comqol-s5

COMPREHENSIVEQUALITYOF LIFE SCALE

– SCHOOL VERSION(Grades 7-12)

FIFTH EDITION(ComQol-S5)

Robert A. CumminsSchool of Psychology

Deakin University

MANUAL

1997

Correspondence to:

Robert A. Cummins Ph.D., F.A.P.S.School of PsychologyDeakin University221 Burwood HighwayBurwood, MelbourneVictoria 3125AUSTRALIA

e-mail: [email protected]

ISBN 0-7300-2726-0

Published by the School of PsychologyDeakin University

3

Contents

1 Introduction1.1 Measuring and defining quality of life1.2 The Scale1.3 Prior editions of the scale1.4 ComQol-S51.5 Summary

2 Administration2.1 General information

3 Comprehensive Quality Of Life Scale (ComQol-S5)

4 Calculation of results4.1 Coding the objective data4.2 Coding the subjective data4.3 Data cleaning4.4 Dealing with data skew4.5 Forms of data analysis

4.5.1 For the practitioner or service provider4.5.2 For the researcher4.5.3 % SM: A standardised comparison statistic

4.6 Individual SQOL diagnostic data tables

5 Theoretical Issues5.1 Why use the ‘Delighted-Terrible’ scale?5.2 Should ‘not important at all’ be scored as 1 or 0?5.3 Should ‘mixed satisfaction/dissatisfaction’ be scored as 0?5.4 Why not score the satisfaction scale from +1 to +7?5.5 Why not use the Ferrans and Powers (1985) scoring system?5.6 Why not use the Raphael et al. (1996) scoring system?

6 Alternative Forms of the Scale6.1 Parallel versions of the scale6.2 Additional domains

4

7 Psychometric Data

Study Codes

7.1 Objective means7.2 Subjective means

7.2.1 Importance sub-scale7.2.2 Satisfaction sub-scale7.2.3 Importance x Satisfaction

7.3 Reliability7.3.1 Cronbach’s alpha7.3.2 Internal reliability7.3.3 Test-retest reliability

7.4 Validity7.4.1 Content validity7.4.2 Construct validity

8 References to the text

APPENDIX A: Psychotropic drug namesAPPENDIX B: Scoring ComQolAPPENDIX C: Author publications

5

Acknowledgements

The author gratefully acknowledges comments and ideas which have contributed to this fifthedition from the following people: Christine Baxter, Eleonora Gullone, Marita McCabe,Shelley Reid, Dale Fogarty, Julie Cochrane.

I am also greatly indebted to the students who have worked with me to produce many of thedata presented in this manual.

I thank Sheryl Monteath for her able assistance with data analysis for studies A6 and A7.

I also thank Rai Sahib, Julie Asquith, Trudy Wallace, Natasha Cho and Betina Gardner forword-processing this document.

6

1 Introduction

1.1 Measuring and defining quality of life

The quality of life (QOL) construct has a complex composition, so it is perhaps not surprisingthat there is neither an agreed definition nor a standard form of measurement. This is not dueto a lack of ideas. Cummins (1996a) has recorded well over 100 instruments which purportto measure life quality in some form, but each one contains an idiosyncratic mixture ofdependent variables.

It is also notable that many QOL instruments have been developed for highly selected groupsin the population; particularly in regard to scales devised to monitor medical conditions orprocedures. Because of this, they are unsuitable for use with the general population.However, even the more general scales which have been devised cannot be used with allsectors of the population. Those created for the general adult population cannot be used withsome population sub-groups such as people with cognitive impairment and children. This isan important limitation since it means that the QOL experienced by such groups cannot benorm-referenced back to the general population.

In order to remedy this situation, the Comprehensive Quality of Life Scale (ComQol) hasbeen developed. This scale has been designed in parallel forms suitable for any populationsub-group. These forms are:

ComQol – A: designed for use with the general adult population.

ComQol – I: designed for use with people who have an intellectual disability or otherform of cognitive impairment.

ComQol – S: designed for use with adolescents 11-18 years who are attending school.

The scale also contains features of construction which reflect contemporary understanding ofthe QOL construct. The details of test development have been published elsewhere(Cummins, 1991; Cummins, McCabe, Romeo and Gullone, 1994).

Definition

The scale that follows is an operationalization of the following definition of quality of life:

Quality of life is both objective and subjective, each axis being theaggregate of seven domains: material well-being, health, productivity,intimacy, safety, community, and emotional well-being. Objective domainscomprise culturally-relevant measures of objective well-being. Subjectivedomains comprise domain satisfaction weighted by their importance to theindividual.

7

1.2 The scale

ComQol incorporates a contemporary understanding of the QOL construct. As such itcontains the following features:

(a) It is multidimensional. There is consensus within the field that the most usefulmeasures of life quality must incorporate the separate components whichcomprise this construct, even though the precise nature of these components aresomewhat conjectural (Felce & Perry, 1995). ComQol defines life quality interms of seven domains which together are intended to be inclusive of all QOLcomponents. These are: Material well-being, health, productivity, intimacy,safety, place in community, and emotional well-being. Evidence for theadoption of these seven domains has been presented by Cummins (1996b,1997a). A discussion of additional domains is provided in 6.2.

(b) It is multi-axial. This takes two forms. The first is in the separate measurementof objective and subjective components. The contemporary literature is quiteconsistent in its determination that, while both of these axes form a part of theQOL construct, they generally have a very poor relationship to one another.For example, physical health and perceived health are poorly correlated (see7.6.1).

The scale is also multi-axial in terms of its subjective measures. Each domain isseparately rated in terms of its importance to the individual as well as on itsperceived satisfaction. It is notable that importance and satisfaction generallyare moderately positively correlated with one another (see 7.6.2). The level ofimportance then provides an individualised weighting factor for each domainsuch that the subjective QOL measurement can be expressed as Importance xSatisfaction.

(c) It can be used with any section of the population. Two parallel versions of theadult ComQol have been developed. ComQol-S is for use with adolescentstudents, while ComQol-I is designed for people who have an intellectualdisability or other form of cognitive impairment.

This latter scale incorporates a pre-testing protocol to determine whether, and towhat level of complexity, respondents are able to use the scale. This pre-testingprogressively moves responding from concrete to abstract. It commences withan ordering task involving differently sized printed blocks, progresses to a taskinvolving block size matching to a printed ladder scale (e.g. the largest blockcorresponds with the top of a printed ladder), and ends with the use of a Likertscale involving the abstract conception of ‘importance’. At each stage of thistesting, people commence with a task involving choice between two types ofresponse (e.g. one large and one small block) and can progress to a maximum offive. The number of response choices successfully negotiated in the finalabstract task is then used to determine the Likert Scale complexity to be usedwith ComQol-ID. For example, if a respondent is only able to manage theabstract task as a choice between two levels of importance, then they will beprovided with a version of ComQol-ID where Likert scales are presented as abinary choice.

8

The use of this process eliminates those respondents who do not have thecognitive capacity to respond validly to the scale. This is crucial given ourunderstanding that people who are placed in a forced-choice situation, wherethey do not understand the task, will often respond either at random or in amanner they consider will please the interviewer. Pre-testing therefore ensuresthat each respondent is provided with a level of Likert scale complexity whichapproximates their discriminative capacity.

(d) The scale is psychometrically sound. It is reliable, stable, valid and sensitive(see Section 7).

(e) The sum of the domain scores for satisfaction can be referenced to the ‘gold-standard’ of 75 ± 2.5% SM (Cummins, 1996b).

1.3 Prior editions of the scale

Fourth EditionWhile the wording of the subjective items is essentially the same as in the thirdedition, more substantial changes were made to the wording of a number of theobjective items. In all cases this involved a clarification or elaboration of the itemrather than a total change in content. The rating scale for each objective item alsoincreased from three-point to five-point. This meant that each objective domainscore was now free to vary from 3 to 15 instead of 3 to 9 as in the third edition. Thepurpose of this change was to increase the discriminative power of the objectivedomain scores.

Fifth EditionThe wording of subjective items has remained unchanged. Within the objectivescale, the wording of several items has been simplified but the sense of the items hasremained. Item 3c has been completely changed. This was the only item that failedto display a significant intra-domain correlation in the fourth edition. No data areavailable on how this new item 3c performs. The satisfaction questions use a 7-point, rather than a 5-point response scale, since we have found this age group toexperience no difficulty in using the more complex version.

1.4 ComQol-S5

This is a parallel version of ComQol-A5.

9

Some items have been modified to make them appropriate for the target group. These areindicated by an asterisk (*)

1.5 Summary

This instrument is based on the following propositions:

• Quality of life (QOL) can be described in both objective (O) and subjective (S)terms.

• Each objective (OQOL) and subjective (SQOL) axis is composed of 7 domains.

1. Material well-being2. Health3. Productivity4. Intimacy5. Safety6. Place in community7. Emotional well-being

• The measurement of each SQOL domain is achieved by obtaining a satisfactionscore of that domain which is weighted by the perceived importance of the domainfor the individual. Thus,

SQOL = Σ (Domain satisfaction x Domain importance).

NoteThis fifth edition of ComQol can be viewed both as a research instrument and a standardised test. Thefirst edition was compiled in 1991. It is anticipated that several further editions will be produced as newdata and ideas indicate ways that the scale can be improved. To this end comments are welcomed.

10

2 Administration

2.1 General Information

The scale is intended to be self-administered.

It should be noted that the instrument exists in two parts, as objective and subjective. Undernormal conditions these two parts show little relationship to one another. This is consistentwith the broader literature on QOL which clearly indicates the independence of objective andsubjective variables. Hence, the full administration of the scale yields two measures of lifequality which are quite separate from one another.

Whether one, or both parts of the scale should be administered is a decision to be madedepending on the purpose of scale administration. If only subjective QOL is of interest thenonly this part of the scale needs to be used.

If it is given in a group situation, it should be emphasised that there is no time limit and thatpeople may ask for assistance with any question that they do not fully understand.

ComQol takes about 15-20 minutes to complete. The subjective section alone takes aboutfive minutes to complete.

11

3 ComQol-S5

This scale has three sections. The first will ask you for some factual information. The nexttwo will ask how you feel about various aspects of your life.

To answer each question put a (√) in the appropriate box. Please ask for assistance if there isanything you do not understand.

Please answer all the questions and do not spend too much time on any one item.

What is your date of birth? ______/______/______ day month year

What is your sex? (circle one) Male Female

12

Section 1

This section asks for information about various aspects of your life. Please tick the box thatmost accurately describes your situation.

*1(a) Where do you live?

A house Do your parents own the placewhere you live or do they pay rent?

A flat or apartmentOwn

A room (e.g. in a hostel)or caravan Rent

*b) How many clothes and toys do you have compared with other people ofyour age?

More thanalmost anyone

More thanmost people

Aboutaverage

Less thanmost people

Less than almostanyone

*c) If either of your parents has paid work, please give the name of their job.

Father _______________________________________________________

Mother _______________________________________________________

2a) How many times have you seen a doctor over the past 3 months?

None 1 - 2 3-4 5-7 8 or more(about once (about every (about once a week

a month) two weeks) or more)

13

b) Do you have any on-going medical problems? (e.g. visual, hearing,physical, health, etc.).

Yes No

If yes please specify:

Name of medical Extent of medicalcondition condition

e.g. Visual Require glasses for readingDiabetes Require daily injectionsEpilepsy Requires daily medication

_______________________ ___________________________________

_______________________ ___________________________________

_______________________ ___________________________________

(c) What regular medication do you take each day?

If none tick box

or

Name(s) of medication (don’t worry if you get the spelling wrong)

_________________________________________________________________

_________________________________________________________________

_________________________________________________________________

_________________________________________________________________

14

3(a) How many hours do you spend on the following each week? (Averageover past 3 months)

Hours work for pay 0 1-10 11-20 21-30 31-40+(not counting pocket money)

Hours at school or college 0 1-10 11-20 21-30 31-40+

Hours unpaid child care 0 1-10 11-20 21-30 31-40+

(b) In your spare time, how often do you have nothing much to do?

Almost always Usually Sometimes Not Usually Almost never

(c) On average, how many hours TV do you watch each day?

Hours per day

None 1 – 2 3 – 5 6 – 9 10 or more

4(a) How often do you talk with a close friend?

Daily Several timesa week

Once a week Once a month Less thanonce a month

(b) If you are feeling sad or depressed, how often does someone show theycare for you?


15

(c) If you want to do something special, how often does someone else wantto do it with you?


5(a) How often do you sleep well?


(b) Are you safe at home?


(c) How often are you worried or anxious during the day?


16

*6(a) Below is a list of leisure activities. Indicate how often in an averagemonth you attend or do each one for your enjoyment (not employment).

Activity Number of times per month

(1) Go to a club/group/society __________

(2) Meet with friend(s) __________

(3) Watch live sporting events(Not on TV) __________

(4) Go to a place of worship __________

(5) Chat with neighbours __________

(6) Eat out __________

(7) Go to a movie __________

(8) Visit family __________

(9) Play sport or go to a gym __________

(10) Other (please describe) ___________________________________

(b) Do you hold an unpaid position of responsibility in relation to anyteam, club, group, or society?

Yes No If no, go to question (c)

If ‘yes’, please indicate the highest level of responsibility held:

Committee Member

Committee Chairperson/Convenor

Secretary/Treasurer/Team Vice-captain

Captain, Group President, Chairperson or Convenor

17

(c) How often do people outside your home ask for your help or advice?

Almost every day Quite often Sometimes Not often Almost never

7(a) How often can you do the things you really want to do?


(b) When you wake up in the morning, how often do you wish you couldstay in bed all day


(c) How often do you have wishes that cannot come true?


18

Section 2

How Important are each of the following life areas to you?

Please answer by placing a (√) in the appropriate box for each question.

There are no right or wrong answers. Please choose the box that best describes howimportant each area is to you. Do not spend too much time on any one question.

1. How important to you ARE THE THINGS YOU OWN?

Could not bemore important

Very important Somewhatimportant

Slightlyimportant

Not importantat all

2. How important to you is YOUR HEALTH?



Slightlyimportant

Not importantat all

3. How important to you is WHAT YOU ACHIEVE IN LIFE?



Slightlyimportant

Not importantat all

19

4. How important to you are CLOSE RELATIONSHIPS WITH YOUR FAMILYOR FRIENDS?



Slightlyimportant

Not importantat all

5. How important to you is HOW SAFE YOU FEEL?



Slightlyimportant

Not importantat all

6. How important to you is DOING THINGS WITH PEOPLE OUTSIDE YOURHOME?



Slightlyimportant

Not importantat all

7. How important to you is YOUR OWN HAPPINESS?



Slightlyimportant

Not importantat all

20

Section 3

How satisfied are you with each of the following life areas?

There are no right or wrong answers. Please ( √ ) the box that best describes how satisfiedyou are with each area.

1. How satisfied are you with the THINGS YOU OWN?

Delighted Pleased Mostlysatisfied

Mixed Mostlydissatisfied

Unhappy Terrible

2. How satisfied are you with your HEALTH?



Unhappy Terrible

3. How satisfied are you with what you ACHIEVE IN LIFE?



Unhappy Terrible

4. How satisfied are you with your CLOSE RELATIONSHIPS WITH FAMILYOR FRIENDS?



Unhappy Terrible

21

5. How satisfied are you with HOW SAFE YOU FEEL?



Unhappy Terrible

6. How satisfied are you with DOING THINGS WITH PEOPLE OUTSIDEYOUR HOME?



Unhappy Terrible

7. How satisfied are you with YOUR OWN HAPPINESS?



Unhappy Terrible

22

4 Calculation of results

4.1 Coding the objective data

The following information is relevant to the scoring procedures:

Missing values: Score as 9 (then get the computer to recognise ë9í as denoting a missingvalue).

Estimated income: The average adult Australian full-time wage in February 1997 was$38,063 per year. Users in other countries will need to modify the scoring of income on apro rata basis.

MATERIAL WELL-BEING

1(a) Accommodation:house + own = 5flat/apartment + own = 4house + rent = 3flat/apartment + rent = 2Room + either = 1

(b) Possessions:More than almost anyone = 5 Less than most people = 2More than most people = 4 Less than almost anyone = 1About average = 3

(c) Estimated income:More than $56,000 = 5 $11,000 - $25,999 = 2$41,000 - $55,999 = 4 Below $10,999 = 1$26,000 - $40,999 = 3

An estimation of the above categories can be obtained from the following occupationalgroupings obtained from the Year Book Australia (1994) and Castles (1992).

Below $10,999StudentsPeople who are unemployed

$11,000 - $25,999Laborers and related workers

23

$26,000-$40,999School teachers Farmers & farm managersParaprofessionals Managing supervisorsClerks Artists & related professionalsDrivers Technical officersPersonal service workers NursesSalespersons PoliceTradespersons Plant & machine operators/driversJunior academics

$41,000-$55,999Legislators & government appoint officialsManagers and administratorsSchool principalsProfessionalsEngineers & building professionalsSocial professionalsBusiness professionals

56,000+Managing directors/General managersMedical doctorsSenior academics

HEALTH

2(a) Doctor None = 5 5-7 = 21-2 = 4 8 or more visits = 13-4 = 3

b) Disability or medical condition5 = No disability4 = Minor disability (e.g. eyeglasses) not likely to interfere with normal life

activities or routines3 = Constant, chronic condition that interferes to some extent with daily life (e.g.

diabetes, heart condition, Alzheimer's disease, migraines, infertility, asthmawhen nothing is recorded under medication, arthritis when nothing is recordedunder medication)

2 = Disability likely to restrict social activities (e.g. profound deafness, blindness,significant physical disability, depression, schizophrenia, arthritis, Parkinson’sDisease, paraplegia, asthma needing regular medication, arthritis needingregular medication, limb missing)

24

1 = Major disability likely to require daily assistance with personal care (e.g.severe psychiatric condition, advanced multiple sclerosis, severe cognitive orphysical impairment, quadriplegia)

NoteIt is sometimes difficult to choose between categories, eg. multiple sclerosis or Alzheimers in theearly stages would probably score 3, but in the latter stages score 2. Put them into these categoriesunless there is some information that tells otherwise. Eg. Assume that a person who has Alzheimers,but is able to answer the questionnaire scores 3, because once social activities become markedlyrestricted they would probably not be capable of completing the questionnaire. If a person has milddeafness, score 3, but if they are completely deaf, score 2.

c) MedicationNo regular medication = 5Single non-psychotropic medication = 4Multiple non-psychotropic medication = 3Psychotropic medication = 2Psychotropic plus non psychotropic medication = 1

NotePsychotropic medication indicates drugs for the control of epilepsy, psychoses, and other abnormalmental states. They include tranquilisers, sedatives, barbiturates and a host of others. Some of thesedrug names are provided in Appendix A.

PRODUCTIVITY

3a) Number of hours31-40+ work, education or child care = 521-30 hours combined work/education/child care = 411-20 hours combined work/education/child care = 31-10 hours combined work/education/child care = 2Neither work nor education nor child care = 1

b) Spare time (Note reverse score)Almost always = 1 Not usually = 4Usually = 2 Almost never = 5Sometimes = 3

c) Hours TV each dayNone = 5 6-9 hours = 21-2 hours = 4 10+ hours = 13-5 hours = 3

25

INTIMACY

4a) TalkDaily = 5 Once a month = 2Several = 4 Less than once a month = 1Once a week = 3

b) CareAlmost always = 5 Not usually = 2Usually = 4 Almost never = 1Sometimes = 3

c) ActivityAlmost always = 5 Not usually = 2Usually = 4 Almost never = 1Sometimes = 3

SAFETY

5a) SleepAlmost always = 5 Not usually = 2Usually = 4 Almost never = 1Sometimes = 3

b) SafeAlmost always = 5 Not usually = 2Usually = 4 Almost never = 1Sometimes = 3

c) Anxiety (Note reverse score)Almost always = 1 Not usually = 4Usually = 2 Almost never = 5Sometimes = 3

PLACE IN COMMUNITY

6a) Activity(i) For each separate activity calculate 0.2 + (0.2 x frequency) for each activity up

to a maximum frequency of 4/month. i.e. Each activity is scored to amaximum of 1.0.

26

(ii) Aggregate the total scores across all activities up to a maximum of 5 activities.Round all fractions to the nearest integer, i.e. the maximum score possible is 5

Additional Comments

(6) eat out “take aways” - exclude(7) movies “watched videos” - exclude(8) other people sometimes write something that should come

under one of the previous categories, [eg. tennis clubor yacht club should come under (i)] put them underthe category that seems most appropriate.

If rather than writing how many times in last month, people write:occasionally record “1”numerous “4”sometimes “1”seldom “9” (i.e. missing value)weekends “4”

b) ResponsibilityChairperson/ President/ Convenor e.g: captain of basketball team, convenor ofa social group = 5

Treasurer/ Secretary or other title denoting specific major area ofresponsibility eg: Immediate past-president, vice-captain = 4

Sub-committee chairperson or other indication of minor area of responsibilityor active involvement eg: Responsible for catering arrangements = 3

Committee or team member = 2

If they say they hold a position but do not state what the position is = 1

None = 1

c) AdviceAlmost every day = 5 Not often = 2Quite often = 4 Almost never = 1Sometimes = 3

27

EMOTIONAL WELL-BEING

7a) Can doAlmost always = 5 Not usually = 2Usually = 4 Almost never = 1Sometimes = 3

b) Bed (reversed scored)Almost always = 1 Not usually = 4Usually = 2 Almost never = 5Sometimes = 3

c) Wishes (reversed scored)Almost always = 1 Not usually = 4Usually = 2 Almost never = 5Sometimes = 3

28

4.2 Coding the subjective data

IMPORTANCE


Veryimportant

Somewhatimportant

Slightlyimportant

Not at allimportant

Missingvalue

SATIS

Deligh

NoteWe useneeds t

In ordefor eac

Delig

+4

Followoverallthis pr

4.3

7

5

FACTION

ted Pleased

the score of 9 to be taken that th

r to calculate ah domain, the s

hted Please

+3

ing this recodi SQOL = (ocedure the SQ

Data clean

It is recommenimplementatio

(a) An andeterm

(b) If a tascannethe imComQ

4

MostlySatisfied

M

o allow computer ese ‘9’ values are

IMPORT

meaningful subatisfaction data

d MostlySatisfied

+2

ng procedure eaIxS). See AppeOL obtained for

ing

ded that the rawn of analytic pro

alysis of responination that the cble of raw data d to detect patteportance or satisfol A5 Section

3

ixed MoDissat

identification of recognized as e

ANCE x SATI

jective QOL (Sneed to be re-co

Mixed

+1

ch SQOL domndix B for a many domain ran

data files be cacedures. In par

se frequency omputer is recis examined, thrns of responseaction scales o

7) which inclu

2

stlyisfied

Unhapp

missing values. Ixcluded values, an

SFACTION

QOL) score (Imded as follows:

Mostlydissatisfied

-2

ain score is calcore detailed desges between -20

refully examineticular;

data for each ognising '9' as a e data for each that are consistver the seven doded 794 adults

1

y Terrible M

f this scheme is usd not included as

portance x Satis

Unhappy Te

-3

ulated as (IxS), cription. As a r and +20.

d prior to the

variable will allmissing variable. subject can be ently at the top omains. In study , 1 percent of

9

issingvalue

6
5 4 3 2 1
e d

fa

r

ae

o

vfAsu

9

d, careata.

ction)

rible

-4

nd thesult of

w the

isually either6 (seebjects

29

recorded this type of response. Such data are excluded prior to analysis sincethey provide no variance and likely reflect a response set.

4.4 Dealing with the data skew

Both the importance, satisfaction, and I X S data are typically moderately negativelyskewed. To restore normality, the most appropriate transformation is reflect and squareroot.

Opinion is divided among statisticians as to whether this procedure is appropriate. Irecommend that the data not be transformed for the following reasons:

1. Authorities such as Tabachnick and Fidell (1996) recommend againsttransforming data which are known to be naturally skewed.

2. In my experience of checking the effects of transformation when applyingmultivariate statistics, the influence is very small.

A related issue concerns the increased intra-group variance that is created by forminga product of importance and satisfaction. We have attempted to reduce this by adding(I + S), both with and without transformation. While this procedure does achieve asomewhat reduced coefficient of variation (mean/standard deviation) it is notrecommended for two reasons as: (a) It does not seem to improve the data sensitivityto between-group differences, and (b) the power of importance to weight thesatisfaction scores is reduced.

4.5 Forms of data analysis

Note: For a step-by-step scoring procedure see Appendix B.

4.5.1 For the practitioner or service provider

The most useful level of analysis is in terms of domain scores. For theobjective QOL this involves a sum of the three component scores for eachdomain. Reference data are available in section 7.1

For subjective QOL the domain scores are obtained by (Importance xSatisfaction) following the recoding of Satisfaction as described in 4.2.Reference data are provided in Section 7.2 and the construction of individualdiagnostic tables are described in 4.5.

30

4.5.2 For the Researcher

The most useful level of analysis may be at the level of the domain for all threeaxes; objective, importance and satisfaction. It has been found that importanceand satisfaction are often fairly independent and their separate variation islikely to be of interest to any investigation of the QOL construct (see 7.6.2).

Note: Group I x S statistics must be based on individually calculated I x Sscores.

4.5.3 % SM: A standardised comparison statistic

In some circumstances it may be useful to compare the relative extent ofimportance and satisfaction which has been expressed in relation to a domain.This cannot be made directly since importance has been scored on a 5-pointscale and satisfaction on a 7-point scale. The comparison can be achieved byconverting each to a statistic which reflects the extent to which a scoreapproximates the maximum score which could be obtained. The formula is asfollows:

% of scale maximum = (Score -1) x 100/(number of scale points -1)

EXAMPLES

Importance coded +1 to +5For example, with an importance score of 4.0 and a 5-point scale% scale max = (4-1) x 100/(5-1)

= 75%

Satisfaction coded +1 to +7For a satisfaction score of 5.2 and a 7-point scale% scale max = (5.2-1) x 100/(7-1)

= 70%

Satisfaction coded -4 to +4 (see 4.2)The calculation here requires a modified formula as:

(a) POSITIVE S scores use:

% sm = [ (score -1) +3] x 100/6

(b) NEGATIVE S scores use:

% sm = [ (score +1) + 3] x 100/6

31

Importance x Satisfaction scores

For an I x S score to be expressed in this way:

a) The S score must be recoded on a +4 to -4 scale (see 4.2)

b) Any I x S score of -1 to +1 is converted to +1 (note this is the mid-point of the recoded satisfaction scale).

c) POSITIVE I x S scores use the formula:% scale maximum = [(score -1) + 19] x 100/38

d) NEGATIVE I x S scores use the formula:% scale maximum = [(score +1) + 19] x 100/38.

An interpretation of this statistic can be made using the Cummins (1995a)paper which brought together previously published studies on overall lifesatisfaction. It reported an average 75.0 ± 2.5 % sm. Section 7.2 reports lifesatisfaction data using ComQol.

4.6 Individual SQOL diagnostic data tables

The following table is an example of how ComQol can be used as a diagnosticinstrument for the individual. Each importance and satisfaction rating has been scoredaccording to Section 4.2. The I x S score is then a standardised measure of domainquality for each client with a range +20 to -20.

32

Table 1 Example of a client diagnostic table

Material well-being

ClientImportance (I)

(Coded +1 to +5)Satisfaction (S)(Coded -4 to +4) I x S

1 2 4 82 3 1 33 5 4 204 1 3 35 5 -4 -206 5 4 207 5 1 58 3 -3 -9 9 3 2 610 5 -4 -20

Mean 3.70 0.80 1.6

An interpretation of the above data can be made in relation to the following studies (see 7.in ComQol A5 for study descriptions). In summary form the I x S results were as follows:

DomainIMPORTANCE X SATISFACTION

(Coded +1 to (Coded -4 to +4)Mean + S.D.

% of Negative I x S Scores

Study A1 + A2 A6 A1 + A2 A6Material 8.04 + 4.23 8.44 + 4.65 2.6 2.1Health 7.62 + 6.31 8.78 + 6.92 8.8 7.4Productivity 7.97 + 5.66 7.45 + 5.50 2.6 5.6Intimacy 9.85 + 7.35 11.51 + 6.34 4.9 3.1Safety 9.81 + 4.04 10.00 + 5.04 2.0 1.9Community 4.90 + 4.29 7.14 + 5.09 3.9 4.8Emotion 8.36 + 6.82 8.98 + 6.09 7.8 5.5

From the above table it can be seen that the reference group had a strong tendency to scoreabove the scale mean of zero. Consequently, the presence of a negative IxS score isindicative of a poor quality domain for that person. It can be seen that fewer than 9% of thesamples obtained a negative IxS on any domain.

33

5 Theoretical issues

5.1 Why use the ‘Delighted-Terrible’ scale?

A major problem with QOL data is their tendency to cluster at the favorable end ofany scale. Andrews & Withey (1976) have reported that the D-T scale creates a morepronounced spread of upper-end results than does the more conventional scale of'Extremely satisfied' to 'Extremely dissatisfied'.

5.2 “Should ‘not important at all’ be scored as 1 or 0”

The importance score is used as a weighting for satisfaction (I x S) as indicatedbelow. Consequently if 'not important at all' was scored as '0' then the product withany S score would also be zero.

In logical terms this could be reasonable; if a domain really does have no importancethen the level of satisfaction is irrelevant and the I x S product should always be zero.However, people who respond 'no importance at all' do still respond to the satisfactionscale. This indicates that levels of satisfaction may be experienced even though therespondent regards the domain as having no importance.

There are several possible reasons for this findings as:

1. Importance and Satisfaction are largely independent constructs. Certainly theycan be independently experienced. Thus, even if one's material things are 'notimportant at all', this will not prevent the experience of being satisfied ordissatisfied in this domain.

2. It might be argued that no domain could actually have a zero importance sinceall domains form a part of each person's continuous life experience. The realmeaning, then of 'not important at all' is of very low importance. This notionwould be consistent with the very gross nature of the scales, where only fivepoints span the continuum of 'importance'.

3. If this argument is accepted, then, for the purposes of using importance as aweighting factor, it would be preferable to score 'no importance at all' in aneutral way rather than in a canceling way. Thus, a score of '1' is preferable toa score of '0'. This allows the lowest rating of importance to have noweighting influence on the measure of satisfaction. It also has the advantageof allowing a greater range of I x S scores in the lowest range.

34

5.3 “Should ‘mixed satisfaction/dissatisfaction’ be scored as ‘O’?”

If this was adopted, then the scale on either side could be scored (+1, -1), (+2, -2),(+3, -3). This system would create a more reasonable interval scale around the mid-point i.e. +1, 0,-1. However, it would have the disadvantage of creating a zero I x Scombination whenever a 'mixed' level of satisfaction was recorded.

The meaning of a zero I x S score in this context is unclear. It would mean thenegation of any assigned score of importance and, thereby, the loss of discriminativedata. It also loses the meaning of the data; a response of ‘mixed’ satisfaction does notimply zero satisfaction. Rather it implies a neutral level of satisfaction/dissatisfactionfor which a score of ‘1’ would be more appropriate. In this way the importance‘weighting’ score would remain uninfluenced by the rating on satisfaction.

One problem with this approach may be seen in terms of the mathematical logic of theresultant interval scale around the point of neutrality i.e. +2, +1, -2. It is clear that thiscannot be a true interval scale. However, Likert scales cannot conform to the strictlogic of interval scales; they are contrived approximations which are dependent on theassumption of an equal degree of psychological separation between the responsepoints. In this case the point of neutrality (scale mid-point) is conceived asencompassing the range (+1-1), and the response point on either side depicts theaddition of one additional unit. See also the scoring of I x S (section 4.4.3, IxS, b) inwhich scores of -1 and +1 are treated as being equivalent and if substituted informulae c and d, yield 50% scale maximum.

5.4 “Why not score the satisfaction scale from +1 (Terrible) to +7(Delighted)?”

If this system was to be adopted then the I x S score interpretation would be ambiguous. Forexample, a score of +4 could be the combination of either low I and high S, or high I andlow S (dissatisfaction). By constructing the scale as it is, these ambiguities have beenreduced, but not entirely eliminated. For example, a score of +4 could be generated by either'Not important at all' x "Pleased" (1 x 4), or by 'A little bit important' x 'Somewhat happy' ( 2x 2). While these could be distinguished by examining the raw data, for most purposes ofthe scale this would not be necessary. More importantly, the distinction between responsesatisfaction and dissatisfaction is made unequivocal by the adopted scoring system of +4 to -4.

35

The possible combinations of I and S scores are as follows:

I x S Possible Combinations I x S Possible CombinationsScore Positive Negative Score Positive Negative20 5 x 4 5 x -4 10 5 x 2 5 x -219 - - 9 3 x 3 3 x -318 - - 8 4 x 2, 2 x 4 4 x -2, 2 x -417 - - 7 - -16 4 x 4 4 x -4 6 3 x 2, 2 x 3 3 x -2, 2 x -315 5 x 3 5 x -3 5 5 x 1 -14 - - 4 4 x 1, 1 x 4, 2 x 2 1 x -4, 2 x -213 - - 3 3 x 1, 1 x 3 1 x -312 4 x 3, 3 x 4 4 x -3, 3 x -4 2 2 x 1, 1 x 2 1 x -211 - - 1 1 x 1 -

Note(a) Possible importance scores are +1, +2, +3, +4, +5(b) Possible satisfaction scores are +4, +3, +2, +1,-2, -3, -4(c) The above table is essentially symmetrical between the positive and negative combinations with the

exception of I x S scores of 4, 3, 2 and 1 which demonstrate a reduced number of negativecombinations.

5.5 “Why not score the satisfaction scale according to the procedureof Ferrans and Powers (1985)?”

These authors use an alternative system to ComQol, but it seems to have noadvantages.Their procedure is as follows:

1. Importance is scored +1 to +6.

2. Satisfaction is scored +1 to +6.

3. The Satisfaction scores are recoded by subtracting 3.5 from each one.

ie. Original: +1 +2 +3 +4 +5 +6Recode: -2.5 -1.5 -.5 +.5 +1.5 +2.5

4. Importance is then multiplied by recoded S and 15 added to the product. Theyclaim: "This adjustment (recode) produces the highest score for items thathave high satisfaction/high importance responses, and the lowest score forhigh dissatisfaction/high importance responses. --- If scores were not recoded,a person who was very dissatisfied with an area of high importance would

36

receive the same item score as a person who was very satisfied with an area oflow importance." (p. 18)

In fact, however, their recoding procedure does not eliminate this problem.The possible I X S scores (recoded -2.5 to 2.5) obtained through the use of theFerrans & Powers formula are as follows:

I x S I x SScore Possible combinations Score Possible combinations

30 6 x 2.5 14.5 1 x -.5 27.5 5 x 2.5 14 2 x -.5 25 4 x 2.5 13.5 1 x -1.5, 3 x -.5 24 6 x 1.5 13 4 x -.5 22.5 3 x 2.5, 5 x 1.5 12.5 1 x -2.5, 5 x -.5 21 4 x 1.5 12 2 x -1.5, 6 x -.5 20 2 x 2.5 10.5 3 x -1.5 19.5 3 x 1.5 10 2 x -2.518 2 x 1.5, 6 x .5 9 4 x -1.517.5 1 x 2.5, 5 x .5 7.5 3 x -2.5, 5 x -1.5 17 4 x .5 6 6 x -1.5 16.5 1 x 1.5, 3 x .5 5 4 x -2.5 16 2 x .5 2.5 5 x -2.5 15.5 1 x .5 0 6 x -2.5

A comparison of the above distribution with that previously provided for ComQolyields the following observations.

1. The form of each distribution is roughly equivalent, with combinationsbunching around the mid-point.

2. The F & P distribution is symmetrical around the mid-range score of 15, whilethe ComQol distribution shows a reduced number of negative combinationsdue to the absence of -1 as a recoded satisfaction score. Thus, the I x S scorecombinations lying just below the mid-range are less ambiguous in the case ofComQol.

3. The area of maximum I x S score ambiguity, in terms of their composition, isgreatest in both distributions just above the mid-range. For example, over arange of three integer units in the F & P distributions, from scores of 12 to 14,eight I x S combinations are represented. The ComQol distribution is

37

marginally less ambiguous with seven I x S combinations over an equivalentscore-range of 2 to 4.

4. Both distributions produce a few I x S combinations which are veryambiguous indeed. For example, an I x S score of 17.5 in the F & Pdistribution could be the combination of either 'lowest I x highest S' or'second-highest I x mid-range S'. An equivalent degree of confusion isprovided by the ComQol I x S score of 4.

ConclusionThe Ferrans and Powers formula is not superior to the simpler ComQol recoding procedure.

5.6 Why not use the Raphael et al. (1996) scoring system?

These authors have devised the 54-item Quality of Life Profile which also usessatisfaction (scored 1-5) weighted by importance (scored 1-5). Their formula is:

QOL = (Importance score/3) x (satisfaction score -3)

Thus, the possible scores are:

5 4 3 2 1

Importance: 1.67, 1.33, 1, 0.67, 0.33

Satisfaction: 2, 1, 0, -1, -2

The following observations can be made:

1 The differential weighting of adjacent items is reduced from ‘1’ in ComQol to0.33. The relative weighting by importance is thus reduced.

2 The possible range of values is +3.33 to -3.33. This seems an awkward rangeto deal with.

3 A satisfaction score of 3 leads to a QOL = 0 regardless of the importancescore. The authors state “This is consistent with our conceptual thinking thatmoderate enjoyment of an aspect of life should result in a ‘neutral’ QOL score,whether it is important or not.” (p.369). See 5.3 for comment.

ConclusionThis dual transformation of both importance and satisfaction data has no advantagesover the ComQol procedure.

38

6 Alternative forms of the scale

6.1 Parallel versions of the scale

ComQol has been designed to have three parallel versions. In addition to ComQol-S5there are:

ComQol-A5. This has been designed for the general adult population. It is availableas a separate manual.

ComQol-I5. This has been designed for people who have an intellectual disability orother form of cognitive impairment. It is available as a separate manual.

6.2 Additional domains

The ComQol scale can be modified through the addition of other domains. To datethree modifications of this type have been considered using ComQol-A and these willbe described. However, a major consideration in this regard is the amount ofadditional unique variance accounted for by the addition of new domains.

Using an internal, stepwise multiple regression procedure where each (IxS) domain isregressed against the total subjective (IxS) score it has been determined that, using theseven standard domains, around 80% of the variance is shared and each domaincontributes only 1 to 3% of unique variance (Cummins, in preparation). Thus, theseven domains are probably adequate to measure overall subjective QOL. Whileother domains are able to contribute unique variance, their addition would be for thepurpose of investigating the specific domain in question rather than contributingvariance to the total subjective QOL score.

The additional domains that have been considered so far are as follows:

1. USEFULNESS: In some ways this is a better term for Productivity. Itcertainly includes all of the terms listed under Productivity. However, it alsoseverely overlaps with Place in Community, in that most, if not all activitiesinvolving others may be considered useful. It is concluded that it combineselements of both productivity and place in community.

2. LEISURE: This is a slippery concept. Unlike the other domains wheregenerally ‘more is better’, this does not apply to leisure. And as soon as somequalifier is introduced, such as ‘quality leisure’, it immediately overlaps withother domains (e.g. productivity). It is concluded that leisure is subsumedwithin emotional well-being.

39

3. SPIRITUAL WELL-BEING: Four studies have experimented with theinclusion of this domain (A7, A9, A10, A14). In each case only the subjectiveaxis has been explored using “How important to you are your religious orspiritual beliefs?”, and, “How satisfied are you with your religion orspirituality?”. The main issues and findings to emerge are as follows:

(a) About one-third of Australian people have neither religion orspirituality. Thus, the ‘satisfaction’ question must be preceded by astatement which gives respondents a choice of answering the item ornot. As a consequence of this, the resultant data set for satisfactioncomprises a mixture of 7- and 8- item responses. This introduces acomplication into the subsequent analysis.

(b) As with the other domains (see above), the domain of spiritual well-being adds only a small amount of unique variance when added to theusual seven domains. The data obtained are as follows:

Study Sample Importance (% sm) Satisfaction (% sm) I X S (%sm) Unique1

Variance

A9 FarmerEx farmerMetro

2.8 + 1.33.4 + 1.42.7 + 1.4

59.371.362.0

5.2 + 1.15.4 + 1.05.2 + 0.9

71.972.870.4

7.6 + 5.8 9.6 + 5.4 8.0 + 5.0

69.273.169.3

TOTAL 4%

A7 YoungMid-ageOlder

3.2 + 1.43.7 + 1.33.8 + 1.3

55.867.369.0

5.2 + 1.35.6 + 1.15.8 + 1.0

70.077.379.5

8.8 + 7.111.1 + 6.511.6 + 5.8

70.676.677.8

TOTAL 4%

1 These estimates of unique variance have been calculated on combined sample data using an 'internal'form of multiple regression, where each domain is regressed against the total [ (IxS)] score. Thesquare of the semi-partial correlation for each domain yields its unique variance.

ConclusionThe seven original domains are sufficient to measure subjective QOL for most purposes.Spiritual well-being may be usefully added as an eighth domain if the population underinvestigation is highly spiritual/religious or if this particular aspect of the QOL construct is tobe examined.

40

4. FAMILY AND FRIENDS: The fourth domain is normally stated as acombined source of intimacy involving ‘family and friends’. This isrecommended for normal use. However, under some circumstances it may bedesirable to obtain separate ratings for each component. Our data on thisseparation are as follows:

Study Sample Importance % sm Satisfaction (% sm) I X S (%sm)

A10 Youth

Elderly

(Family)(Friends)(Family)(Friends)

4.1 + 0.84.2 + 0.64.3 + 0.73.7 + 0.7

77.579.782.068.5

5.5 + 1.25.7 + 0.85.9 + 1.15.6 + 0.8

75.678.381.376.8

10.9 + 6.811.4 + 4.212.7 + 6.010.1 + 4.1

76.077.480.673.7

A11 Using an internal multiple regression, the amount of unique variance contributed by eachdomain was: Family = 2.9%, Friends = 1.5%.

NoteThe intimacy domain is normally rated as higher than the other domains in terms of both importance andsatisfaction. Consequently, the use of two 'intimacy' domains as family and friends will bias the aggregate [ (I xS)] such that it may be higher than the normative value of 75 ± 2.5 %SM (Cummins, 1996b). It is thereforerecommended that, when two separate domains of family and friends are employed, their average combinedscore is used in combination with the other six domains when calculating SQOL.

41

7 Psychometric data

Study Codes

The data to be reported have been drawn from the following studies:

NoteAdditional psychometric data on adults is available from the ComQol-A5 manual.

Study S1: (Gullone & Cummins, 1998, Third edition). 264 adolescents aged 14.6 years(range 12 to 18 years) completed ComQol-S3. The sample was drawn from 7schools in metropolitan Melbourne and comprised 52.9% females.

Study S2: (Bearsley, 1997, Fourth edition). 524 adolescents aged 15.8 years (range 14-17y), 57.2% female comprised three groups as: (a) homeless and ‘at risk’ ofhomelessness (N=105), ‘community school’ students with a high frequency ofemotional, behavioral, learning, or family problems (N=82), and non-homelesssecondary school students (N=337).

7.1 Objective means

Study S1 % sm

MaterialHealthProductivityIntimacySafetyCommunityEmotion

8.43 + 1.01 8.42 + 1.02 6.02 + 1.1612.09 + 2.2911.73 + 2.24 6.16 + 1.19 8.80 + 2.18

45.3 ± 5.445.2 ± 5.525.2 ± 4.975.8 ± 14.472.8 ± 13.926.3 ± 5.148.3 ± 12.0

MEAN 9.34 + 0.68 52.8 ± 3.8

42

7.2 Subjective means

7.2.1 Importance (coded +1 to +5)

Study S1 Study S2Raw % sm % sm


3.65 + 0.783.91 + 0.823.61 + 0.844.16 + 0.803.80 + 0.843.33 + 0.984.06 + 0.85

66.3 ± 14.272.8 ± 15.365.3 ± 15.279.0 ± 15.270.0 ± 15.558.3 ± 17.276.5 ± 16.0

MaterialHealthProductivityFamilyFriendsSafetyCommunityEmotion

66.9± 18.377.1 ± 17.378.4± 16.572.7 ± 22.772.4± 19.470.7± 21.854.4 ± 24.678.7 ± 18.1

MEAN 3.79 + 0.54 69.8 ± 9.9 MEAN 71.2 ± 12.1

7.2.2 Satisfaction (coded -4 to +4)

Study S1 Study S2Raw % sm % sm


2.33 ± 0.751.89 ± 0.892.16 ± 0.762.59 ± 0.852.29 ± 0.742.22 ± 0.842.18 ± 0.89

72.2 ± 10.264.8 ± 11.869.3 ± 10.276.5 ± 11.671.5 ± 10.070.3 ± 11.369.7 ± 12.0


78.9 ± 13.971.9 ± 18.872.2 ± 17.473.4 ± 19.377.4 ± 16.273.8 ± 14.570.5 ± 16.172.2 ± 18.2

MEAN 2.25 ± 0.55 70.8 ± 7.4 MEAN 73.6 ± 10.9

7.2.3 Importance x Satisfaction

Study 1 Study 2Raw % sm % sm


9.34 ± 6.29 7.74 ± 8.89 9.17 ± 6.2411.67 ± 7.16 8.89 ± 6.69 8.06 ± 6.3510.12 ± 7.48

71.9 ± 16.567.7 ± 23.471.5 ± 16.478.1 ± 18.870.8 ± 17.668.6 ± 16.774.0 ± 19.7


74.4 ± 10.971.6 ± 16.872.7 ± 15.572.9 ± 15.575.0 ± 13.572.2 ± 10.466.3 ± 11.872.3 ± 15.6

MEAN 9.59 ± 4.79 72.6 ± 12.6 MEAN 71.9 ± 9.1

43

7.3 Reliability

7.3.1 Cronbach’s alpha

Boyle (1991) and Cortina (1993) strongly condemn the ‘classical’ psychometric belief thathigh alphas are better in terms of intra-scale reliability. Boyle quotes Hattie (1985) as “alphacan be high even if there is no general factor, since (1) it is influenced by the number of itemsand parallel repetitions of items, (2) it increases as the number of factors pertaining to eachitem increases, and (3) it decreases moderately as the item communalities increase.” (pp. 157-8). He concludes that there is an optimum range of internal consistency/item homogeneity ifsignificant item redundancy is to be avoided. According to Kline (1979, p. 3), with itemintercorrelations lower than about 0.3 “each part of the test must be measuring somethingdifferent... A higher correlation than (0.7), on the other hand, suggests that the test is toonarrow and too specific... If one constructs items that are virtually paraphrases of each other,the results would be high internal consistency and very low validity.” Kline also states“maximum validity... is obtained where test items do not all correlate with each other, butwhere each correlates positively with the criterion. Such a test would have only low internal-consistency reliability.” (p.3)

For the purpose of evaluating ComQol, sub-scale alphas will be sought in the range 0.3 to 0.7.

7.3.2 Internal Reliabilities (Cronbach alpha)

SUBJECTIVE Study S1Importance .76Satisfaction .80Importance x Satisfaction .81

44

OBJECTIVE

Alpha Domain vs totalobjective sub-scale correlations

Intra-domaincorrelations

Item vs totaldomain scorecorrelations

vsItem (1)

vsItem (2)

A. MATERIAL .09 .43***a(1) Income .91***a(2) Number of possessions .02 .36***a(3) Standard of accommod. .09 .07 .27***B. HEALTH .47 .18**b(1) Number visits to doctor .73***b(2) Extent of disability .23*** .66***b(3) Severity of medication .17*** .43** .61***C. PRODUCTIVITY .05 .33***c(1) Hours per week .48***c(2) Time on desired goals .03 .38***c(3) Things made, etc. .02 .13** .79***D. INTIMACY .62 .32***d(1) Freq. talk to friends .58***d(2) Freq. joined in activities .23*** .67***d(3) Freq. others care when

depressed.09 .36*** .68***

E. SAFETY .52 .29***e(1) Freq. easily fall asleep .42***e(2) Freq. anxious duríg day .19*** .77***e(3) Freq. feel safe at home .06 .07 .13**F. COMMUNITY .30 .41***f(1) Extent comm. activities .23***f(2) Extent comm. resp. .14* .76***f(3) Extent valued by comm. .04 .24*** .77***G. EMOTION .17 .34***g(1) Freq. choose activities .38***g(2) Freq. impossible wishes .04 .73***g(3) Freq. wish to stay in bed .04 .25*** .67***

NoteAlpha for the objective sub-scale comprising the seven domain scores is .39. Alpha for the sub-scale comprising the21 individual objective items is 0.47* p < 0.05 ** p < 0.01 *** p < 0.001

45

7.3.3 Test-retest Reliability

Study SI: One weekObjective Imp Sat I x S


.85

.88

.60

.76

.73

.66

.56

.51

.62

.40

.56

.66

.50

.58

.59

.40

.60

.54

.48

.45

.70

.64

.40

.63

.68

.60

.49

.74

TOTAL .83 .74 .73 .79

7.4 Validity

7.4.1 Content validity

The major validity data have been published as follows:

Cummins (1995a) combined the data from 16 studies that measured ‘life satisfaction’among large samples drawn from the general population. It was found that theircombination could be described by 75 + 2.5%SM. It was proposed that this statistic,and the implied normative range of 70 - 80%SM could be considered as the goldstandard for Western populations. This statistic has subsequently been elaborated toinclude non-Western populations (Cummins, 1998).

Cummins (1996b) demonstrated that, of 173 different domain-satisfaction namesdrawn from the literature, 68% of the names and 83% of the data they representedcould be grouped under the 7 ComQol domains. Moreover, when the data groupedunder the 7 domains were combined they yielded an average of 73.6 + 3.0%SM. Fromthis it is concluded that content validity has been established and also that the dataderived from the 7 domain satisfaction scores can be compared against the normativestandard of75 + 2.5%SM.

7.4.2 Construct validity

Study SI: Used the Revised Children’s Manifest Anxiety Scale (RCMAS: Reynolds& Richmond, 1985), and the Fear Survey Schedule for Children - II(FSSC-II; Gullone & King, 1992). The correlations with ComQol domainswere:

46

Objective :Anxiety - inverse correlations (-.15 to -.47) with safety, emotion, health,and intimacy.

:Fear - inverse correlations (-.13 to -.28) with safety and emotion.

Imp x Sat :Anxiety - inverse correlations (-.14 to -.32) with material, health, safety,community, emotion, and total score.

:Fear - N.S.

Study S2: Used the Life Attitude Profile - Revised (Reker, 1992) to demonstrate thefollowing relationships with SQOL across the whole sample: Personalmeaning (.61), existential vacuum (-.48), choice and responsibleness (.45),goal seeking (.16), death acceptance (-.01).

7.5 Sensitivity

Study S2: Subjective QOL was compared across the three groups, F (2,280) =10.689, p <0.001. The non-homeless adolescents had a higher mean score (see 7.2.3)than both the homeless and ‘at-risk’ adolescents (60.3 ± 17.1%SM) and the non-homeless community school adolescents (67.2 ± 9.5%SM).

47

8 References to the text

(for references to Cummins, see Appendix C)

Andrews, F.M., & Withey S.B. (1976). Social indicators of well-being: Americans'perceptions of life quality. New York: Plenum Press.

Bearsley, C. (1997). No place called home: Quality of life and meaning in life of homelessyouths. Melbourne: Honours Thesis, School of Psychology, Deakin University.

Boyle, G.J. (1991). Does item homogeneity indicate internal consistency or item redundancyin psychometric scales? Personality and Individual Differences, 12, 291-294.

Castles, I. (1992). 1990 Survey of income and housing costs and amenities Australia:Persons with earned income. Catalogue #6546.0. Canberra: Australian Bureau ofStatistics.

Cortina, J.M. (1993). What is coefficient alpha? An examination of theory and applications.Journal of Applied Psychology, 78, 98 - 104.

Felce, D., & Perry, J. (1995). Quality of life: Its definition and measurement. Research inDevelopmental Disabilities, 16, 51-74.

Ferrans, C.E., & Powers, M.J. (1985). Quality of life index: Development and psychometricproperties. Advances in Nursing Science, 8, 15-24.

Gullone, E., & Cummins, R.A. (1998). The Comprehensive Quality of Life Scale - Schoolversion: A psychometric investigation (submitted).

Gullone, E., & King, N.J. (1992). Psychometric evaluation of a revised fear survey schedulefor children and adolescents. Journal of Child Psychology and Psychiatry, 33, 987-998.

Hattie, J. (1985). Methodology review: assessing unidimensionality of test and items.Applied Psychological Measurement, 9, 139-164.

Kline, P. (1979). Psychometrics and psychology. London: Academic Press.Raphael, D., Rukholm, E., Brown, I., Hill-Bailey, P., & Donato, E. (1996). The quality of

life profile - Adolescent version: Background, description, and initial validation.Journal of Adolescent Health, 19, 366-375.

Recker, G.T. (1992). Life Attitude Profile-Revised. Canada: Trent UniversityReynolds, C.R., & Richmond, B.O. (1985). Revised Children’s Manifest Anxiety Scale

(RCMAS) manual. Los Angeles, CA: Western Psychological Services.

48

Appendix A

Psychotropic drug names(Both generic and trade)

AcetophenazineAdapinAldazineAlprazolamAmitriptylineAnetensolAtaraxAtenoiolAtivanAventyulBlocadrenBusparBuspironeCalmazineCarbamazepineCatapresCelontinCentraxChlordiazepoxideChlorpromazineChlorprothixeneCibalith-sClonazepamClonidineClorazepateCompazineCylertDecanoateDepakeneDesipramineDesyrelDexedrineDextroampetharnineDiazepamDilantinDoxepinDroleptanElavilEndepEquanilEskalithEthosuximideEthotoinFluphenazineHalazepamHaldolHaloperidolHydroxyzineImavateInderal

IsocarboxazidJanamineKlonopinLarquactilLibriumLimbitrolLithaneLithicarbLithobidLithonateLopressorLorazepamLoxapineLoxitaneLudiomilMaprotilineMarplanMebaralMellerilMephenytoinMephobarbitalMeprobamateMesantoinMesoridazineMethsuximideMetroprololMilontinMiltownMobanModecateMolindoneMutabonMysolineNardilNavaneNeulacctilNorpramineNortriptylineNovaneOrapOxazepamPamelorParadioneParamethadioneParnatePaxiparnPeganonePermolinePerphenazinePerrnitil

PertrofranePhenelzinePhenobarbitalPhensuximidePhenytoinPiperacetazinePrazepamPriadelPrimidoneProchloperazineProlixinPropanalolQuideRitalinSeraxSerenaceSerentilSinequanSK-PramineStelazineSurmontilTaractanTegretolTenorminThioridazineThiothixeneThorazineTimololTindalTofranilTranxeneTranylcypromineTrazodoneTriavilTridioneTrifluoperazineTriflupromazineTrilafonTrimethadioneTrimipramineValiumValproic AcidVesprinVistarilVivactilZanaxZarontin

49

Appendix B

Scoring ComQol

1 Recode satisfaction data

+4 +3 +2 +1 -2 -3 -4Delighted Pleased Mostly

satisfiedMixed Mostly

dissatisfiedUnhappy Terrible

2 Obtain individual domain objective scores for each person(a) Following the coding procedure, code items la to 7c.(b) Add the three sub-domain scores (e.g. 1a, 1b, 1c) - This is the TOTAL DOMAIN

SCORE FOR MATERIAL WELL-BEING.(c) Divide the total by 3. Call this score x.(d) Take the score x, and plug into the formula

% scale max = (score x-1) x 100/(5-1)(e) This gives you the objective score for material well-being expressed as %SM.(f) Repeat for the other domains.

3 Obtain overall objective score for each person(a) Sum the scores from all 21 items - this is TOTAL SCORE.(b) Divide total score by 21. Call this Score x.(c) Take Score x and plug into the formula

% scale max = (Score x-1) x 100/(5-1)(d) This is the overall objective score expressed as %SM.

4 Obtain domain importance scores for each person (without satisfaction)(a) Take importance score for each domain. Call this score x.(b) Use formula % SM = (score x-1) x 100/(5-1).

5 Obtain overall importance score for each person (without satisfaction)(a) Sum the 7 importance scores.(b) Divide total by 7. Call this score x.(c) %SM = (mean score x-1) x 100/(5-1).

6 Obtain domain satisfaction scores for each person (without importance)(a) Use non-recoded data: i.e. Use scores coded 1 (Terrible) to 7 (Delighted).(b) Take the satisfaction score for each domain. Call this score x.(c) Use the formula %SM = (score x-1) x 100/(7-1).

50

7 Obtain overall satisfaction score for each person (without importance)(a) Sum the 7 satisfaction scores (scored 1 to 7).(b) Divide total by 7. Call this score x.(c) Use formula in (5).

8 Obtain an overall (I x S) score for each person(a) Sum the (IxS) domain scores for each person.(b) Divide by 7. Call this score x.(c) If result is positive use formula

% scale maximum = [19 + (Score x-1)] x 100/38(d) If result is negative use formula

% scale maximum = [19 + (Score x+1)] x 100/38(e) This figure is the overall subjective domain score expressed as %SM.

9 Obtain average subjective scores (i x s) for each domain using grouped data(a) Recode the satisfaction score for each person as for (1.).(b) Calculate (IxS) for each domain for each subject.(c) Obtain an average (IxS) score for each domain. This is score x.(d) If result is positive use the following formula

% scale maximum = [(Score x-1) + 19] x 100/38(e) If result is negative use formula

% scale maximum = [(Score x +1) + 19] x 100/38(f) This gives the average domain subjective score across the group for each domain

expressed as %SM.

EXAMPLE

1 Overall objective score

Sum 1a to 7c =5353/21 = 2.52% scale max = (2.52-1) x 100/(5-1)

= 38%

51

2 Overall I x S score

Sum I x S scores (data for IDNUMOO1)

Domain Importance Satisfaction I x S1 3 +1 32 4 -3 -123 4 +2 84 5 +4 205 4 +3 126 4 +3 127 4 -2 8

TOTAL 28 8 35

Divide 35 by 7 = 5

Result is positive so -

% scale maximum = [(5-1) + 19] x 100/38= 60.5

3 Individual importance scores for each domain

e.g. from domain 1 above

% scale maximum = (3-1) x 100(5-1) = 50

from domain 2 above % scale maximum = (4-1) x 100(5-1) = 75

52

Appendix cAuthor publications (from 1991)

(as at 11/8/98)

Parallel versions of the scale

* Cummins, R. A. (1997). Comprehensive Quality of Life Scale - Intellectual Disability:ComQol-15. (Fifth Edition). Melbourne: School of Psychology, Deakin University.This is the version of the scale to be used with people who have intellectual disabilities or acognitive impairment.

* Cummins, R. A. (1997). Comprehensive Quality of Life Scale - Student (Grades 7-12):ComQol-S5. (Fifth Edition). Melbourne: School of Psychology, Deakin University. This isthe version of the scale to be self administered by school students in Grades 7 to 12.

Available translations

* The adult version is available in Greek, Italian, Spanish and Persian from the author.

Publications by the author over the past five years:

Baxter, C., & Cummis, R.A. (1993). Extending network support through shared family care: Impact onchildren with a disability and their families. Proceedings, Fourth Australian Family ResearchConference, Electronic Database, Australian Institute of Family Studies, Document #18951.

Cummins, R.A. (1993). Health promotion and the Comprehensive Quality of Life Scale. HealthPromotion Journal of Australia, 3, 46-47.

Cummins, R.A. (1993). In the Community: An Evaluation of the Community Living Support Service.Melbourne: Community Services Victoria, (pp.1-306).

Cummins, R.A. (1993). On being returned to the community: Imposed ideology vs. quality of life.Australian Disability Review, 2-93, 64-72.

Cummins, R.A. (1993). Tranquilisers and your brain. In: E. Ree (Ed.). A training manual for healthprofessionals in the treatment of benzodiazepine dependence and withdrawal. Melbourne:TRANX, 75-80.

Cummins, R.A., & Baxter, C. (1993). A case for the inclusion of subjective quality of life data inservice-delivery evaluations. Proceedings, Australasian Evaluation Society InternationalConference. Brisbane: Australian Evaluation Society, 201-211.

McCabe, M.P., Cummins, R.A., Hinchy, J., Whitmore,M., & Gee, D. (1993). Changing Attitudestowards Self, Family and Boys Held by Young Adolescent Females. Melbourne:Psychology Research Center, Deakin University, (pp.1-82).

Polak, S., Cummins, R.A., & Baxter, C. (1993). Attitudes of families to the Community Living SupportService. Interaction, 6(1), 25-26.

Cummins, R.A., & Baxter, C. (1994). Choice of outcome measures in service delivery evaluations forpeople with disabilities, Evaluation Journal of Australia, 6, 22-30.

Cummins, R.A., Jauernig, R., Hudson, A., & Baxter, C. (1994). A model system for the construction andevaluation of General Service Plans. Australia and New Zealand Journal of DevelopmentalDisabilities, 19, 221-231.

Cummins, R.A., McCabe, M.P., & Romeo, Y. (1994). The Comprehensive Quality of Life Scale -Intellectual disability: Results from a Victorian survey. Proceedings, 29th National Conferenceof the Australian Society for the Study of Intellectual Disability, 93-98.

53

Cummins, R.A., McCabe, M.P., Romeo, Y., & Gullone, E. (1994). The Comprehensive Quality of LifeScale: Instrument development and psychometric evaluation on College staff and students.Educational and Psychological Measurement, 54, 372-832.

McCabe, M.P., & Cummins, R.A. (1994). Sexual abuse among people with intellectual disabilities: Factor fiction. Proceedings, 29th National Conference of the Australian Society for the Study ofIntellectual Disability, 250-254.

McCabe, M.P., Cummins, R.A., & Reid, S.B. (1994). An empirical study of the sexual abuse of peoplewith intellectual disabilty. Journal of Sexuality and Disability, 12, 297-306.

Parmenter, T., Cummins, R.A., Shaddock, A., & Stancliff, R. (1994). Quality of life for people withdisabilities: The view from Australia. In Goode, D. A. et al . An international perspective onquality of life and disability. New York: Brookline Press, pp.75-102.

Baxter, C., Cummins, R.A., & Polak, S. (1995). A longitudinal study of parental stress and support: Theinfluence of child disability from diagnosis to leaving school. International Journal of Disability,Development and Education, 42, 125-136.

Cummins, R.A. (1995). On the trail of the gold standard for life satisfaction, Social Indicators Research,35, 179-200.

Cummins, R.A. (1995). The Comprehensive Quality of Life Scale: Development and evaluation.Proceedings. Health Outcomes and Quality of Life Measurement Conference. AustralianInstitute of Health and Welfare, pp. 18-24 (reprinted in Health Outcomes Bulletin, 7, 7-14).

Cummins, R.A., Fogarty, D., McCabe, M.P., & Hammond, J. (1995). Using the Comprehensive Qualityof Life Scale: A comparison between elderly Australians and normative data. Proceedings, 12thWorld Congress, International Federation of Physical Medicine and Rehabilitation, 1-10.

Gullone, E., Cummins, R.A., & King, N. (1995). Adaptive behaviour in children and adolescents withand without intellectual disability: Relationships with fear and anxiety. Behaviour Change, 12,227-237.

Cummins, R.A. (1996). Directory of instruments to measure quality of life and cognate areas. Secondedition. Melbourne: School of Psychology, Deakin University.

Cummins, R.A. (1996). The domains of life satisfaction: An attempt to order chaos. Social IndicatorsResearch , 38, 303-332.

Cummins, R.A., Baxter, C., Jauernig, R., & Hudson, A. (1996). A model system for the construction andevaluation of Individual Program Plans. Journal of Intellectual and Developmental Disability,21, 59-70.

Cummins, R.A., & Gullone, E. (1996). Measuring the quality of life of people with an intellectualdisability. Proceedings, integrating health outcomes measurement in routine health careconference. Canberra: Australian Institute of Health and Welfare, pp.148 - 152.

Cummins, R.A., McCabe, M.P., Romeo, Y., Reid, S., & Waters, L. (1997). An initial evaluation of theComprehensive Quality of Life Scale – Intellectual Disability. International Journal ofDisability, Development and Education, 44, 7-19.

Gullone, E., Cummins, R.A., & King, N.J. (1996). Fears of youth with mental retardation: Psychometricevaluation of the Fear Survey Schedule for Children - II (FSSC-II). Research in DevelopmentalDisabilities, 17, 269-284.

Gullone, E., Cummins, R.A., & King, N.J. (1996). Self-reported fears: A comparison study of youthswith and without an intellectual disability. Journal of Intellectual Disability Research, 40, 227-240.

McCabe, M.P., & Cummins, R.A. (1996). An evolutionary perspective on human female sexual desire.Sexual and Marital Therapy, 12, 121-126.

McCabe, M.P., & Cummins, R.A. (1996). The sexual knowledge, experience, feelings and needs ofpeople with mild intellectual disability. Education and Training in Mental Retardation andDevelopmental Disabilities, 31, 13-21.

McCabe, M.P., Cummins, R.A., & Romeo, Y. (1996). Personal relationships, relationship quality andhealth. Journal of Family Studies, 2, 109-120.

54

Cummins, R.A. (1997). Assessing quality of life for people with disabilities. In: R.I. Brown (Ed.).Quality of Life for Handicapped People. Second edition. Cheltenham, England: StanleyThomas, pp.116-150.

Cummins, R.A. (1997). Bibliography on quality of life and cognate areas of study. Third Edition.Melbourne: School of Psychology, Deakin University, (pp.1-97). (ISSN 1326-2173).

Cummins, R.A. (1997). Directory of Instruments to measure quality of life and cognate areas. ThirdEdition. Melbourne: School of Psychology, Deakin University, (pp.1-58). (ISSN 1325-0752).

Cummins, R.A. (1997). Measuring quality of life for people with an intellectual disability: A review ofthe scales. Journal of Applied Research in Intellectual Disability, 10, 199-216.

Cummins, R.A. (1997). Quality of life: Its relevance to disability services. In: P. O'Brien & R. Murray(Eds.). Working in Human Services. Auckland: Dunmore Press, pp.225-268.

Cummins, R.A., & Baxter, C. (1997). The influence of disability on quality of life within families.International Journal of Practical Approaches to Disability, 21, 2-8.

Cummins, R.A. McCabe, M.P., Romeo, Y., Reid, S., & Waters, L. (1997). An initial evaluation of theComprehensive Quality of Life Scale - Intellectual Disability. International Journal ofDisability, Development and Education, 44, 7-19.

Cummins, R.H., McCabe, M.P., & Cummins, R.A. (1997). The Girl’s Guide. Melbourne: Harper Collins(pp.1-161).

Fogarty, G.J., Bramston, P., & Cummins, R.A. (1997). Validation of the Lifestress Inventory for peoplewith a mild intellectual handicap. Research in Developmental Disabilities, 18, 435-456.

Baxter, C., & Cummins, R.A. (1998). An international standard for life satisfaction. Proceedings,First International Conference on Quality of Life in Cities, Singapore, 1-9.

Best, C., Cummins, R.A., & Lo, S.K. (1998). Evidence for domain compensation involving thehomeostatic control of subjective life quality among ex-farmers in Australia. Proceedings,First International Conference on Quality of Life in Cities, Singapore, 10-16.

Cummins, R.A. (1998). The Comprehensive Quality of Life Scale (fifth edition). Proceedings, FirstInternational Conference on Quality of Life in Cities, Singapore, 67-77.

Cummins, R.A. (1998). The second approximation to an international standard of life satisfaction.Social Indicators Research, 43, 307-334.

Cummins, R.A. (1998). Quality of Life Definition and Terminology. Blackburg, Virginia: TheInternational Society for Quality-of-Life Studies. Web address:http://www.cob.vt.edu/market/isqols/bibres.htm (pp.1-111).

In press:Baxter, C., Cummins, R.A., & Yiolitis, L. (1998). Parental stress attributed to disabled family members:

A longitudinal study. International Journal of Disability Research (in press).Best, C., & Cummins, R.A. (1998). The quality of rural and metropolitan life. Proceedings, First

International Conference on Quality of Life in Cities, Singapore, (in press).Bramston, P., & Cummins R.A. (1998). Stress and the move into independent accommodation.

Journal of Intellectual and Developmental Disability (in press).Bramston, P., Fogarty, G., & Cummins, R. A (1998). The nature of stressors experienced by people with

an intellectual disability. Journal of Applied Research in Intellectual Disability (in press).Clements, J., Rapley, M., & Cummins, R.A. (1998). On, to, for, or with – vulnerable people and the

practices of the research community. British Journal of Clinical Psychology (in press).Cummins, R.A. (1998). The measurement of subjective health outcome: Issues of concern. Proceedings,

International Conference on Health Outcome, (in press).McCabe, M.P., & Cummins, R.A. (1998). Sexuality and quality of life among young people. Adolescence

(in press).O'Brien, P., Tuck, B., Elkins, J., & Cummins, R.A. (1998). Exploration of visual behavior in diadic

relationships of people with and without a disability. Journal of Intellectual DisabilityResearch (in press).

http://www.cob.vt.edu/market/isqols/bibres.htm

55

Submitted:Best, C., & Cummins, R.A. The quality of rural and metropolitan life (submitted).Cummins, R.A. Normative life satisfaction: Measurement issues and a homeostatic model (submitted).Cummins, R.A. & Baxter, C. The influence of disability on quality of life within families (submitted).Foroughi, E., & Cummins, R.A. The influence of migration, social support and social integration on

the life quality of Persians in Australia (submitted).Fraid, R., & Cummins, R.A. Spiritual well-being and quality of life (submitted).Golding, D., & Cummins, R.A. The contribution of spiritual well-being to quality of life (submitted).Gullone, E., & Cummins, R.A. Fear, anxiety and quality of life: Adolescent self-reports (submitted).Kelly, J., & Cummins, R.A. The concept and measurement of quality of life in psychology

(submitted).McCabe, M.P., & Cummins, R.A. The changing attitudes of adolescent females towards boys,

themselves and their families over a ten year period (submitted).McCabe, M.P., Cummins, R.A., & Deeks, A.A. Construction and psychometric properties of

sexuality scales (submitted).Mellor, D., Cummins, R.A., & Loquet, C. The gold standard for life satisfaction: Confirmation and

elaboration using an imagined scale (submitted).Misajon, R., & Cummins, R.A. Subjective quality of life among Filipino-Australians (submitted).Nistico, H., & Cummins, R.A. Maintaining subjective well-being and avoiding depression: The role

of cognitive illusions (submitted).Petito, F., & Cummins, R.A. Social integration, stress and life quality among Italian migrants in

Australia (submitted).Verri, A.P., Cummins, R.A. Vallero, E., Monteath, S., Gerosa, E., & Nappi, G. An Italian-Australian

comparison of life quality among intellectually disabled people living in the community(submitted).

Yiolitis, L., & Cummins, R.A. The effects of social interaction and stress on the life quality of Greek-Australians, (submitted).

In preparation:Bearsley, C., & Cummins, R.A. No place called home (in preparation).Cummins, R.A., & Baxter, C. The experience of stress and quality of life among people who are

disabled, (in preparation).Cummins, R.A., & Hastings, M. Quality of Life as a function of participation in physical activity and

self-esteem (in preparation).Germano, D., & Cummins, R. A. Quality of life for people with arthritis (in preparation).Hutton, D.M., & Cummins, R.A. A structural model of organisational commitment and career

satisfaction (in preparation).Mallamace, J., & Cummins, R. A. The effects of social and competitive sport participation on

subjective quality of life (in preparation).Simm, L., & Cummins, R. A. Quality of life, work and retirement (in preparation).

Documents

comqol-s5