the everyday memory questionnaire – revised: development of a 13-item scale

8
RESEARCH PAPER The Everyday Memory Questionnaire – revised: Development of a 13-item scale JANE ROYLE & NADINA B. LINCOLN Sheffield Teaching Hospitals NHS Trust, Sheffield, University of Nottingham, Nottingham, UK Accepted January 2007 Abstract Purpose. The Everyday Memory Questionnaire (EMQ) was developed as a subjective measure of memory failure in everyday life [1]. Previous studies have investigated the factor structure of the EMQ in both healthy participants and people with multiple sclerosis (MS). The aim of the present study was to confirm the factor structure of the EMQ, to determine the internal consistency and criterion validity of the scale and to develop a shortened version. Method. A retrospective design, including participants from a study on MS patients and their carers and a study on stroke patients. Psychometric properties of the EMQ-28 were explored, and the measure was further revised from comparative analyses between the clinical and non-clinical groups. Results. Reliability and factor analysis of the EMQ-28 identified two main factors, general memory and attentional function, showing some concordance with previous research. Further analysis reduced the questionnaire to a 13-item measure (EMQ- R), with two main factors (Retrieval and Attentional tracking), strong internal reliability, and good discriminatory properties between clinical and control groups. Conclusions. The 28-item questionnaire consistently differentiated between two broad systems of memory and attention, with some differentiation of visual and verbal, or language systems. Results showed some consistency with previous findings. The revised, 13-item questionnaire is a valid and reliable tool that has good face validity for use with neurological patients. Further exploration of the revised EMQ is recommended to provide information regarding its psychometric and clinical properties. Keywords: Memory, assessment, cognition Introduction Reliable measurement of the functional impact and report of memory problems continues to present a dilemma to the practitioner. Recent developments in the practice of neuropsychology have led to a shift from a diagnostic to ecological and rehabilitative focus [2]. Traditional neuropsychological tests have consistently been found to have poor ecological validity, which is defined as the relationship between performance on neuropsychological tests and beha- viour in everyday settings [3]. Memory question- naires, therefore, are proposed as a potentially useful adjunct to traditional tests to improve such ecological validity. Ecologically-based questionnaires to identify memory problems have some reliability [4,5] but varying validity [2,6,7]. The poor association between memory tests and performance may result from anomalies between everyday tasks and abilities assessed by memory tests. Validity is higher when memory questionnaires are compared to more ecologically-based tests [8], and correlations between questionnaires and test results have been higher where questionnaires were completed by significant others [1,3,9]. Herrman [10] reviewed 14 memory question- naires, proposing that self-report measures do not measure memory performance per se, but are measures of metamemory, or beliefs about memory performance. He concluded that there is valid utility for memory questionnaires in the assessment of first hand observations of one’s own performance versus second-hand report from others; an individual’s susceptibility to cognitive errors under stress; and Correspondence: Jane Royle, Sheffield Teaching Hospitals NHS Trust, Sheffield, and University of Nottingham, UK. E-mail: [email protected] Disability and Rehabilitation, 2008; 30(2): 114 – 121 ISSN 0963-8288 print/ISSN 1464-5165 online ª 2008 Informa UK Ltd. DOI: 10.1080/09638280701223876 Disabil Rehabil Downloaded from informahealthcare.com by SUNY State University of New York at Stony Brook on 10/27/14 For personal use only.

Upload: nadina-b

Post on 28-Feb-2017

215 views

Category:

Documents


2 download

TRANSCRIPT

Page 1: The Everyday Memory Questionnaire – revised: Development of a 13-item scale

RESEARCH PAPER

The Everyday Memory Questionnaire – revised: Developmentof a 13-item scale

JANE ROYLE & NADINA B. LINCOLN

Sheffield Teaching Hospitals NHS Trust, Sheffield, University of Nottingham, Nottingham, UK

Accepted January 2007

AbstractPurpose. The Everyday Memory Questionnaire (EMQ) was developed as a subjective measure of memory failure ineveryday life [1]. Previous studies have investigated the factor structure of the EMQ in both healthy participants and peoplewith multiple sclerosis (MS). The aim of the present study was to confirm the factor structure of the EMQ, to determine theinternal consistency and criterion validity of the scale and to develop a shortened version.Method. A retrospective design, including participants from a study on MS patients and their carers and a study on strokepatients. Psychometric properties of the EMQ-28 were explored, and the measure was further revised from comparativeanalyses between the clinical and non-clinical groups.Results. Reliability and factor analysis of the EMQ-28 identified two main factors, general memory and attentional function,showing some concordance with previous research. Further analysis reduced the questionnaire to a 13-item measure (EMQ-R), with two main factors (Retrieval and Attentional tracking), strong internal reliability, and good discriminatory propertiesbetween clinical and control groups.Conclusions. The 28-item questionnaire consistently differentiated between two broad systems of memory and attention,with some differentiation of visual and verbal, or language systems. Results showed some consistency with previous findings.The revised, 13-item questionnaire is a valid and reliable tool that has good face validity for use with neurological patients.Further exploration of the revised EMQ is recommended to provide information regarding its psychometric and clinicalproperties.

Keywords: Memory, assessment, cognition

Introduction

Reliable measurement of the functional impact and

report of memory problems continues to present a

dilemma to the practitioner. Recent developments in

the practice of neuropsychology have led to a shift

from a diagnostic to ecological and rehabilitative

focus [2]. Traditional neuropsychological tests have

consistently been found to have poor ecological

validity, which is defined as the relationship between

performance on neuropsychological tests and beha-

viour in everyday settings [3]. Memory question-

naires, therefore, are proposed as a potentially useful

adjunct to traditional tests to improve such ecological

validity.

Ecologically-based questionnaires to identify

memory problems have some reliability [4,5]

but varying validity [2,6,7]. The poor association

between memory tests and performance may result

from anomalies between everyday tasks and abilities

assessed by memory tests. Validity is higher when

memory questionnaires are compared to more

ecologically-based tests [8], and correlations between

questionnaires and test results have been higher

where questionnaires were completed by significant

others [1,3,9].

Herrman [10] reviewed 14 memory question-

naires, proposing that self-report measures do not

measure memory performance per se, but are

measures of metamemory, or beliefs about memory

performance. He concluded that there is valid utility

for memory questionnaires in the assessment of first

hand observations of one’s own performance versus

second-hand report from others; an individual’s

susceptibility to cognitive errors under stress; and

Correspondence: Jane Royle, Sheffield Teaching Hospitals NHS Trust, Sheffield, and University of Nottingham, UK. E-mail: [email protected]

Disability and Rehabilitation, 2008; 30(2): 114 – 121

ISSN 0963-8288 print/ISSN 1464-5165 online ª 2008 Informa UK Ltd.

DOI: 10.1080/09638280701223876

Dis

abil

Reh

abil

Dow

nloa

ded

from

info

rmah

ealth

care

.com

by

SUN

Y S

tate

Uni

vers

ity o

f N

ew Y

ork

at S

tony

Bro

ok o

n 10

/27/

14Fo

r pe

rson

al u

se o

nly.

Page 2: The Everyday Memory Questionnaire – revised: Development of a 13-item scale

as an assessment of cognitive function for people

with psychological and neurological disorders.

A valid and reliable memory questionnaire has

several uses for people with psychological, neurolo-

gical and physical health problems including: mea-

surement of outcome or process, an ecological tool

to assess everyday memory as an adjunct to standard

clinical tests, comparing metamemory and memory

beliefs for individuals with memory problems, asses-

sing insight and awareness, particularly if used with

both client and family member, and as a heuristic

tool to compare different types of memory com-

plaint. In the clinical setting they also provide an

opening to explore cognitive difficulties in a non-

threatening manner and an aid to engage someone in

a rehabilitation or therapeutic process, by asking for

the client’s personal views and perceptions.

The Everyday Memory Questionnaire (EMQ) was

initially developed for use with survivors of head

injury [1] and has subsequently been further refined

and developed with both non-clinical and clinical

samples. The original questionnaire consisted of 35

items, which has since been altered to a 28-item

questionnaire [11] to increase the measure’s validity

and facilitate self-administration. The 28-item scale

was developed with 22 items representing valid

memory difficulties, and six items representing floor,

or bogus items, representing atypical memory

difficulties as a measure of response validity. The

response format has also been altered over time,

from relative frequencies (‘sometimes’) to absolute

values (e.g., ‘about once a week’), with a 9-point

scoring system simplified to a five-point scale [12]

although there are no reported analyses of the impact

on reliability and validity.

Clinical samples explored using the EMQ include:

Survivors of head injury [1,13,14,15]; the elderly [7,

16]; people with Alzheimer’s disease [17]; survivors

of stroke [12,18,19]; multiple sclerosis [20,21,22]

and epilepsy [14].

Three studies have explored the factor structure of

the EMQ [17,22,23]. Cornish [23] administered the

28-item scale with the 9-point scoring system to a

sample of 277 undergraduates. Five factors ex-

plained 49% of the variance, which were proposed

to reflect underlying memory processes. Prior to this

study, Richardson and Chan [22] had explored the

factor structure of the original 35-item EMQ, using a

postal questionnaire with 115 MS patients. The

authors modified the measure, by including a

‘nuisance’ rating, and included patients’ self-reports

and the ratings by relatives or carers in their analysis,

which somewhat confounds results. They also found

a five-factor structure, explaining 62% of variance.

More recently, Eflikides et al. [17] explored the

factor structure of the 28-item EMQ with a modified

4-point rating scale as part of a larger study exploring

relationships between the EMQ, Wechsler Memory

Scale (WMS) and the Rivermead Behavioural

Memory Test (RBMT). They included 233 non-

clinical individuals (Greek population) and found a

seven-factor structure, explaining 62% of variance.

They postulated that the EMQ measured metamem-

ory, an awareness of episodic memory problems, and

correlated with more concrete performance mea-

sures within the RBMT.

Although Cornish [23] and Richardson and Chan

[22] both identified a five-factor structure, there were

key differences between them. Richardson and Chan

identified their factors as: Receptive communication;

Route-finding; Absent-mindedness; Face recogni-

tion; and Expressive communication. Cornish’s

factors were interpreted as: Retrieval; Task monitor-

ing; Conversational monitoring; Spatial memory;

and Memory for activities. Cornish proposed that

there were similarities between ‘Conversational

monitoring’ and ‘Expressive communication’;

‘Spatial memory’ and ‘Route-finding’; and ‘Task

monitoring’ and ‘Absent-mindedness’, but that the

‘Receptive communication’ and ‘Face recognition’

factors were not replicated in his study. There are

clear links between ‘Conversational monitoring’ and

‘Expressive communication’, with four of five items

in Richardson and Chan’s study overlapping with

Cornish; and links between ‘Spatial memory’ and

‘Route finding’, with two of four items in Richardson

and Chan’s study overlapping with Cornish. It is

unclear where the proposed links between ‘Task

monitoring’ and ‘Absentmindedness’ occur, how-

ever, given that ‘Absentmindedness’ consists of a

combination of factors 1, 2, 4 and 5 in the Cornish

study, with only one item overlapping with Cornish’s

‘Task monitoring’, and a far stronger link with factor

1 ‘Retrieval’.

In comparison, Eflikides et al. [17] defined the

seven factors identified as: Problems in prospective/

general memory; Difficulties in learning and repeti-

tion of responses; Forgetting changes in daily

routines; Visuo-spatial; Semantic memory; Episodic

memory and for faces; and Visual reconstruction

memory. There are clear links between

‘Visuo-spatial’ and the ‘Spatial’ Factor in Cornish’s

[23] study, and ‘Route-finding’ in Richardson and

Chan’s [22] study. There are also links between

‘Forgetting changes in daily routine’ and the

‘Memory for activities’ factor in Cornish’s study.

Otherwise, there are wide-ranging overlaps between

and across the three different studies.

These three studies indicate difficulties in clearly

defining the underlying factor structure of the EMQ,

other than a strong ‘Spatial’ factor, which is con-

sistent across both non-clinical and an MS popula-

tion. Findings are complicated by the different

samples, methodology and type of questionnaires

The Everyday Memory Questionnaire – revised 115

Dis

abil

Reh

abil

Dow

nloa

ded

from

info

rmah

ealth

care

.com

by

SUN

Y S

tate

Uni

vers

ity o

f N

ew Y

ork

at S

tony

Bro

ok o

n 10

/27/

14Fo

r pe

rson

al u

se o

nly.

Page 3: The Everyday Memory Questionnaire – revised: Development of a 13-item scale

used. We have identified no studies that compare

the validity of the EMQ across different groups of

clinical and non-clinical participants. The EMQ has

also been altered over time, simplifying its scoring

system, with no data on the psychometric impact.

Cornish [23] suggested extending the items in the

EMQ to improve its psychometric properties and

diagnostic utility. In clinical practice, however, the

28-item questionnaire is quite lengthy, with anecdotal

evidence that some items can be difficult to interpret,

and feel repetitive. Although there is obvious merit in

extending the questionnaire to improve its psycho-

metric and diagnostic properties, the use of the EMQ

as a clinical tool would suggest a need to reduce the

items, to improve its face validity and ease of use, but

maintain an element of differentiation between

different everyday memory failures.

The aims of this study were:

. To determine the internal consistency and

factor structure of the 5 response category

version of the EMQ in healthy participants;

. To determine the construct validity by com-

paring patients with memory problems with

healthy participants;

. To develop a shortened version of the ques-

tionnaire for clinical use and to determine its

sensitivity to differences between patients with

memory problems and healthy participants;

. To determine the internal consistency and

factor structure of the revised version of the

EMQ in clinical participants (people with

memory problems following MS and stroke)

and healthy participants.

Method

Participants

Data were drawn from two sources for three groups:

A sample of MS patients (n¼ 160) from a study

evaluating cognitive assessment and intervention

[20]; a healthy control sample, (n¼ 98) comprising

relatives of the MS patients from Lincoln et al. [20];

and a sample of stroke patients (n¼ 90) from a drug

treatment trial with memory-impaired stroke patients

(unpublished).

Participants with clinically definite, clinically

probable or laboratory supported MS were recruited

from a multiple sclerosis management clinic at

University Hospital, Nottingham. Although all were

screened for cognitive profile, there were no exclu-

sion or inclusion criteria based on cognitive perfor-

mance. Participants in this study were randomly

allocated to receive either a cognitive assessment or a

cognitive assessment and rehabilitation. The healthy

control sample was recruited from relatives of the

MS group. The stroke patients were recruited from

the register of stroke admissions to the City,

University and General hospitals in Nottingham for

a treatment trial. All had evidence of memory

problems as assessed on the Rivermead Behavioural

Memory Test. Participant characteristics are shown

in Table I. Ethical approval was granted by Notting-

ham Local Research Ethics Committee.

Procedure

The revised version of the Everyday Memory

Questionnaire [24] was administered to all partici-

pants. Those in the healthy control group and stroke

sample were sent the questionnaire by post. Those in

the MS sample were asked to complete the ques-

tionnaire as part of a broader battery of outcome

measures, carried out by interview 4 months after

recruitment to the study by an independent assessor,

blind to group allocation.

The EMQ consisted of 28 items, each describing

everyday activities, which may involve memory

failure. A 5-point scoring was used with a scale of:

0 (once or less in the last month); 1 (more than once

a month but less than once a week); 2 (about once a

week); 3 (more than once a week but less than once a

day); and 4 (once or more in a day) [20].

Statistical analyses

All data were analysed using SPSS version 11 for

Windows statistical package. Reliability analyses

(using Cronbach’s alpha) and factor analyses (using

a principal components analysis and either a direct

oblimin or varimax rotation, depending on results)

were carried out on the non-clinical sample initially.

Items were removed in a stepwise fashion according

to the following criteria:

. Corrected item-total correlations less than 0.3

in reliability analyses;

Table I. Descriptive statistics for three groups.

Healthy

control n¼98

MS

n¼ 160

Stroke

n¼90

Gender

Male 32% 30% 35%

Female 68% 70% 65%

Age

Mean 43 43 68

SD 15 11 10

Range 14 – 76 17 – 71 41 – 88

Mean total EMQ (25 item)

Mean 0.63 0.93 1.27

SD 0.6 0.81 0.7

Range 0 – 73 0 – 89 2 – 85

116 J. Royle & N. B. Lincoln

Dis

abil

Reh

abil

Dow

nloa

ded

from

info

rmah

ealth

care

.com

by

SUN

Y S

tate

Uni

vers

ity o

f N

ew Y

ork

at S

tony

Bro

ok o

n 10

/27/

14Fo

r pe

rson

al u

se o

nly.

Page 4: The Everyday Memory Questionnaire – revised: Development of a 13-item scale

. Items that poorly differentiated the three

samples (healthy controls, MS and Stroke),

defined as not significant or significant only at

the 5% level in one-way analyses of variance

with Bonferroni post hoc test of multiple

comparisons;

. Items with consistent low scores across the

three samples, defined as (a) having item

means lower than the total scale mean for all

3 groups, and (b) less than 20% of each group

scoring 3 – 4 for the item.

Reliability and factor analysis of the non-clinical

sample were then repeated for the revised scale,

using Cronbach’s alpha and principal components

factor analysis with direct oblimin or varimax

rotation, as appropriate [23,25,26], estimating num-

ber of factors to be retained as those with Eigenvalues

over 1.

Comparisons across the three samples were

carried out, using one-way analyses of variance, with

the non-parametric Welch statistic where variances

were not homogenous and the Bonferroni post hoc

test. The mean total for the 13-item scale and each

factor were compared.

Results

Reliability and Factor analyses, using healthy

control sample

Participants’ total scores on the questionnaire are

shown in Table I. Cronbach’s alpha for the scale was

high (0.91) indicating strong internal reliability. On

checking the individual items, three were found to

have low corrected item-total correlations: item 11

(Failing to recognize, by sight, close relatives or

friends that you meet frequently) (r¼ 0.13); item 19

(Forgetting important details about yourself, e.g.,

your birthdate or where you live) (r¼ 0.01); and item

25 (Getting lost or turning in the wrong direction on

a journey, on a walk, or in a building where you have

often been before) (r¼ 0.18). These items were

removed and the analyses repeated resulting in a

Cronbach’s alpha of 0.92, with all items showing a

corrected item-total correlation of at least 0.3,

indicating strong internal reliability.

A factor analysis of the remaining 25 items was

carried out for the healthy controls (n¼ 98). A

principal components analysis using a varimax

rotation produced a 5-factor solution, which ex-

plained 61% of the variance. Results are shown in

Table II.

Comparisons across three groups

Overall scale totals and all remaining 25 items were

compared across the three groups using one-way

analyses of variance (total scores are shown in Table

I). Levene’s test for homogeneity of variance was

violated for all the items, therefore the more

conservative Welch test for significance was used,

and the Bonferroni post hoc test, to allow for

multiple comparisons. There were significant differ-

ences between the three groups for total EMQ score

(p5 0.001) indicating that the scale is sensitive to

differences in memory between the groups. Bonfer-

roni post hoc analysis indicated that 5 items did not

differentiate between the 3 groups (1,7,10,21,26).

Low occurrence was taken into account by

identifying low scoring items across the 3 groups.

Low scoring was defined as (a) having item means

lower than the total scale mean for all 3 groups, and

(b) less than 20% of each group scoring 3 – 4 for the

item. Nine items were identified as low scoring using

these criteria (2,3,4,12,21,22,23,26,27). Both of

these sets of items were removed, leaving a 13-item

scale (see Appendix 1).

Participants’ total scores on the questionnaire,

summed over all 13 items, ranged from 0 – 41, with a

mean total of 9.75 (SD 8.6) (n¼ 98). Cronbach’s

alpha for the scale was high (0.89), and all items

Table II. Loadings from initial factor analysis, healthy-control sample.

Factor 1 Factor 2 Factor 3 Factor 4 Factor 5

Item Loading Item Loading Item Loading Item Loading Item Loading

1 0.46 9 0.48 2 0.65 10 0.73 23 0.79

4 0.53 17 0.43 3 0.71 12 0.77 24 0.54

5 0.67 22 0.47 20 0.48 16 0.50

6 0.73 26 0.72 21 0.48

7 0.77 27 0.85

8 0.81 28 0.58

13 0.63

14 0.77

15 0.73

18 0.58

The Everyday Memory Questionnaire – revised 117

Dis

abil

Reh

abil

Dow

nloa

ded

from

info

rmah

ealth

care

.com

by

SUN

Y S

tate

Uni

vers

ity o

f N

ew Y

ork

at S

tony

Bro

ok o

n 10

/27/

14Fo

r pe

rson

al u

se o

nly.

Page 5: The Everyday Memory Questionnaire – revised: Development of a 13-item scale

showed a corrected item-total correlation of at least

0.3, indicating strong internal reliability.

A principal components analysis estimating factors

by a scree plot with Eigenvalues over 1, using an

oblimin rotation suggested 3 factors explaining 62%

of the variance. Correlations among the 3 factors

were low to moderate, ranging from 0.20 – 0.48.

Table III shows the final factor pattern matrix,

together with means and communalities for indivi-

dual items. All factors have at least two items with

loadings of 0.5 or greater. Based on the items with

the heaviest loadings, the following interpretations

are offered for two of the three factors:

Factor 1: Retrieval. Similar to Cornish’s ‘retrieval’, items

in this factor concern failing to recall recent events

(5,6,8,15) retrieving words from memory (13) and

prospective memory (14,18). Indications are that these

memories are stored and accessed with prompts, but

retrieval without prompts is unreliable.

Factor 2: Attentional tracking. All four items in this factor

(16,17,20,28, and associated links with 18 and 9)

concern losing track in conversation or when reading,

suggesting an attentional or working memory problem.

Items in factor 3 share little in common and there

is no clear interpretation of the processes involved in

both.

Comparisons across 3 groups: 13-item scale

A one-way analysis of variance was carried out

comparing the whole 13-item scale, and the three

factors. Factor scores were calculated by totalling

the items in each factor, weighting them equally.

As variances were not homogeneous, the more

conservative Welch statistic was used. Bonferroni

post hoc tests were used for multiple comparisons

between groups. Results are shown in Table IV.

Results show highly significant differences be-

tween groups for the revised scale, and for each

factor, with stroke patients scoring the highest

overall. Bonferroni tests also showed significant

differences between groups for all combinations

except between MS and stroke patients for Factor

2, and between healthy controls and MS patients for

Factor 3. Generally, Factor 1 represents the highest

scoring items and Factor 3 the lowest. MS patients

scored highest for Factor 1 (retrieval) and Factor 2

(attentional tracking). Stroke patients scored highly

for all three.

Pearson Product Moment correlations were cal-

culated between the original 28-item and the revised

13-item scales across all 3 groups. The two versions

were very highly significantly correlated, for the

whole sample r¼ 0.97, p5 0.001; and each group

(healthy control r¼ 0.97; p5 0.001; MS r¼ 0.98;

p5 0.001; stroke r¼ 0.94; p5 0.001).

Discussion

The main aims of this study were to determine the

psychometric properties of the 28-item EMQ (five

response category version) for healthy participants

and participants with memory problems. These

analyses were then drawn upon to revise the ques-

tionnaire, resulting in a 13-item questionnaire with

potential to be used as a valid and reliable measure of

individuals’ beliefs about memory performance

in everyday life. The psychometric properties of the

28-item questionnaire were explored using the

healthy control group (n¼ 98). Initial reliability

Table III. Mean scores and factor loadings in the three-factor solution (direct oblimin rotation).

Scores

Orig. No. Factor and abridged item Mean SD Communalities

Factor 1: Retrieval

5 Having to check whether you have done something 1.34 1.2 0.59

6 Forgetting when it was that something happened 1.04 1.2 0.66

8 Forgetting that you were told something yesterday 0.96 1.1 0.71

13 Finding that a word is ‘on the tip of your tongue’ 1.36 1.2 0.61

14 Completely forgetting to do things you said you would 0.67 1.0 0.75

15 Forgetting important details of what you did 0.48 1.0 0.68

18 Forgetting to tell somebody something important 0.70 0.9 0.62

Factor 2: Attentional Tracking

16 When talking to someone, forgetting what you just said 0.90 1.1 0.61

17 When reading a paper, being unable to follow the story 0.47 0.7 0.43

20 Getting the details mixed up 0.40 0.7 0.54

28 Repeating to someone what you have just told them 0.60 1.0 0.64

Factor 3:

9 Starting to read something you have read before 0.39 0.7 0.58

24 Forgetting where things are normally kept 0.44 0.9 0.70

118 J. Royle & N. B. Lincoln

Dis

abil

Reh

abil

Dow

nloa

ded

from

info

rmah

ealth

care

.com

by

SUN

Y S

tate

Uni

vers

ity o

f N

ew Y

ork

at S

tony

Bro

ok o

n 10

/27/

14Fo

r pe

rson

al u

se o

nly.

Page 6: The Everyday Memory Questionnaire – revised: Development of a 13-item scale

analyses indicated internal reliability that was im-

proved by removal of three items (11,19,25). Initial

factor analyses on the remaining 25-item scale

yielded results that differed somewhat from previous

findings [17,22,23], although there was some con-

sistency with Cornish’s [23] study with a ‘General

memory’ factor (incorporating a combination of

Cornish’s ‘Retrieval’ and ‘Memory for activities’);

and a ‘Visual attention’ factor (incorporating a com-

bination of Cornish’s ‘Task monitor’ and ‘Spatial

memory’).

It may be that differences in sample, recruitment

and methodologies could explain differences in

factor structures identified. However, if everyday

memory abilities involve different cognitive domains,

consistency would be expected regardless of the

sample studied. These results indicate that the

questionnaire used with a non-clinical population

does fairly reliably and consistently differentiate

between two broad systems of memory and atten-

tion, with some differentiation of visual versus verbal,

or language systems.

The main aim of this study was to develop a

reliable and valid clinical tool. A 13-item scale was

developed. Reliability analyses indicated strong

internal reliability with good correlations between

each item and the item total. Further factor analyses

of the revised scale carried out using the healthy

control group indicated a three-factor solution. The

first, and strongest, factor ‘Retrieval’ showed marked

similarities with the original ‘General memory’ factor

for the full scale, and Cornish’s [23] ‘Retrieval’ factor

(items 6,8,13,14,15). All of these items involve a

distinct memory failure, suggesting a failure of a

retrieval system (i.e., failing to recall without

prompt). The second factor ‘Attentional tracking’

included four items all involving attention, mostly

verbal, with one visual item (item 17). This factor has

some concordance with the original ‘Visual atten-

tion’ factor (items 17 and 28) and Cornish’s

‘Conversational monitor’ (items 20 and 28). The

third factor showed the strongest overlap with

Eflikides et al.’s [17] ‘Visual reconstruction’ (items

9 and 24) but no similarities were found with

Richardson and Chan’s [22] study, despite including

a similar clinical group. It is not clear what the

common elements to this factor are but it could be

retained pending further clarification.

As with the original questionnaire, the factor

analysis of the revised scale broadly differentiated

between memory and attentional systems. A com-

parison of means across the three groups (clinical

and healthy control) showed that the 13-item scale

differentiated between the groups, with the stroke

group scoring highest overall, for the total score and

factor totals. Analyses indicate that ‘Retrieval’ and

‘Attentional tracking’ factors best differentiated

between both clinical and the healthy control group.

Further evidence of the validity of the revised scale

was confirmed by the strong relationships between

the original and revised versions, suggesting that the

revised 13-item questionnaire could provide a valid

and reliable tool for clinical use as a measure of

individual’s attitudes and beliefs about memory

difficulties, with some differentiation between

memory and attentional problems.

It is important to take into account methodological

limitations when considering these results. A retro-

spective analysis was used, including data from two

previous studies. The healthy control sample was

drawn from relatives of patients in the MS study, and

therefore presented a reasonable match for the MS

patients, but it would have been preferable to have an

additional well-matched control sample for the

stroke group. We have little demographic informa-

tion, and there were differences in sample sizes and

significant age differences between the stroke and the

other two groups. The stroke sample was signifi-

cantly older than the healthy control or MS groups.

However, age alone would not be expected to affect

the factor structure of the scale. Formal memory

assessment would have further informed results,

providing information on the nature of any memory

impairment, as the healthy control group, in parti-

cular, may have had previously unrecognised

memory difficulties.

In addition, there were differences in the method

of administering the EMQ. It was sent out as postal

questionnaires to stroke patients, whereas it was

administered by a researcher who was known to MS

patients and their families. However it would be

expected that the items would retain their relation-

ships with each other even though the levels of

impairment may differ.

This analysis has yielded a revised, shortened

version of the EMQ, with sound psychometric

properties. The questionnaire provides a potentially

useful clinical tool, to explore subjective report and

beliefs about memory and attentional difficulties with

a range of clients. It is short and therefore more likely

to be used in clinical practice than the original

Table IV. Comparison of three groups on the 13-item scale and

three factors.

Healthy

control MS Stroke

Mean SD Mean SD Mean SD

Revised scale*** 0.75 0.66 1.11 0.95 1.51 0.86

Factor 1*** 0.94 0.87 1.31 1.1 1.82 0.98

Factor 2*** 0.59 0.65 1.011 1.0 1.281 0.97

Factor 3*** 0.411 0.64 0.611 0.91 1.02 1.02

Note: ***p50.001: 1post hoc tests not significant.

The Everyday Memory Questionnaire – revised 119

Dis

abil

Reh

abil

Dow

nloa

ded

from

info

rmah

ealth

care

.com

by

SUN

Y S

tate

Uni

vers

ity o

f N

ew Y

ork

at S

tony

Bro

ok o

n 10

/27/

14Fo

r pe

rson

al u

se o

nly.

Page 7: The Everyday Memory Questionnaire – revised: Development of a 13-item scale

version. It may also provide a useful aid to engage-

ment in therapy or rehabilitation, as a systematic

method of gathering subjective perceptions and

beliefs about difficulties.

Further research is needed to directly compare

report using the EMQ-R to formal assessment of

memory and attention, using both traditional and

more ecologically valid assessment tools, such as

the RBMT [27] or the Test of Everyday Attention

[28]. Finally, further concurrent validity is needed,

comparing results from the EMQ-R with similar

questionnaires, such as the Cognitive Failures

Questionnaire [9] and the Cognitive Difficulties

Scale [29], which both measure similar constructs.

In addition an assessment of test-retest reliability is

essential if the EMQ-R is used as a measure of

change, for example following rehabilitation The

reliability of the full EMQ has been found to be good

(r¼ 0.85) [17] but further studies are needed.

References

1. Sunderland A, Harris JE, Baddeley AD. Do laboratory

tests predict everyday memory? A neuropsychological study.

J Verbal Learning Verbal Behav 1983;22:727 – 738.

2. Chaytor N, Schmitter-Edgecombe M. The ecological validity

of neuropsychological tests: A review of the literature on

everyday cognitive skills. Neuropsychol Rev 2003;13:181 –

197.

3. Higginson CI, Arnett PA, Voss WD. The ecological validity of

clinical tests of memory and attention in multiple sclerosis.

Arch Clin Neuropsychol 2000;15:185 – 204.

4. Hickox A, Sunderland A. Questionnaire and checklist

approaches to assessment of everyday memory problems. In:

Crawford JR, Parker DM. editors. A Handbook of Neurop-

sychological Assessment. Mahwah, NJ: Lawrence Erlbaum

Associates; 1992. pp 103 – 114.

5. Ruisel I. Memory rating. Studia Psychologica 1991;33:71 –

77.

6. Larrabee GJ, Crook TH. The ecological validity of memory

testing procedures: Developments in the assessment of every-

day memory. In: Sbordone RJ, Long CJ, editors. Ecological

validity of neuropsychological testing. Delray-Beach FL: GR

Press/St Lucie Press; 1996. pp 225 – 242.

7. Sunderland A, Watts K, Baddeley AD, Harris JE. Subjective

memory assessment and test performance in elderly adults.

J Gerontol 1986;41:376 – 384.

8. Schlechter TM, Herrman DJ, Sronach P, Rubenfeld L,

Zenker S. Metamemory Questionnaires. Paper presented at

the American Educational Research Meeting, New York;

1982.

9. Broadbent DE, Cooper PF, Fitzgerald P, Parkes KR. The

cognitive failures questionnaire (CFQ) and its correlates. Br J

Clin Psychol 1982;21:1 – 16.

10. Herrman DJ. Know thy memory: The use of questionnaires

to assess and study memory. Psychological Bull 1982;92:

434 – 452.

11. Sunderland A, Harris JE, Gleave J. Memory failures in

everyday life following severe head injury. J Clin Neuropsy-

chol 1984;6:125 – 141.

12. Tinson DJ, Lincoln NB. Subjective memory impairment after

stroke. Int Disabil Stud 1987;9:6 – 9.

13. Boake C, Freeland JC, Ringholz GM, Nance ML, Edwards

KE. Awareness of memory loss after severe closed head injury.

Brain Inj 1995;9:273 – 283.

14. Goldstein LH, Polkey CE. Everyday memory after unilateral

temporal lobectomy or amygdalo-hippocampectomy. Cortex

1992;28:189 – 201.

15. Quemada JI, Cespedes JMM, Ezkerra J, Ballesteros J, Ibarra

N, Urruticoechea I. Outcome of memory rehabilitation in

traumatic brain injury assessed by neuropsychological

tests and questionnaires. J Head Trauma Rehabil 2003;18:

532 – 540.

16. Seltzer B, Vasterling JJ, Hale MA, Khurana R. Unawareness

of memory deficit in Alzheimer’s disease: Relation to mood

and other disease variables. Neuropsychiatry, Neuropsychol

Behav Neurol 1995;8:176 – 181.

17. Eflikides A, Yiultsi E, Kangellidou T, Kounti F, Dina F,

Tsolaki M. Wechsler Memory Scale, Rivermead Behavioural

Memory Test and Everyday Memory Questionnaire in healthy

adults and Alzheimer’s patients. Eur J Psychological Assess

2002;18:63 – 77.

18. Stewart FM, Sunderland A, Sluman SM. The nature and

prevalence of memory disorder late after stroke. Br J Clin

Psychol 1996;35:369 – 379.

19. Towle D, Edmans J, Lincoln NB. Use of computer presented

games with memory impaired stroke patients. Clin Rehabil

1988;2:303 – 308.

20. Lincoln NB, Dent A, Harding J, Weyman N, Nicholl C,

Blumhardt LD, Playford ED. Evaluation of cognitive assess-

ment and cognitive intervention for people with multiple

sclerosis. J Neurol, Neurosurg Psychiatry 2002;72:93 – 99.

21. Richardson JTE. Memory impairment in multiple sclerosis:

Reports of patients and relatives. Br J Clin Psychol 1996;35:

205 – 219.

22. Richardson JTE, Chan RCB. The constituent structure of

subjective memory questionnaires: Evidence from multiple

sclerosis. Memory 1995;3:187 – 200.

23. Cornish IM. Factor structure of the everyday memory ques-

tionnaire. Br J Psychol 2000;91:427 – 438.

24. Sunderland A, Harris JE, Baddeley AD. Assessing everyday

memory after severe head injury. In: Harris JE, Morris PE,

editors. Everyday memory, actions, and absentmindedness.

London: Academic Press; 1984. pp 193 – 212.

25. Kline P. The handbook of psychological testing. London:

Routledge; 1993.

26. Kline P. An easy guide to factor analysis. London: Routledge;

1994.

27. Wilson BA, Cockburn J, Baddeley AD. The Rivermead

Behavioural Memory Tests. Bury St Edmunds, UK: Thames

Valley Test Company; 1984.

28. Robertson IH, Ward T, Ridgeway V, Nimmo-Smith I. The

test of everyday attention. Bury St Edmunds, UK: Thames

Valley Test Company; 1994.

29. McNair D, Kahn R. The cognitive difficulties scale. In: Crook

T, Ferris S, Bartus R, editors. Assessment in geriatric

psychopharmacology. New Canaan, CT: Mark Powley and

Associates; 1983. pp 137 – 143.

120 J. Royle & N. B. Lincoln

Dis

abil

Reh

abil

Dow

nloa

ded

from

info

rmah

ealth

care

.com

by

SUN

Y S

tate

Uni

vers

ity o

f N

ew Y

ork

at S

tony

Bro

ok o

n 10

/27/

14Fo

r pe

rson

al u

se o

nly.

Page 8: The Everyday Memory Questionnaire – revised: Development of a 13-item scale

Appendix 1

Everyday Memory Questionnaire – Revised

Instructions

Below are listed some examples of things that happen to people in everyday life. Some of them may happen

frequently and some may happen very rarely. We should like to know how often on average you think each one

has happened to you over the past month. Write the appropriate letter in the box beside the item.

A. Once or less in the last month.

B. More than once a month but less than once a week.

C. About once a week.

D. More that once a week or less than once a day.

E. Once or more in a day.

1. Having to check whether you have done something that you should have done.

2. Forgetting when it was that something happened; for example, whether it was

yesterday or last week.

3. Forgetting that you were told something yesterday or a few days ago, and maybe

having to be reminded about it.

4. Starting to read something (a book or an article in a newspaper, or a magazine)

without realizing you have already read it before.

5. Finding that a word is ‘on the tip of your tongue’. You know what it is but cannot

quite find it.

6. Completely forgetting to do things you said you would do, and things you planned

to do.

7. Forgetting important details of what you did or what happened to you the day

before.

8. When talking to someone, forgetting what you have just said. Maybe saying

‘what was I talking about?’

9. When reading a newspaper or magazine, being unable to follow the thread of a story;

losing track of what it is about.

10. Forgetting to tell somebody something important, perhaps forgetting to pass on a

message or remind someone of something.

11. Getting the details of what someone was told you mixed up and confused.

12. Forgetting where things are normally kept or looking for them in the wrong

place.

13. Repeating to someone what you have just told them or asking someone the same

question twice.

Please check that you have put a letter in EVERY box. THANK YOU.

The Everyday Memory Questionnaire – revised 121

Dis

abil

Reh

abil

Dow

nloa

ded

from

info

rmah

ealth

care

.com

by

SUN

Y S

tate

Uni

vers

ity o

f N

ew Y

ork

at S

tony

Bro

ok o

n 10

/27/

14Fo

r pe

rson

al u

se o

nly.