Statistics Flashcards
Power analysis
calculation to determine the minimum sample size needed to reasonably detect a given effect
Definition of null hypothesis
null hypothesis means there is no difference
Power before and after results
before= sample size (a priori) after= effect size (post hoc)
Confidence interval
margin of error
a range in which response is likely to be accurate
Confidence level
how confident you are that something will happen
Confidence level that results can be reproduced
Absolute risk reduction
risk in control minus the risk in a treatment group
Relative risk reduction
percentage of risk removed by treatment
ARR/risk group= % RRR
Reliability
concept that results can be repeated
Validity
results indicate what researcher thinks they indicate
Internal validity
(Design of study) great results are invalidated by poor design
External validity
assessment of the inferences or causal relationships
design and process are good but the inference is not supported
Can
Quantitative relationships
Causal- smoking causes cancer
Correlational- smoking and HPV
Qualitative
Exploratory
ethnography- lived experience
grounded theory- theory presents after study
Prospective
explore going forward
Data are collected and evaluated
Retrospective
existing data is analyzed