Hygiene Atlas

Atlas · Plate XIX · Community Health and Research · 05

Biostatistics

Biostatistics items test whether you can read results. Know mean, median and mode, how skew pulls them apart, and why p under 0.05 is the usual cutoff for significance. Match the test to the data: chi-square for categories, a t-test for two means, ANOVA for three or more. Sensitivity is the true-positive rate and specificity the true-negative rate.

High-yield facts · 11

  1. Measures of central tendency: mean (average, affected by outliers), median (middle value), mode (most frequent).T9
  2. Measures of dispersion: range, variance and standard deviation.T9
  3. In a negatively skewed distribution (tail to the left), mean is lower than median, which is lower than mode.T9
  4. A null hypothesis says there is no difference; p under 0.05 is the usual cutoff for statistical significance.T9
  5. Power (usually set at 0.80) is the chance of detecting a real difference; larger samples increase power.T9
  6. A 95% confidence interval that crosses 1.0 (for a ratio) or 0 (for a difference) is not statistically significant.T9
  7. Correlation (r) shows the strength and direction of a relationship, not causation; r = 0 means no linear relationship.T9
  8. Parametric tests (t-test, ANOVA, Pearson r) need interval or ratio data with a normal distribution; nonparametric tests (chi-square, Mann-Whitney, Spearman) do not.T9
  9. Paired t-test: same group before and after; independent t-test: two different groups.T9
  10. Sensitivity = true positives / (true positives + false negatives); specificity = true negatives / (true negatives + false positives).T9
  11. Relative risk compares incidence in exposed vs unexposed (cohort studies); odds ratio is used in case-control studies.T9

Facts are written in our own words; each tag links to its source. Values marked "sources vary" differ between references, so check the one your program uses.