Key Concepts: Research Methods - Last-Minute Review

Research process essentials

Questions drive learning: start with a question, search for answers, and relate to your topic.
Scientific method workflow: ask question → review literature (often 30–40 articles) → synthesize findings → craft organized thesis sections.
Methodology planning: specify design, participants, procedures, measurements, and how findings will be reported; these become core thesis components.

Two big data categories in research:
- Quantitative: numerical data from scales or measurements.
- Qualitative: non-numerical data from open-ended responses, transcripts, observations.
Psychological measurement examples: depressive symptoms scales, personality inventories, etc.
Data kinds: scales and measurements can be nominal, ordinal, interval, or ratio; examples include mood scales, test scores.
Open-ended data examples: interview transcripts, survey free-text responses.

Quantitative data uses numbers and scales; common analyses include looking at relationships between numeric variables.
Correlation concept:
- Measures linear relationship between two numerical variables.
- Math representation: r = \frac{\operatorname{cov}(X,Y)}{\sigmaX \sigmaY}, with r \in [-1,1].
- A scatterplot helps judge linearity and strength of the relationship.
Linear relationships and correlation: strong linear patterns show high magnitude of r; non-linear patterns may have weak or misleading r even if there is a relationship.
Causation vs correlation: correlation alone does not establish cause; third variables may influence both variables.
Experimental design basics (to infer causality):
- Independent Variable (IV): the variable manipulated by the researcher.
- Levels/groups: e.g., heavy smoke, moderate smoke, no smoke.
- Dependent Variable (DV): the outcome measured.
- Random assignment: participants randomly assigned to IV groups to ensure equivalence.
- Manipulation of IV should not involve altering DV data; DV is observed as it occurs.
Example structure for an experiment:
- IV: smoking exposure with levels {\text{heavy}, \text{moderate}, \text{none}}
- DV: lung cancer indicators or related outcomes.
Interpreting results: if groups differ on DV, infer possible effect of IV; must control for confounds and consider alternative explanations.

Qualitative data: words, transcripts, observations, discourses.
Data collection: interviews, discussions, ethnographic fieldnotes.
Analysis focuses on themes, patterns, and meanings rather than numerical summaries.
Common ethnographic approach: long-term immersion in a culture to understand everyday life; example: extended fieldwork (e.g., years) in a specific setting.

Observational/Correlation studies: explore relationships without manipulating IV; cannot definitively claim causation.
Experimental studies: researcher manipulates IV, uses random assignment, and observes DV to infer causality.
Key terminology:
- IV: independent variable (causing variable).
- DV: dependent variable (outcome).
- Random assignment: distributing participants to groups by chance.
- Levels/groups: different conditions of the IV.
Important cautions:
- Third-variable problem: a third factor may influence both IV and DV.
- Manipulation of DV is inappropriate; researchers only observe DV.

Demographic variables: gender, education, etc. (often used for description, not primary variables in some studies).
Psychological variables (often central): depressive symptoms, attachment style, parenting style, mood, etc.
For surveys: include at least two psychological variables if studying psychological processes.
For experiments: typically one IV with multiple levels and a DV measuring cognitive/behavioral outcomes.

Topics should involve at least two variables (two-variable relationships are common in studies).
Examples of potential topics:
- Attachment style and romantic relationships (also possible extensions to friendship or family).
- Memory recall and digital media use or music exposure.
- Sleep quality and eating habits.
- Parenting style and academic performance.
- Child abuse effects on development (with focus on specific, concrete outcomes).
If using surveys, select psychological variables (e.g., depressive symptoms, attachment style, parenting style) rather than only demographics.
If opting for an experiment, define a clear IV and a relevant DV (e.g., sleep quality as IV and memory performance as DV).
Final topic selection can evolve; start with a topic and two variables, then adjust with team input.

Question-driven knowledge and literature synthesis are foundational.
Quantitative vs Qualitative data: numbers vs words/transcripts.
Correlation describes association; causation requires experimental manipulation and randomization.
IV, DV, random assignment, and experimental levels are core to design.
Ethnography emphasizes long-term, immersive fieldwork to understand culture.
For last-minute prep, remember two-variable topics and the distinction between survey vs experimental methods.