Please enable JavaScript.
Coggle requires JavaScript to display documents.
Reliability and Validity in Quantitative Research - Coggle Diagram
Reliability and Validity in Quantitative Research
Internal Validity:
In experimental research, the extent to which the observed effect can be attributed to the manipulation of the IV
Concern of avoiding confounding variables
Vincent and Lewycky's 'Better Sleep' Trial:
A randomised controlled trial for insomnia treatment
Highlighted challenges like attrition/mortality, history, sampling, maturation, testing issues and participant expectancy
Demand Characteristics:
Cues in the study that help participants understand expectations, potentially leading to social desirability bias
Observer/Experimenter Expectancy Effect:
Researchers' expectations influencing participant treatment and behaviour
Ways of Improving Validity:
Standardised procedures, counterbalancing, blinding (single or double)
External Validity:
The degree to which results generalise beyond the experimental context to different subjects, groups, settings and periods
Population Validity:
Current research often relies on WEIRD populations, potentially limiting generalisability
Ecological Validity:
Concerns the extent to which research findings apply to real-world, everyday life situations
Construct Validity:
Focuses on measuring abstract psychological constructs accurately
Reliability: The extent to which a measurement is reproducible or consistent over time
Reliability Over Time:
Test/measure should yield consistent results under the same conditions
Internal Reliability:
Consistency of scores across multiple items of a construct, often measured by Cronbach's alpha
External Reliability:
Test-retest reliability and inter-observer reliability
Importance of Reliability and Validity: Ensures the credibility of research, scientific rigor, and accurate clinical applications
Experiments and RCTs:
Laboratory based, fully controlled, with a focus on both internal and external validity
Quasi Experiments:
Similar to experiments but may lack random assignment or full control over the IV
Correlational Studies:
Used to determine relationships between non-manipulated variables, with considerations for internal and external validity
Questionnaires:
Frequently used data collection method with potential issues related to multiple measures
Publication Bias: Researchers tend to only publish positive effects
Validity:
The extent to which a demonstrated effect in research is genuine, not influenced by extraneous variables and not limited to a specific context
Emphasises accurate conclusions, distinguishing between systematic and unsystematic errors
Confounding Variables:
Unrelated factors that may impact results, highlighting the importance of identifying and controlling them
Some may have valuable outcomes
Elizabeth Loftus' Car Accident Experiment:
Illustrates how changes in wording impact participant recall, emphasising the importance of control
Participant Expectancy, Hawthorne Effects, and Demand Characteristics:
Explores how participant expectations and reactions to experimental settings can influence behaviour
Meta-Analyses:
Help to improve reliability and validity by comparing multiple datasets
Type I Error: false positive result
Type II Error: false negative result