This repository contains Rmd and html files for hands-on practicums given at CRG in 2016-2017.
MODULE I. Descriptive Statistics & Intro to Probability.
- Descriptive statistics
- Explore data
- Exercise with table(), cut(), and quantile()
- Looking at subset of data
- Missing (NA) values
- Exercise on missing values
- Summary statistics
- Outliers
- Plots
- Box-plot
- Histogram
- Exercise with hist()
- Scatterplot
- Combining plots
- Empirical cumulative distribution functions (eCDFs)
- Exercise with eCDF
- Distributions
- Normal distribution
- Probability Density function (PDF)
- Z-transformation of a random sample from a normal distribution
- Cumulative Distribution Function (CDF)
- The 68-95-99.7 rule
- Quantile function, or How to obtain the critical values of -z and z for a specified area under the standard normal curve.
MODULE II. Statistical Inference. Parametric tests.
- Parametric tests
- One-sample test on the sample mean for the random sample drawn from the normally distributed population with known variance: z-test
- One-sample test on the sample mean for the sample with unknown variance: t-test
- Two-sample paired and unpaired tests with unknown variance: t-test
- Test for proportions: prop.test()
- Fisher's exact test on proportions
- Confidence intervals and t-distribution
Module III. Statistical Inference. False Discovery Rate. Power analysis. Part 1.
- FWER, FDR
- Power analysis
- Sample size estimation
- Sample size case study 1: Central Tendency (means) difference
- Sample size case study 2: Central Tendency (means) difference, less noisy
- Sample size case study 3: Proportions
Module III. Statistical Inference. Power analysis. Part 2. (re-iteration of Part 1 with more exercises)
- Sample size estimation
- Power of tests
- Types of errors
- Power of the one-sample t test
- Power of the two-sample t test
- Calculating power using the package “pwr”
MODULE IV. Statistical inference. Non-parametric tests. Data transformation.
- QQ-plot
- Tests on normality
- Data transformation
- Non-parametric tests
MODULE V. Statistical modeling. Regression. ANOVA.
- Problems on linear regression
- ANOVA
- Examine the data
- One-way ANOVA
- Post-hoc tests
- Test assumptions
- Two-way ANOVA
- Data transformation
- Normality
- Variance stabilization
- Box-Cox
- Regression
- Plots of residuals vs. fitted values
- Regression: interpretation
- Comparison of methods
- Coordinates transformation