MCRS: Correlation, research design & content anal

Unit non-response Certain people do not participate in study

Consequences of unit non-response Need larger sample. External population validity lower. May not have reached specific subgroups

Cross-sectional survey Capturing opinion at one point in time (pizza-slice)

Trend (Longitudinal survey) Examines change in population over time - samples different individuals

Cohort Follows cohort over time but samples from different individuals

Panel Examines change in individuals across time; Same individuals from same population

Correlational design We can only say something about relationships, not make causal claims

Quasi-experiment Two-group pretest posttest design

Between subjects design Compares scores from two groups

Within subjects design Compares scores from same group (Z-tests)

Random assignment Ruling out alternative explanations, cancelling out variations in participant characteristics.

Factorial design Several independent variables presented in combination. 2by2 design > 4 groups

Internal validity experimental design Only experiments are suited to test causality

External validity experimental design "Is the causal relationship we found applicable to other situations?"

Cross-lagged survey Only survey that can assess causality by measuring dependent and independent variables at 2 points in time

Example leading question "When did you first start plagiarizing your papers?"

Example double-barreled question "Do you think building X is nice or would you rather want it demolished?"

Funnel format question Starts with broad questions and progresses to more specific questions

Ecological isomorphism When the experimental condition is not the same as condition in outside world - Difficult to replicate

X as an expression Manipulation of a variable, what is done in the experimental/controlgroups

R as an expression Random assignment of individuals into groups (experimental design)

O1 and O2 etc as an expression Before and after measures in pretest-posttest design

Selection bias Experimental groups are not comparable

Attrition People drop out of a study

Diffusion Treatment effect spreads across groups

Manipulation check Check how the participants interpreted the manipulated variable

Content analysis Research technique that aims to quantify content in terms of predetermined categories in a systematic and replicable manner

RQ & Hypothesis in content analysis Comparing one to another, not hypothesizing about humans

Universe Population in CA

Checklist of a good Content Analysis variable 1 Variables and categories must be clearly defined and explained

Checklist of a good Content Analysis variable 2 Exhaustive answer categories (make sure all possible categories are accounted for)

Checklist of a good Content Analysis variable 3 Mutually exclusive answer categories (Must not overlap)

Coding unit Level where we make coding decisions

Context unit How much info we use when coding

Stratified sampling in CA "Layers" of data

Codebook Provides rules of a coding scheme

Reiterative process (think evaluation part in empirical cycle) Creating a codebook

Klicka