Controlling Variance

1
Controlling Variance

• Variance can cause misinterpretation of the
  results
• Variance is found in the population under study
• 4 ways to control variance
• Randomization
• Building in factors as IVs
• Holding factors constant
• Statistical control

2
Randomization
Random selection, or assignment, of subjects
spreads the effect of whatever is causing the
variance, evenly across the participating groups.
Study comparing curricula A, B, C
3
Factors as Independent Variables
Include factors causing variance as independent
variables in the study.
Study comparing curricula A, B, C
4
Holding Factors Constant
Reduce the influencing factor (or variable) to a
constant.
Curriculum A
50 Low Ability students ran- domly assigned to
Group 1
Performance on a geography test
Curriculum B
50 Low Ability students ran- domly assigned to
Group 2
Performance on a geography test
Curriculum C
50 Low Ability students ran- domly assigned to
Group 3
Performance on a geography test
Study comparing curricula A, B, C
5
Lets Try One
The researcher is seeking to determine the
effects of two different cohort programs on a
heterogeneous sample of educational
administration students at California State
University.

• Sample 100 students
• 1st 2nd year
• White, African American, Southeast Asian,
  Latino/a
• All have 3 5 years teaching experience

• Cohort Programs
• Urban Leadership Emphasis
• Bilingual Leadership

Apply the Holding Factors Constant strategy to
the study.
6
Now You Try It!

• What is the best way to teach 5th grade students
  who are reading below grade level in Sacramento
  schools?
• Describe the population.
• Identify one factor that may cause variance.
• Describe 3 samples, using each of the first 3
  strategies to control variance.

Post your responses to WebCT under the forum, HW
7.