Please enable JavaScript.
Coggle requires JavaScript to display documents.
TASK 5 (GREGORY (Test bias: (Test not biased = comparisons across…
TASK 5
GREGORY
Test bias: tests are claimed to be culturally and sexually biased so that they discriminate unfairly against racial and ethnic minorities, women and the poor
- Question of bias is objective and empirical, not a matter of personal judgement
Test fairness: broad value concept, recognises the importance of social values in test usage
- Even a test unbiased might be deemed unfair because of the social consequences of using it for selection decisions
Isuue separate from bias
- Unqualified individualism: without exception, the best qualified candidates should be selected for employmen
- Quotas - company's popuation the same as the regions
- Qualified individualism: ethical imperative to refuse to use race, sex etc. as predictor even if it were in fact scientifically valid to do so
Test bias:
- Test is deemed biased, if it is differentially valid for different subgroups - differential validity
Bias in content validity
- Content bias: item/subscale of a test is considered to be biased in content when it is demonstrated to be relatively more difficult for members of one group than another
Bias in predictive / criterion-related validity: unbiased test will predict future performance equally well for persons from different subpopulations
- Predictive validity bias: test is considered biased if the inference drawn from the test score is not mage with the smallest feasible random error or if there is constant error in an inference/ prediction as a function of membership in a particular group
Criterion of homogenous regression: test is unbiased if the results for all relevant subpopulations cluster equally well around a single regression line
-
-
Bias in construct validity: bias exists when test is shown to measure different hypothetical traits for one group than for another,
- Test not biased = comparisons across subpopulations should reveal a high degree of similarity for
- The factorial structure of the test - Important to restrict the comparison to same age group
- The rank order of item difficulties within the test
Also separately checked for each age group
Tested by computing p-value for each subpopulation and compare relative difficulties across same age-subgroups
- Can be gauged objectively by correlation coefficient the closer r =1, the more similar the rank ordering of difficulty for the groups
-
Sprangers – Integrating response shift into health-related quality of life research: a theoretical model
-
Theoretical model: addresses how response shift may affect health-related QoL as a result of changes in health status
- Catalyst: change in respondent’s health status
- Antecedents: stable/ dispositional characteristics of the individual, e.g. sociodemographics, personality, spiritual identity both direct and indirect effects of potentiating response shift
- Mechanisms: behavioural, cognitive and affective processes to accommodate the catalyst, e.g. coping strategies, social comparisons, seeking social support, reframing expectations
- Response shift: change in meaning of one’s self-evaluation of QoL as a result of changes in internal standards, values and the conceptualisation
- Perceived QoL: multidimensional construct incorporating at least 3 broad domains: physical, psychological, and social functioning suboptimal QoL perception will initiate a feedback loop to change mechanisms
- Response tendencies such as social desirability, dissonance reduction may induce changes in QoL without actually affecting internal standards, values or conceptualisation of QoL
- Social desirability may not be a response artefact but rather a personality characteristic that helps people assimilate illness and enjoy greater level of QoL
Nieuwkerk – Are retrospective measures of change in quality of life more valid than retrospective measures?
-
-