Social research methods final exam

Spreadsheet Data Structure

Variables as columns, research subjects as rows, and observed values as cells.

Missing Data

Blank cells in a dataset resulting from unanswered survey questions or incorrect data entry.

Descriptive Statistics

Statistics that summarize sampled observations.

Inferential Statistics

Statistics that provide predictions about a population based on sampled observations.

Nominal Variables

Categorical variables with three or more attributes that are categorized only.

Ordinal Variables

Variables that are both categorized and ranked.

Interval Variables

Quantitative variables that are categorized, ranked, and evenly spaced with a consistent unit.

Ratio Variables

Variables that are categorized, ranked, evenly spaced, and possess an absolute true zero.

Univariate Descriptive Statistics

Statistics that summarize sample observations for a single variable.

Bivariate Descriptive Statistics

Statistics that summarize the relationship between sample observations for two different variables.

Mode

The most common observation in a distribution; most useful for nominal and ordinal variables.

Median

The midpoint observation when data are ordered from low to high; helpful when outliers are present.

Mean

The arithmetic average observation, calculated by summing all values and dividing by the number of observations.

Measures of Variability

Statistics that tell us how spread out sample observations are, such as range, standard deviation, and IQR.

Interquartile Range (IQR)

The range for the middle 50% of observations in a dataset.

Contingency Tables

Tables that describe categorical observations fitting each possible combination of two variables' values.

Correlation Coefficient (Pearson's r)

A numerical value indicating the strength and direction of a linear relationship between two quantitative variables.

True Experiments

Research designs where participants are randomly selected and randomly assigned to treatment or control groups.

Quasi-Experiments

Experimental designs that lack random selection and assignment due to practical or ethical reasons.

Internal Validity

The likelihood that changes in the dependent variable were caused solely by experimental manipulation of the independent variable.

External Validity

The extent to which study results generalize to the overall target population.

Secondary Data

Data originally collected by other researchers, institutions, or government agencies, also called archival data.

Applied Research

Research focused on examining specific problems, programs, and policies to solve real-world issues.

Needs Assessment

An applied research type that identifies a problem and what can be done about it.

Process Evaluation

An applied research type that determines whether a strategy is being implemented as designed.

Outcome Assessment

An applied research type that evaluates whether an intended solution actually works.

Focus Groups

Planned discussions designed to elicit group interaction as a qualitative research method.

Typical Focus Group Size

Usually 3 to 10 participants.

Moderator Role (Focus Groups)

Facilitates discussion by asking questions, presenting information, and keeping the conversation on track.

Note-taker Role (Focus Groups)

Records participant responses, words, body language, and facial expressions for later analysis.

Focus Group Weaknesses

Requires a skilled moderator, risks groupthink or conformity, and lacks anonymity and confidentiality.

Field Observations

A data collection method where researchers observe and record data in natural social settings ('the field').

Field Notes

Notes taken during field observations including rapid notes, visual sketches, and tally marks.

Nonparticipant Observation

An uninvolved 'fly on the wall' approach that minimizes reactivity in public settings.

Participant Observation

Full immersion in the studied setting, giving a complete picture but increasing reactivity.

Qualitative Research Approach

Focuses on qualities and uses an inductive approach to build theory from the ground up.

Forms of Qualitative Data

Primarily field notes and audio/video recordings that require transcription.

Coding (Qualitative Analysis)

The process of applying shorthand labels or codes to organize qualitative data into themes.

Open Coding

Going through files line-by-line to note emerging categories, themes, and patterns.

Focused Coding

Merging themes and categories from open coding into a more focused set of codes.

Codebook

A document containing brief definitions and descriptions of each qualitative code.

Interviews

A one-on-one conversation data collection method between an interviewer and an interviewee.

Structured Interviews

A rigidly organized interview format with a clear set of pre-planned, non-deviating questions.

Semi-structured Interviews

A flexible interview format with prepared questions that allows deviation and follow-ups.

Unstructured Interviews

A free-flowing conversation guided by topics and themes without pre-planned questions.

Open-ended Questions

Free-response questions allowing detailed answers without pre-determined choices.

Closed-ended Questions

Questions with predefined answers like multiple-choice, true/false, or Likert scales for easier analysis.

Surveys

A method to collect qualitative and quantitative data using questionnaires.

Contingency Questions

Follow-up survey questions whose appearance relies on answers given to previous questions.

Non-response bias

Bias occurring when low response rates skew results toward people with stronger opinions.

Survey Fatigue

Respondent attrition and dropouts caused by longer or tedious questionnaires.

Social Desirability Bias

Pressure felt by respondents to give answers they think researchers want to hear.

Recall Bias

The inability of respondents to fully or accurately remember past events.

Cross-sectional Research

Data collected at a single point in time, acting like a photo of a process.

Longitudinal Research

Data collected at multiple points in time to identify trends, acting like a video of a process.

Panel Surveys

Longitudinal studies where data are collected from the exact same sample over time.

Inductive Approach

Finding patterns in data to create a new theory.

Deductive Approach

Using pre-existing theory to analyze new data.

Operationalization

Defining concepts as specific, measurable variables.

Latent Construct

A concept that cannot be directly measured with just one variable, such as intelligence or wellbeing.

Construct Validity

The extent to which chosen indicator variables accurately represent a latent construct.

Quantitative Research

Research where concepts are operationalized as numerical variables and analyzed using statistics.

Qualitative Research

Research where concepts are operationalized as non-numerical variables and analyzed using thematic coding.

Association

An observed, statistically significant relationship between concepts.

Continuous Variable

A variable with an infinite or practically infinite number of possible values along a number line.

Discrete Variable

A variable with a finite number of separate, distinct, and indivisible categories.

Binary Measurement

Data that are dichotomous with only two attributes, such as smoker status or pregnancy.

Nominal Measurement

Categorized data with no inherent rank or order, such as race or major.

Ordinal Measurement

Data that are both categorized and ranked, such as Likert scales or crime severity.

Interval Measurement

Data that are categorized, ranked, and evenly spaced without a true zero, such as IQ or Celsius temperature.

Ratio Measurement

Data that are categorized, ranked, evenly spaced, and possess an absolute, true zero, such as annual income.

Reliability

The precision and consistency of a measurement.

Validity

The accuracy and relevance of a measurement.

Criteria for Causality

Empirical association, proper time-ordering, and non-spuriousness.

Longitudinal Data

Data collected at multiple points in time.

Cross-Sectional Data

Data collected at just one point in time.

Unit of Analysis

The specific entity from which data are collected and compared, such as individuals or neighborhoods.

Research Population

The entire, theoretical group of research subjects that a researcher wants to study.

Sampling Frame

An actual, concrete list of potential research subjects from which a sample is drawn.

Sample

The actual research subjects that provide data in a study.

Parameters vs. Statistics

Parameters are true values for populations; statistics are values calculated from samples.

Simple Random Sample

A probability sampling method where each member of the sampling frame has an equal and independent chance of selection.

Stratified Sample

Subdividing a sampling frame into distinct groups (strata) and taking simple random samples of each.

Cluster Sample

Taking a random sample of broader units (like agencies) and then sampling within those groups.

Purposive Sampling

A nonprobability method that intentionally creates a biased sample by picking specific edge cases or outliers.

Snowball Sampling

A nonprobability method starting with a key informant who recommends other participants, ideal for hidden populations.

Institutional Review Board (IRB)

An impartial reviewer of ethical research design that protects potential subjects from harm and researchers from legal risk.

Belmont Report

A foundational document for research ethics establishing the principles of respect for persons, beneficence, and justice.

Informed Consent

The process of ensuring potential subjects understand a study's purpose, risks, and benefits before willingly choosing to participate.

Null Hypothesis

The default assumption in research stating that there is no relationship between the studied variables.

Independent Variable

The predictor or cause variable (x) theorized to produce changes in another variable.

Dependent Variable

The outcome variable (y) theorized to be affected by changes in the independent variable.

Scientific research projects have 1+ of these purposes

Explore, describe, explain, apply

empirical association

must be an observable relationship between the independent and dependent variable

time-ordering

Changes in the independent variable must precede changes in the dependent variable

Non-spuriousness

there is no other variable responsible for the relationship between the independent and dependent variable

quota

set quotas-minimum #'s of elements that meet criteria and sample until you meet that #

convenience sample

just sample whoever/whatever is most convenient

systematic sample

skip a pre-determined number of elements before picking another

periodicity

a biased sample due to non-random patterns