Lectures (Midterm I) Flashcards
Confirmation Bias
Noticing supporting evidence and disregarding disconfirming evidence
Experimental control
There must be an appropriate control condition in order to avoid confounds
Bias
Measurement error in a particular direction - expectancy effects. This is why blind and double blind studies important
Reliability
Replicability of a result. High correlation if you retry experiment with same measure. Inter-rater reliability.
Validity
How legitimate a measurement is for measuring the thing you want it to. Both internal validity (experimental cohesiveness) and external validity (generalizability) important
Internal validity
How cohesive an experiment is - whether it avoids confounding (more than one possible independent variable acting at the same time). Less confounding -> higher internal validity
Statistical significance
Unlikely to occur by chance
Correlation coefficients
Pearson’s correlation coefficient - “r”
-1 to +1
Descriptive statistics
Summarize/describe data: mean, median, mode, etc.
Regression analysis
Analysis: how the of the dependent variable changes when any one of the independent variables is varied, while the other independent variables are held fixed.
Median split analysis
Dividing the sample in two categories via the median for analysis - can compare means using t-tst
Mediator variables
The independent variable influences the (non-observable) mediator variable, which in turn influences the dependent variable.
MS vs. MTS
MS measure is multidimensional, MTS is unidimensional. MTS developed because of concerns of problematic questions in MS.
t-tests
Look at whether there is significant difference between means of different groups. Can use in median split analysis. Different from correlations, which look at relationships between variables.
p-value
How likely are we to obtain the pattern of data by chance? Before p-value test is performed, a threshold (alpha value) of usually 5% is chosen. A small p-value (p
Alpha level
Usually set at 5%. If p-value is less than alpha then there is a 5% or less likelihood that could have gotten results indicating relationship if no relationship really present.
Operational definition
Defining something according to the process that is used to measure it.
Null hypothesis
Default position that there is no relationship between the variables
Effect size
How strong the relationship between the variables is
Motivated reasoning
Motivation can cause confirmation bias. If hypothesis would be good for you if true, look for positive outcomes and confirmation. If negative, then either more objective or look for disconfirming evidence.
Possible reasons for confirmation bias
Reduce cognitive dissonance, ease of processing positives, motivations/meeting goals
Dual task paradigm
Two independent variables - a primary task and a secondary task. There may be interactions between the two - a specific condition in one may lead to a specific condition in the other. Adjust parameters in both conditions (2x2) to determine effects. DVs for both the primary and secondary tasks
Accuracy
How well a measuring instrument is able to measure - calibration, decimal places, etc.
Face validity
Extent to which a test is subjectively viewed as covering the concept it purports to measure.
Content validity
Extent to which a measure represents all facets of a given construct.
Criterion related validity
Concurrent, predictive: Extent to which a measure is related to an outcome.
Construct validity
The degree to which a test measures what it claims, or purports, to be measuring