Skip to main content

Table 1 Summary of psychometric properties and criteria used to review measures.

From: Measuring the psychosocial health of adolescent and young adult (AYA) cancer survivors: a critical review

Psychometric Property Criteria
Reliability  
Internal consistency
  degree to which responses to all items on a scale are consistent [43]
Calculated correlations for total scale and domains [44]
  - Cronbach's alpha (α) > 0.70 [42, 44]
  - Kuder-Richardson 20 (KR-20) > 0.70 [42, 44]
Test-retest
  reproducibility of scores on a scale over repeated administrations [44]
Second administration within 2-14 days [46]
Calculated correlations for total scale, domains and items [47]
  - Cohen's kappa coefficient (κ) > 0.60 [44]
  - Pearson correlation coefficient (r) > 0.70 [42, 44]
  - Intraclass correlation coefficient (ICC) > 0.70 [42, 44]
Validity  
Face
    subjective assessment of whether a scale 'appears' to measure what it is designed to measure [43]
Assessed as reasonable by those who administer/complete it [43]
Content
    degree to which the content of a scale is representative of the issue being measured [43]
Reported item selection process [42, 44]
Content assessed by experts [42, 44]
Reported which aspects of the measure were revised [42, 44]
Construct
    way in which the internal structure of a scale relates to other conceptual constructs [44]
Stated hypothesis about correlations between measures [44]
    - Convergent (r) > 0.40 or Divergent (r) < 0.30 [48]
Calculated correlations between known-groups [42]
Performed factor analysis [44]
    - Eigenvalues > 1 [49]
Criterion
  how well a scale agrees with existing "gold standard" measurement of the same issue [44]
Provided rationale for "gold standard" measure [44]
Stated type of criterion validity (concurrent or predictive) [43]
Reported proportions [44, 50]
    - Sensitivity - % with issue correctly classified [44, 50]
    - Specificity - % without issue correctly classified [44, 50]
Responsiveness
  sensitivity of a scale to detect clinically important change in an outcome or behaviour over time [42, 50]
Reported floor/ceiling effects [51]
- < 5% of respondents have highest or lowest score [51]
Reported magnitude of change [42]
- Effect size > 0.5 [42, 44, 50]
Acceptability
  level of burden placed on those who complete the measure [42]
Reported response rate, missing items, reading level, time to complete [42]
Feasibility
  level of burden placed on those who administer the measure [42]
Reported perceived time to administer, score, interpret [42]
Cross-cultural adaptation
  conceptually, linguistically equivalent and display similar psychometric properties to the original form [42]
Confirmed reliability and validity reflects the original version [42]