Critical Appraisal Flashcards
Hawthorne Effect
When the participant responds in a way they think is desirable because they realise they are a participant in a study
Actor-Observer Bias
When you attribute the behaviour of others to internal factors but the same thing in yourself to external factors: an attribution bias
Confirmation Bias
When the experimenters own beliefs result in them being more prone to reporting findings in their beliefs favour
Information / measurement biases
Where one of the key variables is inaccurately measured
Recall Bias
When participants are asked to (self-report) recall events from the past. This is biased because recall will lean towards more unusual info than more routine info
Observer Bias
The tendency for the experimenter to observe what they expect to happen rather than what’s actually happening
You can reduce observer bias by….
…using single or double blinding
Recall Bias can be reduced by…
…ensuring the Control group has the same level of Recall Bias
Performance Bias
Where partipant pre-knowledge of their assigned intervention affects how they perform
Performance Bias can be reduced by…
Blinding
John Henry effect
When participants alter their behaviour because they are aware they are being compared to the experimental group
Regression to the Mean
When an initial measurement of the variable is extremely far from the mean and a second measurement is closer to the mean: so you wrong infer that change has happened.
What’s a
Interviewer Bias
When either 1) the method of interviewing or 2) characteristics of the interview could bias the data captured
Publication Bias
When studies tend to only report statistically significant results
Researcher bias
When the researchers beliefs or expectations impact (however consciously or not) the data collection + whole study
Pygmalion (Or Rosenthal) Effect
Where the researchers high expectations for participant performance leads to better performance
Response Biases
Where respondents tend to provide inaccurate or false answers because of something in the experiment impacting on how they Will respond.
Acquiescence bias (“yes-pleasing”)
The tendency for participants to to agree with a “yes or no” response
Demand Characteristic
When behaviour of the experimenter signals to the participant what they would like to see (I.e. what the ‘right answers’ are
Social desirability bias
Tendency to provide answers that they feel are more socially desirable
Courtesy Bias
The tendency to avoid giving the questioner negative feedback out of politeness
Question Order Bias
When the order of questions impacts upon responses given
Extreme responding
The tendency to report extreme answers when given an array of options
Extreme responding can be countered by…
….using multiple types of outcome measures
Selection Bias
When the method of participant selection isn’t unbiased
Attrition Bias
When the people that drop out of the study ends up skewing the data you Still have
Volunteer or self-selection bias
When self-volunteers have characteristics that matter to the study
Nonresponse Bias
When those who did not respond to a survey differ from those who did(the data could.have been different had they replied)
Reduce nonresponding bias by….
Offering the survey/questionnaire in multiple formats
Undercoverage Bias
When you only sample from a subset of your population (e.g. only offering a survey to pensioners online. This excludes this with no access to the Internet)
Validity
How accurate the research was
Reliability
How consistent the research is: how replicatable is it
Is a Reliable research always Valid?
No. The results might be reproducible, but not necessarily correct
