Lesson 3: Audit Spreadsheet Evidence

A spreadsheet can look correct and still answer the wrong question.

Today you will audit your spreadsheet instead of adding more decoration. The job is to prove that the records, formulas, and chart match the claim you are making.

Audit the records first

Before checking formulas, check the records.

Ask:

• Does each row describe one item?
• Do the columns keep one kind of information each?
• Are category and status values consistent?
• Are blank cells intentional?
• Is private information removed?

A formula cannot rescue unclear records.

Formula ranges are evidence

A formula has a rule and a range.

Example:

Rule: count every artifact marked Ready for review
Range: Status cells for the six artifact rows

If the range skips a row, the answer may be wrong.

If the range includes the header, the answer may be wrong.

If the status names are inconsistent, the answer may be wrong.

Chart claims need limits

A chart turns selected data into a visual pattern.

The chart does not automatically prove quality, effort, or mastery.

A strong chart explanation identifies:

• the question;
• the selected data range;
• the chart type;
• the pattern shown;
• the supported claim; and
• the limitation.

Lab 3: Spreadsheet audit table

Create an audit table for your spreadsheet.

Use this structure:

You need at least five audit rows.

Make one repair

Choose one issue and repair it.

Good repairs include:

• correcting a formula range;
• standardizing status values;
• renaming a chart title;
• changing the chart type;
• adding a limitation to the claim;
• removing private information; or
• splitting a mixed record into two rows.

Record the before and after.

Avoid fake precision

Numbers can make weak evidence sound stronger than it is.

If your spreadsheet has six records, do not make claims about the whole class, the whole school, or every portfolio.

Say what the data actually covers.

Supported: In my six-record portfolio inventory, website artifacts appear most often.
Too strong: Students mostly make website artifacts.

The second claim uses one small dataset to speak for a group it does not measure.

Evidence checkpoint

Save:

• the spreadsheet audit table;
• one repaired spreadsheet or chart issue;
• the corrected formula or chart note;
• one supported claim; and
• one too-strong claim you avoided.

Good data work is not just making charts. It is knowing what the chart can responsibly say.