Introduction to Biostatistics

• Biostatistics involves applying statistics to biological fields, including pharmacy, medicine, other medical areas, and environmental studies.
• The research process includes a research question, identification of variables, study design, data collection, data analysis, and interpretation of results.

Statistics Process

• Collect, organize, present, summarize, analyze, and interpret data.
• Data includes variables which is a measurable characteristic that can assume different values.
• Data includes Values, which is a quantitative or qualitative measure associated with a variable.

Uses of Statistics

• Statistics can describe large datasets.
• Statistics can generalize small sample results to a larger population.
• Statistics can compare different variables and test underlying hypotheses.
• Data may come from an organization or individual, a designed experiment, a survey, or an observational study.

Types of Variables

• Variables can be Qualitative which is typically categorical.
• Variables can be Quantitative which is on a numerical scale.

Qualitative Variables

• Qualitative variables are data grouped into categories like treatment, exposure, or disease status.
• Examples are gender (male, female), color (blue, orange, yellow), and blood group (A, B, AB, O).
• Dichotomous variables are binary with 2 levels such as dead/alive, treatment/placebo, disease/no disease etc
• Nominal variables are named categories where order doesn't matter; examples include blood type (O, A, B, AB), marital status, and occupation.
• Ordinal variables are ordered categories where order matters, examples include cancer staging (I, II, III, IV), birth order (1st, 2nd, 3rd), letter grades (A, B, C, D, F) etc

Quantitative Variables

• Quantitative variables use numerical data.
• Discrete numbers are a limited set of distinct values, such as whole numbers such as number of new AIDS cases in CA in a year, years of school completed.
• Count data are finite such as number of children in a family.
• Continuous variables can take on any value within a defined range, such as time-to-event, age, blood pressure, serum insulin, speed of a car, and income.
• Measurements data are infinite such as height.

Sample vs Population

• Statistical analyses typically use samples instead of entire populations to save time, money, and effort.
• For example, if you want to analyze tablets for the concentration of active ingredients you work with a sample for destructive analysis rather than the whole population.
• Samples must be representative of the population to avoid misleading results, where random selection is crucial.
• Measures computed from sample data are known as statistics
• Measures used to describe a population are known as parameters

Types of Samples

• Types of Non-Probability samples include judgement samples and convenience samples
• Types of Probability samples include Simple Random, Systematic, Stratified, and Cluster samples

Nonprobability Samples

• A nonprobability sample involves choosing items without considering their probability of occurrence.
• Convenience sampling selects items based on ease, cost-effectiveness, or convenience.
• Judgment sampling is based on the opinions of preselected experts.

Probability Samples

• Items in a probability sample are selected based on known probabilities.
• In a simple random sample, every individual or item has an equal chance of being selected, using random number tables or computer generators.
• Stratified sampling divides the population into subgroups (strata) based on a common characteristic and then selects samples from each subgroup proportional to the strata sizes.
• Systematic sampling involves randomly selecting one individual from the first group and then selecting every kth individual thereafter.
• Cluster sampling divides the population into clusters, each representative of the population, and then selects a random sample of clusters.

Comparing Sampling Methods

• Simple random and Systematic samples are easy to use, but may not accurately represent the population's underlying characteristics.
• Stratified samples ensure representation across the population.
• Cluster samples are cost-effective but less efficient, needing larger samples for precision.