Exporting your data to Excel for research & analysis

Cesura can export everything from a simulation — every run, with scores, answers, and full transcripts — to Excel. It’s a rich dataset for analysis and research. This article covers how to export, the three formats, and what each contains.

How to export

Open the Space that contains the simulation and go to the Content tab.
On the activity’s card, open its menu and choose Export Data. (The activity’s results report also has an Export Data button.)
In the Export Results dialog, pick a format (below) and click the export button, which shows how many runs will be exported. An .xlsx file downloads to your computer.

This exports one activity at a time. To export the whole space in one file, see Export a whole space. If your space uses groups, you can scope the export to a single group. Under Include you can also leave out chat transcripts, evaluations, or form answers. When you export from the activity’s results report, a fourth format, Gradebook (CSV), is offered for importing grades into your LMS (except for challenge activities and simulations with Advanced Privacy). (If the simulation has Advanced Privacy enabled, conversation transcripts and free text are left out; counts and scores remain.)

The three formats

Simple Table — quick analysis

One sheet, one row per run, with every answer as readable text: participant and timing details, form and quiz answers, message counts, the overall grade and feedback, and per-criterion feedback. Open it in Excel or Google Sheets and start filtering and charting right away.

Choices appear as the text the author wrote, in the form’s language. An “Other” answer prints as the Other label, a colon, and what the participant typed, for example Other: my own answer. Tag-box, matrix, and dynamic-choice answers are included. If two questions share a header, the headers are numbered so you can tell them apart. An answer that no longer fits any question, for example one from a question whose type changed, is listed on an Unmapped Answers sheet instead of being dropped.

Coded Forms — statistical analysis (SPSS, R, Python)

The same data, prepared for statistics. Answers are converted to numeric codes with labeled variables, across these sheets:

Columns — a data dictionary: each variable’s name, label, type, and source.
Codes and Labels — a codebook mapping every numeric code back to its original answer.
Coded Data — one row per run with the coded values.
Unmapped Answers — added only when there is something to report: any answer that had no place in its column, with the participant, the question and the reason. Typical causes are an “other” free-text item or a choice from a retired version of the form.

Numeric scales are exported as real numbers rather than text, so they are ready to average and model. An answer that cannot be coded leaves the cell empty and is listed on the Unmapped Answers sheet, rather than being written in as text that would break the analysis.

Multiple-answer questions get one 0/1 variable per choice. That includes tag-box questions: a tag-box question Qn becomes Qn_<value>, one column per choice. A checkbox or tag-box question whose choices come from another question or source gets a single text column of the chosen answers instead, because its choices exist only while the form runs.

⚠️ Check saved SPSS syntax. Tag-box questions used to export as a single variable. If you saved syntax or scripts that reference the old variable, update them to the per-choice variables.

Use this for SPSS, R, Python/pandas, or any tool that expects coded variables and a codebook.

Full Report — the complete record

A multi-sheet workbook (up to nine sheets) with everything, including the detail the other formats summarise:

Summary Dashboard — one row per run with all the key metrics, including timing for each stage of the run (see below).
All Conversations — the full chat transcript of every run.
All Evaluations and Criteria Details — overall and per-criterion scores and feedback text.
Evaluator Chat and Scenario Build — the feedback chat and (for build challenges) the learner’s authoring conversation. Each Evaluator Chat message carries its timestamp.
Quiz Results and Form Results — every quiz and form submission, one row per question (see How Quiz Results Appear in Excel Exports and How Form Results Appear in Excel Exports).
Analytics Events — the raw event log (starts, completions, report views, downloads…).

Choose this when you want the complete dataset — especially the transcripts and per-criterion feedback for qualitative work.

Stage timing in the Summary Dashboard

After Last Interaction, the Summary Dashboard times each stage of the run:

Conversation Ended and Conversation Duration (seconds)
Debrief Submitted and Debrief Time (seconds)
Evaluator Chat Start, Evaluator Chat End, Evaluator Chat Duration (seconds) and Evaluator Chat First Reply (seconds)
Evaluator Chat Messages, Evaluator Chat User Messages and Evaluator Chat Assistant Messages

A stage the run never reached is left blank. Runs from before Cesura recorded these times also leave the timing blank rather than 0; their message counts still fill in. The Coded Forms format carries the same timing as numeric variables, with missing stages left empty.

Which format should I use?

You want to…	Use
Quickly browse, filter, or chart results	Simple Table
Run statistics (SPSS / R / Python)	Coded Forms
Analyse transcripts, per-criterion feedback, or the full activity record	Full Report

Export a whole space

To export every activity in a space at once, open the space’s menu and choose Export space data. The workbook opens with a Participants sheet: one row per participant, with a block of columns for each activity:

Status
Conversation ended
Debrief written: Yes if the learner submitted a debrief
Evaluation completed: Yes if the feedback report was produced
Run ID

A participant who never opened an activity has blank cells in that activity’s block, not No.

💡 Debrief written replaces the earlier Debrief completed column, which reported whether an evaluation existed. Debrief and evaluation are separate columns because a simulation without a debrief still gets an evaluation.

What’s in the data

Across the formats you’ll find a participant identifier (their name, or any ID passed via an embed parameter or intake form), preferred language, group/cohort, start & end time and duration, status, the AI model used, message counts, form and quiz answers, the overall grade, and per-criterion scores and feedback — plus, in Full Report, the full transcripts and the analytics event log.

💡 Anonymous runs show as "anonymous." To attach names or IDs to results, add a short intake form at the start of the simulation — see Identifying and tracking individual participants.