Cultural Issues in Testing

1
Cultural Issues in Testing

• Culture here used pretty broadly, including any
  group separate from normative sample
• Any time tests used on group different from
  normative sample, problems can arise
• Many tests (MMPI, WAIS, NEO-PIR) use US
  representative sampling and still are accused
  of cultural bias

2
Cultural bias

• Test demonstrates mean difference in scores
  between 2 cultural groups, but this score has no
  practical significance
• Score can be misinterpreted as having practical
  significance
• Not all group differences are biased
• Ex. diagnostic rates between males and females
• The Question to be asked Does score difference
  indicate a real difference between 2 groups, or
  is this an artifact

3
Examples

• Religious conservatives score higher on
  schizophrenia scales
• Does not indicate real/practical significance
• Males score higher on measures of aggression than
  females
• Does indicate real/practical significance
• Political correctness can intrude on this issue
• No one wants their group to show any
  real/practical negative significance, but
  embrace real/practical positive significance
• In-group philosophy

4
Another example

• The Ferguson test of history knowledge
• Who was the second president of the US?
• Who was president of the US during the civil war?
• In which year did the US enter WWII?
• The first permanent settlement in British North
  American colonies was
• The current US president is

5
Version 2

• The Ferguson test of history knowledge v2.0
• The Spanish defeated this native-American culture
  in conquering Mexico
• Mexican independence from Spain was achieved in
• During the most of the1860s much of Mexico was
  occupied by this nation
• Laredo was once capital of this short-lived
  republic
• The current president of Mexico is

6
Construct vs. predictive validity

• Moving a test from a group it was normed on to
  one that it was not may result in loss of
  construct validity
• May retain predictive validity however
• Ex. IQ tests with non-native English speakers in
  US
• Test-givers must be careful to note this issue in
  reports

7
Moving tests to new population

• Sometimes need arises for testing with new
  population
• Ex. inmates, children, individuals with
  disabilities, elderly, medically ill populations
• Is test valid for that populations
• Must be empirically tested for reliability/validit
  y
• Must be renormed
• Reliability/validity/normative data only useful
  for population represented by normative sample

8
This can really start to get tricky

• Populations can break apart into smaller and
  smaller groups
• Ex. medically ill with emphysema vs. those with
  CHF vs. those with brain injury
• minimum security vs. maximum security inmates
• Ideally each subgroup needs to be examined
  separately.
• Same for research
• If possible researchers should report
  reliability, means and SD (at least) of their
  instruments with their sample, even for
  empirically validated measures.

9
Culture and personality

• Different cultures can tend to produce certain
  personality traits
• Does this means these traits are due to
  environment?
• Can result in stereotypes when cultural
  differences are exaggerated
• Each culture tends to judge the unique traits of
  other cultures negatively
• Terror Management Theory and genocide
• This is common to all cultures

10
Then how to represent culture

• Tendency has been to focus on large groups
• Over concern with revising labels
• Ex. Hispanics vs. Latinos/Latinas vs. Chicanos
• African-Americans vs. blacks.
• Probably large-group labels are misleading and
  unhelpful
• People within those groups likely to be very
  distinct from each other.
• Ex. Caucasian Cubans vs. South-American tribal
  members
• Celts vs. Italians vs. Poles vs. Germans vs.
  Slavs
• Muslims vs. Arabs

11
Acculturation issues

• The next issue is whether membership in one
  culture is even meaningful
• Mixed-culture individuals in the US
• Census data somewhat misleading
• Practical reasons for endorsing minority
  membership
• Are people who are 3rd generation US-born
  identical to those just immigrated from
  elsewhere?
• Tests of acculturation

12
Ultimately

• Each test used should be researched for its
  applicability to the individual you are testing
• Any discrepancies should be noted in the report
• You may have to use tests on an individual even
  though they were not originally designed for that
  individual
• Apply critical eye/thinking to the use of each
  test