Lecture 02 Flashcards
What can we find in research papers? Name 4 of them.
- Title, authors, institutions
- Abstract
- Introduction
- Method
- Results
- Discussion
- References
Statistical power:
1. What is it?
2. When should it be calculated?
3. What is its range?
4. How much -> interpret findings that are not statistically significant with reasonable confidence?
- Power estimates the probability of detecting a statistically significant difference in the sample if no one exists in the population.
- Planning stage
- 0.00-1.00
- 0.80
What does it mean when p < 0.05?
Statistically significant, true effect may not be zero
Other than p-values, what else can we use to report research significance?
Effect size and confidence intervals
ES: estimates the average size of treatment, measures the average difference between sample groups
Comparison of independent means?
Small?
Medium?
Large?
d
Small: 0.20
Medium: 0.50
Large: 0.80
Define validity.
Whether a test accurately reflects the CONSTRUCT that it was designed to measure.
度到要度嘅野？
What is the difference between internal and external validity?
Internal: evidence -> reflection of patients/procedures/settings that were observed
External: not observed
Controlled + Experimental?
Controlled only?
Experimental only?
Neither?
Controlled + experimental: controlled trial, multiple baseline
Controlled: cohort, case-control, cross-sectional
Experimental only: uncontrolled trials
Neither: case report, case series
What are the differences between a prospective and a retrospective study?
Prospective: collect data
Retrospective: analyse existing data
Prospective: identify and recruit participants
Retrospective: no control to participants
What do Type I and Type II error mean?
Type I: concluded difference exists but no difference is true (rejected H0)
Type II: conclude H0 as correct but have difference in fact (X reject H0)
