or by hard
copy. Between 1 week and 1 month following the experimenters’ receipt of these,
the participant was sent a package containing A3 size (42 cm ⫻29.7 cm) sheets of
photocopied, randomized diary records and dream reports with instructions for the
correspondence task. Each sheet comprised one diary record and one dream report.
In addition to the order of the dream reports for each diary day being randomized
(up to the maximum 14 dream reports per diary day), the order of the diary day
bundles (maximum of 14 bundles of 14 diary record/dream report sheets per
participant) was also randomized. The maximum possible number of diary
record/dream report sheets for any one participant was 196.
Participants were instructed to underline or highlight on each diary
record/dream report sheet any text in the diary record and text in the dream report
that constituted a correspondence, and to give each such pairing of texts a unique
code. Participants were encouraged to identify as many or as few correspondences
as they thought were in evidence, and were told that it would be acceptable for
many sheets to be marked to indicate there were no correspondences. They then
gave an intensity score for each correspondence in answer to the following
question, using an 8-point scale: What is the extent of correspondence between this
This document is copyrighted by the American Psychological Association or one of its allied publishers.
This article is intended solely for the personal use of the individual user and is not to be disseminated broadly.
76 Henley-Einion and Blagrove

part of your dream and the part of your diary record for the day? 0 ⫽none;2⫽
weak;4⫽moderate;6⫽high; 8⫽extremely high.
Participants were given training for this task by instructions that asked them to
consider different types of correspondence, including literal, weak, personal, and
symbolic correspondences. Explanations and examples of each of these were given
with the instructions. They were also given an example of a completed diary
record/dream report sheet with underlined texts and correspondences and an
example of a completed score sheet for the ratings of the correspondence
intensities. Participants were advised to take as much time as needed, and to have
as many breaks as required, as it was expected that this would be a large and
time-consuming task.
Length of Time Between Diary Days and Dream Occurrence
Each diary record/dream report sheet was categorized in terms of the number
of days between the diary day and the occurrence of the dream; this could vary from
⫺13 (the dream occurred 13 days before the diary day) and ⫹14 (the dream
occurred 14 days after the diary day). In this terminology Day ⫺1 means the dream
occurs during the night before or on the morning of the diary day, and Day 1 means
that the dream occurs on the night or morning after the diary day. (There is no Day
0.) Summary correspondence variables for each of the diary day to dream periods
(Day ⫺13 to Day ⫹14) were computed for each of the 14 participants. As the
period between diary day and dream occurrence increases there are fewer possible
diary record to dream report combinations that can be used to calculate the
summary variables for that period, with a maximum of 14 combinations contribut-
ing to the diary to dream period of 1 day, and a maximum of one combination for
the period ⫺13 days and one combination for the period 14 days. As some
participants returned fewer than 14 dream reports, some of the periods at the
extremes had no diary record/dream report sheets to derive those values. The low
number of possible combinations for matching diary records to dream reports when
these are about 2 weeks apart resulted in some participants not having a score for
the end points of ⫹14 days (i.e., the period where the diary day was 14 days before
the dream) and ⫺13 days (i.e., the period where the diary day was 13 days after the
dream). It was decided to use the periods from ⫺10 to ⫹10 days for timecourse
analyses, as the mean calculated for each day would be derived from several sheets,
thus avoiding low sample sizes for the number of sheets.
Statistical Analyses
Four methods were used to derive a single score that summarizes the level of
correspondence between diary record and dream reports for each diary day to
dream occurrence period:
1. The sum of correspondence intensity scores between daily diary and dream
report per diary record/dream report sheet.
This document is copyrighted by the American Psychological Association or one of its allied publishers.
This article is intended solely for the personal use of the individual user and is not to be disseminated broadly.
Comparing Dream Reports and Daily Diaries 77

2. The number of correspondences identified between daily diary and dream
report, irrespective of intensity score, per diary record/dream report sheet.
3. The number of words in each dream report identified as incorporating
elements from the diary day, expressed as a percentage of the total number
of words in the dream reports for that diary day to dream period.
4. The sum of the individual products of correspondence intensity score and
number of words in the dream expressing that correspondence, divided by
the total number of words in the dream reports for that diary day to dream
period.
These four variables are then expressed as a function of time between diary
day and dream occurrence. For descriptive purposes the data are presented for all
the separate periods where a dream follows the comparison waking day (Day 1 to
Day 10), with the days where dreams precede the comparison day grouped into a
baseline period of Days ⫺10 to ⫺1. For inferential statistics Wilcoxon tests are used
to test for the day-residue effect (comparing Day 1 with the mean of Days ⫺10 to
⫺1, and with the mean of Days 2 to 4) and to test for the dream-lag effect
(comparing the mean of Days 5 to 7 with the mean of Days 2 to 4). As the direction
of the day-residue effect and the dream-lag effect are predicted in advance,
one-tailed tests are used throughout. Although the study is predicated upon the
incorporation of past events into dreams it is accepted that correspondences
between dream reports and diary records of subsequent days can occur due to
dream content being related to, or even causing or priming future behavior
(Selterman, Apetroaia, Riela, & Aron, 2014). Such a predictive effect of dream
content would increase the baseline, Days ⫺10 to ⫺1, control value, and thus work
against the finding of a significant day-residue in comparison with that control.
As participants were not given a limit on the number of correspondences that
could be recorded on each diary record/dream report sheet, there was considerable
variation in the number of correspondences that participants identified. To take
account of individual differences in the identifying of correspondences, after
Wilcoxon tests were performed on the whole sample participants were divided into
two groups based on a median split around the median of the total number of
correspondences identified by each participant, per their total number of daily
diary/dream report sheets. Wilcoxon tests were then repeated as described above,
on the two median split groups separately.
RESULTS
One-hundred and 66 dreams were collected from the 14 participants (mini-
mum from a participant ⫽8, maximum ⫽14). The mean word length of the mean
of each participants’ dreams ⫽183.67; SD ⫽113.68; minimum ⫽97.50, maxi-
mum ⫽520.25. All participants returned 14 daily diary entries, except for one
participant who mistakenly reproduced a diary entry of diary day 5 for diary day 6;
correspondences with the diary for that day (diary day 6) are excluded from the
data for that participant. This yielded a total number of dream report to diary
record combinations of 2,308, on which participants altogether identified a total of
This document is copyrighted by the American Psychological Association or one of its allied publishers.
This article is intended solely for the personal use of the individual user and is not to be disseminated broadly.
78 Henley-Einion and Blagrove

2,217 correspondences (many diary record/dream report sheets resulted in a zero
match, which is not counted as a correspondence). The identification of such a high
number of matches resulted in a considerable workload for participants.
Excerpts from a diary record and dream report, the comparison of which was
rated as 8 by the participant (extremely high), are as follows:
Participant 8, Day 1
Diary record day 8: “...My elder son Angus...was tidying up and ready to
move out.”
Dream report dream 8: “...I was at the school—Angus’s old housemaster had
left and his belongings were all in a plastic box.”
Excerpts from a diary record and dream report, the comparison of which was
rated as 2 by the participant (weak), are as follows:
Participant 12, Day 9
Diary record day 6: “...Had coffee outside.”
Dream report dream 14: “...making a cup of coffee.”
The mean levels of correspondence between diary records and dream reports
(and SDs), calculated as the sum of correspondence intensities for each diary day
to dream occurrence period, averaged per diary record/dream report sheet that
compose that period, as a function of time between diary day and dream
occurrence, are presented in Figure 1, top panel. The baseline is the mean of this
variable for the periods Days ⫺10 to ⫺1, that is, periods where the dream occurred
prior to the comparison diary day.
Wilcoxon tests were used to test for the day-residue effect (comparing Day 1
with Days ⫺10 to ⫺1, and with Days 2 to 4) and to test for the dream-lag effect
(comparing Days 5 to 7 with Days 2 to 4). The day-residue effect was shown by a
significant difference between Day 1 and mean of Days 2 to 4 (z⫽1.915, p⫽.028),
with a trend for the comparison of Day 1 and mean of Days ⫺10 to ⫺1(z⫽1.538,
p⫽.062). No dream-lag effect was found (comparing mean of Days 5 to 7 with
mean of Days 2 to 4, z⫽.220).
As described above, the sample was then divided into two groups, split around the
median total number of correspondences per diary record/dream report sheet. The
total number of correspondences identified by each participant ranged from 50 to 336
matches, with a mean of 158.36. As the number of diary record/dream report sheets
differed between participants, ranging from 112 to the maximum of 196, the number of
correspondences per diary record/dream report sheet for each participant was calcu-
lated. For the 14 participants the median number of correspondences per diary
record/dream report sheet was 0.87 (min ⫽0.29, max ⫽1.85). The high correspon-
dences group consisted of those who identified more than the median number of
correspondences per sheet (M⫽1.35; SD ⫽0.29; min ⫽0.92 max ⫽1.85; n⫽7; six
female) and the low correspondences group consisted of those who identified less than
the median number of correspondences per sheet (M⫽0.54; SD ⫽0.21; min ⫽0.29,
max ⫽0.81; n⫽7; five female). The mean (SD) ages were 52.29 (9.67) and 48.67
(13.35) years for the high and low correspondences groups respectively. To assess
whether number of correspondences identified per sheet may be a function of number
of sheets to be scored by each participant, mean number of sheets scored by the two
groups were calculated: means (SDs) were 162 (26.63) and 167.71 (30.02) for the high
and low correspondences groups, respectively, indicating that identification of corre-
This document is copyrighted by the American Psychological Association or one of its allied publishers.
This article is intended solely for the personal use of the individual user and is not to be disseminated broadly.
Comparing Dream Reports and Daily Diaries 79

Figure 1. Sum of correspondence intensity scores between daily diary records and dream reports as a
function of time between diary day and dream occurrence. Top panel, all participants, n⫽14. Bottom panel,
participants with below median total number of correspondences between daily diaries and dream reports,
n⫽7.
*
p⬍.05 (one-tailed). Note: Sum of intensity scores is averaged by the number of diary record/dream
report sheets for that diary to dream occurrence period, as there are more diary/dream sheet combinations
as the number of days between diary day and dream occurrence becomes smaller. Day ⫺10 to ⫺1 refers to
periods in which the dream report was compared with diary records completed on days after the dream
occurred. The mean of these 10 periods is represented here as a baseline measure.
This document is copyrighted by the American Psychological Association or one of its allied publishers.
This article is intended solely for the personal use of the individual user and is not to be disseminated broadly.
80 Henley-Einion and Blagrove

spondences was not subject to a fatigue effect due to number of sheets. The association
between correspondences group (high vs. low) and recruitment method (Mensa,
writers, Web sites) was found to be nonsignificant (Fisher’s exact test, p⫽.388).
The sum of correspondence intensity scores, averaged per diary record/dream
report sheet, was calculated separately for the two groups. For inferential statistics
these means were again combined into the two experimental periods, Day 1 and
Days 5 to 7, and control periods of Days ⫺10 to ⫺1 and Days 2 to 4. For both
groups there was no significant dream-lag effect (high correspondences group, z⫽
0.169; low correspondences group, z⫽0.338). The high correspondences group did
not show a day-residue effect (both zs⬍1.36). In contrast, the low correspondences
group showed a significant day-residue effect, with a significant difference between
Day 1 and Days ⫺10 to ⫺1(z⫽1.690, p⫽.046) and a trend for the comparison
ofDay1andDays2to4(z⫽1.521, p⫽.064). Figure 1, bottom panel, shows the
sum of correspondence intensity scores for the low correspondences group.
The second method of calculating a summary correspondence variable used
the count of the number of correspondences per diary day to dream occurrence
period, irrespective of intensity score, per the number of diary record/dream report
sheets for that period. In this variable a correspondence intensity score of 1 is
therefore given equal weight to an intensity score of 8. Figure 2, top panel, presents
the means of this variable for the baseline period and for periods of 1 to 10 days
between daily diary and dream occurrence. For the 14 participants, the comparison
of Day 1 with Days 2 to 4 was significant (z⫽1.789, p⫽.037), but the comparison
of Day 1 with Days ⫺10 to ⫺1 was not (z⫽1.161). The comparison of Days 5 to
7 with Days 2 to 4 was not significant (z⫽0.220).
For participants with above median total number of correspondences between
daily diaries and dream reports, all three predicted comparisons were nonsignificant
(all zs⬍0.51), and the Day 1 mean was lower than the Days ⫺10 to ⫺1 baseline.
In contrast, for the low correspondences group there was a significant day-residue
effect, with Day 1 having significantly more correspondences than Days ⫺10 to ⫺1
(z⫽2.028, p⫽.022) and Days 2 to 4 (z⫽1.859, p⫽.032). However, there was no
significant difference between Days 5 to 7 and Days 2 to 4 (z ⫽0.000). The means
of this variable for participants with below median total number of correspon-
dences between daily diaries and dream reports are presented in Figure 2, bottom
panel. A further analysis was conducted to assess whether the day-residue effect
was more significant if weak correspondences (intensity scores ⬍4) were not
included in the total number of correspondences measure. When this was per-
formed for the whole sample both day-residue comparisons became nonsignificant
(both zs⬍1.61); for the low correspondences group, zs for both day-residue
comparisons decreased and one became nonsignificant, and for the high correspon-
dences group both day-residue comparisons remained nonsignificant. Overall the
removal of weaker correspondences from the total number of correspondences
algorithm diminished the day-residue effect.
The third method of calculating a summary correspondence variable used the
number of words of dream text that had been highlighted by the participant as
corresponding to elements from the diary records, expressed as a percentage of the
total number of words of the dream reports for that diary day to dream occurrence
period. This variable was calculated for each period of time between diary day and
dream occurrence. Figure 3, top panel, presents the means of this variable for the
This document is copyrighted by the American Psychological Association or one of its allied publishers.
This article is intended solely for the personal use of the individual user and is not to be disseminated broadly.
Comparing Dream Reports and Daily Diaries 81

baseline period and for periods of 1 to 10 days between daily diary and dream
occurrence. For the 14 participants, the comparisons of Day 1 with Days 2 to 4 (z⫽
⫺1.161) and Days ⫺10 to ⫺1(z⫽1.601) were nonsignificant. The comparison of
Days 5 to 7 with Days 2 to 4 was also nonsignificant (z⫽.345).
Figure 2. Number of correspondences between daily diary records and dream reports as a function of
time between diary day and dream occurrence. Top panel, all participants, n⫽14. Bottom panel,
participants with below median total number of correspondences between daily diaries and dream
reports, n⫽7.
*
p⬍.05 (one-tailed).
**
p⬍.025 (one-tailed). Note: Number of correspondences is
averaged by the number of diary record/dream report sheets for that diary to dream occurrence period,
as there are more diary/dream sheet combinations as the number of days between diary day and dream
occurrence becomes smaller.
This document is copyrighted by the American Psychological Association or one of its allied publishers.
This article is intended solely for the personal use of the individual user and is not to be disseminated broadly.
82 Henley-Einion and Blagrove

For participants with above median total number of correspondences between
daily diary records and dream reports, all hypothesized comparisons for this
percentage incorporated words variable were nonsignificant (all zs⬍1.53), and the
Day 1 mean was lower than the Days ⫺10 to ⫺1 baseline. In contrast, for the low
correspondences group there was a significant day-residue effect, with Day 1 having
a significantly higher percentage of dream text highlighted as corresponding to
elements from the diary records than did Days ⫺10 to ⫺1(z⫽2.197, p⫽.014). The
other hypothesized comparisons were nonsignificant (both zs⬍1.20). The means of
this variable for participants with below median total number of correspondences
between daily diary records and dream reports are presented in Figure 3, bottom
panel.
Figure 3. Percentage of dream text highlighted by each participant as corresponding to elements from
the diary records, as a function of time between diary day and dream occurrence. Top panel, all
participants, n⫽14. Bottom panel, participants with below median total number of correspondences
between daily diaries and dream reports, n⫽7.
*
p⬍.025 (one-tailed).
This document is copyrighted by the American Psychological Association or one of its allied publishers.
This article is intended solely for the personal use of the individual user and is not to be disseminated broadly.
Comparing Dream Reports and Daily Diaries 83

The fourth method of deriving a summary correspondence variable first
calculated for each diary record/dream report sheet the product of each correspon-
dence intensity score and the number of words of the dream text highlighted by the
participant for that correspondence. These products were then summed for each
period between diary day and dream occurrence, and a mean score for each period
calculated by dividing by the total number of words in the dream reports for that
period. Figure 4, top panel, presents the means of this weighted correspondence
intensity variable for the baseline period and for periods of 1 to 10 days between
daily diary and dream occurrence. For the 14 participants, the comparison of Day
1 with Days ⫺10 to ⫺1 was significant (z⫽2.229, p⫽.013), but the other
hypothesized comparisons were nonsignificant (both zs⬍1.30).
For the high correspondences group, all three predicted comparisons were
nonsignificant (all zs⬍.68). In contrast, for the low correspondences group there
was a significant day-residue effect, with the Day 1 mean being significantly higher
than the mean of Days ⫺10 to ⫺1(z⫽2.336, p⫽.009), although not Days 2 to 4
(z⫽1.187). The means of this weighted correspondence intensity variable for
participants with below median total number of correspondences between daily
diary records and dream reports are presented in Figure 4, bottom panel.
DISCUSSION
The dream-lag effect has previously been found in designs where a record of
a single event is compared with a dream report, and a single correspondence score
produced, or where a single score is required as a summary of the comparison of a
dream report with a diary record that could refer to several events during a day.
The current design used diary records of what could be several events from a day,
but with scoring by participants of multiple correspondences, between parts of the
diary record and parts of the dream report. Four methods of calculating summary
scores from the sets of correspondence scores were then applied. Although this
design uses features of the previous two types of designs, which had found evidence
for the dream-lag, and although all four methods of computing summary corre-
spondence variables gave evidence for the day-residue effect, no evidence for the
dream-lag effect was obtained. It is necessary to address why this design, allowing
for multiple correspondences to be identified, had different results regarding the
dream-lag from previous studies. One possible explanation is that it may be that the
emphasis on comparing parts of diary records to parts of dream reports led to a
literalness in the comparisons, such that participants could “not see the wood for
the trees.” This literalness may not have been problematic for the identification of
the day-residue, but may be problematic if, as suggested by Blagrove, Henley-
Einion, et al. (2011), the manner of incorporation of waking life events differs
between the day-residue and dream-lag, and is more abstract for dream-lag
correspondences. It may thus be that the design caused participants to give
correspondence responses akin to those of independent judges, who were found, in
Blagrove, Henley-Einion, et al. (2011), to be able to identify the day-residue effect,
but not the dream-lag effect.
Another possibility is that the use of an hourly time template for the diary may
have resulted in too little space to be available for describing all the details of
This document is copyrighted by the American Psychological Association or one of its allied publishers.
This article is intended solely for the personal use of the individual user and is not to be disseminated broadly.
84 Henley-Einion and Blagrove

personally significant events. The template did have the aim of making the
record-keeping more thorough, but may instead have resulted in the day being seen
in terms of activities and times of day that things occurred, rather than eliciting
responses according to their personal significance. There was also a difference
between the participants in this study and those in the previous studies, who were
Figure 4. Sum of the products of each correspondence intensity score and length in words of the dream
report text that the correspondence score refers to, divided by total word length of dream reports for
that period, as a function of time between diary day and dream occurrence. Top panel, all participants,
n⫽14. Bottom panel, participants with below median total number of correspondences between daily
diaries and dream reports, n⫽7.
*
p⬍.025 (one-tailed).
**
p⬍.01 (one-tailed).
This document is copyrighted by the American Psychological Association or one of its allied publishers.
This article is intended solely for the personal use of the individual user and is not to be disseminated broadly.
Comparing Dream Reports and Daily Diaries 85

mainly students. That the participants in the current study were older than for
previous studies may have resulted in daytime activities being more ordered and
regular than for students, and especially students on Christmas vacation, as were
studied in Blagrove, Henley-Einion, et al. (2011). Considering all these differences
between the current and previous designs, a recommendation for future work is to
not use a diary structured by hours of the day, and to instead elicit records based
on personal significance of events. Alternatively, it may of course be that the
current study was a valid and accurate testing of the dream-lag effect, and that the
dream-lag effect does not exist, despite previous literature supporting it.
Partial evidence for the day-residue effect was found using each of the four
methods of calculating an overall correspondence score for each diary day to dream
occurrence period: sum of correspondence intensity scores, number of correspon-
dences, percentage of dream text highlighted, and sum of the products of correspon-
dence intensity scores and amount of dream text highlighted. However, the 1 day
period was only significantly higher than both the baseline and the 2 to 4 day period
for the low correspondences group using the number of correspondences method.
It is necessary to consider why the sum of correspondence intensity scores
method did not identify the day-residue effect as well as did the number of
incorporations method, given that the latter method loses some of the information
provided by participants, that is, the intensity scores. In answer to this, it may be
that participants are not able to accurately, validly, or reliably assess intensity of
correspondences, a concern also raised by Marquardt et al. (1996). The possibility
of unreliable scoring, in addition to the results here, means that we do not
recommend that the algorithm based on sum of intensity scores be used in future
to calculate a single score that summarizes a set of correspondences between diary
records and dream reports.
The question is also raised of why the low correspondences group evidenced
the day-residue effect better than did the high correspondences group. Previous
work (e.g., Blagrove, Fouquet, et al., 2011;Blagrove, Henley-Einion, et al., 2011;
Nielsen et al., 2004) required participants or judges to give one score to summarize
the relationship between a diary record and a dream report, although the length of
the diary record varied greatly between studies due to design differences. It appears
that for participants identifying a large number of correspondences, the day-residue
effect is diluted, due to an increase in number of correspondences for all diary day
to dream occurrence periods, whereas those participants identifying correspon-
dences more sparingly continue to evidence the day-residue effect.
Aside from the important methodological implication of this finding for future
work, there is the further question of why individuals differ in the number of
correspondences they identify. This may be a dream production effect, in that some
individuals may differ in their incorporation of items from waking life into dreams,
or it may be that this is a reporting or judging effect, in which some individuals are
better able to identify, or confabulate, links between dream reports and diary
records than are others. Source of recruitment in the current study, including the
lifestyle or waking occupation characteristic of being a Mensa member or writer,
was not significantly associated with membership of the low versus high correspon-
dences group. Number of correspondences identified may instead be a function of
personality. Relevant here is the notion of overinclusion, which is related to
creativity, psychoticism, and the failure to filter out or inhibit irrelevant stimuli
This document is copyrighted by the American Psychological Association or one of its allied publishers.
This article is intended solely for the personal use of the individual user and is not to be disseminated broadly.
86 Henley-Einion and Blagrove

(Eysenck, 1994). Such overinclusion could cause an individual difference in the
ability to identify links between texts, or even to confabulate such links. Blackmore
(1997) proposed that some individuals have an excessive use of inclusive categories,
which can appear as an affirmative bias, or, as we would say here, a bias to affirm
multiple links between two texts. Bressan (2002) termed these “weak and strong
pattern finders” who differ in “associative readiness.” It may be that individual
differences in overinclusion could affect judgments about the extent of relationship
between diary records and dream content. Support for this is given by Blagrove,
French, and Jones (2006), who used a set of questions devised by Blackmore (1997),
“Do you have a scar on your left knee?,” “Do you have back pain at the moment?,”
and “Do you have a cat?,” that participants were asked to answer. Blagrove et al.
(2006) found that individuals who claimed to have had a precognitive dream gave
affirmative answers to these three questions more often than did individuals who
said they had not had such an experience.
Individuals who tend to connect the content of dreams and the content of
events happening later in time thus answer these seemingly factual questions
affirmatively, with the number of these questions answered affirmatively being
positively and significantly related to the number of precognitive dreams that
participants claim to have experienced. Blagrove et al. (2006) state that there may
be a similar ability to connect the content of horoscopes with one’s own life. Of
relevance here is the “Barnum Effect,” which refers to a tendency for people to
endorse as accurately describing themselves a general personality description, even
when the same description is given to many people, who also see it as describing
themselves. Indeed, Snyder and Shenkel (1976) found that subjects were more
inclined to accept Barnum descriptions when they believed that these had been
produced uniquely for themselves, a feature that may well be seen as a character-
istic of one’s own dreams. Individual differences in the Barnum Effect have been
related to positive schizotypy (Claridge, Clark, Powney, & Hassan, 2008), which is
itself related to an increased tendency to generate (false) hypotheses about random
and illusory contingencies (Brugger et al., 1993;Fyfe, Williams, Pickup, & Mason,
2008). We consider that it is important to address why individuals differ in the
number of correspondences they identify between waking life records and dream
reports, and especially to investigate whether these differences occur at dream
production and/or when dream reports and daily life records are compared. We
accept, however, that the above suggestions of associations between personality
and number of correspondences identified are speculative, as personality measures
were not included in the current study. Furthermore, we acknowledge that the
number of participants in the study was low and that there will have been
uncontrolled interindividual variation in the study as a result of factors such as
recruitment from different sources and a wide age range of participants.
We suggest that the use of a median split around the total number of correspon-
dences found by each individual between diary records and dream reports acts to
remove from the analysis individuals who have an excessive propensity to connect
events. The day-residue is better evidenced when this median split is used to identify
the low correspondences participants. Obviously, for designs where only one score is
elicited from each participant for each diary record and dream report combination, this
individual difference confound does not occur, but for designs that allow multiple
correspondences to be identified, this subsample analysis may be necessary.
This document is copyrighted by the American Psychological Association or one of its allied publishers.
This article is intended solely for the personal use of the individual user and is not to be disseminated broadly.
Comparing Dream Reports and Daily Diaries 87

To summarize, when a subsample of participants who had been more sparing
in their identification of correspondences between diary records and dream reports
was analyzed, and with summary variables for each diary day to dream occurrence
period calculated using the number of correspondences between diary records and
dream reports, the day-residue effect was more readily identified. This number of
correspondences algorithm, which is the simplest of the four methods investigated
here, and the median split analysis to derive a low incorporator subsample, are thus
recommended for future work on the time-course of the incorporation of waking
life events into dreams, where designs allow for multiple correspondences to be
identified.
REFERENCES
Blackmore, S. J. (1997). Probability misjudgement and belief in the paranormal: A newspaper survey.
British Journal of Psychology, 88, 683–689. doi:10.1111/j.2044-8295.1997.tb02665.x
Blagrove, M., Fouquet, N. C., Henley-Einion, J. A., Pace-Schott, E. F., Davies, A. C., Neuschaffer, J. L.,
& Turnbull, O. H. (2011). Assessing the dream-lag effect for REM and NREM stage 2 dreams.
PLoS ONE, 6, e26708. doi:10.1371/journal.pone.0026708
Blagrove, M., French, C. C., & Jones, G. (2006). Probabilistic reasoning, affirmative bias and belief in
precognitive dreams. Applied Cognitive Psychology, 20, 65–83. doi:10.1002/acp.1165
Blagrove, M., Henley-Einion, J., Barnett, A., Edwards, D., & Seage, H. C. (2011). A replication of the
5–7 day dream-lag effect with comparison of dreams to future events as control for baseline
matching. Consciousness and Cognition, 20, 384–391. doi:10.1016/j.concog.2010.07.006
Bressan, P. (2002). The connection between random sequences, everyday coincidences, and belief in the
paranormal. Applied Cognitive Psychology, 16, 17–34. doi:10.1002/acp.754
Brugger, P., Regard, M., Landis, T., Cook, C., Krebs, D., & Niederberger, J. (1993). Meaningful patterns
in visual noise: Effects of lateral stimulation and the observer’s belief in ESP. Psychopathology, 26,
261–265. doi:10.1159/000284831
Claridge, G., Clark, K., Powney, E., & Hassan, E. (2008). Schizotypy and the Barnum effect. Personality
and Individual Differences, 44, 436–444. doi:10.1016/j.paid.2007.09.006
Edwards, C. L., Ruby, P. M., Malinowski, J. E., Bennett, P. D., & Blagrove, M. T. (2013). Dreaming and
insight. Frontiers in Psychology, 4, 979. doi:10.3389/fpsyg.2013.00979
Eysenck, H. J. (1994). The measurement of creativity. In M. A. Boden (Ed.), Dimensions of creativity
(pp. 199–242). Cambridge, MA: MIT Press.
Fyfe, S., Williams, C., Pickup, G., & Mason, O. (2008). Apophenia, theory of mind and schizotypy: Perceiving
meaning and intentionality in randomness. Cortex, 44, 1316–1325. doi:10.1016/j.cortex.2007.07.009
Marquardt, C. J. G., Bonato, R. A., & Hoffmann, R. F. (1996). An empirical investigation into the
day-residue and dream-lag effects. Dreaming, 6, 57–65. doi:10.1037/h0094446
Nielsen, T. A., Kuiken, D., Alain, G., Stenstrom, P., & Powell, R. A. (2004). Immediate and delayed
incorporations of events into dreams: Further replication and implications for dream function.
Journal of Sleep Research, 13, 327–336. doi:10.1111/j.1365-2869.2004.00421.x
Nielsen, T. A., & Powell, R. A. (1989). The dream-lag effect: A 6-day temporal delay in dream content
incorporation. Psychiatric Journal of the University of Ottawa, 14, 561–565.
Nielsen, T. A., & Powell, R. A. (1992). The day-residue and dream-lag effects: A review and limited
replication. Dreaming, 2, 67–77. doi:10.1037/h0094348
Nielsen, T. A., & Stenstrom, P. (2005). What are the memory sources of dreaming? Nature, 437,
1286–1289. doi:10.1038/nature04288
Nikles, C. D., II, Brecht, D. L., Klinger, E., & Bursell, A. L. (1998). The effects of current concern- and
nonconcern-related waking suggestions on nocturnal dream content. Journal of Personality and
Social Psychology, 75, 242–255. doi:10.1037/0022-3514.75.1.242
Saredi, R., Baylor, G. W., Meier, B., & Strauch, I. (1997). Current concerns and REM-dreams: A
laboratory study of dream incubation. Dreaming, 7, 195–208. doi:10.1037/h0094474
Selterman, D. F., Apetroaia, A. I., Riela, S., & Aron, A. (2014). Dreaming of you: Behavior and emotion
in dreams of significant others predict subsequent relational behaviour. Social Psychological and
Personality Science, 5, 111–118. doi:10.1177/1948550613486678
Snyder, C. R., & Shenkel, R. J. (1976). Effects of favourability, modality, relevance on acceptance of
general personality interpretations prior to and after receiving diagnostic feedback. Journal of
Consulting and Clinical Psychology, 44, 34–41. doi:10.1037/0022-006X.44.1.34
This document is copyrighted by the American Psychological Association or one of its allied publishers.
This article is intended solely for the personal use of the individual user and is not to be disseminated broadly.
88 Henley-Einion and Blagrove