Statistics for Experimental Research Lecture Notes

Jul 23, 2024

Notes on Statistics for Experimental Research

Introduction

  • Purpose: Today's lecture focuses on the role of statistics in experimental research.
  • Core Idea: Experiments are fundamental to laboratory research, while statistics help in making sense of experimental work.
  • Target Audience: Postgraduates and scientists in medical, social, life, and natural sciences.

Definition of Experimental Research

  • Systematic inquiry to describe, explain, predict, and control observed phenomena.
  • Key Principles: Integrity, reproducibility, and reducing external factors.

Objectives of the Presentation

  1. Enhance capacity to conceptualize and design experimental research.
  2. Plan statistical analysis in early research stages.

Key Concepts in Statistics

  • Variables: Independent and dependent variables, analysis of variance, correlation, regression analysis, causality.
  • Experimental Design: Statistical analysis should be integrated during the design phase of the experiment, not just post-experiment.

Key Steps in Experimental Research

  1. Knowledge of statistics.
  2. Design the experiment.
  3. Measure and perform analytics.
  4. Analyze and visualize data.

Three Principles of Experimental Research

  • Randomization
  • Replication
  • Reduction of noise (control)

Importance of Sampling

  • Population: Entire aggregation of cases meeting criteria (eligibility, inclusion, exclusion).
  • Sampling: Selecting a portion of the population to represent the overall group.
  • Target Population: The group to which researchers want to generalize findings.

Key Terms in Statistics

  • Effect Size: (e.g., Cohen's d, Pearson's r)
  • Sample Size: Report precise numbers for all experimental groups.
  • Randomness of Sampling: Essential for reporting.

Types of Sampling Techniques

  1. Non-probability Sampling: Convenience, snowball, quota, purpose sampling (less representative).
  2. Probability Sampling: Simple random, systematic, stratified, cluster, and multi-stage sampling (more representative).

Conducting Simple Random Sampling

  1. Define the population.
  2. Decide sample size.
  3. Randomly select the sample.
  4. Collect data.

Statistical Models Inclusion

  • Independent Variable: Affects the dependent variable; significant for analysis.
  • Dependent Variable: Outcome of the experimental test.

Starting Statistical Analysis

  • Key Steps:
    1. Conduct literature review.
    2. Operationalize variables.
    3. Data collection based on the hypotheses.

Hypothesis Testing

  • Null Hypothesis (H0): No significant difference between group means.
  • Alternative Hypothesis (H1): Indicates a true population difference.
  • Report p-values, degrees of freedom, and effect size.

Logic of Statistics

  1. Describe sample/target population.
  2. Detect variance from interventions.
  3. Compare and test hypotheses.
  4. Establish correlation and predict.
  5. Prove causality.

Types of Statistical Analysis

  1. Descriptive Statistics: Summarize and organize data (mean, standard deviation).
  2. Inferential Statistics: Make generalizations about population parameters based on sample data.

Major Statistical Tests and Their Functions

  1. Descriptive Statistics: Standard deviation, Kolmogorov-Smirnov test for distribution.
  2. Variance Detection: Chi-square test, Wilcoxon signed-rank test, ANOVA.
  3. Comparison and Testing: ANOVA, t-tests.
  4. Correlation Establishment: Spearman, Pearson tests.
  5. Regression Analysis: Predict relationships between variables.