Experimental Design

  • We undertake experiments to identify if there is a difference or relationship
  • Data which is collected is either:
    • Quantitative (numeric)
    • Qualitative (non-numeric or opinions)
  • There are five main experiment designs which I have listed in rank order based on robustness
  • Computing students my have to use a range if they are testing people and computers
  • You might write about your experiment design like this…

This study will use a Quasi-Experimental Design as the participants are already divided into groups which does not allow for the randomisation of the participants. The strength of this experiment design is….

The issues with using a non-random sample is …..

The implementation of the following control measures will reduce this effect:

1) ….
2) ….
3) ….

 

Key Information

Computing

Single-Case Experiment Design (A-B-A Design)

  • Used with individual computers (laptop, raspberry pi) or one group of computers (network)
  • A base-line assessment is completed to benchmark system (speed, range, bandwidth etc…)
  • Apply the changes (implement your artefact)
  • Complete the base-line assessment again
  • If there is a difference then the intervention then it worked (had an effect)
  • This is called ‘related samples,’ as the data comes from the same sample but at different times
  • Non-Experimental Design (Descriptive Study)
    • Non-experimental designs has no control (testing or randomness)
    • Examples of non-experimental designs could be monitoring the computer system to understand what is happening
    • Create a idea (hypothesis) base on the monitoring data
    • Move a different experiment design

Social Science

  • Meta-Analysis
    • A review all the data collected and presented by different researchers over a period of time
  • Randomised Control Trials (RCTs)
    • Is where the participants are randomly assigned to groups
    • This is done to reduce bias (established groups, personality clustering etc…)
    • Group 1 undertake the intervention (testing group) and group 2 does not (control group)
    • If there is a difference between the two groups then the intervention worked
  • Quasi-Experimental Design (non-random design)
    • Uses pre-existing groups (a class, a football team etc…)
    • Often the researcher (teacher, coach) already know the participants
    • The researcher highlights how the two groups are equal, based on key factors and therefore can be compared e.g.
      • The same number of participants
      • Same number of students working at the same level
      • Same amount of contact time
    • If there is a difference between the two groups then the intervention worked
    • This is called ‘independent sample,’ as the data comes from different groups (male vs. female)
  • Single-Case Experiment Design (A-B-A Design)
    • Used with individuals or one group
    • A pre-intervention assessment is completed based on the variables the intervention will target
    • Intervention is provided
    • The pre-intervention assessment is completed again
    • If there is a difference then the intervention worked
    • This is called ‘related samples,’ as the data comes from the same sample but at different times
  • Non-Experimental Design (Descriptive Study)
    • Non-experimental designs has no control (testing or randomness)
    • Examples of non-experimental designs are:
      • Interviews and focus groups
      • Observations
      • Feedback forms, personal accounts

Leave a Comment