Seminar: Problems in Applied Data Analysis


A consideration of complications in the analysis of non-experimental social science data by traditional inferential and descriptive methods. Included are issues such as null hypothesis testing versus effect size, "clinical" significance versus statistical significance, confidence intervals, the effective use of graphical methods, the interpretation of explained variance, measurement error, sample size and power, the treatment of outliers and missing data, the use of ratios and residuals for statistical "control," transformations, significant figures, repeatability, the assumption of linearity, conditional probability, accuracy and precision, sensitivity and specificity, predictive validity, regression to the mean, ecological correlation, Simpson's Paradox, Lord's Paradox. This course is open to graduate level students only.
Course Attributes:

Section 01

Seminar: Problems in Applied Data Analysis
View Course Listing - SP2022
View Course Listing - SP2023