1Why do we need statistical inference?
Because we usually observe a sample rather than the entire population, so statistics help estimate unknown population values.
2What is a population parameter?
The true but usually unobserved value of a measurement in the entire population.
3What is a sample estimate?
A value calculated from sample data that is used to estimate the true population parameter.
4What is the relationship between statistical and scientific inference?
Statistical inference is a subset of scientific inference.
5Why must we quantify uncertainty in an estimate?
Because a sample estimate is not exactly equal to the true population value.
6What is the null hypothesis (H₀)?
The assumption that there is no difference or association between the groups being compared.
7What is the purpose of hypothesis testing?
To determine how plausible the observed data would be if the null hypothesis were true.
8Do we “accept” the null hypothesis?
No. We either reject or fail to reject the null hypothesis.
9What does a p-value measure?
The probability of obtaining the observed data, or something more extreme, assuming the null hypothesis is true.
10What does p ≤ 0.05 generally indicate? low
The result is considered statistically significant and provides evidence to reject the null hypothesis.
11What does p > 0.05 generally indicate? high
The data are compatible with random variation, so we fail to reject the null hypothesis.
12An NSAID study in arthritic dogs finds p = 0.03. What does this mean?
If the drug truly had no effect, a difference this large or larger would occur in about 3% of similar studies by chance.
13If p = 0.03, does this mean there is a 3% chance the drug does not work?
No. A p-value does NOT give the probability that the null hypothesis is true.
14If p = 0.03, does this mean there is a 97% chance the drug works?
No.
15Does a small p-value tell you that an effect is clinically important?
No. Statistical significance does not automatically mean clinical significance.
16What two questions should you always ask when evaluating a study result?
1) Is it statistically significant? 2) Is the effect clinically meaningful?
17What is statistical significance?
Evidence that the observed result is unlikely under the null hypothesis according to the chosen significance threshold.
18What is clinical significance?
Whether the size of an observed effect is large enough to actually matter clinically.
19Why can a tiny effect become statistically significant in a very large study?
Large sample sizes can detect very small differences, even if those differences are biologically or clinically trivial.
20Why should effect size be reported along with a p-value?
Effect size tells you the magnitude of the difference or association, while the p-value addresses statistical evidence.
21What does a confidence interval help describe?
The uncertainty or margin of error surrounding an estimate. Lecture63.pdf
22A study reports P = 0.15. Can you say “there was no difference between groups”?
No. You should say there was no statistically significant difference; non-significance does not prove the groups are equivalent.
23What does failing to reject H₀ mean?
There is insufficient evidence to reject the null hypothesis; it does NOT prove H₀ is true. Lecture63.pdf
24What is a Type I error?
A false positive—rejecting H₀ when there is actually no real effect.
25What is a Type II error?
A false negative—failing to reject H₀ when a real effect actually exists.
26Type I error
False positive.
27Type II error
False negative.
28What is statistical power?
The ability to detect a real effect; power = 1 − β. Lecture63.pdf
29What measure of association is directly calculated from a cohort study?
Risk ratio (RR).
30What does a cohort study compare?
The probabilityisk of developing disease in exposed versus unexposed animals.
31What measure of association is typically calculated from a case-control study?
Odds ratio (OR).
32What does a case-control study compare?
The odds of exposure among cases versus controls.
33Cohort = ?
Risk → Risk Ratio.
34Case-control = ?
Odds → Odds Ratio. Lecture63.pdf
35Why can’t you directly calculate disease risk from a case-control study?
Researchers predetermine how many cases and controls are included, so their proportions do not represent natural disease incidence in the population.
36What is the odds ratio formula for a 2×2 table?
OR = (A × D) ÷ (B × C).
37What is the null value for an odds ratio?
OR = 1.
38What does OR = 1 mean?
There is no association between the exposure and disease.
39What does OR > 1 mean?
There is a positive association; the exposure is associated with increased odds of disease.
40What does OR < 1 mean?
There is a negative association; the exposure may have a protective effect by reducing the odds of disease.
41Does OR > 1 prove that the exposure caused the disease?
No. Association is NOT causation. Lecture63.pdf
42In the leptospirosis example, what exposure was investigated?
Exposure of dogs to rainwater puddles.
43In the leptospirosis case-control study, what were the 2×2 values?
Lepto+ dogs: 60 exposed and 40 unexposed; Lepto− controls: 30 exposed and 70 unexposed.
44What was the odds ratio for rainwater puddle exposure and leptospirosis?
OR = (60 × 70) ÷ (40 × 30) = 3.5.
45How should OR = 3.5 be interpreted in the leptospirosis study?
Dogs with rainwater puddle exposure had 3.5 times the odds of being a leptospirosis case compared with dogs without that exposure.
46Why is it WRONG to say the exposed dogs had “3.5 times the risk” in this case-control study?
Case-control studies directly estimate odds, not disease risk.
47What is the relationship between probability and odds?
Odds = p/(1−p).
48When can an odds ratio approximate a risk ratio?
When the disease is rare.
49What rare-disease cutoff is given in this lecture for OR to approximate RR?
Disease probability/incidence less than about 5%.
50Why are case-control studies efficient for rare diseases?
A cohort study might require following thousands or millions of animals to obtain enough cases, while a case-control study deliberately selects cases and controls.
51BIG EXAM RULE: Rare disease + case-control study
Use the odds ratio; when disease is rare (<5%), OR can approximate RR.
52BIG EXAM RULE: Cohort vs. case-control
Cohort studies directly estimate disease risk and RR; case-control studies estimate exposure odds and OR. Lecture63.pdf
53What is the overall scientific inference process?
Estimate an unknown population parameter from sample data → quantify uncertainty → test H₀ using p-values and confidence intervals → reject or fail to reject H₀ → consider Type I and Type II errors.
54What is the most important rule when interpreting scientific evidence?
Do not rely only on statistical significance; consider effect size, uncertainty, clinical importance, study design, and whether the association actually supports a meaningful conclusion.