Experimental vs non-experimental Methods Flashcards
Experimental Design
Tests effect of manipulation of IV on DV
Based on hypothesis testing; Karl Popper, falsification
Experimental Manipulation
IV is categorical variable
Different groups or distinct interventions e.g. before & after
Reasons variables may be beyond experimental control
Ethics
Impractical/tricky
Too many variables or levels to meaningfully control
Pros of non-experimental methods
Can ask fundamentally different questions
Can look at wider range of contributing factors; can try & control quantitative variables statistically if not experimentally
Produces far deeper & richer data
Confirmatory Data Analysis
Have an idea you want to prove/disprove: top-down
Deductive approach
Exploratory Data Analysis
Driven by patterns that emerge from data, not necessarily guided by pre-existing theory
Inductive approach
Bottom-up
Regression
Typically used in non-experimental studies
Goes beyond simple correlation
Types of Regression
Simple Linear; one predictor, one outcome (both continuous variables)
Multiple; multiple predictors e.g. anxiety, gender, one outcome e.g. depression score (predictors usually continuous but can include categorical variables)
Logistic; multiple predictors, one outcome (typically continuous predictors, outcome is categorical)
Mediation & moderation
Extension to regression analysis, shows relationships between predictors
Moderation; is listening to country music predicts happiness, but only if you’re a country music fan
Mediation; if liking & playing fortnight both predict how good you are at it, but liking fortnight more means you play it more
Qualitative Analysis
Asks fundamentally different questions
More focused on human experience & subjectivity
Thematic Analysis
One kind of analysis within qualitative methods
Looks at transcripts or any other text & tries to identify themes being discussed
Factor Analysis
Way of chunking data into theoretically relevant subscales or factors
Looks at underlying structure of (questionnaire) data
