Research Methods in Social Science

Research is needed to answer questions, examine common sense assumptions, understand observations in the real world, solve problems, and understand how things work.
Facilitated communication is an example of a research topic where the need for research arose from the way an interaction between a child with autism and an adult led to interpretations and conclusions.

Steps in doing research include determining the research question and hypothesis, operationalizing variables, identifying study participants, creating study materials, obtaining ethics approval, and doing research.

Research questions should be based on common sense assumptions, observations from the real world, problem-solving, and understanding how things work.

A population is the entire group of people of interest.
A sample is a smaller group of people, drawn from the population, who participate in the study.
Random selection is important to ensure the sample accurately represents the population and is pertinent to experiments.

Operational definitions translate research questions into specific, testable procedures that can be measured and observed.
Examples of concepts to operationalize include aggression in children, stress levels in college students, online harassment, and sleep quality.

Research methods include naturalistic observation, case studies, self-report measures/surveys, correlational designs, and experimental designs.
Naturalistic observation involves watching behavior in natural settings.
Advantages include high external validity and rich detail, and can be the only option in some scenarios.
Disadvantages include lack of control, time-consuming data gathering, observer bias, and the inability to determine cause and effect.
Case studies involve in-depth analysis of a single person or setting.
Advantages include detailed descriptions and can be the only method possible.
Disadvantages include low external validity and researcher bias.
Self-report measures/surveys involve asking participants to describe their own behaviors, attitudes, views, and perceptions.
Advantages include affordability, large samples, and a variety of variables.
Disadvantages include participants' honesty and potential for bias.
Correlational designs examine the strength of relationships between variables, where variables are observed, but not manipulated.
A higher correlation value indicates a stronger relationship.
Experimental designs look at the causal influence between variables whereby at least one variable is manipulated, another is measured, and participants are randomly assigned to experimental or control groups.

Survey research provides a "snapshot" of how people think and behave at a specific time. Survey methods are affordable and efficient and can include large numbers of participants.
Self-report measures like questionnaires and surveys can be used when collecting data about people's behaviors, thoughts, and beliefs.
Questionnaires and surveys can be employed in both experimental and correlational research..

People may answer untruthfully in surveys out of a desire to present themselves positively.
Approaches to avoid this include embedding reverse-scored items in questionnaires and incorporating social desirability scales.
The Marlowe Crowne Scale is an example of a social desirability scale.

Considerations when choosing research measures include the operational variable and research question, cost (monetary and practical), and quality (reliability and validity).

Reliability refers to the consistency of measurement. Reliability can be examined through test-retest reliability, where participants’ responses are examined at two different times. Alternatively, intarrater reliability can be assessed through the agreement between two or more observers in scoring.

Validity refers to how well a measure assesses what it intends to measure. A valid measure is reliable but a reliable measure may not always be valid.

Correlational methods explore relationships between variables.
Variables aren't manipulated; they are observed.
Correlation coefficients range from -1.0 to +1.0, with higher values indicating stronger relationships.
Scatter plots visually represent relationships among variables. Relationships between variables can be positive, zero, negative, or curvilinear.
Third variables/confounds can influence the relationship between two variables.

Correlation does not imply causation. Many factors can explain relationships between variables.
Experimental designs are required to determine causation.

Third variables/confounds are external factors that may influence the study variables, potentially creating a false association.
Alternative explanations for findings should always be considered.

Manipulating an independent variable and observing its effect on a dependent variable, along with random assignment of participants to groups, are essential elements in this approach.
The independent variable (IV) is manipulated, while the dependent variable (DV) is measured. Different levels of the IV represent different experimental conditions.

The independent variable in experimental research should have at least two levels (conditions) - for instance, treatment groups vs. placebo groups, a high vs low manipulation.
Common examples include no sleep, little sleep and sufficient sleep in examining test scores to represent the independent variables.

A control condition in an experiment lacks any treatment or manipulation, providing a basis for comparison with the experimental conditions.

Placebo and nocebo effects (positive and negative outcomes due to expectations) are examples of a potential issue in research design
One element to be aware of is experimental bias and include blind test and double blind test designs to mitigate the bias.

The replication crisis in the field of psychology highlights the difficulty of replicating some findings, indicating potential issues with study design, sample size, and other aspects of the research.

Best practices to increase the replicability of research include publicizing data, allowing others to investigate using the same data for verification purposes and pre-registering the expected outcomes.

Ethical considerations in research include informed consent, protection from harm, and deception/debriefing (full disclosure of the study’s purpose following its execution).
Examples to elucidate ethical issues include the Tuskegee Syphilis Study.

Case Study Example: The Mask of Sanity by Hervey M. Cleckley: This example illustrates a case study and explains aspects of a case study example.
Case Study Example: Russell Williams: This example serves as another example to further understand case studies.
Naturalistic Research Example: This example demonstrates how researchers can apply their methods of observation in a real world setting and illustrates how it can work.