Chapter 15: Statistical Evaluation of Data

Descriptive statistics form the first layer of data analysis, using measures of central tendency such as mean, median, and mode to identify where data cluster, while measures of variability including standard deviation and variance reveal how dispersed scores are around that center. Visual representation techniques including frequency distributions, histograms, scatter plots, and bar graphs translate raw numerical data into interpretable patterns that reveal both central tendency and relationships among variables. The chapter then transitions to inferential statistics, which allows researchers to draw conclusions about populations based on sample evidence while accounting for sampling error and standard error. Hypothesis testing operates through a logical framework involving null hypothesis formulation, alpha level selection, and p-value computation to determine whether observed findings reflect genuine population effects or random variation. Understanding Type I errors, which occur when researchers reject true null hypotheses, and Type II errors, which occur when they fail to reject false null hypotheses, becomes critical for evaluating research validity. The chapter introduces specific statistical procedures including t-tests for comparing group means, correlation analyses using both Pearson and Spearman coefficients to measure linear and rank-order relationships, chi-square tests for categorical data, and regression analysis including multiple regression models for predicting outcomes from multiple predictors. A key emphasis throughout is that statistical significance alone does not determine practical importance, necessitating effect size measures such as Cohen's d, r-squared, and eta-squared alongside p-values. Confidence intervals provide additional perspective by establishing ranges of plausible population values. The chapter concludes with practical guidance on selecting appropriate statistical procedures based on research design, measurement scales, and data characteristics, ensuring that analytical choices match research questions and data structures.