Research Methods: Key Concepts and Definitions

Study Design and Variables

Recordkeeping is fundamental to good science because it supports integrity and accountability in research studies. Records must be understandable to others and contain sufficient detail that another skilled researcher can repeat the same work and obtain the same results. Research integrity standards emphasize that high-quality documentation is essential for replicability.

Antecedent Variable in statistics and social science research is a variable that helps explain the relationship between other variables that have some degree of cause-and-effect relationship. In regression analysis, an antecedent variable provides influence on both the dependent and independent variables.

Causal vs. Spurious Relationships differ fundamentally. A spurious relationship occurs when variable X causes both Y and Z, creating an apparent relationship between Y and Z that is not direct. A causal relationship, by contrast, occurs when the first event directly results in the second event's occurrence.

Characteristics of a Good Hypothesis include the ability to be tested using the scientific method. A hypothesis proposes an explanation for a phenomenon and must be formulated in a way that allows for empirical testing.

Characteristics of a Good Research Question include asking for information that can be answered within the scope of the research methodology. The question should be specific, measurable, and achievable given available resources.

Characteristics of Scientific Knowledge consist of knowledge formulated through testable explanations. Scientific knowledge must be grounded in observable phenomena and subject to verification or refutation.

Characteristics of Small-N Designs include testing only one or very few subjects who are treated as individuals rather than assigned to groups. Data collection extends over long periods to enable stabilization of performance over time. Data from subjects are handled separately, analyzed visually, and subjected to minimal inferential statistical analysis.

Control Group in a scientific study is the group that receives no treatment, whereas the experimental group does receive treatment. The control group eliminates alternate explanations that might arise from the experiment's results.

Correlation vs. Causality represent distinct concepts. Correlation measures the relationship between two variables or data values, while causality refers to the relationship where an effect arises from the physical consequence of a cause.

Deduction and Induction are opposing reasoning approaches. Deduction is an approach based on what is observed naturally, while induction moves from specifics to generalization.

Dependent Variable (also called the response variable) is the outcome variable that researchers measure to assess the effect of their manipulation or treatment.

Measurement and Analysis

Determinants of Causality are variables that either increase or bring about a decrease in risk.

Directional Hypothesis is more specific than a non-directional hypothesis because the researcher not only predicts the existence of a relationship but also specifies the direction that relationship will take.

Ecological Fallacy is a logical fallacy that occurs when statistical data regarding groups is interpreted to make inferences about the nature of individuals within those groups.

Ecological Inference is the process of using ecological data to make conclusions about behavior at the individual level.

Empirical refers to science-based knowledge or research grounded in observation and measurement.

Empirical Generalization involves making a statement about regularities that are observed, without attempting to provide an explanation for those patterns.

Independent Variable is the variable that the researcher varies or manipulates during the study.

Unit of Analysis is that which undergoes analysis in a study, whether individuals, groups, social interactions, or organizations.

Level of Measurement (nominal, ordinal, interval, ratio) describes how variables are categorized and measured. Nominal measurement uses numerical values to label attributes uniquely. Ordinal measurement rank-orders attributes. Interval measurement measures the distance between two attributes. Ratio measurement involves an absolute zero with meaningful interpretation, and fractions are meaningful when used with ratio variables.

Likert Scale is a scale in which participants rate their answers on a continuum, typically from minus one (completely disagree) to plus five (completely agree), allowing researchers to quantify subjective responses.

Validity, Reliability, and Quality Assurance

Multi-Item Measures (indexes and scales) ask more than one question about a subject. For example, a measure might ask: (1) How likely are you to purchase accessories for clothing you own in the next month? and (2) How likely are you to have alterations done to clothing you own in the next month? An index is a compiled measure that provides a summary and rank-order to specific observations, while a scale is representative of a more general dimension compiled from scored individual items.

Operationalize means to express the terms or methods used to determine and measure research outcomes.

Normative vs. Empirical distinguish between what is standard or normative and what is science-based or empirical.

Mill's Method of Difference uses five responses to an experience. All participants rate their experience and answer yes or no to five questions; the degree of differences in answers determines the methodology of analysis.

Steps in Content Analysis involve ten key steps: (1) Copy and read the transcript, making notes about pertinent information; (2) Review notes and list the various types of information located; (3) Read the list and categorize each item, providing descriptions of identified information; (4) Identify whether categories are potentially linked and list primary and minor categories; (5) Compare and contrast primary and minor categories; (6) If multiple transcripts exist, repeat the first five steps for each additional transcript and collect themes or categories for detailed examination; (7) After categorizing transcript data, review to ensure correct categorization; (8) Review all categories and determine if any should be combined or require subcategorization; and (9) Return to the original transcript to ensure all information needing categorization has been included.

Falsifiability is the degree to which data could potentially be falsified or proven false, an essential criterion for scientific claims.

Tautology occurs when the same thing is stated twice using different words.

Parsimony in research refers to the principle of not making unnecessary use of available resources and avoiding overcomplication of explanations.

Data Collection and Ethical Considerations

Validity takes multiple forms. Construct validity is the approximated truth gained by a conclusion based on how well the operationalization matches the construct. Face validity is the validity that is immediately apparent or "on the face of it." Content validity involves checking the operationalization against relevant content for the construct.

Validity: Internal vs. External describes two essential dimensions. Internal validity refers to how rigorously the study was conducted, ensuring that observed effects genuinely result from the independent variable. External validity is the degree to which results from the study are generalizable to other contexts, populations, or settings.

Reliability encompasses several types. Test-retest reliability measures stability by giving the same test twice to a group and correlating the scores to establish consistency across time. Parallel forms of reliability measures consistency by giving varying versions of a test to the same group and correlating the scores. Inter-rater reliability determines the degree to which different raters agree in assessment decisions.

The Importance of Theory lies in its role as a guide for research direction and the researcher's processes and methodology. Theory frames the research questions, informs hypotheses, and provides the conceptual foundation for interpretation.

Theory itself is rational and contemplative, involving generalization or abstract thinking. It describes different contexts of how nature or people operate or work.

Characteristics of Scientific Knowledge consist of knowledge comprised by testable explanations that can be verified or falsified through empirical observation.

Episodic Record differs from systematic record-keeping. Episodic records are preserved on a personal, casual, or accidental level rather than through structured methods.

Research Types and Methodologies

Running Record is a systematic method of record-keeping during research and also an informal method that occurs naturally during observation.

Ethical Considerations in Interviewing require both informed consent and confidentiality. Researchers must obtain permission from participants and protect the privacy and confidentiality of information provided.

Examples of Observational Studies include several types: case studies examine an individual, group, or other subject in depth; comparative studies focus on comparing two or more groups, individuals, or phenomena; surveys are formulated and administered in person or distributed; cross-sectional studies examine groups at a single point in time; aggregate studies examine group-level data; and longitudinal time-series studies take place over lengthy time periods.

Field Experiment is an experiment conducted in the natural setting of the subject or subjects, as opposed to in a controlled laboratory environment.

Steps in a Classical Randomized Experiment include the following: (1) Randomly assign individuals to treatment or control groups; (2) Administer a pre-test to both groups; (3) Ensure both groups have the same experience and conditions, with the experimental group receiving the treatment; (4) Administer a post-test to both groups; and (5) Assess the amount of change in the dependent variable from pre-test to post-test for each group, analyzed separately.

Policy/Intervention Analysis is conducted prior to beginning a push for changes in policy matters or intervention methods, helping to predict outcomes and identify potential consequences.

Transmissible means that findings or methods can be spread or applied to other contexts, populations, or researchers.

Hypothesis is a statement of what the research expects to find, grounded in theory and prior research.

Explanatory Research is focused on providing an explanation for some phenomenon, going beyond description to identify causes and mechanisms.

Qualitative vs. Quantitative Methodology represent distinct approaches. Qualitative methodology uses social research methods such as surveys, interviews, focus group studies, and observation, presenting results in descriptive form. Quantitative research, by contrast, uses statistical testing and scientific formulas to measure results numerically.