Chapter 8 & 9 Flashcards
Experiments & Evaluation Research (38 cards)
Experiments are a procedure designed to test the _______ of some treatment on some outcome
effect
what is an effect
language of causality (relationship; neg. association, pos. association)
what is a treatment and what variable is it?
focus on 1 cause
Treatment is an independent variable
what is an outcome and what variable is it?
focus on 1 effect
outcome is dependent variable
what are the three causality conditions
correlation
time order
non-spurious
what does correlation mean?
must have relationship
what does time-order mean?
cause must be before effect
what does non-spurious mean?
relationship between IV and DV (NO 3rd influencing factor)
What are the advantages of experiments and what it does?
study the same person/ group
- eliminates all confounders (3rd spurious factor)
- best way to confirm non-spuriousness
- relationship
- groups must be similar in other characteristics to ensure no bias
Experiments are the gold standard of ________ and has high ___________.
causality; internal validity
Steps to create an experiment
- create environment
- Manipulate IV (design treatment/ how to operationalize)
- hold other variables constant (random assignment)
- compare & measure outcomes
- debrief
how do you create an environment for an experiment and what’s a fault?
- participants should be actively engaged (purpose/ rationale)
- use cover story to keep participants engaged without revealing study’s true purpose
- social desirability bias
how do you manipulate the independent variable?
design the treatment (IV)
- must be salient to participants (debrief)
- shouldn’t be strong enough for people to guess study’s hypothesis
- confederates (pretends to be participants)
- between-subject design
- within subject design
- mixed
what is between-subject design
randomly assigned to different levels of IV (split people into 2 groups, each group assigned condition)
what is within-subject design
receive all levels of IV (same person under 2 conditions)
how do you hold other variables constant?
- participants randomly assigned to treatment/ experimental group or control group
- distribute differences equally
- randomly assign study participants to groups (cross-group equivalence)
what are the types of random assignment?
- random assignment
- matched pairs (match on things associated with outcome like age, religion, randomly assign treatment within each pair)
- block design/ block-group design (groups generally similar people together instead of pairs)
how do you use a block design/ block-group design?
randomly assign treatment and control within each block
- group similar people together
what are the types of measures you can use?
- behavioral (actions, tripping/ falling)
- attitudinal (self-reported, social desirability bias)
- physiological (biological, heart rate)
how do you debrief?
- explain the deception
- ensure they understand it wasn’t real
- give participants opportunity to learn
types of experiments
- field experiment
- audit studies
- survey experiments
- quasi-experiments
what is a field experiment?
- takes place in natural setting
- policy interventions to improve educational or health outcomes
- double-blind study (research & participant doesn’t know which condition they’re in)
what are audit studies?
- assess if characteristics like gender/ race
- factor design (2 or more independent variables)
what are survey experiments?
- survey methods
- representative sample
- participants read description of scenario, then answer questions on how they would react
