Statistics for Research - L3 - What is SPSS and When to use it?

SPSS (Statistical Package for the Social Sciences) is positioned as a go-to software tool for running statistical analysis in social science research, especially when researchers need more than basic summaries. It supports a broad workflow—from descriptive statistics and inferential tests to data visualization—making it useful for turning raw survey or behavioral data into tables, charts, and test results that can be communicated clearly.

SPSS becomes particularly relevant when projects involve large datasets and require complex analysis paired with clean presentation. The transcript highlights common tasks where SPSS fits naturally: producing tables and charts, running statistical tests such as t tests, ANOVA, and chi-square tests, and performing methods like correlation/regression and exploratory factor analysis. In other words, SPSS is framed as a single environment where both analysis and reporting outputs can be generated.

For summarizing data, SPSS is described as supporting descriptive statistics, which organize and present key features of a dataset in easy-to-read forms. These summaries can appear as bar charts, histograms, pie charts, and other visual formats. The specific descriptive measures mentioned include central tendency and spread—mean, median, mode, range, variance—as well as distribution shape indicators like skewness and kurtosis, along with minimum and maximum values.

When the goal shifts from “what the data looks like” to “how variables relate,” the transcript points to correlation analysis for examining relationships between variables. A concrete example is offered: testing whether increasing stress corresponds to changes in organizational or employee performance. For explaining how one variable changes with others, regression analysis is presented as the next step—such as assessing how organizational culture and organizational commitment influence employee performance by estimating how much variation in performance is explained by one or more predictors.

SPSS is also presented as a toolkit for comparing groups, with the choice of test depending on distribution assumptions and the number of groups. For two groups, the transcript distinguishes between an independent samples t test for normally distributed variables and the Mann–Whitney U test as a non-parametric alternative when distributions are non-normal. For more than two groups, it recommends one-way ANOVA under normality and Kruskal–Wallis when normality does not hold. For intervention or pre/post designs, it differentiates paired-sample t tests for normal distributions from Wilcoxon signed-rank tests for non-normal distributions.

Finally, the transcript explains the basic workflow inside SPSS: opening the software reveals a data view that resembles an Excel-style grid, but before entering data, researchers must define variables. The session closes by teeing up a deeper look at how to define and enter variables in later sessions.