Biostat Methods I - Introduction

Study Units/Experimental Units

The smallest unit we are interested in studying, often individual people in biostatistics

Variables

Things we can measure about each of our study units, often a column of a dataset

Observation

A row of a dataset often corresponding to a study unit

Quantitative Variable

variable consisting of meaningfully interpretable numbers

Qualitative Variable

variable best described by categories (categorical), often strings or characters

Discrete Data

Type of quantitative variable, assumes whole number values

Continuous Data

Type of quantitative variable, can take on any value in a range

Nominal Data

Type of qualitative data, variable where the order of categories doesn’t matter

Ordinal Data

Type of qualitative data, variable where the order of categories DOES matter

Population

Abstract group of people (or other units) about whom we want to learn

Sample

A group of subjects from a population that we can actually observe

Inference

Process of formally learning about a population from a sample

Sampling Bias

Occurs when a sample is unrepresentative of the population of interest

Survivorship Bias

Sampling subjects after some selection process removes certain subjects from the observable population

Self-selection/Non-response bias

Sampling via a process that overemphasizes individuals with particular characteristics

Selection Bias

Choosing subjects from unrepresentative clusters of the population

Experiment

Study in which the investigator determines at least one of the variables, offers more control over how sample is drawn from population, assignment of treatment groups

Observational Study

Study where the investigator does not influence the process of interest, provide evidence of associations, using pre-existing data

Endpoint

The outcome of a study

Temporal

Means of classifying a study design, when the sample was collected relative to the phenomenon of interest

Selection

Means of classifying a study design, how the observations were chosen to be in the study

Cohort Studies

Observational studies in which the study units are members of a cohort

Cohort

Group of people with an identifying characteristic

Case-Control Studies

Study with at least two groups: case and control

Prospective Studies

Studies that look forward in time, subjects are followed over time and monitored for certain characteristics

Incidence

Number or rate of new cases

Prevalence

Proportion of population with a condition

Cross-Sectional Studies

Study that considers a sample from the population at a particular time point, establishing prevalence, employing existing data

Retrospective Studies

Studies that identify participants after data has been generated and outcomes occurred

Baseline Characteristics

Info on subjects at the beginning of the study or designation starting point

Inclusion and Exclusion Criteria

Rules for who to include/exclude when enrolling subjects or selecting them from a data source

Control

Describes something we’re comparing another thing to, doesn’t receive the treatment, might receive placebo

Placebo

Inactive treatment typically given to the control group of a case-control study

Blinding

Making it unknown to the study units which treatment (control or case) they are receiving to not influence the outcome

Confounding

Underlying variables/factors that might be influencing the study outcomes like age, sEx, etc, can control for confounding variables