Lecture 5 Flashcards
Measurement Error
Parallax (sight), Calibration, Zero, Damage, Limit of reading of the measurement device
Types of Errors:
Gross (human), Systematic (Instrumental, environmental & observational), Random (caused by unknown disturbances)
Significance Level
= alpha = error rate = probability of rejecting null hypothesis when it’s true
P-value
Probability of obtaining an effect as extreme as the one in the sample data assuming the null hypothesis is true
T1 Error
Study reveals a difference when there is no actual difference
T2 Error
Study reveals no difference when there actually is one
Bias
Factors that operate on a sample that make it unrepresentative of the population –> Large enough samples will eliminate unknown factors that cause bias
Confirmation Bias
Finding what you’re looking for
Recording Biases
Recall bias, only remember first or last things said
Halo effect
Non-experimental variables affect experimental measures
Social Desirability Bias
Participant selectively reports positive information to the experimenter
Behavioural Bias
Expectency can cause participants to behave differently
4 types of participant
Good, Bad, Faithful, Apprehensive
Good Participant
Confirms experimenters hypothesis
Bad Participant
Disconfirms Ha
Faithful Participant
Follows instructions to the t
Apprehensive Participant
Unusually concerned with experimenter’s evaluation of him/her
Pygmalion Effect
Experimenter causes real changes in participants due to changes in his/her behavior during the experiment.
Hawthorne Effect
Refers to any scenario in which participants alter their behavior because they’re being watched
Halo Effect
Uncontrolled novelty of treatment (demonstrate sig improvements in the short term (8 months))
Placebo effect
Spontaneous remission, non-specific effects & re-interpretation of ourcome measures
Biosocial Experimenter cues
Age, sex & attractiveness –> for age & sex, can be determined by congruence, or it can be a constant effect
Psychosocial Experimenter cues
Warmth, status, etc.
Reducing Expectancy Effects
Standardize experimenter-participant interaction, use blinding techniques, use deception, convince participants that you can detect lying.
Safeguards against misleading studies
Competition for funding, disseminated results in a peer reviewed journal & replication guards from T2 errors and invisible bias
