stats
Chapter 1: Introduction
- Overview of Topics Covered
- Discussion of moderate statistics focusing on samples
- Mention of sample size calculation as a critical skill in biology and nursing research
- Assumptions of statistical tests briefly introduced - Importance of understanding variable types for selecting statistical tests
- Common statistical tests:
- Example of selecting a t-test based on variable types
- Awareness that some datasets may not conform to normal distribution
- Alternative tests available when normality does not hold, e.g., Wilcoxon signed-rank test
- Concept of nonparametric tests introduced, emphasized their independence from normality assumptions - Mention of power analysis in the context of hypothesis testing
Chapter 2: Youtube Tutorial on Power Calculations
- Usage of R for power calculations
- Recommendation to download G*Power, a free software for statistical power analysis
- Illustration of user-friendly interface with buttons for calculations
- Availability of YouTube tutorials for learning its functionality - R as an alternative for power calculations, with emphasis on convenience and cost-saving
Chapter 3: Normal Distribution Curve
- Description of the bell curve
- Representation of likelihood probability distribution - Key concepts associated with the normal distribution:
- Mean, median, mode should ideally align at the center of the curve - Note on potential distortions of the distribution:
- Possible occurrences of skewness in data
Chapter 4: Nonparametric Test Considerations
- Procedure after running t-tests in R
- Importance of checking the assumption of normality before interpreting test results
- Tests for normality to be utilized based on sample size
- For sample sizes less than 50: Shapiro-Wilk test recommended
- For sample sizes greater than 50: Kolmogorov-Smirnov (KS) test suggested
Chapter 5: Running The Shapiro-Wilk Test
- Explanation of what to do if data does not conform to the normality assumption
- Implications of p-values obtained from the Shapiro-Wilk test:
- Existence of significance in the results may determine test choices - Mention of the KS test and procedure to conduct this test in R
Chapter 6: Paired T-Test
- Conditions for conducting a t-test based on p-values
- If p-value is statistically significant, reevaluation of normality is necessary - Introduction to nonparametric alternatives:
- Wilcoxon signed-rank test as a method to analyze differences in ranking, rather than actual values
Chapter 7: Independent T-Test
- Discussion of independent t-tests and the Mann-Whitney U test as its nonparametric alternative
- Reference to the paired t-test alternative: Wilcoxon signed-rank test
- Mention of linear regression alternatives:
- Spearman correlation introduced, noted as an older method
- Explanation of robust regression methods in R, with emphasis on correctly estimating values
Chapter 8: Linear Analysis Overview
- Clarification that linear regression provides associations, not predictions, underlined for student understanding
- Introduction to nonparametric equivalent of ANOVA: Kruskal-Wallis test
- Outlook on covering linear regression in upcoming lessons - Instruction to refer to Brightspace for accessing assumption testing files
- Exercise practices advised for hands-on understanding of concepts
Chapter 9: Testing the Vector
- Instructions for creating a vector and conducting tests based on sample sizes
- Emphasis on examining results to decide on performing a t-test or alternatives - Introduction of the QQ plot as a tool for detecting outliers in datasets
- QQ plot functions explained:
- Alignment along the line indicates a good fit
- Outliers or deviations suggest skewness, classifying data inconsistencies
Chapter 10: Conclusion
- Recap of assessing distribution and appropriate testing based on normality
- Mention of article listings as supplementary materials for deeper exploration of topics covered.