Lecture 3: Basics of Inferential Statistics i Flashcards
What are the Methods of inferential statistics?
Estimation
Hypothesis testing
What is Estimation in inferential statistics?
Point and interval estimation
What is Hypothesis testing inferential statistics?
Hypothesis testing in statistics is a way to test the results of a survey or experiment to see if you have meaningful results.
You’re basically testing whether your results are valid by figuring out the probability that your results have happened by chance.
•The smaller the probability: the more convinced we are that our results reflect the truth and is statistically significant
In Hypothesis testing, the smaller the probability: the more convinced we are that our results reflect the truth and is statistically significant
True or false?
True
What are the 5 steps of Hypothesis testing?
1- State the null hypothesis (H0)
2- State the alternative hypothesis (H1 or HA )
3- Choose your significance level, what kind of test you need to perform
4- Find the p-value
5- Reach a conclusion
What is the null Hypothesis?
What does it state?
H0
It states: There is no difference things are the same as each other or the same as a theoretical expectation
What is the alternative hypothesis?
What does it state?
Ha or H1
It states that things are different from each other, or different from a theoretical expectation.
What is the significance level?
α (alpha)
Is the probability that results could have occurred by chance.
When do we say the event is significant?
If the level is quite low, that is, the probability of occurring by chance is quite small, we say the event is significant.
This is a threshold of the amount of error we are willing to accept.
Usually below 0.05 (5%)
What are P values?
- P values are used to determine statistical significance in a hypothesis test.
- Probability that an effect/difference is due to chance
- Probability of Type I error or α (alpha) error
What does it mean If P < significance value?
It means obtaining that result purely by chance is small
So there is a significant difference
What does it mean If P value > 0.05?
The probability that results are purely do to chance is greater than 5%, so conclude no significance
So no significant difference
What is Type I error?
False positive results - α error
Falsely concluding a statistically significant relationship exists when in fact it does not.
Acceptable error at 0.05
What causes type I error?
- Type I alpha error is due to ‘random’ chance associated with sampling.
- “Fishing Expedition”: If you perform multiple tests, the chance of random error increases.
What is Fishing Expedition ?
If you perform multiple tests, the chance of random error increases.
What is Type II error?
FALSE NEGATIVE RESULTS - β error
Set at 0.2
What is the Power of a test?
POWER= 1 - β
= 0.8 or 80%
What cause Type II error?
•Sample size (total and within group)
–Small groups have lower power
–Unbalanced groups have lower power
•The effect size
–It’s easier to find a LARGE effect.
What is the Power at a test ?
The ability to reject the null hypothesis (H0) when its false
The ability to correctly to find a true effect
Power = 1-𝜷 (it is the inverse of the likelihood of making a Type II error)
If there is a statistically significant difference, is it clinically relevant
True or false?
True
This is particularly important if the sample size is very large
Statistical significance implies clinical importance
True or false?
False
Statistical significance does not imply clinical importance
Statistical significance ≠ clinical significance
True or false?
True
If there is no statistically significant difference it doesn’t necessarily mean there isn’t a difference
True or false?
True, consider the sample size in conjunction with the p-value.
It might just mean there wasn’t enough evidence
What is the concento of Statistical significance ≠ clinical significance?
- Is there a difference vs. does this difference matter?
* A statistically significant effect does not always imply a clinical and public health significance
What are the 3 things you need to do when reading am academic paper?
- Always consider what the research question is
- Estimate: The measure of effect (mean, mean difference, risk ratio, odds ratio)
•Finally, what is the P value?
–Significant - yes/no
–Due to chance - yes/no
–Significance level: is it 0.05 or otherwise?
P value indicates statistical significance but not clinical significance
True or false?
True
Type I or type II error occurs when we are wrong in our conclusions
True or false
True
