Biostatistics, Chapters I & II
Sampling
- Population: complete collection of all measurements or data that are being considered.
- Sample: sub-collecion of members selected from a population
- Simple Random Sample: each member of the population has the same change of being included, and samples are chosen independently
- Cluster Sampling: dividing the population into groups by a category. All of the individuals within the single group are the sample.
- Stratified Random Sampling: divide the population into groups (strata) based on one+ classification criteria. Then perform a simple random sample within each strata
- Sampling Bias: some members of the population have a higher chance to be selected than others.
Variables
- Categorical Variables: two+ categories, but no intrinsic ordering (ex: blood type)
- Ordinal Variable: categorical variables but with a clear ordering (small/medium/large)
- Numeric Variables * Discrete Variables: a numeric variable for which we can list the possible values (think: integers) * Continuous Variable: a numeric variable that is measured on a continuous scale (temperature, height)
- Bar Charts: frequency distribution for categorical variables
- Histograms: frequency distribution but no spaces
Frequency Variables
- Mean, denoted by ȳ * Mean: The average of the observations * Only for discrete or continuous data * ȳ = (Σ yi)/(n) * Sensitive to outliers
- Median, denoted by ỹ * N is odd: (n + 1)th largest value * N is even: average of (n/2)th largest value and (n/(2) + 1)th
- Symmetric and Unimodal Curve
*

- Symmetric and Multimodal Curve
*

Box Plots
- Quartiles * Q1 = 25th Percentile * Q2 = 50th Percentile (Median) * Q3 = 75th Percentile
- Fences * LF = Q1 - h * UF = Q3 + h * h = 1.5(Q3 - Q1) * Outliers are any points that lie outside of the LF and UF
- Drawing a Box Plot * Central box from Q1 to Q3 * Line in the middle is Q2 * Whiskers extend to the point CLOSEST to the LF & UF (not the actual values of the fences) * Outliers are marked by small circles
Label y axis
Variance
- Sample variance * s^2 = Σ(yi - ȳ)^2 / n - 1 * Remember to subtract one from n
- Simple Standard deviation * Sqrt(s^2) * Same unit as the original data value
- \
\