Exam 1 Flashcards
Scientific Method
Means of gathering information, making conclusions and testing those conclusions
Types of questions: descriptions
what happens
Types of questions: prediction
when does it happen
Types of questions: explanation
why does it happen
Theory
- general explanation
- must be falsifiable
Hypothesis
specific testable prediction
Operalization
Specific concrete way of thinking about a psychological construct
(there are many ways to operationalize the same psychological construct
Construct validity
the degree to which the operalization used to reflect the psychological construct of interest
Types of measures
Self reports
Observational
Psychological
Archival
Self reports
- Could be an open-ended or a fixed response
- must watch out for social desirability concerns
- sentiment override
Observational measures
inter-rater reliability is key
Sample
People from whom data are collected
Population
group about which the researcher wants to draw conclusions from
Obtaining participants
Convenience sample
Representative sample
Correlation is necessary for (but does not imply) causation meaning
X may cause Y
Y may cause X
OR some other variable may cause X and Y
Longitudanal design
- correlation
- daily diary
- experience sampling
experimental design requirements
Dependent variable
independent variable
control
random assignment
External Validity
-Extent to which results can be generalized
Internal validity
extent to which it is possible to draw conlusions about the hypotheses from the data
Extraneous variable
variable capable of explaining findings
Threats to internal validity
- social desirability
- demand characteristics
- organismic variables
- selective or distorted memory
Invalid operational definitions
Item overlap problem
Item overlap problem (content contamination)
In psychology constructs of depression anxiety and neuroticism covary
Item overlap problem (content contamination) : depression
Sadness
- i don’t feel sad
- i feel sad much of the time
- i feel sad all of the time
Item overlap problem (content contamination) : Anxiety
i feel sad
Item overlap problem (content contamination) : neuroticism
i am seldom sad or depressed
Statistical Significance
The researcher assumes the null hypothesis is true, collects data, and examines how likely it would be to get those data is the null hypothesis is true
P value
Risk of mistakenly rejecting null hypothesis
if P < .05 the null hypothesis is rejected
—-If this happens the research hypothesis is supported and the result is statistically significant
Limitations to statistical significance : Publication Bias
publish only statistically significant results (rather than null results)
Clinical/Practical significance : Effect Size
- Degree to which the phenomenon is present in the population (degree to which the null hypothesis is false)
- Used to estimate power (probability that is a statistiacal test will lead to rejection of the null hypothesis)
R(rsquared = coefficient of determinations: Cohen’s d guidelines for interpretations
Small = .10 medium = .30 Large = . 50
R(rsquared = coefficient of determinations: Gignac and Szodorai’s (2016) guidelines for interpretation
- Recommended guidlines based on empirical evaluation of 708 correlations in meta-analyses published in 6 journals
- small (25th percentile ) : .10
- medium (50th percentile ): .20
- Large (75th percentile): .30
Importance of replication :
- Researchers test aspects of theories through many different studies
- Contradictory and initially stronger effects in Highly Cited Clinical research
- Estimating the reproductability Psychological Science
Importance of replication : Researchers test aspects of theories through many different studies
- The more studies there are that demonstrate results supporting the theory the more confidence researcher have in theory
- If study results contradicts a theory the theory is modified and the modified theory is tested
Importance of replication : Contradictory initially stronger effects in Highly Cited Clinical research
- Evaluated highly-cited (>1000) original clinical research studies
- Of 49 studies, 45 claimed an intervention was effective
- —16% contradicted the subsequent studies
- —16% found effects stronger than subsequent studies
- —44% were replicated
- —24% remained largely unchallenged
Importance of replication: Estimating the reproductability Psychological Science
- Replication of 100 experimental and correlational studies in social and cognitive psychology using highly powered designs and original material when availabe
- mean effect size of replication effects ( r= .20) was half magnitude of original effects (r =.20)
- -97% of original studies had significant results (p< .05) comparted to 36% of replication studies
Meta analysis definition
Statistical methodof individual studies
— By combining results across studies, we (a) increase power, and (b) improve the precision of our estimate of effect size