• TABLE OF CONTENTS
HIDE
 Title Page
 Acknowledgement
 Table of Contents
 Abstract
 Introduction
 Method
 Results
 Discussion
 Summary
 Appendices
 References
 Biographical sketch














Title: Sleep length and variability
CITATION PDF VIEWER THUMBNAILS PAGE IMAGE ZOOMABLE
Full Citation
STANDARD VIEW MARC VIEW
Permanent Link: http://ufdc.ufl.edu/UF00097541/00001
 Material Information
Title: Sleep length and variability measurement and interrelationships
Physical Description: vii, 98 leaves : ; 28cm.
Language: English
Creator: White, Royce Marvin, 1945-
Publication Date: 1975
Copyright Date: 1975
 Subjects
Subject: Sleep   ( lcsh )
Psychology thesis Ph. D   ( lcsh )
Dissertations, Academic -- Psychology -- UF   ( lcsh )
Genre: bibliography   ( marcgt )
non-fiction   ( marcgt )
 Notes
Thesis: Thesis--University of Florida.
Bibliography: Bibliography: leaves 93-97.
Additional Physical Form: Also available on World Wide Web
Statement of Responsibility: by Royce M. White, Jr.
General Note: Typescript.
General Note: Vita.
 Record Information
Bibliographic ID: UF00097541
Volume ID: VID00001
Source Institution: University of Florida
Holding Location: University of Florida
Rights Management: All rights reserved by the source institution and holding location.
Resource Identifier: alephbibnum - 000166536
oclc - 02830797
notis - AAT2919

Downloads

This item has the following downloads:

PDF ( 3 MBs ) ( PDF )


Table of Contents
    Title Page
        Page i
        Page i-a
    Acknowledgement
        Page ii
    Table of Contents
        Page iii
        Page iv
    Abstract
        Page v
        Page vi
        Page vii
    Introduction
        Page 1
        Page 2
        Page 3
        Page 4
        Page 5
        Page 6
        Page 7
        Page 8
        Page 9
        Page 10
        Page 11
        Page 12
        Page 13
        Page 14
    Method
        Page 15
        Page 16
        Page 17
    Results
        Page 18
        Page 19
        Page 20
        Page 21
        Page 22
        Page 23
        Page 24
        Page 25
        Page 26
        Page 27
        Page 28
        Page 29
        Page 30
        Page 31
        Page 32
        Page 33
        Page 34
        Page 35
        Page 36
        Page 37
        Page 38
        Page 39
        Page 40
        Page 41
        Page 42
        Page 43
        Page 44
        Page 45
        Page 46
        Page 47
        Page 48
        Page 49
        Page 50
        Page 51
        Page 52
        Page 53
        Page 54
        Page 55
        Page 56
        Page 57
    Discussion
        Page 58
        Page 59
        Page 60
        Page 61
        Page 62
        Page 63
        Page 64
        Page 65
        Page 66
        Page 67
        Page 68
        Page 69
        Page 70
        Page 71
        Page 72
        Page 73
        Page 74
        Page 75
        Page 76
        Page 77
        Page 78
        Page 79
    Summary
        Page 80
        Page 81
        Page 82
    Appendices
        Page 83
        Page 84
        Page 85
        Page 86
        Page 87
        Page 88
        Page 89
        Page 90
        Page 91
        Page 92
    References
        Page 93
        Page 94
        Page 95
        Page 96
        Page 97
    Biographical sketch
        Page 98
        Page 99
        Page 100
Full Text











SLEEP LENGTH AND VARIABILITY: MEASUREMENT
AND INTERRELATIONSHIPS













BY

ROYCE M. WHITE, JR.


A DISSERTATION PRESENTED TO THE GRADUATE COUNCIL OF
THE UNIVERSITY OF FLORIDA IN PARTIAL FULFILLMENT
OF THE REQUIREMENTS FOR THE DEGREE OF
DOCTOR OF PHILOSOPHY





UNIVERSITY OF FLORIDA


1975














ACKNOWLEDGMENTS


The members of the author's committee and all of the

other faculty members who have shared their knowledge and

perspectives are greatly appreciated. They have made the

pursuit of this degree a pleasure and a challenge. Special

thanks is given to Dr. Wilse B. Webb, for his expert guid-

ance as committee chairman, and to Dr. Richard J. Anderson

for his friendship and counsel over many years. Most

importantly, gratitude is expressed to Valerie for her

support and assistance in all phases of this project.

















CONTENTS


ACKNOWLEDGMENTS


ABSTRACT . . . . . . . . .

SECTION

I. INTRODUCTION . . . . . .

Uses of Sleep-Length Measures .
Methods of Measuring Sleep Length
Comparison of Methods . . .
Needed Research . . . . .

II. METHOD . . . . . . . .

III. RESULTS . . . . . . .

Group Sleep Behaviors . . .
Intra-Individual Comparisons . .
Interrelationships . . . .

IV. DISCUSSION . . . . . . .

Group Data . . . . . .
Individual Data . . . . .
Interrelationships . . . .

V. SUMMARY . . . . . . .


APPENDIX

1.

2.

3.


:X

SLEEP QUESTIONNAIRE (PRE-LOG)

SLEEP LOG . . . . . .

SLEEP QUESTIONNAIRE (POST-LOG) .


4. FREQUENCY DISTRIBUTION OF ANSWER CATEGORIES FOR
SELECTED PRE-LOG QUESTIONS . . . . .

5. FREQUENCY DISTRIBUTION OF ANSWER CATEGORIES FOR
SELECTED POST-LOG QUESTIONS . . . . .


. . . . . . . . . . ii










REFERENCES . . . . . . . . . 93

BIOGRAPHICAL SKETCH . . . . . . ... .98










Abstract of Dissertation Presented to the Graduate Council
of the University of Florida in Partial Fulfillment of
Requirements for the Degree of
Doctor of Philosophy


SLEEP LENGTH AND VARIABILITY: MEASUREMENT
AND INTERRELATIONSHIPS

By

Royce M. White, Jr.


June, 1975


Chairman: Wilse B. Webb
Major Department: Psychology

Length of sleep in humans has been used for a variety

of purposes, including a desire to compare or establish

norms for various groups. Sleep length has also been used

as both a dependent and independent variable, e.g., to

evaluate the influence of variables such as intelligence on

sleep or of social variables such as stress on sleep length.

The two most common methods of measuring sleep length

in groups and individuals have been the sleep questionnaire

and the daily diary or log. The present study has investi-

gated the relationship between these two methods, interrela-

tionship among sleep measures, and relations to selected

nonsleep measures.

A total of 102 college students took a Pre-Log Sleep

Questionnaire about usual sleep behaviors, including sleep

length, bedtime, wake-up time, and naps. They then kept a









Sleep Log for 2 to 4 weeks, followed by a Post-Log Sleep

Questionnaire which asked for estimates of various sleep

behaviors during the log recording period.

Group means and variances of both Questionnaires give

results within a few tenths of an hour of the Sleep Log

measures on most variables. In addition to the usual data

on average sleep length per night (7.5 hours) and nap time

per week (2.6 hours), use of the Sleep Log allowed the

measurement of within-subject night-to-night variability.

The average student is quite variable in both bedtime (SD =

1.28 hours) and wake-up time (SD = 1.29 hours). The result-

ing sleep length standard deviation about his own mean for

the average student is 1.55 hours. Other sleep characteris-

tics of the population are detailed and discussed.

Individual data is not as stable as the group data in

comparisons between the Sleep Log and the Questionnaires.

The comparison of Sleep Log to Post-Log Questionnaire esti-

mates of sleep length per night results in a correlation of

r = +.72, with 53% of the students having a difference of

.5 hour or more and 15% having a difference of 1.0 hour

or more. The Sleep Log vs. Pre-Log Questionnaire comparisons

produce results similar to the above. The relationship of

several Questionnaire questions on sleep variability to

Sleep Log measures of variability (SD) is very limited or

nonexistent. While there are large day-to-day differences










in sleep length for most students, biweekly sleep length

means correlate r = +.88.

The interrelationships among the various Sleep Log

measures were investigated. A short average sleep length

is significantly related to several measures, including a

high variability (SD) of sleep length (r = -.48). Relation-

ships between sleep variables and nonsleep measures are not

generally significant; e.g., academic achievement and a

weight/height ratio are not related to sleep length. Only

one MMPI scale (MA) correlates with sleep length (r = -.28).

The variability of one's sleep may be more related to per-

sonality measures; 4 of 13 MMPI scales correlate with the

standard deviation of sleep length. The relationship of

sleep and personality are discussed in terms of previous

findings.














SECTION I

INTRODUCTION


Uses of Sleep-Length Measures

Everyone spends some amount of time sleeping. Various

types of studies have been undertaken to measure the length

of this sleep time in various individuals and groups under

a variety of conditions. There have been many reasons for

studying sleep length. One of the earliest and most common

was the gathering of normative data on average sleep length

of various groups (e.g., age groups) simply to satisfy

curiosity about the average sleep length of people. Of

course, this type of data was often then used by others to

establish a norm of expected or desirable sleep length.

Thus, if the average 2-year-old was found to get 12 hours

of sleep per day, many would argue that any child getting

less was not getting his proper rest.

Another reason for collecting sleep-length data on

large groups has been to relate sleep length to other vari-

ables of interest such as intelligence, body-build, geo-

graphical setting, culture, etc. Such studies have often

focused on sleep either as a dependent or independent vari-

able as they have attempted to discover the determinants of

average sleep length or the consequences of sleeping a given










average length of time (Johns, Bruce, & Masterton, 1974;

Terman & Hocking, 1913; Tune, 1969a).

Sleep-length measures have also often been desired by

sleep researchers who were screening potential subjects for

use in sleep laboratory experiments. Often these re-

searchers have attempted to select "normal" subjects with

typical sleep lengths for their age group in order to study

the effects of such manipulations as sleep deprivation or

drugs on the sleep process. On occasion, the selection

process has been aimed at finding persons who have unusual

sleep lengths or patterns in order to evaluate these dif-

ferences (Hartmann, 1973a, 1973b; Hartmann, Baekeland, &

Zwilling, 1972; Hartmann, Baekeland, Zwilling, & Hoy, 1971;

Webb & Friel, 1970, 1971).

Average sleep length and its change over time has

sometimes been used to plan and evaluate treatment programs

for persons with medical problems in general or sleep prob-

lems in particular. Sleep length has been related not only

to some types of insomnia, but also to many problems

associated with shift work and jet lag.

Finally, sleep length has recently been used as a vari-

able in studies of the general function of sleep. This area

of research has emphasized theories about the need for sleep

in general, as well as questions about possible changes in

one's need for given amounts of sleep (Hartmann, 1973a,

1973b; Meddis, Pearson, & Langford, 1973; Webb, 1974).









Methods of Measuring Sleep Length

In the pursuit of these diverse goals, several dif-

ferent methods have been used to measure sleep length and

the associated variables of bedtime, wake-up time, sleep-

onset latency, and awakenings in the night. The systematic

methods have included (a) the sleep questionnaire; (b) the

sleep diary or log kept for a period of days, weeks, or

months; and (c) the electroencephalograph (EEG).


Sleep Questionnaires

Sleep questionnaires have asked such questions as "How

long do you usually sleep each night?" There have been many

variations on both the exact questions asked and the format

for answering. Some questionnaires have not provided any

structure or guidance, with the person being allowed to

write any answer. Others have provided a selection of

answers, usually in 1/2- to 1-hour intervals. Providing

1/2-hour intervals may produce a less discrete and more con-

tinuous scale than would otherwise be used by subjects who

are asked simply to fill in a blank space. Authors often

have not reported on either the nature of the questions or

the format of the answers (e.g., Laird, 1931). This has

made it difficult to interpret results and compare data

across studies.

When answering a question about "usual" or "average"

sleep length, some persons may include weekend sleep, which

is often different from weekday sleep; others probably do










not. In several experiments (Johns et al., 1974; Johns,

Egan, Gay, & Masterton, 1970; Johns, Gay, Goodyear, &

Masterton, 1971), students were asked about their weekday

and weekend sleep separately in order to control for dif-

ferent interpretations of average length. In addition, they

were asked about usual bedtime and wake-up time on weekdays

and weekends instead of simply for an estimate of sleep

length. This procedure allowed the investigators to cal-

culate the overall sleep length (and weekday and weekend

length) and also to obtain a measure of bedtime and wake-up

time apart from the duration of sleep.

When one asks about bedtime and wake-up time rather

than sleep length, the latency to sleep onset is included

in the resulting measure of sleep length. Johns et. al.,

in the studies cited above, asked for an estimate of the

usual amount of time taken to fall asleep and subtracted

this from the time in bed. This reduces the estimate of

sleep time very little in most populations, however, with

latencies usually being on the order of 10 to 15 minutes

(Agnew & Webb, 1971; Johns et al., 1971). For most groups

the amount of time spent awake during the night is also

usually very little. However, for selected individuals or

groups (such as insomniacs and hospital patients), sleep-

onset latency and awakenings during the night can signifi-

cantly affect the estimate of sleep length.

Little attention has been given to the reliability of

answers to questions asking about average sleep length. Of









the few studies reporting on reliability is one by Webb and

Stone (1963), which asked several hundred college students,

"How many hours, on the average, do you sleep per night?"

Two different groups yielded test-retest correlations of

r = +.68 and r = +.64. O'Connor (1964) asked 22 subjects,

"How many hours do you usually sleep at night?" Over a

2-week test-retest period he found a correlation of r =

+.87. It is not known to what extent the variations of

sleep-length questions will affect the reliability of the

answers.


Sleep Diaries

A tool for measuring sleep which has not been used as

much as the questionnaire is the daily sleep diary or log.

In addition to providing an estimate of average sleep

length, sleep logs can provide estimates of average bed-

time, wake-up time, latency to sleep onset, and awakenings

during the night. If a measure of sleep per 24 hours is

desired, the subject can also be asked to record nap

periods. One measure which can probably be obtained with

the sleep log more readily and accurately than with the

questionnaire is the intrasubject variability of sleep

length. Reynolds (1935) and Reynolds and Mallay (1933)

used difference scores between biweekly sleep-length means

obtained from daily observations of children to demonstrate

the stability of these means as compared to wide daily

fluctuations. Several more recent investigations using










sleep logs have reported on individuals' standard deviations

around their own mean sleep length (Baekeland & Hartmann,

1971; Frankel, Buchbinder, & Snyder, 1973, p. 149; Karacan,

Williams, Littell, & Salis, 1973, pp. 102-132).

There have been several different formats and types of

questions used in sleep logs. One type has asked the sub-

ject to record his bedtime and wake-up time (Andress, 1911;

Kleitman, Mullin, Cooper, & Titelbaum, 1937). Others have

asked in addition for an estimate of how long it took to go

to sleep or an estimate of what time the person actually

went to sleep (Baekeland, 1969; Webb & Agnew, 1973, pp.

166-168). This estimate of latency can be subtracted from

the total time in bed. Weiss, McPartland, and Kupfer (1973)

simply asked psychiatric patients every morning for 14 days

how many hours (to the nearest 1/2 hour) they had slept the

night before.

A type of sleep log which does not ask for exact bed-

times and times of waking has been used by Tune (1968, 1969a,

1969b, 1969c) and Shurley, Pierce, Natani, and Brooks

(1970). These subjects were given a card divided into 1/2-

hour intervals across the 24-hour day for each of 14 days.

They drew a line through each 1/2-hour period which con-

tained more sleep than wakefulness. In this type of sleep

log the subject estimates his latency to sleep onset and

begins the line when he thinks he went to sleep. A record

of bedtimes per se and latencies is thus not obtained. Long









awakenings during the night (a break in the line) are not

included in the sleep estimate, and naps lasting more than

15 minutes are included. Several other investigators have

used similar sleep charts divided into 1-hour intervals

(Lewis, H. E., & Masterton, 1957; Masterton, 1965a, pp.

387-397, 1965b; Williams, 1959).

When the subject cannot or will not record his own

daily sleep behavior accurately, the sleep log can be kept

by an observer. Such has been the case with children

(Kleitman & Engelmann, 1953; Parmelee, Schulz, & Disbrow,

1961; Parmelee, Wenner, & Schulz, 1964; Reynolds, 1935;

Reynolds and Mallay, 1933), some elderly persons (Webb &

Swinburne, 1971), and psychiatric patients (Dlin, Rosen,

Dickstein, Lyons, & Fischer, 1971; Erwin & Zung, 1970;

Kupfer, Wyatt, & Snyder, 1970; Weiss et al., 1973). There

have been several types of observational techniques used.

One involves the relatively continuous observation of the

subject. Another, used often with psychiatric patients,

involves time-sampling techniques, a common interval between

observations being 1/2 hour. A third variation is simply

observing the bedtime or sleep-onset time and the awakening

or rising time, with an inference of continuous sleep be-

tween these two events unless the subject leaves his bed or

creates a disturbance.










Other Measures

Although sleep questionnaires and sleep logs have been

the major tools for estimating either the average amount of

sleep obtained or the length of sleep for one night, several

other methods have been occasionally used. One such method

is the electroencephalograph (EEG). The EEG is usually

used to study the normal infrastructure (stages) of sleep

or changes in this structure under various experimental con-

ditions. Such study is possible by observing the systematic

changes in amplitude and frequency of brainwaves throughout

the night. The EEG is unsurpassed in the ability to dis-

tinguish sleep from other states of relative inactivity. It

has not often been used to measure sleep length, however, as

it is much more expensive than sleep questionnaires and

logs, especially for the extended observation needed to

establish an average sleep length. The EEG is very useful,

however, in measuring precisely the latency to sleep onset

and the frequency and duration of awakenings, however brief,

during the night (Agnew & Webb, 1971, 1972).

Another method which has been studied and recommended

by Johns, Cornell, & Masterton (1969) involves the use of

the basal skin resistance (BSR), which is relatively in-

expensive to record when compared to the EEG. The BSR as

recorded between two points on the hand or other site

(similar to the well-known galvanic skin response or GSR)

usually increases with sleep and drops abruptly with any









awakening. Johns et al. have suggested using such a device

for recording the sleep of hospital patients.


Comparison of Methods

Each of the types of measures of sleep length has its

advantages and limitations for various specific uses, some

of which have been mentioned above. A questionnaire is the

quickest and least expensive method of gathering information

on the average duration of sleep in individuals or groups.

The answers to questionnaires are apparently fairly reli-

able. Reliability does not insure criterion validity, how-

ever. Comparison of one's estimate of average sleep length

to one's actual mean or median sleep length may result in

an over- or underestimation. In order to obtain one's true

sleep length average, however, it is necessary to have a

record of daily sleep behavior for an extended period of

time. This is necessary, of course, because most persons

do not sleep exactly the same amount every night and any one

night's sleep may not be representative.

It is surprising that there have been no major studies

reported that compare sleep-questionnaire estimates to a

sleep-length average computed from a daily sleep log. As

a small part of a large questionnaire study, O'Connor (1964)

did have 10 subjects keep a sleep log for 5 days after

answering a question about usual sleep length. The two

estimates of sleep had a correlation of r = +.64. Several

studies have used 2-week sleep logs as a check on potential










subjects' reports of average sleep length. Subjects not

meeting the required criteria on both logs and question-

naires were usually eliminated from further consideration

in the planned sleep experiment. Such studies were not

designed, however, to evaluate the general incidence in a

given population of significant discrepancies between

questionnaire and sleep-log estimates of sleep length.

Although comparisons of estimates from sleep logs and

sleep questionnaires are lacking, one type of comparison

between methods which has been systematically studied is

that between the EEG and the subject's or observer's esti-

mate of sleep the previous night. Such comparisons have

been made for sleep length per se, as well as for latency

to sleep onset and awakenings during the night. Even if a

subject or observer can accurately estimate time in bed,

large errors in latency to sleep onset or in the duration of

awakenings can lead to large errors in estimates of total

sleep length.

Several studies with typical young adults (Baekeland

and Hoy, 1971; Lewis, S. A., 1969; Thornby, Karacan, Beutler,

Booth, Anch, Williams, Salis, & Blackburn, 1974) have re-

sulted in the impression that such persons are fairly accu-

rate in estimating the latency to sleep onset. Most com-

parison data is for group means, however. Many individuals

can have inaccurate estimates and yet the group means can

still be "accurate." The number of brief awakenings in the









night as defined by EEG criteria is sometimes underesti-

mated since subjects do not always remember them in the

morning. Awakenings longer than 4 or 5 minutes seem to be

remembered.

Studies with chronic insomniacs (Bixler, Kales, Leo, &

Slye, 1973, p. 143; Frankel et al., 1973, p. 149) have

found that these persons significantly overestimate the

actual amount of time taken to go to sleep.

Weiss et al. (1973) reported on a study in which the

recorded EEGs of psychiatric patients were compared to both

the patients' own subjective estimates of sleep length

(given to the nearest 1/2 hour) and to the estimates of

nurses who made observations at 30-minute intervals during

the night. Comparisons between nurses' estimates and EEG

values of total sleep length were also made by Kupfer et al.

(1970) and Erwin and Zung (1970) on other groups of psychi-

atric patients. The general results of these three studies

show an overestimation of sleep length and an underestima-

tion of awakenings in the night compared to the EEG. The

unreliability of such estimates by nurses or the patients

themselves is likely due to the large amount of quiet wake-

fulness present during the time in bed. For normal subjects

in home environments, the amount of wakefulness during the

time in bed is not usually large.

A close examination of these studies comparing EEG

measures to subjective reports of sleep latency and










awakenings last night reveals that for normal subjects, the

group means for the two types of estimates are in close

agreement. Each individual is also fairly accurate in

estimating his own sleep last night. For insomniacs or

hospitalized patients, however, self-reports or observa-

tional reports of sleep length should be accepted with

caution.


Needed Research

Several lines of research on sleep length which have

not been adequately pursued are addressed in the present

study. Already mentioned is the fact that the relationship

between sleep questionnaire data and sleep log data has

not been explored. Such comparisons can be made between

either group averages for the two methods or individuals'

values. One can compare sleep log data to questionnaire

questions about usual sleep and to questions about recent

sleep. If the questions about recent sleep are given after

a sleep log has been recorded for several weeks or longer,

one can evaluate the validity of this questionnaire.

Another area that needs to be explored is the vari-

ability of sleep. Variability can be measured both between

and within subjects. When reporting on sleep-length

studies, many investigators report only on the mean sleep

behaviors of subjects and fail to include data on the

standard deviation around that mean. Normative data on

expected variations between subjects needs to be collected









as well as data on the averages. This study also considers

a type of variability seldom studied in human sleep--

variability within subjects. Variability within each person

around his own mean sleep length, bedtime, etc., can be more

accurately measured by a sleep log than a questionnaire.

Two subjects may have the same mean sleep length as measured

by either a questionnaire or sleep log and yet have very

different day-to-day variability in sleep. Knowledge about

the expected average and range of this intrasubject vari-

ability for various groups of persons would contribute to

the designing of studies which could evaluate the causes and

effects of such intra-individual daily variations in sleep

behavior.

A topic closely related to between-subject and within-

subject variability is that of the stability of sleep log

measures over time. For both group means and individuals,

it is important to know how well a 1- or 2-week sample of

sleep behavior will predict the next 1 or 2 weeks.

Finally, the interrelationships among the various sleep

measures obtained from sleep logs are examined. For ex-

ample, some previous work has indicated that short sleepers

tend to be more variable in their sleep length than long

sleepers (Baekeland & Hartmann, 1971). Such relationships

between sleep length, bedtime, etc., and the intra-

individual variability of these measures need to be examined

to see in what way the aspects of molar sleep behavior are










related. Also included in the discussion on interrelation-

ships is some limited data examining the extent to which

sleep behavior may be related to both attitudes and feelings

about one's sleep and to nonsleep measures such as person-

ality.

Sleep-length measures as independent and dependent

variables have many uses. We need, however, more informa-

tion about the above measurement techniques and interrela-

tionships in order to better approach such important

questions as the function of sleep, the determinants of

average sleep length and sleep length for any given night,

and the relationship of the length and variability of sleep

to such variables as personality and sleep problems.














SECTION II

METHOD


A total of 102 subjects, 44 males and 58 females, com-

pleted two sleep questionnaires and kept daily records of

sleep at home for 2 to 4 weeks. The subjects, age 18 to 27

years, were all students at the University of Florida who

volunteered to participate in order to fulfill a psychology

course requirement of participation in a few experiments of

choice. Of the 102 students, 48 participated during the

summer quarter of 1974 and 54 participated during the winter

quarter of 1975. Both groups participated during the middle

of the academic quarter when there were no holidays, vaca-

tions, or final examinations.

Both the summer group and winter group first filled

out the two-page Pre-Log Sleep Questionnaire (Appendix 1)

which asked for estimates of usual sleep behavior, sleep

attitudes, and such information as age, sex, height, and

weight. The students then received a copy of the Sleep Log

(Appendix 2) which was to be filled out at night before re-

tiring and in the morning upon rising. The night portion

asked for time and length of naps that day and the time of

going to bed with the intention of sleeping. The morning

portion asked for time of awakening, an estimate of










sleep-onset latency the night before, and the number and

length of recalled awakenings during the night.

The summer group kept the Sleep Log for 4 weeks (28

days) and the winter group kept it for 2 weeks (14 days).

The Sleep Logs were in one-week booklets. All subjects were

required to return their Sleep Log at the end of each week

in order to maintain contact with the experimenter and

assure that the Sleep Logs were being completed properly.

The students were encouraged to be accurate to the nearest

5 minutes on recorded times. It was stressed, however, that

if the bedtimes or wake-up times could not be recalled to

at least the nearest 1/2 hour (e.g., due to forgetting to

fill out the Sleep Log on a given morning or evening), that

portion should be left blank with an explanatory note. They

were told that missing data was preferable to inaccurate

data. Such missing data was occasionally found in the sub-

sequent analysis.

When the last weekly Sleep Log was returned, subjects

were given the Post-Log Questionnaire (Appendix 3) which

asked for estimates of their sleep behavior during the time

of keeping the Sleep Log (past 2 or 4 weeks). The Question-

naires were not labeled Pre-Log and Post-Log when adminis-

tered, and the subjects were not aware that they would be

given the Post-Log Questionnaire. The winter group then

took the Minnesota Multiphasic Personality Inventory (MMPI).

The winter group also took the Marlowe-Crowne Social-

Desirability Scale (Crowne & Marlowe, 1964) and gave the









experimenter permission to obtain academic information from

the University Registrar.

Only one subject of 105 failed to complete the study;

he was eliminated from the calculations. Also eliminated

were two subjects who were ill during much of the experi-

mental period. The 102 subjects who completed the study

seemed to be quite accurate and conscientious in the re-

cording of their Sleep Logs. Probably as a result of the

procedures used (initially meeting in small groups with the

experimenter and close weekly contact) the majority seemed

to be personally involved and interested. Other conditions

may not produce carefully recorded sleep logs.















SECTION III

RESULTS


The results will be discussed in three parts: (a)

the presentation and comparison of group summary statistics

for Sleep Logs, Pre-Log Questionnaires, and Post-Log Ques-

tionnaires for both summer and winter groups, (b) intra-

individual comparisons of these three methods of measurement

and estimation of the individual stability of Sleep Log

measures over time, and (c) the interrelationships of Sleep

Log parameters with each other, with sleep attitudes, and

with nonsleep measures such as personality. All data for

amounts of time and time of day are presented in tenths of

hours rather than minutes, and a 24-hour clock is used. A

time of 0.8 would refer, for example, to .8 of an hour past

midnight, or about 12:48 a.m.


Group Sleep Behaviors

The group average sleep measures are approached from

various perspectives. These include an analysis of Sleep

Log and Questionnaire differences, sex differences, winter

and summer group differences, and differences between vari-

ous sleep parameters such as sleep on weekdays and weekends.









Comparison of Methods of Measuring Sleep

Tables 1, 2, and 3 present summary data for all 102

students on the Sleep Log, Pre-Log Questionnaire, and Post-

Log Questionnaire. Several comparisons have been done in

order to discover if the particular type of measure used

influences the group mean value of comparable sleep para-

meters.

One type of comparison is between the Pre-Log Ques-

tionnaire estimates of usual sleep and the Post-Log Ques-

tionnaire estimates of sleep the past few weeks. The

differences between the various comparable measures are

generally on the order of .1 to .3 hour. The largest dif-

ference is for number of naps per week, with the Post-Log

estimate .6 hour more than the Pre-Log estimate. Pairs of

comparable measures for the Pre-Log and Post-Log Question-

naires are correlated about r = +.65 to r = +.79. This

high degree of relationship affects t tests for related

means. As a result, all of the differences of .2 hour or

greater (8 of 11 comparisons) are significant at the p < .05

level for n = 102.

Differences between the Pre-Log Questionnaire and com-

parable Sleep Log measures are mostly on the order of .2

hour to .3 hour. The only differences larger than .3 hour

are for weekday bedtime (.6 hour) and weekday wake-up time

(.4 hour). In both cases the Pre-Log estimates earlier

times than the Sleep Log. Again, because of sizable










TABLE 1

Summary Statistics for All 102 Subjects for Sleep Log


Sleep Log variable Mean SDa

Sleep length per night* 7.5 hr .85

Sleep length per weeknight* 7.4 hr .92

Sleep length/weekend night* 7.9 hr 1.07

Sleep length per 24 hours 7.9 hr .82

Bedtime 0.9 am 1.02

Weekday bedtime 0.6 am 1.02

Weekend bedtime 1.6 am 1.34

Wake-up time 8.5 am .89

Weekday wake-uptime 8.1 am .95

Weekend wake-uptime 9.6 am 1.14

Individual's SD of length 1.55 hr .63

Individual's SD of bedtime 1.28 hr .52

Individual's SD of wake-up time 1.29 hr .50

Number of naps per week 1.6 1.32

Hours of naps per week 2.6 hr 2.85

Number of wake-ups/night 0.9 .60

Minutes awake/night (median) < 5 min

Sleep onset latency (median) 5-15 min


*Sleep lengths are here calculated as bedtime to wake-up
time.

aEach standard deviation is calculated from the deviations
of each subject's mean from the grand mean.









Table 2

Summary Statistics for All 102 Subjects for
Pre-Log Questionnaire


Question # Mean SD

1. Sleep length per night 7.4 hr .81

Sleep length per night* 7.9 hr .94

Sleep length/weeknight* 7.7 hr 1.07

Sleep length/weekend night* 8.2 hr 1.30

2. Sleep length needed 7.7 hr .89

3. Sleep length preferred 8.7 hr 1.01

5. Number of naps per week 1.7 1.80

6. Hours of naps per week 2.4 hr 3.06

7. Number of wake-ups/night 0.8 .87

16. Weekday bedtime 24.0 pm .99

17. Weekend bedtime 1.6 am 1.27

18. Weekday wake-up time 7.7 am .90

19. Weekend wake-up time 9.8 am 1.30

Note. All questions asked about usual or average sleep
behavior (see Appendix 1).

*These sleep-length estimates were calculated from question-
naire answers on bedtime and wake-up time for weekdays and
weekends (#16 to #19).










Table 3

Summary Statistics for All 102 Subjects for
Post-Log Questionnaire


Question # Mean SD

1. Sleep length per night 7.3 hr .88

Sleep length per night* 7.7 hr .93

Sleep length/weeknight* 7.6 hr 1.02

Sleep length/weekend night* 8.1 hr 1.08

5. Number of naps per week 2.0 2.02

6. Hours of naps per week 3.0 hr 3.40

7. Number of wake-ups/night 1.1 .87

10. Weekday bedtime 0.4 am 1.08

11. Weekend bedtime 1.4 am 1.21

12. Weekday wake-up time 8.0 am .99

13. Weekend wake-up time 9.5 am 1.08

14. Sleep length of peers 7.5 hr .62


Note. All questions except #14 asked about average sleep
behavior for the past 4 (summer), or 2 (winter) weeks
(see Appendix 3).









correlations between comparable measures, differences of .2

hour or more (7 of 11 comparisons) are generally significant

at the p < .05 level.

Comparisons of group means for the Post-Log Question-

naire and Sleep Log reveal that the Post-Log estimates are

usually closer to the Sleep Log than are those of the Pre-

Log Questionnaire (although 8 of 11 are different by .2 hour

or greater). This is especially true for the two measures

on which the Pre-Log Questionnaire and Sleep Log differ

most; for weekday bedtime, instead of .6 hour, the Post-Log

estimate differs only .2 hour from the Sleep Log, and the

Post-Log weekday wake-up time differs only .1 hour from the

Sleep Log instead of .4 hour.

In summary, if a difference of .3 hour or less between

any of the three tools used to measure sleep behavior is

not considered to be a large one, then the three types of

measures produce similar results for group means, with the

few exceptions noted above. The Pre-Log Questionnaire,

which asks about usual sleep behavior, predicts the group's

actual sleep behavior as measured over a 2- or 4-week

period. The group also fairly accurately estimates the

average sleep behavior of the past 2 or 4 weeks on the Post-

Log Questionnaire, during which time a daily record of

sleep was kept. One interesting finding is that the vari-

ances around the means do not differ between comparable

measures on the Sleep Log and Pre-Log and Post-Log Ques-

tionnaires.










If one did decide to use a sleep log for gathering

group data, a choice would have to be made about the length

of time for which the sleep log should be kept. As will be

seen below, 1 week should be minimum in order to assure the

inclusion of differences between weekday and weekend sleep.

Extension of the sleep log beyond 1 week may, however, not

be necessary for group data. Table 4 is derived totally

from Sleep Log data. This table first lists the group

means for the first 2-week period and the last 2-week period

of the 4-week Sleep Log kept by the summer group. The means

for these two biweekly periods are essentially the same for

all of the sleep parameters. The means for each of the 4

separate weeks (not shown) are also essentially equal. Com-

parable data appear for the winter group, for which the

means for Week 1 and Week 2 of the 2-week Sleep Log are

virtually identical. These results indicate that for group

data, a 1-week sleep log can yield mean values which are

quite stable across time--at least in this particular en-

vironment and population. Of course, large changes in the

environment (final-exam week, vacation time) may alter some

of the mean values.


Comparison of Males and Females

In addition to the analysis of methodological issues

in the measurement of group sleep behavior, several other

types of group analyses were done to examine the sleep of

these college students.











N C 4 E

) C) a) L N a



S N ro CD n (N H l





4-













0 4-J
44
() N m N Ul H cj







ra









4 o
(0



a 0




'r 4-J

a 4 (4
a) 4-) -14 -.a-Ha a)
a (n m(L H &1.0 &-0 I
A . . .m
I LN N0 r- HN H H a
Src4
o e o



i E
0 a) a) . -a a) a) *H
(a H aa) r') '0 a) '








L"-4 4-1

1)
0aj
o a a. H
-H 1- 41 *H I u

(a 4-1 0 a) <
a .4 >1 a a) a) _) 1-4

0 *. 'a) r .0
UCa a0 1 4-4 4-4 a)
H1 4 0 0 0 o
I-4 a) I-4
( a W a)
*H 04 Un co CW)
0-4 UW 4-
> 4-i (a e - U)
1) j 0) -H H H H a)
o) s: a 4-) a) (a (a a)
O (L) "-4 a) 3 a 3H

,- O rHl a) a d ') -

H 0 H ) (a z Z a 4J
U) U) Ba U) M (a H H H-










First, for the 44 males and 58 females, a sex dif-

ference analysis was done for all of the measures on the

Sleep Log, Pre-Log Questionnaire, and Post-Log Questionnaire

that are listed in Tables 1, 2, and 3. As expected, there

are essentially no sex differences in sleep behaviors or

attitudes. For the sleep behaviors, the largest differences

between the sexes are on the order of .3 hour; none of these

differences are significant at the p < .05 level. For the

sleep attitudes, only one significant difference is present;

the females say that they feel they usually need 7.9 hours

of sleep per night vs. 7.5 hours for the males, t (100)

2.22, p < .05. Thus, regardless of the method used to

measure sleep, sleep behavior does not seem to differ between

sexes for this college group.


Comparison of Summer and Winter Groups

Another comparison of group data is between the summer

and winter groups for each variable of the three measurement

tools. A comparison of Sleep Log data between the summer

and winter groups (Table 5) reveals that they are similar

in most respects. The major difference between the two

groups is for weekday wake-up time, with the winter group

rising .7 hour later, t (100) = 3.89, p < .001. This dif-

ference is responsible for the winter group's getting .4

hour more sleep on weeknights than the summer group, t

(100) = 2.22, p < .05. As weekend sleep does not differ

between the groups, however, overall sleep length per night










Table 5

Summary Statistics for Summer and Winter Sleep Logs


Summer Winter
Sleep Log variable Mean SD Mean SD


Sleep length per night* 7.4 hr .92 7.7 hr .77

Sleep length per weeknight* 7.2 hr .96 7.6 hr .82

Sleep length/weekend night* 8.0 hr 1.11 7.9 hr 1.04

Sleep length per 24 hours 7.9 hr .82 8.0 hr .83

Bedtime 0.8 am .97 1.0 am 1.06

Weekday bedtime 0.5 am .95 0.7 am 1.08

Weekend bedtime 1.6 am 1.46 1.6 am 1.24

Wake-up time 8.2 am .78 8.7 am .92

Weekday wake-up time 7.7 am .82 8.4 am .96

Weekend wake-up time 9.5 am 1.09 9.6 am 1.19

Individual's SD of length 1.58 hr .51 1.53 hr .73

Individual's SD of bedtime 1.33 hr .61 1.24 hr .41

Individual's SD of wake time 1.39 hr .48 1.20 hr .51

Number of naps per week 1.9 hr 1.33 1.3 1.25

Hours of naps per week 3.3 hr 2.90 2.1 hr 2.70

Number of wake-ups/night 1.0 .60 0.8 .58

Minutes awake/night (median) < 5 min None

Sleep-onset latency (median) 5-15 min 5-15 min


Note. Summer data is from 4 weeks (n = 48); winter data is from 2 weeks
(n = 54).

*Sleep lengths are here calculated as bedtime to wake-up time.










does not differ significantly. Although the winter group

got slightly more sleep per night overall (.3 hour), they

took significantly fewer naps per week (1.9 vs. 1.3), t

(100) = 2.18, p < .05. The sleep per 24 hours thus differs

by only .1 hour between the summer and winter groups.

Table 6 presents the means and standard deviations of

data from some of the questions on the Pre-Log Questionnaire

about usual or average sleep behavior for the summer and

winter groups. (Questions which allowed only 3 or 4 cate-

gories for answers are presented in Appendix 4 along with

the frequency of answers in each category.) The differences

between the two groups follow the same pattern as those for

the Sleep Log. The winter group claims to usually have an

average weekday wake-up time of .5 hour later than the

summer group, with a weeknight sleep length of .6 hour

longer than the summer group. They also claim to usually

take significantly fewer naps and spend less time napping.

Table 7 presents summary statistics for some of the

questions from the Post-Log Questionnaire. (See Appendix

5 for the frequency distributions of answers with 3 or 4

categories only.) This Questionnaire asked about sleep

behavior for the past 4 weeks (summer) or 2 weeks (winter),

in addition to several other more general questions about

sleep. It was administered on the same day that the last

week's Sleep Log was completed. As with the Sleep Log and

Pre-Log Questionnaire, the summer and winter groups are










Table 6

Summary Statistics for Summer and Winter
Pre-Log Questionnaire


Summer Winter
Question # Mean SD Mean SD


1. Sleep length per night

Sleep length per night*

Sleep length/weeknight*

Sleep length/weekend night*

2. Sleep length needed

3. Sleep length preferred

5. Number of naps per week

6. Hours of naps per week

7. Number of wake-ups/night

16. Weekday bedtime

17. Weekend bedtime

18. Weekday wake-up time

19. Weekend wake-up time

27. Bedtime variability

28. Wake-up time variability


7.4 hr

7.7 hr

7.4 hr

8.2 hr

7.6 hr

8.5 hr

2.1

3.2 hr

1.0

0.1 am

1.5 am

7.5 am

9.6 am


.87 7.5 hr

1.07 8.0 hr

1.18 8.0 hr

1.41 8.3 hr

.86 7.9 hr

.89 8.8 hr

1.86 1.3

3.45 1.6 hr

.82 .6

.92 24.0 pm

1.42 1.7 am

.88 8.0 am

1.20 9.9 am

--- 2.0 da'

-- 1.5 da'


.75

.78

.91

1.20

.91

1.10

1.66

2.45

.88

1.06

1.12

.88

1.38

1.16

1.06


Note. All questions asked about usual or average sleep behavior (see
Appendix 1). Summer n = 48; winter n = 54.

*These sleep length estimates were calculated from Questionnaire answers
on bedtime and wake-up time for weekdays and weekends (#16 to #19).










Table 7

Summary Statistics for Summer and Winter
Post-Log Questionnaire


Summer Winter
Question # Mean SD Mean SD


1. Sleep length per night 7.3 hr .83 7.3 hr .93

Sleep length per night* 7.6 hr .94 7.8 hr .91

Sleep length-weeknight* 7.4 hr 1.02 7.7 hr 1.01

Sleep length/weekend night* 8.2 hr 1.14 8.0 hr 1.02

5. Number of naps per week 2.4 2.14 1.7 1.86

6. Hours of naps per week 3.4 hr 3.47 2.5 hr 3.31

7. Number of wake-ups/night 1.1 .80 1.0 .92

10. Weekday bedtime 0.3 am 1.02 0.5 am 1.13

11. Weekend bedtime 1.2 am 1.27 1.6 am 1.15

12. Weekday wake-up time 7.7 am .85 8.2 am 1.06

13. Weekend wake-up time 9.4 am 1.06 9.6 am 1.09

14. Sleep length of peers 7.4 hr .62 7.5 hr .61


Note. All questions except #14 asked about average sleep behavior for
the past 4 (summer), or 2 (winter) weeks (see Appendix 3). Summer n =
48; winter n = 54.

*These sleep-length estimates were calculated from Questionnaire
answers on bedtime and wake-up time for weekdays and weekends (#10 to
13).









different in their reports of weekday wake-up time and nap

behavior; the winter group claims to have awakened later on

weekdays and to have taken fewer naps.

In brief, the two groups of students do not differ in

most measures, including nighttime sleep length per week

and sleep per 24 hours. The few measures which do differ

are the same for the Sleep Log, Pre-Log Questionnaire, and

Post-Log Questionnaire.


Differences Between Various Sleep Parameters

Examination of the various measures calculated from the

Sleep Log (Table 1) reveals differences between some of the

group averages. The students went to bed 1.0 hour later on

weekends (Friday and Saturday nights) than on weekdays.

They awakened 1.5 hours later on weekends (Saturday and Sun-

day mornings). The weekend sleep length is thus .5 hour

longer than weekday sleep length, t test for related means

(101) = 5.10, p < .001. Another difference is between sleep

length per night and sleep length per 24 hours. Due to

naps, this difference is .4 hour, t (101) = 9.85, p < .001.

These differences between Sleep Log measures are also

present for both the winter and summer groups taken sepa-

rately.

One result worth noting is the short amount of time

spent in going to sleep and in awakenings during the night.

The median and modal category for latency to sleep onset is

5 to 15 minutes, and the amount of time spent awake after










sleep onset is less than 5 minutes. These results are

similar to those found in other studies of normal young

adults.

One comparison of Sleep Log values which does not

produce a difference is that between the means for indi-

viduals' standard deviation of bedtime and standard devia-

tion of wake-up time. These values were obtained by first

finding each person's standard deviation around his mean

bedtime and mean wake-up time over the 2- or 4-week period,

and then finding the average of these standard deviations

across the 102 students. For these college students, there

is no difference between the average variability of bedtime

(SD = 1.28) and average variability of wake-up time (SD =

1.29). One-half (50 of 102) of the students have a wake-up

time standard deviation which is larger than their bedtime

standard deviation. This large a standard deviation of

wake-up time is likely due to the lack of a need to awaken

each day at the same time because of varying schedules for

classes throughout the week.

The Pre-Log Questionnaire results are similar to the

Sleep Log in that the students claim to go to bed and wake

up later on weekends than on weekdays. The Post-Log Ques-

tionnaire also exhibits the expected differences between

weekday and weekend sleep.









Intra-Individual Comparisons

In addition to describing and studying the sleep be-

havior of groups, investigators often attempt to study the

sleep of particular individuals relative to others in the

population and to relate the sleep of individuals to other

measures of interest. This section presents the intra-

individual relationships between the three methods of

measuring sleep which were studied. Also presented is the

stability for individuals of Sleep Log measures over time

for both the 2-week and 4-week Sleep Logs.


Stability of Sleep Log Parameters
and Questionnaire Questions

Before looking at the intra-individual relationships

between the Sleep Log and the Questionnaires, we shall

examine the reliability or stability of these methods over

time. Table 8 presents the correlations of the individuals'

two biweekly values of various Sleep Log parameters for the

summer group. Sleep-length biweekly means have a correla-

tion of r = +.88, which indicates a high degree of pre-

dictability of one's biweekly mean sleep length if the pre-

vious 2-week's mean is known. Of the 48 subjects in the

summer group, 31 (65%) have differences between these two

biweekly means of less than .5 hour. Two (4%) have bi-

weekly differences of 1 hour or more. When naps are added

to the nightly sleep length, the correlation between the

two resulting biweekly means of sleep length per 24 hours












6. o ol \ o\o co mo rn in

+ II 1 CM + + + + + + +

II II II II II II II II
'NJ lC' -4


0C
O-4
-H M
4-T (4 ()
.--I Q)

O-1 41
0 0 -I
UO




0-

-41 00)
(0 0) (U


4 >

Uo


dp o Ln
m a(3+


d4








6 o\-1
m
IN +
.--4 4












ll
















*
ai
(- (4-







S0




in
n C4

-o
,C 4-1



0 0



0 U)
On 0


oo4

rh H-


m D
inI in
N Ln




34 in
II II












H -




















4-1




0' -
0 0


S O
0)








O O
(D 4-1



H a0




0 0
Q -
Enl rl

0 0
0 '0
*H- *H
> >
'0 '0

H H


0 -N
Or



fi 44 01
0 0

0 0-r
u S


-4






O)
0
0





a







0

-,-I
o




4.r
.0 -i

-)H


-rl
-)


-r4


0
i-C



N0





a4

0 4
0









is r = +.75, with only 22 (45%) persons having less than a

.5-hour difference, and 5 (10%) having a 1-hour or more

difference.

The correlation in Table 8 for sleep length per night

is r = +.62 for Week 1 vs. Week 2 of the 2-week winter Sleep

Log, with 22 (41%) persons having less than a .5-hour dif-

ference. This correlation is significantly lower than the

biweekly comparisons of the summer 4-week Sleep Log (r =

+.88), but is comparable to the correlation of Week 1 and

Week 2 on the summer Sleep Log (r = +.68), as seen in the

second column of Table 8. When naps are added to the sleep

length of the winter group the correlation does not change

(r = +.63), and a few more persons (52%) then have less than

a .5-hour difference. For both sleep per night and sleep

per 24 hours, 11 (20%) of the winter subjects have a dif-

ference between weekly means of 1 hour or more.

A separate analysis of the week-to-week stability of

Sleep Log sleep length for those persons above and below

the mean in sleep length and also for those above and below

the mean in variability (SD) of sleep length produces no

differences for both the winter and summer groups. These

variables do not seem to be related to the week-to-week

stability of sleep length.

In contrast to the correlations of r = +.88 and r =

+.68 for the 2-week and 1-week means for the summer group,

an ANOVA correlation of daily sleep lengths over the 28 days










yields a coefficient of r = +.23 as an estimate of the ex-

pected relationship between one day and another.

Bedtime correlations and wake-up time correlations for

biweekly means (summer group) and weekly means (summer and

winter groups) indicate that the stability of these measures

does not change with an increase in the length of measurement

beyond 1 week. These correlations range from r = +.74 to

r = +.87, indicating that the weekly averages for the

placement of sleep are fairly stable.

Table 8 also presents the correlations of the weekly

and biweekly standard deviation of each individual's mean

sleep length, bedtime, and wake-up time. The values in-

dicate that variability within individuals is significantly

correlated from week to week, although not generally as

highly as the individual's means of these parameters. Vari-

ability is, then, to some extent stable.

A 2-week sleep log can, thus, provide fairly reliable

measures for the basic sleep parameters. The test-retest

reliability of the Pre-Log Questionnaire is probably also

fairly good, although no such coefficient is directly avail-

able for these particular students. This is due to the

fact that the Pre-Log Questionnaire and the Post-Log Ques-

tionnaire do not ask exactly the same questions (i.e., one

asks for usual sleep, while the other asks for sleep the

past few weeks). When the Pre-Log Questionnaire and Post-

Log Questionnaire questions asking for similar but not









identical information are correlated, the coefficients

range from r = +.65 to r = +.79. For example, the two ques-

tions on sleep length have a correlation of r = +.67. This

is highly similar to a 2-week test-retest correlation of

r = +.68 found by Webb and Stone (1963) on a question about

usual sleep length answered by several hundred college

students. Thus, a reliability coefficient of r = +.67 is a

reasonable expectation for the stability of a college stu-

dent's answer to a question on usual sleep length.


Correlations Between Methods of Measurement

It may often be of interest to know how accurate indi-

vidual persons are in estimating their sleep parameters.

Table 9 (summer group) and Table 10 (winter group) show the

degree of relationship between each individual's Question-

naire estimates and his corresponding Sleep Log parameters.

There are no significant differences between the summer and

winter groups in the degree of relationship between the

Sleep Log measures and comparable Questionnaire estimates.

The degree of these relationships is thus replicated and is

probably generalizable to other groups--at least college

students. As the two groups of students do not differ,

attention can be focused to Table 11, which presents the

same data for all 102 students combined.

For the total group of 102 students, the correlation

between the Sleep Log mean sleep length and the Pre-Log

question #1 is r = +.70, with 40% of the subjects having a










Table 9

Relationships Between Individual's 4-Week Sleep Log
Measures and Corresponding Questionnaire Estimates:
Summer Group (n = 48)


Comparisons Pre-Log Post-Log


Sleep Log Mean sleep length vs.
question #1 on Questionnaires*

Mean of difference scores between
Sleep Log and Questionnaire

Standard deviation of differences

Number of persons with less than
.5 hr difference

Number of persons with .5 to less
than 1 hr difference

Number of persons with 1 hr or
more difference

Sleep Log mean weekday bedtime
vs. Questionnaire estimate

Sleep Log mean weekend bedtime
vs. Questionnaire estimate

Sleep Log mean weekday wake-up time
vs. Questionnaire estimate

Sleep Log mean weekend wake-up time
vs. Questionnaire estimate

Sleep Log SD of bedtime vs. question
#14 on bedtime regularity

Sleep Log SD of wake-up time vs.
question #15 on wake-up regularity

*Sleep log lengths are here calculated
time.


r = +.74 r = +.73


0.0 hr

.65 hr


0.1 hr

.65 hr


31 (65%) 26 (54%)


11 (23%) 18 (38%)


6 (12%) 4 ( 8%)


r = +.78 r = +.90


r = +.68 r = +.82


r = +.72 r = +.91


r = +.56 r = +.85


r = +.07


r = +.19

as bedtime to wake-up









Table 10

Relationships Between Individual's 2-Week Sleep Log
Measures and Corresponding Questionnaire Estimates:
Winter Group (n = 54)


Comparisons Pre-Log Post-Log

Sleep Log mean sleep length vs.
question #1 on Questionnaires* r = +.66 r = +.74

Mean of difference scores be-
tween Sleep Log and Question-
naire 0.2 hr 0.4 hr

Standard deviation of differ-
ences .62 hr .63 hr

Number of persons with less
than .5 hour difference 30 (55%) 22 (41%)

Number of persons with .5 to
less than 1 hour difference 15 (28%) 21 (39%)

Number of persons with 1 hour
or more difference 9 (17%) 11 (20%)

Sleep Log mean weekday bedtime vs.
Questionnaire estimate r = +.78 r = +.91

Sleep Log mean weekend bedtime vs.
Questionnaire estimate r = +.64 r = +.87

Sleep Log mean weekday wake-up time
vs. Questionnaire estimate r = +.73 r = +.91

Sleep Log mean weekend wake-up time
vs. Questionnaire estimate r = +.59 r = +.76

Sleep Log SD of bedtime vs. question
#14 on bedtime regularity r = +.22

Sleep Log SD of wake-up time vs.
question #15 on wake-up time regu-
larity r = +.20

Sleep Log SD of bedtime vs. question
#27 on bedtime regularity r = +.28











Table 10--(continued)


Comparisons Pre-Log Post-Log

Sleep Log SD of wake-up time vs.
question #28 on wake-up time
regularity r = +.06

*Sleep Log lengths are here calculated as bedtime to wake-up
time.










Table 11

Relationship Between Individual's Sleep Log Measures and
Corresponding Questionnaire Estimates for All 102 Subjects


Comparisons Pre-Log Post-Log


Sleep length: Sleep Log mean vs. length
from Questionnaire question #1 r = +.70 r = +.72

Mean of difference scores (Log Quest.) .1 hr .2 hr

Standard deviation of difference scores .64 hr .65 hr

Persons with less than .5 hr difference 61 (60%) 48 (47%)

From .5 to less than 1.0 hr difference 26 (25%) 39 (38%)

1.0 hr or greater difference 15 (15%) 15 (15%)



Sleep length: Sleep Log mean vs. length
from Questionnaire bed and wake-up times r = +.72 r = +.83

Mean difference scores (Log Quest.) -.3 hr -.1 hr

Standard deviation of difference scores .67 hr .52 hr

Persons with less than .5 hr difference 55 (54%) 67 (66%)

From .5 to less than 1.0 hr difference 27 (26%) 26 (25%)

1.0 hr or greater difference 20 (20%) 9 ( 9%)



Naps: Sleep Log mean hours per week
vs. Questionnaire estimate r = +.80 r = +.85



Naps: Sleep Log mean number per week
vs. Questionnaire estimate r = +.71 r = +.92










Table 11--(continued)


Comparisons Pre-Log Post-Log


Weekday bedtime: Sleep Log mean vs.
Questionnaire estimate r = +.78 r = +.91

Mean of difference scores (Log Quest.) .6 hr .2 hr

Standard deviation of difference scores .67 hr .45 hr

Persons with less than .5 hr difference 45 (44%) 65 (64%)

From .5 to less than 1.0 hr difference 28 (28%) 31 (30%)

1.0 hr or greater difference 29 (28%) 6 ( 6%)



Weekend bedtime: Sleep Log mean vs.
Questionnaire estimate r = +.66 r = +.84

Mean of difference scores (Log Quest.) .0 hr .1 hr

Standard deviation of difference scores 1.07 hr .74 hr

Persons with less than .5 hr difference 40 (39%) 60 (59%)

From .5 to less than 1.0 hr difference 30 (30%) 28 (27%)

1.0 hr or greater difference 32 (31%) 14 (14%)



Regularity of bedtime: Sleep Log SD vs.
Questionnaire question #14 r = +.13









Table 11--(continued)


Comparisons Pre-Log Post-Log


Weekday wake-up time: Sleep Log mean
vs. Questionnaire estimate r = +.74 r = +.91

Mean of difference scores (Log Quest.) .3 hr .1 hr

Standard deviation of difference scores .66 hr .40 hr

Persons with less than .5 hr difference 48 (47%) 77 (75%)

From .5 to less than 1.0 hr difference 37 (36%) 21 (21%)

1.0 hr or greater difference 17 (17%) 4 ( 4%)



Weekend wake-up time: Sleep Log mean
vs. Questionnaire estimate r = +.58 r = +.80

Mean of difference scores (Log Quest.) -.2 hr .1 hr

Standard deviation of difference scores 1.14 hr .70 hr

Persons with less than .5 hr difference 44 (43%) 54 (53%)

From .5 to less than 1.0 hr difference 31 (30%) 33 (32%)

1.0 hr or greater difference 27 (27%) 15 (15%)



Regularity of wake-up time: Sleep Log SD
vs. Questionnaire question #15 r = +.19










difference between the two values of .5 hour or more. It

could be possible that a part of the difference between

individual's Pre-Log estimates of usual sleep and their

Sleep Log mean sleep length is due to the period of the

Sleep Log not being representative of the "usual" sleep

being described on the Pre-Log Questionnaire. Several fac-

tors suggest, however, that this explanation is not suffi-

cient to explain all of the difference between the Sleep

Log and the Pre-Log Questionnaire.

First, the relationship between the Sleep Log mean

sleep length and the Post-Log question #1, which asks about

sleep length during the period of keeping the Sleep Log,

is essentially the same as between the Sleep Log and Pre-

Log Questionnaire means. For all 102 students this corre-

lation is r = +.72, with 53% of the students making errors

of .5 hour or more in estimating sleep length for the past

few weeks. This is after spending 2 to 4 weeks attending

to sleep behavior daily.

Another factor also suggests that a sleep log period

of "unusual" sleep does not, per se, account for the dis-

crepancies between the Pre-Log estimates and the Sleep Log

mean sleep lengths. Subjects were asked on the Post-Log

Questionnaire if their sleep length for the past 2 or 4

weeks was "about the same as usual," "more than usual," or

"less than usual." For all 102 students, of the 52 persons

answering more or less (i.e., different), only 19 (37%)










have a .5-hour or more difference between their Sleep Log

mean and their Pre-Log estimate of average sleep length.

In comparison, of the 50 persons answering "about the same

as usual," 21 (42%) have a .5-hour or more difference be-

tween the Sleep Log mean and Pre-Log estimate. Students

who said that their Sleep Log sleep was more or less than

usual have essentially the same degree of Pre-Log/Sleep Log

discrepancy as those saying that the Sleep Log sleep was

usual. Thus, the amount of discrepancy between the Pre-Log

estimate of usual sleep length and the Sleep Log mean can-

not be explained by merely attributing it to the persons

who said their Sleep Log period did not reflect usual sleep.

Data on the relationship between each individual's

Sleep Log mean bedtime and wake-up time and the corre-

sponding Questionnaire estimates are also presented in

Tables 9, 10, and 11. As seen in Table 11, the correla-

tions between the Pre-Log estimates of usual bedtime and

wake-up time for weeknights and weekends and the Sleep Log

means for these four values are between r = +.58 and r =

+.78. The corresponding correlations between the Post-Log

estimates and the Sleep Log means are between r = +.80 and

r = +.91. In general, the Post-Log estimates of bedtime

and wake-up time are more closely related to the Sleep Log

values than the Pre-Log estimates are to the Sleep Log.

The actual amount of discrepancy between the Sleep

Log and Questionnaire estimates of the placement of sleep










in time can be seen on the second and third pages of Table

11. For example, for weekday bedtime, only 44% of the 102

students have less than a .5-hour difference between the

Sleep Log mean and the Pre-Log question on usual weekday

bedtime. When this Sleep Log mean is compared to the Post-

Log question on weekday bedtime the past few weeks, 64% of

the 102 students have less than a .5-hour discrepancy.

An effort was made to discover if certain subgroups of

persons are more likely to have large discrepancies between

the Sleep Log mean sleep length and the Pre-Log estimate of

sleep length. Each student's discrepancy (absolute value)

was correlated with both his sleep length and his vari-

ability (SD) of sleep length. Neither variable is signifi-

cantly related to the size of the difference between the

Sleep Log and the Pre-Log Questionnaire. It thus seems

that one's sleep length or variability of sleep length is

not a predictor of the size of the discrepancy between the

two methods of measuring sleep length.

Sleep length is, however, related to the direction of

one's difference between one's Pre-Log estimate and Sleep

Log sleep length. The Sleep Log sleep length was correlated

with the Sleep Log length minus the Pre-Log question #1

estimate of sleep length. Persons with Pre-Log estimates

that are less than the Sleep Log parameters are the long

sleepers; the short sleepers tend to overestimate their

sleep length on the Pre-Log Questionnaire (summer, r = +.42,

p < .01; winter, r = +.45, p < .001).









A final comparison between the Questionnaires and the

Sleep Log is seen in the relationships between the indi-

vidual's standard deviation around his mean bedtime and

wake-up time on the Sleep Log and his Pre-Log estimates of

the regularity of bedtime (question #14) and wake-up time

(question #15). As seen in Table 9 (summer group) and

Table 10 (winter group), there is essentially no relation-

ship present. Apparently, when the question is asked in

the form, "How regular are your bedtimes?", persons can-

not correctly evaluate themselves relative to others in

terms of bedtime or wake-up time variability.

The winter group had in addition to questions #14 and

#15 on the regularity of sleep times, two further questions

about regularity of bedtime (question #27) and wake-up time

(question #28). As seen in Appendix 3, these questions

asked for estimates of regularity not in terms of comparing

oneself to others, but in terms of one's own deviations

around one's own mean bedtime and wake-up time. The ques-

tion on wake-up time regularity again shows essentially no

correlation with the standard deviation of Sleep Log wake-up

times. The question on bedtime regularity does correlate

significantly with the standard deviation of Sleep Log bed-

times, but the degree of relationship is not large, r =

+.28, p < .04.










Interrelationships

The present section describes various types of corre-

lations among the sleep parameters themselves and between

sleep measures and other measures taken on each individual

student. One type of relationship examined is between the

various Sleep Log measures. Another is between various

sleep behaviors as measured by the Questionnaires. A third

type of relationship explored is between the Sleep Log and

measures of attitudes and feelings about one's sleep as

obtained from the Questionnaires. Finally, data will be

presented on the relationships of Sleep Log parameters to

such nonsleep measures as personality (MMPI), scholastic

achievement, and a weight/height ratio.


Correlations Among Sleep Log Parameters

Tables 12 and 13 present the relationships among vari-

ous Sleep Log measures. If the summer and winter groups

are viewed as a study with replication, it can be concluded

that sleep length is negatively related to bedtime (late

retirers get less sleep). Sleep length is also negatively

related to the variability of sleep as measured by the per-

son's standard deviation of sleep length, bedtime, and

wake-up time; short sleepers as individuals are more vari-

able in their day-to-day sleep behavior.

As expected from the above, variability itself is posi-

tively related to bedtime; persons with late average bed-

times tend to be those with highly variable bedtimes. These











Table 12

Correlations Among Individual's Means and Standard Deviations on
Various Sleep Log Parameters: Summer Group


Sleep Log variable 1 2 3 4 5 6


1. Sleep length per night

2. Bedtime -.67

3. Wake-up time +.26 +.53

4. Individual's SD of
sleep length -.48 +.48 +.14

5. Individual's SD of
bedtime -.42 +.35 -.03 +.67

6. Individual's SD of
wake-up time -.28 +.23 +.01 +.64 +.73

7. Hours of naps per week -.46 +.39 +.02 +.45 +.35 +.44


Note. The correlated variables are means or standard deviations of these
means from a 28-day sample of sleep behavior; n = 48. The correlation of
-.28 is significant at p < .05. The larger correlations are significant
at p < .01 or better.











Table 13

Correlations Among Individual's Means and Standard Deviations on
Various Sleep Log Parameters: Winter Group


Sleep Log variable 1 2 3 4 5 6


1. Sleep length per night

2. Bedtime -.56

3. Wake-up time +.03 +.78

4. Individual's SD of
sleep length -.48 +.51 +.41

5. Individual's SD of
bedtime -.20 +.29 +.20 +.46

6. Individual's SD of
wakeup time -.29 +.46 +.45 +.64 +.14

7. Hours of naps per week -.07 +.29 +.39 +.38 +.10 +.60


Note. The correlated variables are means or standard deviations of these
means from a 14-day sample of sleep behavior; n = 54. The correlations
of .29 are significant at p < .05. The larger correlations are signifi-
cant at p < .01 or better.









persons with late bedtimes or variable sleep also tend to

spend more time taking naps. Bedtime and wake-up time are

also highly related positively; a later average bedtime is

associated with a later wake-up time.

A correlation not shown in the tables is that between

sleep length on weekdays and weekends. The summer group

has a correlation of r = +.60; the winter group has a

correlation of r = +.45. Thus, the students as individuals

tend to remain either relatively long or short sleepers

throughout the full week, even though as a group they sleep

longer on weekends.

An unusual intra-individual relationship between week-

day and weekend sleep which was calculated by Johns et al.

(1971) was done here also. Johns took each individual's

questionnaire estimate of usual weekday sleep length (WD)

and subtracted it from his estimate of weekend sleep

length (WE). The resulting figure (WE WD) was correlated

with the weekday sleep length (WD). In the present study,

the correlation between (WE WD) and WD, as taken from the

Pre-Log Questionnaire, is r = -.43 for the summer group and

r = -.52 for the winter group (Johns obtained a correlation

of r = -.52). This negative relationship between (WE WD)

and WD suggests that those who claim to sleep more on week-

ends than their own weekday sleep length also claim to be

the relatively short sleepers on weekdays when compared to

the group as a whole. The same relationship is present,










but not as strongly, in the actual Sleep Log data (summer,

r = -.32; winter, r = -.35).


Correlations Between Sleep Log Behavior
and Sleep Attitudes

Several general questions concerning opinions, atti-

tudes, and feelings or impressions concerning one's sleep

were asked on the Pre-Log and the Post-Log Questionnaires.

Most of these were asked of both the summer and winter

groups and can be seen in Appendices 1 and 2; the distribu-

tion of answers to many of these questions are presented in

Appendices 4 and 5. The questions include such things as

whether enough sleep is usually obtained or was obtained

while keeping the Sleep Log, how much trouble the student

has going to sleep, how much he enjoys sleep, whether he

sleeps well, and feelings upon awakening. The answers to

these questions were correlated with the following actual

sleep behaviors calculated from each student's Sleep Log:

mean sleep length, standard deviation of sleep length, mean

bedtime, standard deviation of bedtime, mean wake-up time,

and standard deviation of wake-up time. In brief, almost

no measures are significantly (p < .05) related to Sleep

Log behavior for both the summer and winter groups; a corre-

lation of r = .26 is needed for each group in order to

meet this criterion.

Sleep Log mean sleep length is significantly associated

with only one question for both groups of subjects. This









is Post-Log question #3, which asks about whether enough

sleep was obtained during the past 2 or 4 weeks. The rela-

tionship is an expected one, with short sleepers tending to

say that they did not get enough sleep (summer, r = +.29;

winter, r = +.35).

Sleep Log mean bedtime is also significantly correlated

with only one measure for both groups. This is Pre-Log

question #4 asking if the person usually obtains enough

sleep. The summer group has a correlation of r = -.32 and

the winter group has a correlation of r = -.29. This sug-

gests that persons with late bedtimes tend to complain of

usually not getting enough sleep. The other four Sleep Log

parameters mentioned are not significantly related to any

of the Questionnaire measures for both the summer and winter

groups.

Some data related to these questions concerning getting

enough sleep are seen in a comparison of Sleep Log mean

sleep length to Pre-Log questions on sleep length needed

(question #2) and sleep length preferred (question #3).

The group average for sleep needed is only .2 hour more than

the actual sleep obtained; sleep length preferred is 1.2

hours more than average sleep length obtained. Comparable

findings have been reported by Strauch, Dubral, and Struch-

holz (1973).

Another interesting relationship is that between the

Sleep Log mean sleep length and the mean of the students'










estimates of the sleep length of their peers (Post-Log

question #14). The mean of the Sleep Log sleep lengths is

7.5 hours (SD = .85) and the mean of the estimates of peer

sleep length is also 7.5 hours (SD = .62). Thus, the stu-

dents as a group are quite accurate in their estimation of

the sleep length of their peers.

In addition to the measures taken on both the summer

and winter groups, several were taken only on the winter

subjects. Four of these are questions on sleep (Post-Log

questions #17 to #20); two of these correlate significantly

with some Sleep Log measures. Stating that one had at one

time slept at least 2 hours per night longer than usual for

2 weeks or more (question #18) is related to a large sleep-

length standard deviation (r = +.29) and large wake-up time

standard deviation (r = +.26) as calculated from the Sleep

Log. Usually feeling sleepy upon awakening (question #20)

is also related to a large standard deviation of sleep

length (r = +.37) and wake-up time (r = +.31). The gen-

eralizability of these relationships cannot be judged as

they are available only for the winter group.


Correlations Between Sleep Log Behavior
and Nonsleep Measures

A few nonsleep measures were taken for both the summer

and winter groups. These were college class, height,

weight, and a calculated ratio of height to weight. They

are not significantly related to Sleep Log measures for both

the summer and winter groups.









For the winter group the opportunity was taken to

collect more information on other aspects of each indi-

vidual's life and to relate this to his Sleep Log behaviors.

These additional measures include several "trait" measures:

the MMPI, the Marlowe-Crowne Social-Desirability Scale, the

grade-point averages and college entrance scores obtained

from the University Registrar, and a ratio of these last

two measures. Also included are questions #21 to #27 on

the Post-Log Questionnaire, which are "state" variables

rather than "trait" variables. They ask for a rating of

one's feelings and activities during the past 2 weeks when

compared to usually (more, about the same as usual, less).

When each subject's mean and standard deviation for

Sleep Log sleep length, bedtime, and wake-up time are

correlated with the trait measures, some MMPI scales are

significantly (p < .05) correlated with some of the sleep

measures, as seen in Table 14. These relationships will

have to be confirmed by further research before their gen-

eralizability can be established. The other trait vari-

ables are not significantly correlated with any of the

sleep measure. As seen in Table 14, the variability of

one's sleep length seems to be related to more personality

measures than is one's sleep length itself.

As an additional analysis of interest, for the winter

group, question #2 on the Post-Log Questionnaire was re-

lated to questions #21 through #27 by means of Chi-Square




















a




0 a
14 0
0 04
4cI Ila, i c
4 ) E m (1




0 00
U) 0
00 41



E N r
(0 0 H m m 0
H +J4I 4.1
S O0 m + + + +U

O c4
0 ( 0

10

0 o


) MI 4o1


0 4-


rd
a u

H 4J r- Q O 0 *
1 CQ (r iN m n

; 0 + + + + r

0
5 B)
4(4 0
0 0



0 Ha 4-1 CO Q)
.1l Q) m CN 4
41 (0
0 r0 1 H
H(U


O H

H P aU


14 4H
r4 4.;2'I)









tests. Question #2 asks whether sleep length the past two

weeks was the same as usual, more, or less. The other

questions ask whether various other activities and feelings

were the same as usual, more, or less. As there were not

sufficient expected frequencies in each cell of a 3 X 3

Chi-Square analysis, each variable was collapsed into two

dimensions prior to the analysis, resulting in a 2 X 2 Chi

Square. Significant (p < .05) results were obtained be-

tween question #2 and question #21 (X = 7.36), question

#22 (X2 = 5.28), and question #26 (X2 = 4.10). If these

relationships are confirmed by other research, it would

indicate that less sleep than usual (question #2) is related

to more mental activity (question #21), more depression

(question #22), and eating less food (question #26).















SECTION IV

DISCUSSION


The major thrust of this research has been to examine

the similarities and differences between sleep question-

naires and sleep logs--both for groups and individuals.

Both of these measurement tools have been used for various

purposes by researchers and practitioners. There have been

no reports, however, of comparisons of the two. Few

studies have considered questions about the reliability and

validity of questionnaire estimates of sleep parameters.

Several studies, as mentioned in the Introduction, have

examined the validity of sleep log types of questions (which

ask for subjective estimates of sleep behavior last night)

by comparing these to the EEG. Few studies, however, have

looked at the week-to-week stability of such sleep log

measures. Hopefully, this study will provide some informa-

tion on these topics for those who wish to study the sleep

of groups or individuals with sleep questionnaires or sleep

logs.

Although the main purpose of this research was to

study the measurement of sleep behavior, the opportunity

was taken to relate sleep behavior to other variables.









In the last section of this discussion, the findings in

this area will be related to previous research.

In this investigation of measurement techniques and

interrelationships, the college student sample used is

probably ideal from the viewpoint of providing extremes of

normal sleep behavior. Previous work had suggested that

students would have much between-subject and within-subject

variability, and this was the case for both the summer and

winter groups. The relative freedom in the placement of

sleep allowed by the nature of their daily schedules was

desirable for the discovery of relationships which might

otherwise be muted by the demands of a more regimented

situation. The specific results found are at present

limited, however, to groups with such sleep environments.

Hopefully, the variables and relationships examined here

will be studied in other groups.


Group Data

If one wishes to obtain group means and variances for

various types of sleep measures, a decision must be made

concerning whether to use a sleep questionnaire or a sleep

log. In addition to the obvious differences in cost, time,

and effort required for the two methods, several other fac-

tors must be considered.

First, it must be decided exactly what type of sleep

measure is desired. For example, one may be interested

only in an estimate of usual or average sleep. This may,










of course, be defined as average for the recent past or for

the past 2 years, etc. Or, one may be interested in the

sleep behavior of the group at the present time. The two

variables, usual sleep and present sleep, may not always be

the same. A second consideration is the acceptable toler-

ance for error. A small difference between a sleep log and

questionnaire may be statistically significant but not

meaningfully large for the purposes of the study.

If a measure of present group sleep behavior is de-

sired, a sleep log seems to be the choice which will give

the most accurate results. There are many significant dif-

ferences between the Post-Log Questionnaire which asks

about sleep the past few weeks and the Sleep Log which more

closely represents the actual average sleep of those weeks.

These differences are usually only a few tenths of an hour,

however. If such differences are not critical, a question-

naire about present sleep behavior may often be adequate to

gather information on the current behavior of groups.

An additional finding is that the variance around the

mean will probably not be different if one uses a question-

naire rather than a sleep log. Therefore, questionnaire

estimates of the expected range of average sleep behavior

for a group will be fairly valid.

Studies of usual sleep length do not typically involve

group means that are accurate to more than a few tenths of

an hour. A sleep questionnaire about usual sleep can give









means and standard deviations which are this close to sleep

log values. Furthermore, when one is interested in usual

sleep, a sleep log kept for a short period during unusual

circumstances may actually give results which are not

typical of average sleep. The Sleep Log in this study was

given during a relatively stable period of the academic

quarter. If it had been kept during the week of final

exams, the discrepancies between the Pre-Log Questionnaire

and the Sleep Log may have been greater.

In summary, sleep questionnaires seem to give results

comparable to a 2-week or 4-week sleep log when group means

and variances are the measures of interest. If this con-

clusion is to be applied to other groups, however, several

cautions are in order. First, college students may not be

typical in their ability to answer questionnaires about

their own behavior, or in the accuracy of their recording

of sleep logs for several weeks. Furthermore, these findings

cannot yet be generalized to other types of questionnaires

and sleep logs. For example, this Questionnaire asked for

estimates of sleep length in 1/2-hour intervals. Many

questionnaires use 1-hour intervals or none at all (i.e.,

simply leave a blank space for the answer). Also, the stu-

dents were asked to fill out the times on the Sleep Log to

the nearest 5 minutes. Many sleep logs are accurate only

to the nearest 1/2 hour. These differences may affect the

relationships between sleep questionnaires and logs.










On some occasions, one may wish to use a sleep log

rather than a questionnaire for gathering group data. One

reason may be to obtain a measure of the average day-to-day

variability of sleep length, bedtime, or wake-up time for

the group of interest. If a sleep log is used for gath-

ering group data, a 1-week sleep log may often give the

same results as a longer one if there are no major changes

in the environment that change the sleep of the group as a

whole; in this study the weekly group means of the sleep

measures do not differ. Using less than a 1-week sleep

log is not advisable if a picture of the total sleep be-

havior is desired because of weekday and weekend differ-

ences.

The winter and summer groups basically differ on only

two dimensions--weekday wake-up time and nap behavior. The

winter students awakened .7 hour later on weekdays than the

summer students on the average. The winter students claim

on the Pre-Log Questionnaire to have a usual weekday wake-

up time of .5 hour later than the summer students. The

fact that they claim to usually awaken later on weekdays

may be interpreted in several ways. They may as a group

in fact always awaken later than the summer group, even

during the summer. Or they may awaken later only during

the winter (due to later sunrise, activities, etc.). If

the latter interpretation is correct, then their statement

of "usual" wake-up time is heavily influenced by the present









or recent past, i.e., their winter sleep behavior. To an-

swer this question, longitudinal studies are needed in which

the same subjects' sleep log behavior and sleep question-

naire estimates are sampled several times throughout the

year. If some aspects of sleep do change with changing

seasons or circumstances, it is expected that the interpre-

tation of "usual" sleep will be heavily influenced by the

recent past rather than an average of the past several

years or more. One longitudinal study using sleep logs

only was reported by Kleitman et al. (1937). A daily record

of sleep was kept by 26 adults for an average of 8 months

each. Average wake-up times were not reported. Average

sleep length for the various seasons did not differ.

The present study provides much group data on the sleep

of college students. Using similar methods of measurement,

data on other groups can be collected and compared. The

average sleep length per night for this sample is about

7 1/2 hours according to both the Sleep Log and the Ques-

tionnaires. The standard deviation around this mean is .85

hour. This suggests a range of between 5 hours and 10 hours

average sleep per night for 99.7% of the college population.

Naps are quite common in this group, however, and average

2.6 hours per week or .4 hour per day. This results in a

sleep length per 24 hours of 7.9 hours.

Sleep length on weekends is .5 hour longer than week-

nights. This is a significant difference, but is not as










large as expected. Averaging the Sleep Log weeknight sleep

length (7.4 hours) with weekend sleep length (7.9 hours)

results in an overall average of 7.5 hours. This overall

average is only .1 hour (1%) more than weeknight sleep (due

to the weighing of weeknight sleep by a factor of 5 and

weekend sleep by a factor of 2). The placement of sleep is

influenced more by the day of the week than is the length

of sleep. Average bedtime is 1.0 hour later on weekends,

and wake-up time is 1.5 hours later.

A measure of variability of one's sleep is possible

only with the Sleep Log. The average standard deviation of

bedtime over all 7 days of the week is 1.28 hours. The

standard deviation of weeknight sleep only (5 weekdays) is

1.09 hours. These figures suggest that the typical stu-

dent has widely varying bedtimes, with 1/3 of his bedtimes

occurring more than 1.3 hours earlier or later than his 7-

day mean bedtime. Several students have a bedtime standard

deviation of over 3 hours.

The equally high standard deviation of wake-up times,

1.29 hours, is an unexpected finding. It has generally

been believed that most persons' wake-up times are more

stable than their bedtimes. This may be true for many

groups with fixed weekly schedules which demand that one

arise each morning at about the same time. The use of sleep

logs to measure bedtime and wake-up time variability in

other groups of persons could answer this question. We may









find, however, that these other groups not only have less

variability in wake-up time than college students, but also

less variance in bedtime. Thus, they may still have equal

variance in bedtime and wake-up time. In any case, one

area for future research is the measurement of this vari-

ability in various groups, along with an analysis of the

causes (social and biological) and results of this vari-

ability.


Individual Data

Stability of group means over time does not assure

individual stability of sleep measures. In order to pro-

vide some information to those who are interested in

studying or predicting individual sleep behavior rather

than, or in addition to, group characteristics, several

types of data have been presented. First, the week-to-

week stability of the Sleep Log was examined. Correlations

of weekly sleep-length means for the summer group (Week

1 and Week 2) and winter group are r = +.68 and r = +.62

respectively. The correlation of biweekly sleep length

means for the summer group is r = +.88. As the group means

for these 2 biweekly periods do not differ, the 2-week

length mean is quite stable for this group of students

under the conditions stated. It is likely that extending

the length of the Sleep Log beyond 2 weeks would not appre-

ciably increase the stability of the sleep length measure.

The Spearman-Brown formula predicts a reliability









coefficient of r = +.93 for a 4-week mean sleep length,

given a correlation of r = +.88 for 2 weeks. This is an

increase of 9% in the variance accounted for.

Adding naps to sleep per night does not increase the

biweekly stability of sleep length. In fact, sleep length

per 24 hours is somewhat less stable (r = +.75) than sleep

length per night (r = +.88), although the variability (SD)

around the group mean does not differ for sleep length per

night (.85 hour) and sleep length per 24 hours (.82 hour).

This finding does not lend support to the idea that naps

are necessarily a compensatory mechanism to even out the

day-to-day differences in one's sleep length.

In addition to the fairly high reliability of Sleep

Log sleep length, for the summer group the biweekly sta-

bility of bedtime (r = +.87) and wake-up time (r = +.79)

is also good. The biweekly stability of bedtime variability

(SD) is also fairly high (r = +.75). Variability of sleep

length and wake-up time is not as stable across time, how-

ever, with the two biweekly values correlating r = +.48

for each of these measures. These latter two correlations

are significant however (p < .001), and one can conclude

that inconsistency in one's sleep behavior is somewhat con-

sistent over time.

All of the questions on the Pre-Log Questionnaire which

ask about usual sleep behavior correlate with the corres-

ponding Post-Log Questionnaire questions between r = +.65









and r = +.79. Other research suggests that a 2-week test-

retest correlation with the same questionnaire would yield

coefficients this high or higher (O'Connor, 1964; Webb &

Stone, 1963).

After looking at the similar group means for the Sleep

Questionnaires and Sleep Log, one may get the impression

that an individual's Questionnaire estimates of sleep be-

havior may be quite similar to his Sleep Log values. In

fact, several investigators have made this claim after

looking at group means for other types of sleep measurement

comparisons. It is possible, however, to have a negative

correlation between one's sleep questionnaire and sleep log

and still have equal group means. This is not the case for

the present groups; the Questionnaires and Sleep Log are

significantly positively related.

The summer and winter groups do not differ signifi-

cantly in the degree of correlation between the Sleep Log

mean sleep length and Post-Log question #1 on sleep length

for the past few weeks. The correlation for all 102 stu-

dents is r = +.72. A number of students have rather large

differences between the two sleep-length measures. Of the

102 students, 53% have a difference of .5 hour or more,

while 15% have a difference of 1 hour or more.

Sleep Log sleep length was also compared to Post-Log

Questionnaire sleep length as calculated from questions on

bedtime and wake-up time on weekdays and weekends (questions









#10 to #13). Using this definition of sleep length, the

correlation of Sleep Log to Post-Log Questionnaire is

r = +.83. This is significantly higher than the correlation

of r = +.72 between the Sleep Log and Post-Log question #1

as the correlations are not independent (the two methods of

estimating sleep length from the Post-Log Questionnaire

correlate r = +.67 with each other). Fewer of the 102 stu-

dents (34%) now differ from the Sleep Log by .5 hour or

more. This is still a rather large number, however, con-

sidering the fact they had been paying special attention to

their sleep for the past few weeks.

It should be noted that a high correlation, even

r = 1.0, would not imply a lack of individual discrepancies

between the Sleep Log and a Questionnaire estimate of sleep

behavior. All students could have a large difference be-

tween the two measures in the same direction and to the

same degree. In such a case, a high correlation coefficient

would simply allow accurate prediction of one measure from

the other, even if the two measures were different. This

is not the case in the present study, however; there is no

large consistent over- or underestimation. Some subjects

underestimated and others overestimated the sleep measures.

The inability of many students to estimate sleep length

accurately for the past few weeks offers a partial explana-

tion for the individual discrepancies between the Sleep Log

and the Pre-Log question about one's usual sleep length.










Again, the winter and summer groups do not differ signifi-

cantly in the correlations between Sleep Log sleep length

and Pre-Log estimates of sleep length. For the total group

of 102 students the correlation is r = +.70, which is about

the same as the r = +.72 for the Sleep Log and Post-Log

Questionnaire relationship. A total of 41 (40%) of the

students have a difference of .5 hour or more between the

Sleep Log and Pre-Log Questionnaire. Although some of this

difference is probably due to the fact that the Sleep Log

period may not be representative of "usual" sleep, a sub-

stantial portion of the difference is likely due to the

students' lack of accurate knowledge regarding their own

sleep behavior.

Anecdotal evidence supports the view that many stu-

dents' estimates of usual sleep on the Pre-Log Question-

naire are not valid. Upon returning the first week's Sleep

Log, several students volunteered the unsolicited informa-

tion that their Sleep Log experience had made them realize

their Questionnaire estimates were incorrect. In an

apparent attempt to help the experimenter, some of the stu-

dents even asked to be allowed to change their Question-

naire so that it would be "correct."

Although the correspondence between the Sleep Log and

Questionnaires is about the same for Pre-Log and Post-Log

estimates of sleep length, Questionnaire estimates of bed-

time and wake-up time are better for the Post-Log










Questionnaire than the Pre-Log Questionnaire. Correlations

between the Sleep Log and Post-Log Questionnaire are higher

than those between the Sleep Log and Pre-Log Questionnaire

(p < .001), with more students having less than a .5 hour

difference between the Post-Log Questionnaire and the Sleep

Log than between the Pre-Log Questionnaire and Sleep Log.

The apparent ability of students to improve in their

estimation of bedtime and wake-up time but not in their

estimation of sleep length may have a simple explanation.

During the daily recording of the Sleep Log, students were

forced to pay attention to their bedtimes and wake-up times

They were never asked, however, to provide a daily record

of the duration of sleep. Many of the students thus

probably never calculated daily sleep length and were not

attending to this aspect of their sleep behavior. Perhaps

if they calculated and recorded their daily sleep lengths,

they would give better estimates on the Post-Log Question-

naire.

A final observation on the relationships of Sleep Log

parameters to Questionnaire estimates concerns the apparent

inability of the students to estimate their variability of

bedtime and wake-up time. Questions #14 and #15 on the

Pre-Log Questionnaire for both the summer and winter groups

and questions #27 and #28 for the winter group asked per-

sons to estimate the variability of bedtime and wake-up

time in two different ways. The correlations between the









answers to these questions and the individual's standard

deviation of bedtime or wake-up time from the Sleep Log

range from r = +.06 to r = +.28. The validity of question-

naire estimates of the variability of sleep is thus in

doubt. Perhaps other ways of wording such questions will

be found which will enable subjects to tell us about the

variability of their sleep parameters. In the meantime,

if a measure of individual variability is desired, one

should rely on a sleep log for estimating this measure.

To summarize the data on individuals, there are often

large discrepancies between an individual's Sleep Log

measures and comparable Questionnaire estimates, whether

Pre-Log or Post-Log. If the sleep behavior of individuals

is the measure of interest, one may do well to obtain the

estimates of sleep by using a 2-week sleep log. This has

often been done by investigators who use a questionnaire

for preliminary screening of subjects to serve in laboratory

experiments, followed by a sleep log for further elimina-

tion of subjects. Sleep logs have not been typically used,

however, when gathering sleep data, along with other data,

on large groups of persons in order to correlate sleep be-

havior with personality, intelligence, etc. Thus, many such

studies are relating variables of interest to sleep behavior

as estimated by the subject on a questionnaire. As this

information is often not the same as the sleep behavior as

measured by a sleep log, the relationships found (or not










found) may be based on the subject's guesses about his

usual sleep behavior, rather than on the sleep behavior it-

self.

Although a sleep log may be more accurate, its use may

not always be possible because of a lack of time, money,

or cooperative subjects. In such a case, a sleep question-

naire may be the only recourse. If so, close attention

should be paid to the nature of the questions and the for-

mat provided for the answers. Research is needed to deter-

mine the way to phrase questions on usual sleep length so

that an individual's answer will most closely match his

actual sleep behavior.


Interrelationships

In these two groups of college students, sleep be-

haviors are definitely related to each other. For example,

one not unexpected relationship is that short sleepers go

to bed later than their peers. Wake-up time, however, is

not highly related to one's sleep length. Another rela-

tionship found is that a short average sleep length is

significantly correlated with a high variability (SD) of

sleep length. For both the summer and winter groups, this

correlation is r = -.48. This relationship was also found

by Baekeland and Hartmann (1971) for a large group of non-

students of all ages who were either long or short sleepers.

The reasons for this relationship are not clear; perhaps

the long sleepers have less choice in the placement of









their sleep (there are only so many hours in a day), and

thus have a smaller standard deviation of sleep length.

The examination of relationships between Sleep Log

variables and other measures, both sleep and nonsleep, does

not produce many relationships which are either significant

or large. This may be due to many reasons. Perhaps these

factors are simply not highly related to sleep behaviors.

Or, perhaps the relatively simple scales used to measure

them are not of adequate sensitivity to differentiate be-

tween individuals. Another possibility is that the rela-

tionships are not the same during the winter and summer

seasons or school terms, or for the two groups of students

measured. If this is the case, the generalizability of any

statement about the relationships which were found only for

one group or the other is limited to that particular group

or time of measurement.

Several ideas presented by Hartmann (1973a, 1973b) can

be related to the present findings on the relationships be-

tween one's statement of obtaining more or less sleep than

usual during the past few weeks of keeping the Sleep Log

(Post-Log question #2) and Post-Log questions #21 to #27

concerning changes in mood, amount of stress, etc., during

this same period. Hartmann's work has led him to believe

that one needs more sleep than usual when under stress or

depressed and less sleep than usual when things are going

well. He emphasizes changes in the need for sleep rather










than possible changes in the actual amount of sleep ob-

tained. In contrast, the present study asked for a state-

ment of whether the student obtained more or less sleep,

rather than whether he felt he needed more or less, during

the past few weeks.

The Chi-Square analyses between a perceived change in

one's sleep amount and in one's feelings and activity re-

sult in two interesting relationships. A statement of

getting less sleep than usual is related to more perceived

mental activity and more depression. These findings are

opposite to those expected from Hartmann's conclusions.

Assuming that both the present findings and Hartmann's are

correct, one might conclude that during periods of stress,

depression, or increased mental activity, people need more

sleep but get less. Or, perhaps they only feel that they

obtain less sleep than usual because they need more but

are getting the same amount of sleep. Such questions can-

not be resolved here. However, in future research of this

nature a definite distinction needs to be maintained be-

tween sleep amount and the more elusive and difficult to

measure sleep need. At present, there are no adequate ways

to distinguish the two.

The relationship of sleep variables to the MMPI scales

is best put into perspective by comparing the present re-

sults to those in a comparable study by Johns et al.

(1974). Johns used sleep questionnaires to obtain estimates









of the sleep behavior of 104 male students. His question-

naire asked for bedtime and wake-up time for weekdays and

weekends (among other things). He calculated sleep length

per week from these values, apparently after correcting for

time in bed not spent sleeping. He then correlated various

sleep parameters with the raw scale scores from the MMPI.

For eight sleep variables (bedtime on weeknights and week-

ends, wake-up time on weeknights and weekends, sleep per

week, naps per week, latency to sleep onset, and amount of

time spent awake during the night), Johns obtained a total

of 13 correlations (p < .05 or better) with the 13 MMPI

scales. These coefficients ranged from .20 to .30 in mag-

nitude.

In general, correlations in the present study between

the Pre-Log Questionnaire and the MMPI for the winter group

do not confirm the correlations which Johns found, either

in magnitude or significance levels. Only two scales are

significantly related to sleep measures in both Johns'

study and the present study: scale 4, PD, correlates with

weekend bedtime r = +.25 in Johns' study and r = +.27 in

this study; scale F correlates with weekday wake-up time

r = +.28 in Johns' study and r = +.29 here.

As Johns used only males in his study, the above

analyses were repeated using only the male subjects in the

present study. As only 20 males took the MMPI, the level

of correlation required to reach statistical significance










at the p < .05 level is much higher; no pair of sleep meas-

ures and MMPI scales reach significance for these males.

Looking, however, at the magnitude of the correlations for

males, only 3 of the 13 correlations found significant by

Johns have here a correlation of r = +.20 or greater, which

is the smallest significant correlation in Johns' study.

Thus, whether for all subjects or for males only, the

present data do not seem to confirm the majority of Johns'

findings.

Of particular importance for studies of sleep length

per se are two findings. First, the one variable found by

Johns to be significantly related to sleep length (scale

3, HY, r = -.24, p < .05) is not related to sleep length

derived in a similar manner here with the Pre-Log ques-

tions on bedtime and wake-up time. This MMPI scale is also

not related to sleep length as calculated from the Sleep

Log or the Questionnaire questions on either usual sleep

length (Pre-Log Questionnaire) or sleep length the past 2

weeks (Post-Log Questionnaire). Second, in the present

study, sleep length as calculated by Johns (from question-

naire bedtime and wake-up time) is not significantly re-

lated to any of the other MMPI scales either. As defined

by both the Sleep Log and Pre-Log question #1 on usual

sleep length, one's sleep length is correlated with only

one MMPI measure, scale 9, MA, r = -.27 and r = -.28,

p < .05.









In addition to asking about sleep behavior per se,

both Johns' study and the present study asked about the

usual quality of sleep. Johns found that this single ques-

tion correlated with 7 of the 13 MMPI scales at the p < .05

level of significance or better. The present study found

4 of the MMPI scales related to a similar question about

one's usual quality of sleep (Pre-Log question #20). Three

of these four are the same as those found significant by

Johns (scale K, r = -.39; scale 2, D, r = +.32; scale O,

SI, r = +.36). A hypothesis for further study is that one's

feelings about sleep may be more related to personality than

one's sleep length is related to personality. Also, one's

variability of sleep may be more related to personality

than is sleep length; in this study, four MMPI scales

correlated significantly with the standard deviation of

sleep length.

Several previous studies of the relationship between

sleep length and personality were reported by Hartmann

(1973a, 1973b), Hartmann et al. (1972), and Webb and Friel

(1970, 1971). These studies used various selection tech-

niques to find and study the characteristics of only long

(greater than 9 hours) sleepers and short (less than 6

hours) sleepers. Only one used a sleep log as a check on

the subjects' initial statements of usual sleep length.

Webb and Friel found no consistent differences between long

and short sleepers on the MMPI or other trait variables for












several groups of college students. Hartmann and his co-

workers found that their long and short sleepers differed

on only 2 of 13 MMPI scales at the p < .05 level of sig-

nificance. Their psychiatric interviews resulted in the

impression that the short sleepers were more psychologically

"healthy" than the long sleepers. Because of the selection

procedures used, however, these impressions are likely not

representative of long and short sleepers in general.

Johns' study and the present study both used persons

with a wide range of average sleep lengths: long, medium,

and short. Johns' study used only a questionnaire to esti-

mate usual sleep, and he found no large relationships be-

tween sleep length and MMPI scales. In addition to using

several questionnaire definitions of average sleep length,

the present study also calculated sleep length from daily

sleep records. The MMPI scales not only were not related

to one's statement about usual sleep length, but were also

not related to average sleep length as calculated from

daily records. When considering either long or short

sleepers only, or the entire range of sleep lengths, one's

usual duration of sleep as measured by either questionnaire

or sleep log does not seem to be highly related to various

measures of personality. If sleep length is related to

personality, the relationships are likely to be subtle

ones.









When considering the overall lack of strong relation-

ships between sleep and other measures, including person-

ality, it must be remembered that this study was concerned

with "normal" persons who are "normal" sleepers. Extremes

of sleep behavior may well often be related to extremes of

attitudes or personality, as well as other facets of the

individual's life. Furthermore, sleep behavior may be more

stably related to a combination of various personality

measures or, alternatively, a combination of sleep meas-

ures may be related to a given personality trait. A large

study in which several hundred persons are given the MMPI

and a 2-week sleep log could use factor analysis techniques

to explore these possibilities. It would, of course, need

to be replicated if significant results were found.















SECTION V

SUMMARY


The uses of sleep length measures include studies of

both groups and individuals. If group data is desired, a

sleep questionnaire asking about usual sleep behavior or

recent sleep behavior seems to be adequate. For these

college students there was no large over- or underestimation

of most sleep behaviors, including sleep length itself.

Similar studies need to be done with other groups, however,

in order to evaluate their sleep log/questionnaire com-

parisons. One type of group data apparently not obtainable

by questionnaire is a measure of the expected variability

of persons in their day-to-day sleep behavior; the use of a

sleep log enables one to measure this average within-subject

variability and compare it to that of other groups.

Individual data is not as stable as group data in com-

parisons between the Sleep Log and Questionnaires. The

Pre-Log question on average sleep length correlates r = +.70

with the Sleep Log mean sleep length; 40% of the students

have a .5 hour or greater difference between these two meas-

ures of sleep length. Comparison of the Sleep Log mean

sleep length to the Post-Log estimate of sleep length for

the past few weeks produces results comparable to those

above.









Comparisons of Pre-Log bedtime and wake-up time esti-

mates for weeknights and weekends to comparable Sleep Log

measures produce correlations of r = +.58 to r = +.78 with

50% to 60% of the students having a .5 hour or greater dif-

ference between the two measures. Comparisons of these

variables for the Post-Log Questionnaire and Sleep Log are

somewhat better, with correlations of r = +.80 to r = +.91;

25% to 47% of the students have a .5 hour or greater dif-

ference between comparable measures.

Thus, even under the best of circumstances, many per-

sons seem to be inaccurate in their estimations of average

sleep behaviors. The best estimates are of average weekday

wake-up time during the period of recording the Sleep Log;

however, 25% of the students were not able to do this with

less than a .5 hour discrepancy with the Sleep Log record.

Such inaccuracy on estimates of recent sleep behavior after

recording sleep daily makes one suspect even more the

accuracy of estimates of average sleep behavior. These

findings suggest that when one wishes to either study the

relationships of sleep to other aspects of life or to

select individual subjects for sleep experiments, a sleep

log should definitely be included in the determination of

sleep length or other sleep behaviors.

If a sleep log is used for measuring individuals'

sleep parameters, a 2-week sleep log is probably adequate;

biweekly means are quite stable, even though many of the






82



persons in this sample have high night-to-night variability.

It is suggested that this within-subject variability of

sleep be studied as a potentially important independent

and dependent variable. It is also suggested that we avoid

being evaluative or prescriptive about both between-subject

and within-subject variability until we know more about the

causes and effects of such variations.
















APPENDIX 1

SLEEP QUESTIONNAIRE (PRE-LOG)


1. How many hours, on the average, do you usually sleep per night?

5 5 1/2 6 6 1/2 7 7 1/2 8 8 1/2 9 9 1/2
less
10
more

2. How many hours of sleep per night do you think that you need to
feel well and function adequately?

3. How many hours of sleep per night do you prefer to have when you
have a chance?

4. In general, the amount of sleep that you usually get is

not enough, about enough too much

5. On the average, about how many naps do you usually take each week?

0, 1, 2, 3, 4, 5, 6, 7, 8 or more

6. About how many total hours do you usually spend each week taking
naps?

7. About how many times on the average do you wake up each night?

0, 1, 2, 3, 4 or more

8. About how much total time do you spend awake after going to sleep?

none, less than 5 min., 5-15, 16-30, 31-60, more

9. About how many minutes does it typically take for you to go to
sleep?

less than 5, 5-15, 16-30, 31-60, more than 1 hr.

10. How often do you have trouble getting to sleep as quickly as you
would like?

almost never, occasionally, often, almost always











11. How important a role does sleep play in your daily life?

very somewhat fairly very
important important important unimportant

12. How much do you enjoy sleep?

not at all, a little, moderately, much

13. When you were a child, did your parents usually regulate or
enforce your sleep times? Yes No Don't know

14. How regular are your bedtimes?

very regular, somewhat regular, somewhat irregular,very irregular

15. How regular are your wake-up times?

very regular, somewhat regular, somewhat irregular, very irregular

16. What is your average bedtime during weekdays?
(to nearest 15 min.)

17. What is your average bedtime on weekends?
(to nearest 15 min.)

18. What is your average wake-up time on weekdays?
(to nearest 15 min.)

19. What is your average wake-up time on weekends?
(to nearest 15 min.)

20. How well do you usually sleep at night?

very well, satisfactorily, some problems, poorly

21. Do you usually feel well rested when you wake up, or soon there-
after?

almost never, occasionally, often, almost always

22. Usually what time is your first scheduled class (or work,
appointment, etc.) on these days? (Beneath each day put a time
or the word NONE.)

MON TUE WED THUR FRI SAT SUN
AM AM AM AM AM AM AM
PM PM PM PM PM PM PM

23. What is your; Age Sex











24. What is your college class?

Freshman, Sophomore, Junior, Senior, Grad. Student

25. What is your: Height Weight

26. How many other persons usually sleep in the same room with you?

None, 1, 2, 3, 4 or more

27.* On the average how many days per week do you usually go to bed
more than one (1) hour earlier or later than your average bedtime?

0, 1, 2, 3, 4, 5, 6, 7

28.* On the average, how many days per week do you usually wake up
more than one (1) hour earlier or later than your usual wake-up
time?

0, 1, 2, 3, 4, 5, 6, 7


*Appeared only for the Winter group.

















APPENDIX 2

SLEEP LOG


(Complete immediately prior to going to bed with the intention of
sleeping.)

1. How many naps did you take today? (if none, put "0")
If you took any naps, record here the times you went to sleep and
woke up.

1st nap: from to AM 2nd nap: from to A
_PM ----PM

AM
2. Time you are going to bed to try to sleep
PM



(Complete immediately upon awakening.)

3. What was your final wake-up time this morning?

4. How many minutes would you estimate it took you to go to sleep
last night?

less than 5, 5-15, 16-30, 31-60, more

5. Do you remember waking up any during your sleep last night?

Yes No

6. About how much total time did you spend awake after going to sleep?

none, less than 5 min., 5-15, 16-30, 31-60, more
















APPENDIX 3

SLEEP QUESTIONNAIRE (POST-LOG)


1. For the past 2 weeks, how many hours on the average did you sleep
per night?

5 5 1/2 6 6 1/2 7 7 1/2 8 8 1/2 9 9 1/2

10
more

2. For the past 2 weeks, how much did you sleep compared to your
usual amount?

less, about the same as usual, more

3. For the past 2 weeks, the amount of sleep you got was on the
average

not enough, about enough, too much

4. When compared to your average sleep, how well did you sleep during
the past 2 weeks?

worse, same as usual, better

5. During the past 2 weeks, about how many naps per week did you take?

0, 1, 2, 3, 4, 5, 6, 7, 8 or more

6. About how many total hours did you spend each week taking naps
for the past 2 weeks?

7. For the past 2 weeks, about how many times on the average did you
wake up each night?

0, 1, 2, 3, 4 or more

8. About how much total time did you spend awake each night (on the
average) during your sleep the past 2 weeks?

none, less than 5 min., 5-15, 16-30, 31-60, more than 1 hr.











9. During the past 2 weeks, about how many minutes on the average
did it take you to go to sleep each night?

less than 5, 5-15, 16-30, 31-90, more than 1 hr.

10. For the past 2 weeks, what was your approximate average bed-
time during weekdays (to the nearest 15 minutes)?

AM PM

11. For the past 2 weeks, what was your approximate average bedtime
on weekends (to the nearest 15 minutes)?

AM PM

12. For the past 2 weeks, what was your approximate average wake-up
time on weekdays (to the nearest 15 minutes)?

AM PM

13. For the past 2 weeks, what was your approximate average wake-up
time on weekends (to the nearest 15 minutes):

AM PM

14. About how many hours of sleep per night do you think that the
average person your age usually gets (to the nearest 1/2 hour)?

Please rate the following two questions according to this scale:
(1) strongly agree, (2) moderately agree, (3) neither agree or
disagree, (4) moderately disagree, (5) strongly disagree

15. Always sleeping 10 hours or more per night is approved of by
people in general for a person your age. 1, 2, 3, 4, 5

16. Always sleeping 5 hours or less per night is approved of by
people in general for a person your age. 1, 2, 3, 4, 5

17.* Have you ever spent at least several straight weeks sleeping at
least 2 hours less per night than usual? Yes No Don't know

If yes, for how long a period of time?

18.* Have you ever spent at least several striaght weeks sleeping at
least 2 hours more per night than usual? Yes No Don't know

If yes, for how long a period of time?

19.* If you get 2 hours less sleep than usual, do you generally feel
tired or irratable? Yes No Don't know











20.* How do you generally feel immediately upon awakening from an
average night's sleep?

very alert, fairly alert, somewhat sleepy, very sleepy

*Rate the following phrases in terms of any change during the past 2 weeks:

a. Amount of mental activity

More, About same as usual, Less

b. Depressed or upset mood

More, About same as usual, Less

c. Stressful period of time

More About same as usual, Less

d. Everything going well

More, About same as usual, Less

e. Amount of physical work or exercise

More, About same as usual, Less

f. Amount of food eaten per day

More, About same as usual, Less

g. Amount of general activity

More, About same as usual, Less



Note. The summer group was asked about sleep for the past 4, not 2,
weeks.

*Appeared only for winter group.




















(I--
0
1 ,



O

H
0














0
0

C-
p<





















P U








0
U










H









0
r a
O

>1
H















CC:
P I
<:


z







H


CC







H
CC


a

I



0 0


C 0
uO 3) a


c) ( H
Hq C) En i Hj 1



O) 0 ( r.
0i *H I U 0






0 0 r
0) a ) 1) 0)





0 0 (C) m C

S0 HM0 H


o r ) o co -


S4 4J

5H -H1
0 4J (2)




S 0 0

C H >1 >1 ()
S 4J 4J H -C 4-
-r 4 -l (U C

4J Cd H4 H4

0 Q ( C) C1)
- a -4 A




-n A r in O H.


I O 0 I I I I I I I I
I I I I I I I I


I in [- N O m I N r-H O L
N I N

CD 1o 01 0) m1 1 D -i LH n lo



e -H N r-o m 0, N -

0) m in I' O O C N H- C0 N





N N N N N H N N H- r




I 0 0 I I I I I I


I 01 co 01 H N I N N O N
1 H I H'


o ai 01 C' 0 in (N o N ( i n
H H N N N

o N n m 0 N O (C) CH 0r k (0
S N N N H N N rN

co 0 '0 Wo irn H CD Ln C' co C
H H H H H




N N N N N 0 H N N N 01









S I I C H m a I I I I
n I I I r-l CM I l l I

I I I C (N I I I I


in O i- (N 0 co in Lri 0 r-

Z II IU
O
E in .N IN m ( N H-. m N





0 *H

I

m 0
I
0. 4 i-- C-I
o 4
H0 C i 0 m C ( CC





UN N e H N 0H
I *
O O Cd i Cl (N (N H 0 (N W ( 0 CO











a4









n 0 H N (N 0 N (N N H
W
U)
SI I I i









O i l l-










0
0i Cd in
H 1-0U(N (N m Cl (N (N
X H




















0 0 O H 0in O(N
SO 0 1 c1 4 0n o O O r C
w UW r0














*H







S(N 0 w c a 0 ( N () N
in
0 W O i


S0 0 0








O H C
0 f0 '0 C1 C 0
H 4 U Cd > Cd Cd Cd Gd

H H dm 6 0 Ca a 0 Cd
S0 4 4 H 0. rHl C d .
H e o 0Ba 0. 4
0. *H Cd I-0 C- d C

C.) C3 d Cd 0C aCd Cd H 1- C



Cd Cd C0 Cd +l C Cd 0 Cd Cd Cd






3 (N Cl O CO CO in ^ N CO cO 0

























Cm' t0 Cm t 0 '







i-l
N H H H


) )
in 0
0 *4
-rH U
0






>1
C



> .
w 0


(NN NqNNq "


cl






4



0





.










Cde






,O
41
0

(I



Cd H
>

4J -

0




Cd3
"o


S(1)

ina










4-)
'0C
i-d
-r




0-






C)


0




University of Florida Home Page
© 2004 - 2010 University of Florida George A. Smathers Libraries.
All rights reserved.

Acceptable Use, Copyright, and Disclaimer Statement
Last updated October 10, 2010 - - mvs