Quantitative Research Designs

1
Quantitative Research Designs

• V. Darleen Opfer

2
Conditions for Establishing Cause-Effect
Relationships

• Covariation
• Temporal Precedence
• No Plausible Alternative Explanations

3
Types of Alternative Explanations

• History
• Testing
• Instrumentation
• Regression
• Mortality
• Maturation
• Selection
• Selection by maturation interaction

• Ambiguity about causal direction
• Diffusion of treatments
• Compensatory equalization of treatments
• Compensatory rivalry

4
Ways to Minimize Threats to Validity

• By argument
• By measurement or observation
• By analysis
• By preventive action
• By design

5
(No Transcript)
6
Basic Design Elements

• Time
• Program(s) or Treatment (s) X
• Observation(s) or Measure(s) O
• Groups or Individuals
• R Randomly assigned groups

7
An Example
R O X O
R O O
8
Expanding a DesignX O

• Expanding across time
• O O X O
• O X O O X O
• Expanding across programs
• O X1 O
• O X2 O
• Expanding across observations
• O1O2 X O1O2
• Expanding across groups
• R O X O
• R O O
• N O O
• ______________________________
• R O X O
• R O O
• N O O

9
(No Transcript)
10
A Strategy for Design Construction

• Depict the hypothesized causal relationship
• Identify the possible alternative explanation
  threats
• Over-expand the basic design by expanding across
  time, program observations, and groups accounting
  for as many alternative explanations as possible
• Scale back the design to a manageable plan by
  considering the effect of eliminating each design
  component

11
The Nature of Good Design

• Theory-Grounded
• Situational
• Feasible
• Redundant
• Efficient

12
Design Task

• A principal/head teacher wants to know whether
  the provision of professional development will
  improve the instructional practices of teachers.
• Design a study for answering her question.
• Address the following questions
• What are your variables? (What are will you be
  measuring?)
• What are potential validity/alternative
  explanation issues with this study?
• What are the validity issues accounted for in
  your approach? And how would you minimize those
  not accounted for in the design?