This bench runs the lab handouts live. You order a dataset the way the AI Data Protocol asks, with the truth built in. The bench draws a random sample from that truth, you verify it, choose the test, read SPSS-style output, and write the APA sentence. Then draw again and watch what sampling error does to an effect you know is real.
Your data spec, the truth statement, and an honest log of what the AI got wrong.
Count, range, structure, descriptives. Before any analysis, or not at all.
The right procedure, correctly specified, with output and syntax.
An APA results statement plus a plain-language sentence a non-expert gets.
Everything on this page runs on data the bench made up. Your lab runs on data you ordered, verified and logged, and on an interpretation you wrote. Use the bench to rehearse a procedure, to see what a result should look like, and to check your reasoning after you have done your own.
Do not use it to produce your submission. Copying a sentence from this page into your write-up, or using its worked answers in place of your own thinking, is the prohibited use of AI and technology described in your syllabus. It offloads the thinking, and the thinking is the assignment. Bench numbers are never your numbers, and your log is graded against your own truth statement.
In this course you don't download data, you order it. Here the bench plays the AI, and it fails the way chat tools really fail: wrong row counts, values out of range, a table instead of CSV, data that is too clean, output that cuts off. Write the spec, send the prompt, catch what is wrong, fix it, and walk away with a log that earns its two points.
Software will run whatever you tell it to run, including the wrong thing, without complaint. The comprehensive practical hands you a dataset and research questions and nobody tells you which tests to run. This is the decision tree from class: find the DV, find the IVs, count levels, decide between or within, then name the test and say why.
Each lab begins with a short frame explaining the research situation and why the analysis fits it, so you can open whichever one you need. The order below is the order the course builds the skills, from setting up a file to predicting with a regression line. Thirteen labs at 10 points each are 65% of the grade, so every point is half a percent.