Robert A. Cummins Australian Centre on Quality of Life Deakin University Measurement scales and depression .

Robert A. CumminsAustralian Centre on Quality of Life

Deakin University

Measurement scales and depression

http://www.deakin.edu.au/research/acqol

What are the issues under investigation? (a) The problem of sub-optimal response scales(b) The problem of sub-optimal depression sales

Why are these issues important?(a) Likert scales are blunt instruments(b) We seem not to understand what depression actually is

What are the implications? (a) Our response scales may be compromising our measurements(b) We may be misdiagnosing depression

Overview

Freyd, M. (1923). The graphic rating scale. Journal of Educational Psychology, 14, 83-102.

“When you have satisfied yourself on the standing of this person in the trait on which you are rating him, place a check at the appropriate point on the horizontal line. You do not have to place your check directly above a descriptive phrase. You may place your check at any point on the line.” (p.88).

Extremely neat and clean. Almost

a dude.

Appropriately and neatly dressed.

Inconspicuous in dress.

Somewhat careless in his

dress.

Very slovenly and unkempt.

For job interviewsDoes he appear neat or slovenly in his dress?

[then standardized to 0-10]

1903 - 1981

Rensis Likert

HeadSurvey Research CenterUniversity of Michigan

Why only five levels of choice?

Likert, R. (1932) A technique for the measurement of attitudes. Archives of Psychology, No.14, New York.

Strongly

disapprove

DisapproveUndecidedApproveStrongly

approve

1 2 3 4 5

1. People can make more than five points of discrimination. They are therefore blunt instruments, not capturing the full extent of discrimination

Problem #1

Increasing the number of choice points above 5 increases scale sensitivity

• Diefenbach, M.A., Weinstein, N.D., & O’Reilly, J. (1993). Scales for assessing perceptions of health hazard susceptibility. Health Education Research, 8, 181-192.

• Russell, C., & Bobko, P. (1992). Moderated regression analysis and Likert scales: Too coarse for comfort. Journal of Applied Psychology, 77, 336-342.

• Jaeschke, R., & Guyatt, G.H. (1990). How to develop and validate a new quality of life instrument. In: B. Spilker (Ed.) Quality of life assessment in clinical trials (pp.47-57). New York: Raven Press.

1. People can make more than five points of discrimination. They are therefore blunt instruments

2. The number of choice points cannot easily be expanded because we do not have the necessary adjectives

Problems with Likert scales

Roy Morgan Research (1993)

Roy Morgan Research (1993). International values audit, 22/23 May. Melbourne: Roy Morgan Research Centre.

DelightedVery

pleased PleasedMostlysatisfied

Mixed feelings Unhappy

Mostlydissatisfied

Veryunhappy Terrible

1. People can make more than five points of discrimination. They are therefore blunt instruments

2. The number of choice points cannot easily be expanded because we do not have the necessary adjectives

3. The psychometric distance between the named adjectives does not accord with the interval nature of the scale

Problems with Likert scales

Ware and Gandek (1994) used the Thurstone method of equal-appearing intervals to

calculate the following distances between category labels used in the SF-36

5.04.33.42.31.0Actual

psychometric separation

Poor Fair GoodVerygood Excellent

Ware, J. E., & Gandek, B. (1994) The SF-36 Health Survey: Development and use in mental health research and the IQOLA project. International Journal of Mental Health, 23, 49-73.

Louis Leon Thurstone

(1887 -1955)

Dept PsychologyUniversity of Chicago

Jones, L.V., & Thurstone, L.L. (1955) The psychophysics of semantics: An experimental investigation. The Journal of Applied Psychology, 39(1), 31-36.

11-point, end-defined scale

109876543210

CompletelyDissatisfied

CompletelySatisfiedMixed

How satisfied are you with your life as a whole?

0.5 0.3 0.6 1.2 1.7

7.05.5

17.9

33.1

17.115.1

0

5

10

15

20

25

30

35

40

0 1 2 3 4 5 6 7 8 9 10

Frequency

Response scale 0 - 10

Can people reliably use 11-points of discrimination?

N≈30,000

The relationship between SWB and depression

Can the Subjective Wellbeing, or Positive Affect, be used as a measure of depression?

Critical issue #2

Dominant Source of SWB Control

Challengingconditions

SetPointrange

SWB

50

UpperThreshold

Lower Threshold

Nochallenge

Set point

Homeostasis

Very strongchallengeStrength of challenging agent

70

Defensive range

Strong homeostatic defense

80

a

bc

Lower Threshold

Theoretical proposition

Positive wellbeing is controlled by a homeostatic process

Homeostatic defeat means positive wellbeing is lost

Depression is the loss of positive wellbeing

The measurement of positive wellbeing should be THE measure of depression

UpperThreshold

Lower Threshold

Strong homeostatic defense

a

b

c

Lower Threshold

Loss of positivewellbeing= depression

Is this idea consistent with (DSM-IV) ?

Symptoms of depression include the following:(a) depressed mood (such as feelings of sadness or

emptiness)

(b) reduced interest in activities that used to be enjoyed, sleep disturbances (either not being able to sleep well or sleeping to much)

(c) loss of energy or a significant reduction in energy level

(d) difficulty concentrating, holding a conversation, paying attention, or making decisions that used to be made fairly easily

(e) suicidal thoughts or intentions.

Depression is---

(a) Loss of positive affect due to homeostatic failure

(b)reduced interest

(c) loss of energy

(d)difficulty concentrating

(e) suicidal thoughts

These are just the consequential symptomscaused by the loss of positive affect

How do the measures of depression and SWB relate to each other?

r ≈ .7

How do we establish that loss of positive wellbeing = depression?

4.4%

0 0.1 0.41.1

2.8

7.3

19.1

35.3

26.5

7.4

0

5

10

15

20

25

30

35

40

0-10 11-20 21-30 31-40 41-50 51-60 61-70 71-80 81-90 91-100

Frequency

Percentage points of SWB

How do the distributions of population incidence match?

Commonwealth Department of Health and Aged Care (2000) 5.8%1997 National Survey of Mental Health and Wellbeing 6.0%

Incidence of depression in Australia

Depression scores (DASS)

Normal Mild Moderate

79.7

77.7

76.0

74.4

72.070.9 71.0

65.0

63.3

60

65

70

75

80

0 1-2 3-4 5-6 7-8 9-10 11-12 13-14 15-16

PWI

The relationship between SWB and depression (symptoms) follows the theoretical pattern prescribed by homeostasis.

Conclusions

(a) 11-point end-defined scales are superior to Likert scales

(b) Depression should be defined, and measured, as a loss of positive affect.