Prepare the Support for Students evidence pack

This guide pulls together the board report and the underlying CSVs for the annual Support for Students evidence pack, then spends its second half on framing, because the framing is what stops a strong report from becoming a fragile one when an auditor reads it.

The two framing rules to internalise before you write a line: morale data is a leading indicator (it tells you what students feel now, which informs what retention might look like), it is not retention proof. Support-activity data is evidence that help was offered and used (requests raised, answered, resolved), it is not evidence of an outcome. The difference is small in words and very large in defensibility.

Run this end-to-end once per academic year, close to your submission deadline. If you export the numbers a week before you write the narrative, the totals in your text will drift from the totals in the export, and the auditor will spot it.

About 15 min · Reference guide.

Pull the board report

The narrative + the charts your board will read.

Open Insights and read what's there

Insights at /insights is a single Pulse view: the week-by-week morale report, what's weighing on students, and the cohort comparison lens on one page. Read it before you start exporting. The narrative comes from here, not from staring at CSVs.

What you're looking for in Insights: how morale moved week-to-week, where the check-in response count thins out or drops below the small-N threshold (which means you've got a thin signal and should say so in the report), and any patterns across audiences (program, campus, first-in-family).

Walk through it

• Open Insights.
• Read the pulse. Note the highest and lowest weeks for morale.
• Read the Morale report. Note any small-N suppressions. These limit what you can claim.

In the console: Open Insights.

Download the board report PDF

The Morale report on /insights has a Download PDF action that renders a board-ready document: cover page with cohort N, week-by-week pulse with annotations, the morale chart with the small-N wall annotated where it applies.

Use this PDF directly in the evidence pack. It already carries the framing rules (small-N suppression, leading-indicator language) so you don't have to re-do them in your own document.

Walk through it

• Open the Morale section on Insights.
• Click Download PDF.
• Save with a clear filename (institution + term + "morale-report").

In the console: Open Morale report.

Export the underlying numbers

What the auditors will ask for.

Export your data

Export (in the sidebar) → Export your data downloads one CSV per dataset. The help-requests CSV contains student names, emails and message notes: it is identifying, so share only the aggregate morale CSV with an auditor and handle the rest under your data policy.

Run the export close to when you're finalising the evidence pack so the numbers match what's in your written narrative. If the export and the narrative were pulled days apart, the auditor will spot the drift.

Walk through it

• Open Export → Export your data.
• Download the CSVs you need.
• Cross-check the totals against what you wrote.

In the console: Open Export.

Pull the Cohort morale CSV

The Cohort morale CSV from Export gives per-week counts for each band (thriving / finding their feet / wobbling / needs a hand). It contains raw counts with no suppression, so apply the 5-response rule yourself before quoting a segment.

The suppression threshold is fixed at 5 responses per segment. It is not configurable, which is the stronger guarantee. When you quote from this CSV, treat any segment with fewer than 5 responses as "not enough students answered to show a real signal", not as a number to publish.

Walk through it

• Open Export → Export your data.
• Download the Cohort morale CSV.
• Apply the 5-response rule before quoting any segment.

In the console: Open Export.

Frame the narrative

Leading indicators, not retention claims.

Frame morale as a leading indicator

The temptation in a board report is to claim morale data as retention proof. "morale was high, so retention will be high." This is wrong, and an auditor will catch it. Morale is a leading indicator: it tells you what students are feeling now, which informs what retention might look like next term once enrolment data is in. Frame it as such.

Good framing: "Morale data shows X% positive responses in Week 3, suggesting most students felt supported during the census-week stress window."

Bad framing: "Morale data demonstrates we will retain X% of students."

The difference is small in words and large in defensibility.

Walk through it

• Read your morale narrative. Find every retention-adjacent claim.
• Rewrite each as a leading-indicator statement.

In the console: Open Morale report.

Frame support activity as support offered, not outcomes achieved

Support-activity metrics measure what happened on the platform (requests raised, answered, resolved, and how quickly). They don't measure learning or retention outcomes. Frame them as evidence support was offered and used: "X help requests were raised and resolved, with a median first response of Y hours," not "X students were retained because of this support."

Benefit / outcome claims need outcome data (academic results, follow-up surveys, retention rates), which First Six doesn't claim to measure. Keep the lanes separate.

Walk through it

• Read your support-activity narrative. Find every "benefit" or "outcome" claim.
• Rewrite each as a support-offered statement.

In the console: Open Inbox.

What to leave out of the evidence pack

The evidence pack is cohort-level. Do not include per-student check-in scores, individual reflection text, individual help-request content, or any aggregate where the cohort segment is below the small-N suppression threshold. The platform will export per-student help-request content (names, emails, notes), so deliberately exclude that CSV from the evidence pack and hold the aggregate-only line yourself.

Aggregate-only is also the privacy posture students were promised when they signed in. Breaking it for a board report is the fastest way to lose trust in the entire signal.

Walk through it

• Scan the evidence pack for any per-student names, ids, or quotes.
• Remove them.
• Confirm every aggregate is above the small-N suppression threshold for its cohort segment.

In the console: Open Insights.

Related guides