Research Methods in Education

Research Design and Methods

• Descriptive designs: describe what is, observational, no control of variables, e.g., cross-sectional, comparative, and longitudinal designs.
• Correlational designs: investigate relationships between two or more variables, no control of variables, e.g., predictive, descriptive correlation, and model testing designs.
• Quasi-experimental designs: test cause-effect with no control, independent variable(s) not manipulated, e.g., post-test only, pre-post-test, and test-retest designs.
• Experimental designs: test cause-effect with control, independent variable(s) manipulated, e.g., classic pre-post-test, control group, and randomized control trials (RCTs) designs.

Research Questions and Hypotheses

• Research questions: investigate a topic/problem, inquisitive, written in question form, clear, concise, and sharp.

• Types of research questions:

  • Causal: comparing two or more conditions, e.g., "Does the number of hours spent studying affect BEd in ECS subject achievement scores?"
  • Descriptive: how much (change), how often of a phenomenon, e.g., "How often do BEd Hons in STEM/ECS Education students use Blackboard?"
  • Comparative: establishing difference between two or more groups, e.g., "What is the difference in ICT use of BEd Hons students and PGCE students?"

• Research hypotheses:

  • Predictive in nature
  • Data collection meant to support or reject the hypothesis
  • Hypothesis to include variables, the population, and the predicted relationship between variables
  • Null hypothesis (H0): e.g., "There is no difference in Blackboard use among PGCE and BEd Hons in STEM/ECS students" or "µ1=µ2"
  • Alternative hypothesis (H1): e.g., "There is a difference in achievement, there is a relationship, etc." or "µ1≠µ2"

Instrument Validity and Reliability

• Validity: an instrument must measure what it is supposed to measure, e.g., relationship between age-race-religion-language-residential area and academic achievement.

Quasi-experimental Research Design

• Tentatively investigates cause-effect relationships between two or more variables
• Researcher does not assign groups apriori (beforehand)
• No random assignment
• Research does not actively manipulate the independent variable(s)
• Experimental groups are identified and exposed non-randomly to the variable
• Experimental group results are compared with results from (control) groups that are NOT random