#ChatGPT for #Research: How to use ChatGPT with #SPSS for Descriptive Statistics?

Using ChatGPT to write and format APA-ready results from SPSS descriptive statistics can save time—especially when the workflow is clear: run the right SPSS procedure for each variable type, copy the output table, then prompt ChatGPT to interpret and report it in APA style.

The session starts with a practical setup using three demographic variables: age, gender, and experience. Age is treated as a continuous variable, so SPSS should use **Descriptives** (not frequencies). Gender is handled as a nominal variable, so SPSS should use **Frequencies**. Experience is described as a categorical/ordinal variable, which also signals that the correct SPSS output depends on how the variable is coded. A key reassurance is that recoding values for understanding (e.g., changing how categories are labeled) doesn’t alter the analysis as long as the underlying meaning remains consistent.

For age, the workflow is straightforward: run **Analyze → Descriptive Statistics → Descriptives**, place the age variable into the variable box, and (optionally) adjust options like skewness or kurtosis—though the example keeps it simple. After SPSS produces the descriptive table, the output is copied and pasted into ChatGPT with a prompt asking for interpretation and APA-style reporting, including a properly labeled table. The presenter also demonstrates that ChatGPT may not always format tables perfectly, so manual formatting steps in a document editor (e.g., adding table titles like “Table 1. Descriptive Statistics for Age,” adjusting borders, and ensuring APA table styling) may still be required.

For gender, the session shifts to **Frequencies** because nominal data are summarized by counts and percentages rather than means and standard deviations. The SPSS frequency output is copied into ChatGPT and prompted for APA-style interpretation and table reporting. Even when ChatGPT’s table formatting is imperfect, the core content—what the distribution looks like and how to report it—can still be extracted and then polished manually.

The final example scales up from demographics to measurement items for a construct: servant leadership measured by multiple items (SL1 through SL7). Here, SPSS again uses **Analyze → Descriptive Statistics → Descriptives** for the item set. A crucial improvement is to use SPSS variable labels so ChatGPT understands what each item represents (e.g., SL1 corresponds to a specific statement about a leader addressing work-related problems). ChatGPT is then prompted not only to interpret the overall descriptive statistics in APA style, but also to identify which item has the highest and which has the lowest score—an example of targeted follow-up prompting that turns a generic table interpretation into a more decision-ready summary.

The session closes with a warning that ChatGPT can’t compensate for missing statistical fundamentals. If someone doesn’t know whether to run Descriptives versus Frequencies—or what those outputs mean—then even the best prompt will produce shaky results. The practical takeaway is to combine basic descriptive-statistics knowledge with a repeatable SPSS-to-ChatGPT workflow, then finish with human APA formatting checks.