6: Statistical Hypothesis Testing & P-value ✅ Flashcards
What do we look at when assessing association?
The presence and magnitude of association
What is it important to state with high certainty?
Whether an association exists in the source population
->based on sample estimates
2 possibilities with association
Association doesn’t exist in the population
Association does exist in the population
Null hypothesis always states..
That there is no association between the 2 variables in the population
Alternative hypothesis always states..
There IS an association between the 2 variables in the population
Hypothesis testing: formal process
- Define the Null and Alternatives hypotheses
- Initially assume that NO association exists in the pop
- Define the significance level (what is sufficient evidence against Ho)
- Collect sample data from population (evidence)
- Does the sample estimate provide evidence against Ho?
- Could the sample estimate be explained by random error
- Calculate the test statistic
- Derive the probability that quantifies our belief against Ho (p-value)
- Interpret p-value
Lower p-value means
The lower the p-value, the more likely we are to reject Ho as the association is stronger
What significance level do we use for p-value?
Significance level of 5%
So: when p-value is LESS than 5%, we reject Ho
What greatly affects p-value?
Sample size and magnitude of association
-larger sample sizes = smaller p-values
-estimates of large magnitudes = smaller p-values
Mean difference
If the 95% confidence interval includes 0, Ho cannot be rejected
-> as 0 is likely a value in the source pop
Regression coefficient and correlation coefficient
If the 95% confidence interval includes 0, then Ho cannot be rejected as 0 is a likely value in the source population
Odds ratio/risk ratio/rate ratio
if the 95% confidence interval includes 1 then Ho cannot be rejected as it is a likely value in the source population
