Research Methods Quiz

• Research minimizes bias and ensures decisions are based on facts.
• It follows rigorous methods, allowing broad application of information.
• A scientific approach involves an interplay between theory and data.
• The scientific process includes theory, data, hypothesis, and conclusion.

Theory

• A theory is a narrative explaining phenomena in the world.

Data

• Data consists of observations of what is happening.

Hypothesis

• A hypothesis is an unconfirmed expectation about the relationship between phenomena.

Verifiable data

• Data must be observable to researchers and the scientific community.
• Scientific collection and analysis of verifiable data follow a process.

Empirical Pattern

• Represents general tendencies based on data.

Logical Reasoning

• A process of moving from information to conclusions.

Inductive Reasoning

• Starts with specific observations, pattern recognition, forms general conclusions (data → patterns → theory; "bottom-up").

Deductive Reasoning

• Progresses from general ideas to specific conclusions (theory → hypothesis → data; "top-down").
• A scientific approach is useful especially when questions are important, when conflicting information arises, or when prioritization is needed.
• Science consumes time, money, and resources.
• Some questions are not answerable using scientific research due to their ambiguous nature.
• Inductive reasoning leads to weaker conclusions because the conclusion's certainty is less even with true evidence.
• Deductive research leads to stronger conclusions backed by evidence and is theory focused.
• Inductive research focuses on developing new theories.

Research Design: Setting the Stage

• Unit of analysis defines the level of social life a research question focuses on (what or whom is being studied).
• Examples of units of analysis: individuals, groups, organizations, interactions, artifacts.

Ecological Fallacy

• An error where conclusions about individual processes are drawn from group data.
• Example: Assuming Suzi is more likely to commit crime because her neighborhood's crime rate is higher.

Cross-Sectional Research Design

• Employs one sample taken at one point in time.

Repeated Cross-Sectional Research Design

• Samples at multiple time points to observe changes over time.

Panel Data Research Design

• Aims to follow the same sample over time, though attrition is an issue.
• Longitudinal designs are stronger for detecting causal effects, especially panel designs.
• Cross-sectional designs are often chosen for measuring things like employee satisfaction.
• Longitudinal designs are useful for placebo or medication trials.

Qualitative Research

• Qualitative research gains a deep understanding of knowledge.
• It uses questions about social processes or the meaning and cultural significance of people's actions.
• It aims to establish context-specific meaning or contribute to the development of theory.
• Example research question: how do nurses make sense of unexpected patient deaths?

Quantitative research

• Quantitative research uncovers general relationships, questions that are empirical between 2+ variables.
• It aims to test a hypothesis or establish causality.
• Examples: Do real estate agents that send birthday cards get more repeat business?

Variable

• A measured concept that varies across cases or over time.
• Independent variable (or predictor): A factor manipulated in an experiment, such as perfect attendance.
• Dependent variable (or outcome): A factor that depends on another controlled factor, like a grade in a class.

Causation

• Requires association, direction of influence, and the absence of spuriousness.
• Association: A change in one variable is associated with a change in another (e.g., perfect attendance → A+ in class).
• Direction of influence: Changes in the predictor (X) cause changes in the outcome (Y), not the other way around.
• Non-spuriousness: Association between variables should have no common cause.
• Antecedent variables are common causes that precede the relationship being assessed (e.g., age, gender, socio-economic status).
• Intervening variables act as both an effect of the predictor and the cause of the outcome (e.g., attendance → better notes → A+).
• Qualitative data collection is systematic, though it is not as standardized as Quantitative.

Qualities of an Appropriate Social Science Research Question

• Considers who, what, where, when, why, or how of social life.
• Requires data obtainable through the senses.
• Is important to examine, ethical, and sensitive to participants.
• Is explorable given practicalities.

Properties of a Good Research Question

• Relevant, well-articulated, doable, and approved by experts.
• Two Rules: Use question words/marks and allow answers through systemic empirical observation.

Research Ethics

• Based on the Belmont Report:
• Respect for persons: Acknowledge autonomy.
• Beneficence: Responsibility to do good.
• Justice: Fairly distribute research benefits and risks.

Researcher concerns for human subjects

• Potential harm (mitigated via informed consent, screening).
• Informed consent (freedom of choice, no coercion).
• Deception (researchers must be transparent).
• Invasion of privacy (ensure anonymity and confidentiality).
• Deception and informed consent
• Ethical research needs to yield an understandable conclusion.
• Invalid research is unethical due to wasted resources and potential risk to participants.

Rosenhan experiment

• Demonstrated ethics are needed in research.
• Rosenhan's study involved graduate students feigning schizophrenia
• Posed ethical issues due to potential harm and deception towards the public.

Measurement and Measurement in Action

• Conceptualization: define and clarifying the concept .
• Operationalization: Identify measures to measure data.
• Operational Definiton: Specify researcher activities.
• Indicators points to, provide evidence of, or measure a concept.
• Level of measurement affects how one describes information.
• Good measures are reliable (consistent) and valid (accurate).
• Reliability includes test-retest, internal consistency, and inter-rater reliability.
• Validity assesses whether the measure accurately captures what it's intended to.

US Census

• US Census is a data set that represents the demographics of the US

Sampling

• Involves population vs. sample and money and time.
• Random selection is the best strategy because it has equal chances of being selected.
• Sampling frame : the list of units composing a population from which a sample.
• Sampling error, chance variation, measurement error, and bias all have uncertainties in drawing conclusions about a pop.
• Large sample sizes decrease the chance of outliers affecting the value.
• Standard error means average distance between sample and population.
• Simple random means every case has equal probability.
• Stratified means some people are selected from each group, reducing standard error and cheaper, but complicated.
• Proportionate includes the sample size of each group.
• Disproportionate Stratified means some groups are more/less.
• Multistage Cluster means divide up pop and mini-pop

Sampling in Action

• Conflict over sampling means no random sampling.
• Not Random - It is what you conclude from your study with no choice

Probability

• It is guided by researcher and uses snowball
• They test the data through statistical and theoretical
• Only provides a basis for
• Saturation means reinforcements already.

Test your insights on appropriate research methods for various scenarios, hypothesis formulation, inductive reasoning, and data selection. Questions cover correlation studies, qualitative research, longitudinal studies, and sampling techniques.