Research Methods Review
Terms
undefined, object
copy deck
- Validity
- truth and legitimacy
- Threats to validity
- barriers between us and the truth
- Construct validity
- labeling
- Measurement validity
- accuracy of measurement
- Internal Validity
- how well the relationship found in the study reflects the truth of the sample of participants
- external validity
- how well the results derived from a sample reflect the truth about the target population
- history
- a threat to external and internal validity; events that occur during the time of the study
- attrition
- participants leaving the study before the study was completed
- regression to the mean
- extreme versions of phenomenon tend to become less extremeover time; internal validity
- maturation
- participants matured during the study; internal validity
- Selection
- some people may influence their chances of being selected.
- control
- any design feature or procedures used to counteract the threats to validity
- control is directed to what 3 aspects of a study
-
1. Sampling - on that is true/valid of the target population
2. Assignment - test the efficacy of a new form of treatment
3. Measurement - contraint
- the amount of autonomy a participant has during the course of of the study
- designs
- the blueprint of a study
- Quasi-experimental
- test/describe presumed cause and effect relationships between naturally occurring phenomena.
- Experimental
- test/describe presumed cause and effect relationships between manipulated variables and outcomes
- Random Sampling
- how your draw the people for your study; threat to external validity
- Random Assignment
- how you randomly assigned your selection to groups; threat to internal validity
- Scale of Measurement
- No Oils In Rivers
- Nominal Scale
- responses are names or categories; hard to make comparisons; is not in order
- Ordinal Scale
- ordered; responses ranged from least to most satisfied; allows comparisons; fail to capture important info
- Interval
- order is numerical; intervals are on a scale; zero does not mean the absence of something
- Ratio
- order is numerical; intervals are on a scale; sero means the absent of something
- Type 1 error
- rejected the null hypothesis when it was true; p-value is set in advanced; not affected by sample size
- alpha level
- cut off used to judge the p-value
- Type 2 error
- failed to reject the null hypothesis tested when a given alternative hypothesis was true
- p-value
- used to determine if there is a difference; the probability of wrongly rejecting the null hypothesis if is in fact true
- small p-values
- suggests that the null hypothesis is unlikely to be true; p<.05 will reject the null hypothesis
- high p-values
- p>.05 we fail to reject the null hypothesis *The difference is significantly significant
- central tendency
- estimate of the center of a distribution of values
- mean
- the average
- mode
- the number that you see the most
- median
- the number in the middle
- standard deviation
- desribes how the standard amount variables differ from the mean; variability
- frequency distribution
- variability
- nominal
- mode; frequency; bar chart and stem & leaf plot
- ordinal
- mode; range (variance); stem & leaf plot
- T-test
- assesses whether the mean of two groups are statistically different from each other (interval and ratio variable); group difference
- ANOVA
- same as t-test but with more than 2 groups; measures the difference between control and experimental group simultaneously; measures the different correlations; employ a nominal scale and predict a interval and ratio outcome
- correlation
- indicates the strength and direction of a linear relationship between two random variables; linear association
- chi-test
- use to test the association between two or more nominal scale variables; no cause and effect