Absence of Bias

1
Absence of Bias

• Chapter 4
• Popham (2005)

2
Assessment Bias

• Qualities of an assessment instrument that
  offend or unfairly penalize a group of students
  because of the students gender, ethnicity,
  socioeconomic status, religion, or other such
  group-defining characteristics (p. 77)
• Offensiveness and Unfair Penalization

3
Offensive

• A students performance on a test is distorted
  because the content (test items) of the test is
  offensive to a student because of his/her
  gender, culture, ethnicity, etc.
• Offensive negative stereotypes of subgroups

4
Unfair Penalization

• A students performance on a test is distorted
  because the content (test items) of the test puts
  that student at a disadvantage because of that
  students gender, culture, ethnicity, etc.

5
Disparate Impact

• If test scores of different groups (for example),
  different ethnic of religious groups) are
  decidedly different, this is described as an
  assessment procedure having a disparate impact on
  test takers.
• However, disparate impact does not always mean
  the assessment was bias, but it does warrant
  further investigation (or scrutiny) to determine
  possible reasons for the disparity.

6
Judgmental Bias-detection

• Review Panels
• 15-25 reviewers knowledgeable of the content and
  representative (ethnically and culturally) of
  test takers
• with a mix of male and female reviewers
• Given review of assessment bias and guidelines to
  follow for the review
• Per Item judgment

7
Judgmental Bias Detection (cont)

• After orientation, panelists review each item of
  the test by
• asking if each item of the text might be
  offensive or unfairly penalize groups of students
  based on gender, ethnicity, religion or race.
• Respond yes or no
• Percentages of yes/no responses are computed (the
  fewer No responses, the less bias the text)
• Overall judgments would look at the test as a
  whole (in toto)-looking at test questions
  collectively might reveal bias-such as a series
  of math word problems always depicting Males in
  their questions.

8
Empirical Approaches

• Differential item functioning (DIF)
• For high stakes tests to be given to a larger
  number if students, it is possible to gather
  evidence about the performance of different
  groups of students on individual test items
• Items are flagged and subject to further scrutiny
  by bias reviewers
• At this point, items may be eliminated

9
-
10
Assessment Accommodations

• An attempt to alter conditions associated with
  test administration Ex. Special lighting or large
  print for the visually impaired
• Audio-taped directions
• Additional time
• Test read aloud to students
• Alternative Assessment
• To Note Issues related LEP or ELL students (can
  not accurately test an ELL student in English in
  content areas if they are not proficient in
  Englishdoes not give an accurate representation
  of knowledge of content areas

11
Special Needs

• NCLB and Special Needs students
• TO BE REVISITED

12
What Teachers need to Know

• Teachers need to know that
• Assessment bias exists
• Teachers often unknowingly create tests or test
  items that are biased
• What do teachers need to do
• Review tests and test items (peer review)
• Be aware individual students cultural, ethnic,
  linguistic, etc. backgrounds