Research Methods in Biostatistics

Research Methods and Biostatistics

• The lecture is for undergraduate students
• Dr. Walhan ALSHAER is the instructor
• He is the Director of Pharmacological and Diagnostic Research Center at Al-Ahliyya Amman University (AAU)
• He is also a Senior Research Scientist at the Cell Therapy Center, The University of Jordan

Introduction to Biostatistics

• Basic Statistical Concepts
  • Descriptive statistics: Measures of central tendency and variability
  • Inferential statistics: Population vs. sample, hypothesis testing

Descriptive Statistics vs. Inferential Statistics

• Descriptive Statistics: Summarizing data (e.g., mean, median, mode)
• Inferential Statistics: Drawing conclusions or making predictions about a population based on sample data
• Importance of Inferential Statistics: Helps make evidence-based decisions; essential in research, clinical trials, quality control

Key Concepts in Inferential Statistics

• Population: The entire group of interest in a study
• Sample: A smaller group selected from the population
• Example:
  • Population: All diabetic patients in a city
  • Sample: 150 diabetic patients chosen randomly

Inferential Statistical Methods

• Estimation and Confidence Intervals
  • Point Estimate: Single value used to estimate a population parameter (e.g., sample mean)
  • Confidence Interval (CI): A range of values derived from a sample that likely contains the population parameter
  • Formula for CI for the Mean: CI = x̄ ± z*(s/√n)
    • x̄ = sample mean
    • z = confidence level value
    • s = sample standard deviation
    • n = sample size
• Example: If the sample mean of exam scores is 70 with a standard deviation of 5 (n=100), a 95% CI is: 70±1.96.5/√100 = [69.02,70.98]

Hypothesis Testing

• Definition: A method to test claims or hypotheses about population parameters

• Steps of Hypothesis Testing:

  • State the null (H₀) and alternative hypotheses (H₁)
  • Set the significance level (α = 0.05)
  • Select the test statistic
  • Compute the test statistic and p-value
  • Compare the p-value with α
    • If p < α, reject H₀
  • Interpret the result

• Example: Testing if a new drug reduces blood pressure

  • H₀: The drug has no effect
  • H₁: The drug reduces blood pressure

Applications of Inferential Statistics

• Medical Research: Clinical trials to test new vaccines
• Business Analytics: Estimating customer satisfaction from survey data
• Environmental Science: Checking pollution levels in a city
• Quality Control: Testing if a factory produces defect-free products

Advantages and Limitations of Inferential Statistics

• Advantages: Allows decision-making with incomplete data; provides estimates and confidence about population characteristics
• Limitations: Accuracy depends on sampling methods; sampling errors and biases can mislead results

Discussion

• Interpreting p-values:
  • p-value of 0.03: Statistically significant
  • p-value of 0.06: Not statistically significant

Statistical Analysis Techniques

• Parametric vs. Non-Parametric Tests:
  • Parametric: T-tests, ANOVA, regression, assumes data follows a specific distribution (usually normal)
  • Non-Parametric: Chi-square tests, Mann-Whitney U test, Wilcoxon signed-rank test, Kruskal-Wallis test, Spearman correlation, no assumptions about data distribution
  • Use parametric when assumptions are met and non-parametric when they are not. Non-parametric are suitable for small samples, skewed data, or ordinal variables.
• T-tests:
  • Independent t-test: Compares means of two independent samples
  • Paired t-test: Compares means within the same group (before and after)
• ANOVA: Compares means across three or more groups
  • One-Way ANOVA: One independent variable
  • Two-Way ANOVA: Two independent variables
• Regression Analysis: Examines relationships between variables and predicts outcomes
  • Simple Regression: One independent variable
  • Multiple Regression: Two or more independent variables

Non-Parametric Alternatives

• Mann-Whitney U test: Alternative to the independent t-test
• Wilcoxon Signed-Rank test: Alternative to the paired t-test
• Kruskal-Wallis test: Alternative to ANOVA
• Spearman correlation: Non-parametric alternative to Pearson correlation

Correlation and Regression Analysis

• Correlation: Measures the strength and direction of the relationship between two variables
  • Pearson Correlation Coefficient (r): Used for continuous data (interval/ratio) that meets normality assumptions.
    • Range: -1 to +1
    • Positive r: Positive relationship
    • Negative r: Negative relationship
    • r = 0: No correlation
  • Spearman Rank Correlation: Used for ordinal data or when assumptions of normality are violated; Based on ranks of data rather than raw values.
• Regression Analysis:
  • Simple Linear Regression: Models the relationship between one independent variable (X) and one dependent variable (Y)
    • Equation: Y = β₀ + β₁X + ε
  • Multiple Regression: Models the relationship between multiple independent variables (X1, X2,...) and one dependent variable (Y)
    • Equation: Y = β₀ + β₁X₁ + β₂X₂ + ... + ε

Practical Applications of Statistical Techniques

• Medical Research: Using t-tests and regression for clinical outcomes
• Social Sciences: Chi-square tests for survey analysis
• Business Analytics: Multiple regression for sales predictions
• Environmental Studies: Correlation analysis for pollutant levels and health outcomes

Discussion (Q&A)

• Assumptions for a t-test
• When to use Spearman instead of Pearson correlation

Next Lecture: Interpreting and Presenting Data

• Statistical software applications
  • Introduction to software
  • Data entry and management
• Data Visualization
  • Creating graphs and charts
  • Effective presentation of statistical findings
• Interpreting Results Understanding
  • p-values and confidence intervals
  • Making inferences from data