Brainscape's Knowledge GenomeTM

Browse over 1 million classes created by top students, professors, publishers, and experts.


See full index

2206 RESEARCH METHODS > Week 9 : Experiments > Flashcards

Week 9 : Experiments Flashcards

1
Q

An experiment is…

A
  • A research method where the researcher manipulates one or more independent variable(s) to determine the effect(s) on a dependent variable
  • A procedure designed to test the EFFECT of some treatment on some outcome
How well did you know this?
1
Not at all
2
3
4
5
Perfectly
2
Q

Basic things about experiments…

A
  • quantitative method
  • non-probability sample
  • controlled settings
  • very few variables
How well did you know this?
1
Not at all
2
3
4
5
Perfectly
3
Q

All experiments share 3 key features…

A
  1. Manipulation of the independent variable
  2. Random assignment of participants to experimental & control conditions (essential for causality) - but is subject to random error
  3. experimental control of other factors that could influence the outcome of the experiment
How well did you know this?
1
Not at all
2
3
4
5
Perfectly
4
Q

Variables in Experiments

A
  • Independent variable - treatment, cause, predictor
  • Dependent variable - outcome, effect
How well did you know this?
1
Not at all
2
3
4
5
Perfectly
5
Q

3 Advantages of Experiments

A
  1. Experimental is the best research method to establish causality
  2. Experiments can uncover mechanisms that produce an outcome (quantify the effect)
  3. Experiments can be used to evaluate abstract theories about the workings of the social world (internal validity)
How well did you know this?
1
Not at all
2
3
4
5
Perfectly
6
Q

Causality - 3 Conditions

A
  1. The two variables must be correlated (positively/negatively)
  2. time order - the cause (IV) must precede the effect (DV)
  3. the relationship between the independent variable & dependent variable must not be spurious (caused by some other factor/confound) - hardest to satisfy
  • experiments are well suited for removing spuriousness/confounders because of controlled setting & random assignment
How well did you know this?
1
Not at all
2
3
4
5
Perfectly
7
Q

Key weaknesses in experiments

A
  • non-probability sample… target population is abstract (ppl), systematic errors & errors with generalizability
  • realness (treatment vs real life decisions)
  • small sample size (relative to surveys)
  • external validity
How well did you know this?
1
Not at all
2
3
4
5
Perfectly
8
Q

4 stages in designing a laboratory (‘true’) experiment

A
  1. create a setting that’s engaging & makes sense to the participant
  2. design treatment / manipulate independent variable while holding other factors constant & randomly assign
  3. Valid & reliable measure of dependent variable
  4. debrief & assess quality
How well did you know this?
1
Not at all
2
3
4
5
Perfectly
9
Q

4 stages in designing a laboratory (‘true’) experiment

1 - Creating the setting

A
  • Keep participants actively enaged in the experiment (a sense of purpose & rationale for thier participation
  • participants need to follow experimental procedures & pay attention to what’s being asked of them
  • participants should be comfortable & not distracted
  • Should not be suspicious about the study’s hypothesis (use a cover story)
How well did you know this?
1
Not at all
2
3
4
5
Perfectly
10
Q

4 stages in designing a laboratory (‘true’) experiment

2 - Manipulate independent variable

A
  • maniuplation must be salient to participants
  • but should not be so strong that participants can guess the study’s hypothesis
  • Use confederates - individuals who are trained to pretend to be study participants & interact with the participant (e.g. soccer fans & helping study)
How well did you know this?
1
Not at all
2
3
4
5
Perfectly
11
Q

4 stages in designing a laboratory (‘true’) experiment

2 - between-subject & within-subject designs

A
  1. between-subject design… participants are randomly assigned to different levels of the independent variable
  2. within-subject design… all the participants receive all levels of the independent variable (preferred cuz easier to control & require fewer participants)
  3. sometimes researchers use a mixture
How well did you know this?
1
Not at all
2
3
4
5
Perfectly
12
Q

4 stages in designing a laboratory (‘true’) experiment

3 - Measuring the dependent variable (3 ways)

A
  1. behavioural measures - observation of overt & observable actions (Commonly used to measure whether stereotypes have a negative effect)
  2. attitudinal measures - Self-reports in which participants respond to questions about their attitudes, opinions & beliefs (susceptible to social desireability bias)
  3. Physiological measures - biological responses to stimuli (hard & expensive)
How well did you know this?
1
Not at all
2
3
4
5
Perfectly
13
Q

4 stages in designing a laboratory (‘true’) experiment

4 - wrapping up the experiment with debriefing

A
  • explain to the participants that the decetion served a scientific purpose
  • ensure that the participants understand the deception was not real
  • give the participants an opportunity to learn about the research
How well did you know this?
1
Not at all
2
3
4
5
Perfectly
14
Q

Random assignment

A
  • participants randomly assigned to treatment/control groups
  • random assignment removes systematic differences between the two groups
  • on average leads to identical groups, but this is subject to random error
How well did you know this?
1
Not at all
2
3
4
5
Perfectly
15
Q

3 kinds of random assignment

1 - true random assignment

A
  • random numner generator ideal
  • but for example, each person takes their turn to flip a coin to get seperated into the treatment/control groups
How well did you know this?
1
Not at all
2
3
4
5
Perfectly
16
Q

3 kinds of random assignment

2 - matched pairs

A
  • A matched pairs design is an experimental design where researchers match participants by characteristics and assign them to different groups.
  • match on things that might be associated with outcome/effect (e.g. drinking)
  • randomly assign treatment within each pair
17
Q

3 kinds of random assignment

3 - Block group design

A
  • divide participants into homogeneous subsets called blocks based on specific characteristics, such as age or gender, to ensure balanced representation across experimental conditions
  • randomly assign to treatment/control within each block
  • e.g. in a study evaluating the effectiveness of a new teaching method, participants could be grouped into blocks based on their initial academic performance levels to ensure that each teaching method condition has an equal representation of high, medium, and low-performing students.
18
Q

5 newer types of experiments that go outside of the lab…

A
  1. online experiments
  2. field experiments
  3. audit studies
  4. survye experiments
  5. quasi-experiments
19
Q

1 - online experiments

A
  • similar to laboratory experiments but conducted online
  • bigger sample size
  • have to ensure to make the treatment salient
20
Q

2 - field experiments

A
  • implement the treatment in the real world
  • requires lots of resources, organization, logistics & time
  • ethical considerations
  • Often used to evaluate the success of interventions to improve educational & health outcomes
  • can have a double-blind design
21
Q

3 - audit studies

A
  • study discrimination in labour & housing markets
  • e.g. send fake applications to job/house postings, (treatment - characteristics of the job applicants) (outcome - successful application)
  • factorial design - can have 2 or more independent variables (e.g. combine race (black/white) with criminal record (yes/no) so that’s 4 different groups)
  • READING THIS WEEK WAS THIS!! (Pager et al.)
22
Q

4 - population based survey experiments

A
  • Called ‘survey’ experiments cuz they rely on survey methods to recruit a representative sample of participants
  • probability sample
  • treatment - participants read a description of a scenario
  • outcome - then they answer questions about how they would react to the given situation
  • still treamtnet & control group with different scenarios given
  • higher degree of external validity than other types (& lower internal)
23
Q

5 - quasi-experiments (natural)

A
  • experiment without random assignment
  • commonly used to provide evidences to advocate for policy changes
  • sometimes the treatment occurs naturally (like the orchestra blind auditions with more women hired OR natural disasters for example)
  • main task is to find the treatment & control groups (a difference-in-difference design)
24
Q

Overall strengths of experiments

A
  • internal validity
  • working with abstract theory
  • possible to test specific mechanisms
25
Q

overall limitations of experiments

A
  • ethical issues… consider costs vs benefits, consent in field experiments, etc.
  • reproductibility… replication crisis, open data!
  • external validity… fix w/ triangulation & replication with different samples
26
Q

When to use experiements

A
  • explanatory research questions
  • deductive
  • abstract theory
  • known empirical association
  • knowable time order
  • focus on a single dependent variable (effect)
  • focus on a single independent variable (cause)

2206 RESEARCH METHODS (11 decks)

Brainscape helps you reach your goals faster, through stronger study habits.
© 2024 Bold Learning Solutions. Terms and Conditions