(6) Descriptive Methods & Research Methods (Video Notes)

Descriptive methods describe behavior and do not necessarily establish causation.
Three broad ways to test hypotheses include descriptive methods, correlational methods, and experimental methods:
- Descriptive methods describe behaviors, often by using case studies, surveys, or naturalistic observations.
- Correlational methods examine associations between variables; a variable is anything that contributes to a result.
- Experimental methods manipulate variables to discover their effects.

Quantitative Data (Numerical): numbers-based information gathered from surveys, tests, or experiments. Helps identify patterns and relationships in a precise way.
- Q for Quantity: Quantitative data is about the Quantity of numbers.
Qualitative Data (Non-Numerical): non-numeric information that gives deeper insights into topics.
- Q for Quality: Qualitative data is about the Quality of experiences and observations.
- Collected through methods like interviews or observations, focusing on people’s experiences and behaviors.

Descriptive methods describe behaviors (e.g., case studies, surveys, naturalistic observations).
Correlational methods identify associations between factors/variables.
Experimental methods manipulate variables to determine causal effects.

Case Study: One person studied in depth to reveal underlying truths about people.
Key idea: Rich, detailed data about a single case can illuminate broader principles but may not generalize.
Typical critique: Limited generalizability due to single-subject focus.

Definition: Observing behavior in people’s natural environments without interference.
Examples from slides:
- Observing and recording animal behavior in the wild.
- Recording seating patterns in a multiracial school lunchroom.
Purpose: Capture behavior as it occurs naturally, without experimental manipulation.
Cautions:
- Observer effects: behavior may change when watched.
- Ethical considerations about privacy and consent in some settings.

EarthCam Las Vegas link provided as an example of live-lens data collection (illustrative).
Twitter mood graph (Figure 0.3-2):
- Demonstrates how researchers can study human behavior at scale using anonymized data.
- Can correlate mood with location, weather, and information flow through social networks.
Implication: Big data enables patterns across large populations beyond traditional samples.

Survey: Self-reporting of behaviors or attitudes.
Example data note: A sample indicates that about one-half of people across 24 countries believe in life outside of earth.
Common caveat: Self-report data can be biased by wording, social desirability, and recall issues.

Wording can change survey results significantly.
Example 1 (Q1 vs Q2):
- Q1: "Do you believe we should be providing aid to the needy?"
- Q2: "Do you believe we should give people welfare?"
Result: Different phrasing can lead to different levels of approval because of connotations.
Concept name: Wording Effect.

Common phrases that increase approval:
- "aid to those in need";
- "undocumented workers";
- "gun safety laws";
Phrases that decrease approval:
- "welfare";
- "illegal aliens";
- "gun control laws";
- "revenue enhancers";
- "taxes";
- "enhanced interrogation";
- "torture";
- "pre-owned"; "used".
Purpose: Demonstrate how language influences attitudes and responses.

Likert Scales: A measurement tool used in surveys to assess attitudes or opinions.
Structure: Respondents indicate agreement on a 5-point scale (commonly):
- Strongly Agree, Agree, Neutral, Disagree, Strongly Disagree.
Example prompts:
- "I enjoy spending time with friends." (response options listed above)
- "I feel confident in my ability to succeed in challenging tasks." (response options listed above)
Note: 5-point scale is typical, but other variants exist (e.g., 7-point).

Structured Interview: Predetermined questions are asked to all participants in the same order.
Benefit: Ensures consistency and enables comparisons.
Example prompt (outdoor activities):
- Question: "How often do you engage in outdoor activities such as hiking, camping, or picnics?"
- Response options: a) Daily b) Several times a week c) Once a week d) Occasionally e) Rarely or never
Follow-up prompts: "What factors influence your decision to participate in outdoor activities?" with multiple choice or open-ended options.

Social desirability bias: respondents answer in a way they think will please the researcher rather than reflect true beliefs.
Example prompts illustrating bias:
- Are you a smoker? (Respondents may deny or minimize smoking.)
- Am I a hard worker? (Ambiguity can lead to varying responses.)
Consequence: Data may overrepresent socially accepted responses and underrepresent stigmatized behaviors.

Sampling Bias: Generalizing from a few vivid but unrepresentative samples can mislead conclusions.
Random Sampling: Each member of the population should have an equal chance of being included to be unbiased and representative.
If the sample is biased, its results are not valid.
Analogy: The fastest way to know marble color ratio is to blindly transfer a few marbles into a smaller jar and count them (random sampling).

Case Study
- Purpose: Study one person (or small group) in depth.
- Strength: A lot of detail.
- Weakness: Generalizability may be limited.
Survey
- Purpose: Gather self-reported data from a population.
- Strength: Fast; large data sets.
- Weakness: Response biases; wording effects; false consensus effect.
Naturalistic Observation
- Purpose: Observe behavior in natural settings.
- Strength: Real-world behavior; less artificial context.
- Weakness: Observer effects; less control over variables.

Prompt: Pick one of the following and explain a PRO and a CON:
1. Survey
2. Case Study
3. Naturalistic Observation

Throughout these methods, ethical considerations include privacy, consent, and minimizing harm.
Foundational principle: Descriptive methods provide the groundwork for understanding phenomena, often informing the design of correlational and experimental studies.
Real-world relevance: Surveys and big-data analyses enable scale and speed, whereas case studies and naturalistic observations provide depth and ecological validity.
Key terms to remember: replication, generalization, variables, correlation, causation, sampling bias, random sampling, social desirability bias, wording effects, Likert scale, structured interview.