A Reliability Generalization of the Life Satisfaction Index

1
A Reliability Generalization of the Life
Satisfaction Index

• K. A. Wallace J. C. Caruso
• University of Montana

Presented at the Annual Meeting of The
Gerontological Society of America, November 2002.
2
Purpose

• To examine score reliability for a widely used
  measure of life satisfaction, the Life
  Satisfaction Index (LSI Neugarten et al., 1961)

• Average score reliability
• Variation in score reliability as a function of
  sample characteristics (e.g., gender mean age
  scale length etc.)

3
Developing Ratings of Life Satisfaction

• Crafted to assess well-being using a subjective
  evaluation of ones own present and/or past life
• Part of the Kansas City Study of Adult Life
  (Neugarten et al., 1961)
• Thematic analysis of measures of adjustment and
  morale
• 5 components of well-being
• Zest versus apathy
• Resolution and fortitude
• Congruence between desired and achieved goals
• Self-concept
• Positive mood tone

4
The Scales

• Life Satisfaction Index A (LSIA)
• Life Satisfaction Index B (LSIB)
• Life Satisfaction Index Z (LSIZ Wood et al.,
  1969)
• 18-item version (Adams, 1969)
• 8-item version (LSIW James et al., 1986)
• 30-item version (Maynard, 1993)

5
Reliability Generalization

• Meta-analytic technique
• Examines average score reliability (e.g.,
  Vacha-Haase, 1998)
• Examines relationships between study
  characteristics and score reliability

6
Method Data Collection

• PsycINFO literature search
• Life satisfaction index and LSI
• 157 possible articles
• 59.87 no mention of reliability
• 9.56 indicated LSI reliable test, no data
• 6.37 cited reliability from previous work
• 3.18 reported reliability in unusable form
• .64 not empirical
• 1.27 could not be obtained
• 19.11 (30) provided usable reliability
  information
• Total of 34 samples used

7
Descriptive Statistics for Sample Characteristics
(N34)
Variable M
SD Range Sample size 235.53 328.46 20-1571 Scale
length 17.09 4.21 8-30 Mean age 61.79 17.17 20.2-8
3.3 Std dev age 6.89 3.38 3-15.7 Proportion
female .63 .28 0-1.00 Mean LSI 15.56 13.01 3.08-65
.5 Std dev LSI 3.40 2.34 .52-11.83 Lang of
admin .78 .42 0-1.00 Sample type .26 .51 0-1.00
8
Results

• Average Score Reliability
• Mean .79 (SD .10)
• Median .79
• Range of .56 (.42 to .98)
• Bivariate correlations
• Score reliability was not significantly related
  to scale length, mean age, standard deviation of
  age, proportion female, sample size, mean LSI, or
  standard deviation LSI
• t tests
• No difference in score reliability as function of
  language of administration or sample type

9
Discussion

• Adequate average score reliability for the LSI
• File drawer problem?
• No relationship between score reliability and
  sample characteristics
• Preliminary evidence for adequate reliability of
  scores generated with LSI across various sample
  characteristics
• Limitations?

10
Future Research Implications

• Replication
• Conceptualization of reliability as property of
  scores (e.g., Wilkinson APA Task Force on
  Statistical Inference, 1999)
• Inclusion of more detailed demographic and
  reliability information

11
Method Data Analysis

• Sample characteristics recorded
• Scale length
• Mean age, standard deviation of age
• Gender (proportion female)
• Sample size
• Language of administration (Eng1 Other0)
• Sample type (Instit/Mixed1 CommResid0)
• Mean and standard deviation of scores