Understanding Biostatistics and Its Importance

Sep 1, 2024

Lecture Notes on Biostatistics

Introduction to Biostatistics

  • Definition: The application of statistics to the analysis of biological data.
  • Philosophy: Considered as a search for truth.
  • Branches:
    • Statistical genetics
    • Clinical trials
    • Exposure methods
  • Limitations: Cannot prove hypotheses, only provides weight of evidence.

Importance of Biostatistics

  • Responsibility: Scientists must accurately explain scientific findings to the public.
  • Public Understanding: Most of the public does not understand statistics but wants clear answers.
  • Application: Interpreting data, explaining study results, and evaluating evidence.
  • Misinformation: Addressing poor use of statistics and poor study designs.
  • Career Relevance: Essential for public health policy and interpreting scientific studies.

Role in Scientific Research

  • Inference: Making inferences about populations based on samples.
  • Study Comparison: Comparing results across multiple studies to form conclusions.
  • Sample Size: Estimating sufficient sample size for studies, including power calculations.

Example: Gallup Poll

  • Context: Gallup poll on congressional approval in 2013.
  • Sample Size: 1,039 adults with a 95% confidence margin of ±4 percentage points.
  • Population Size: Approximately 316 million U.S. adults.
  • Sampling Validity: Discussed validity of the small sample size representing a large population.

Population vs. Sample

  • Definitions:
    • Population: Totality of subjects under study.
    • Sample: Subset of the population used for measurements.
  • Sampling Error: Differences in study methods can introduce biases or errors.

Statistical Descriptions

  • Parameters: Describes populations, often using Greek symbols.
  • Sample Statistics: Estimates population parameters, often using Roman letters.
  • Sampling Impact: How sampling can affect research outcomes.

Sampling Types

  • Random Sample: Eliminates bias, most reliable.
  • Stratified Random Sample: Ensures balance across strata, such as gender.
  • Sample of Convenience: Risk of bias, not representative of the population.

Homework and Discussion

  • Assignments:
    • Watch John Oliver's episode "Science" and a Colbert Report episode on the 2012 election.
    • Familiarize with papers by Shikuma, Zhang, and Arshad for discussion.