Answer a real question with a spreadsheet

WHAT YOU WILL BE ABLE TO DO

Build a small spreadsheet that summarizes records by category, checks its totals and states a conclusion with an honest limit.

A spreadsheet is useful when its structure matches the question. Decide what one row represents, keep numbers and categories consistent, and calculate only what you need. Then check the answer by hand on a small example. A neat table can still mislead if it counts duplicates or compares the wrong quantities.

THE METHOD

Step by step

1. 01

   Write the question and define the row

   Ask a question that the records can answer: which workshop time had more attendees per session in our six recorded sessions? One row will represent one completed session, not one attendee. Use columns A: Session ID, B: Time group and C: Attendees. Define attendees consistently, including whether helpers count. Record the period covered so the conclusion does not silently expand to every future event.

2. 02

   Enter a plain rectangular dataset

   Put headers in row 1 and the six records in rows 2–7. Keep a separate value in each cell, no merged cells and no totals mixed among the sessions. Enter attendee counts as numbers, not strings such as '8 people.' Use exactly Morning or Evening for the group. Each session needs its own unique ID so a copied row can be recognized as a possible duplicate.

3. 03

   Check completeness and categories before calculating

   Compare each row with the source note. Look for blank IDs, negative counts, repeated IDs and category spelling differences. Use =COUNT(C2:C7) to check that all six attendee cells are numeric; a result below six means something needs inspection. Zero is an observed count, while a blank is missing information. Do not replace an unknown count with zero just to complete the table.

4. 04

   Build a small summary beside the records

   Put Time group, Total attendees, Sessions and Attendees per session in F1:I1. Enter Morning in F2 and Evening in F3. In G2 use =SUMIF($B$2:$B$7,F2,$C$2:$C$7). It tests the group cells and adds the corresponding attendee cells. In H2 use =COUNTIF($B$2:$B$7,F2). Copy those two formulas down to row 3. The dollar signs hold the dataset references fixed while F2 becomes F3.

5. 05

   Calculate the quantity that actually answers the question

   When the session count is greater than zero, enter =G2/H2 in I2 and copy it to I3. This divides attendee visits by recorded sessions. If a category has no sessions, leave its average unreported and explain why; dividing by zero is not an average of zero. Total attendance alone would favor a category with more sessions, even if each session was smaller.

6. 06

   Reconcile and inspect the formulas

   Calculate =SUM(C2:C7) separately and compare it with =SUM(G2:G3). Compare the six data rows with =SUM(H2:H3). These equalities check whether your listed categories cover the records. Hand-add one group and count its rows. Click each summary cell to inspect its formula, including range endpoints. A formula that stops at row 6 can look plausible while omitting the final session.

7. 07

   State the answer and test a controlled update

   Write a sentence using the metric and period, plus the limit of the small record set. Save the working version, then change one practice count and predict the effect before checking the formulas. Restore it afterward. When adding a seventh session, extend every relevant range or use the application's table feature correctly. A calculation is repeatable only if new records are actually included.

SEE IT IN PRACTICE

If it does not go to plan

Numbers include words or arrive as text

Check the numeric count and source values. Formatting a text cell to look like a number is not always the same as converting its stored value.

Only one column is sorted

Select the complete dataset or use the application's whole-table sort so IDs, groups and counts remain attached to the same session.

A total is treated as an average

Name the unit and denominator. Total attendees, attendees per session and unique people answer different questions.

New rows sit outside old formulas

Inspect and extend the ranges, then repeat the reconciliation. A new visible record is not automatically included in every formula.

BUILD THE SKILL

A little more capable each time

1. First attempt

   Recreate the six fictional sessions and hand-check both groups before relying on the formulas.

2. Use it for real

   In a copy, introduce one text count, one category typo and one duplicate ID. Find and explain each problem rather than merely deleting anything unusual.

3. Take it further

   Use a small permitted real dataset. Define the row, calculate one useful summary and give someone the original records and reconciliation so they can reproduce your conclusion.

CHECK YOUR LEARNING

Can you do it?

Judge the result against these signs. If one is missing, return to the relevant step and try again.

• One row has a consistent meaning and each numeric column has a clear unit.
• The grouped counts and totals reconcile with all included records.
• The formulas use aligned ranges and include controlled updates as expected.
• The conclusion answers the stated question without turning a small observation into a general causal claim.

Explain it to yourself

Try answering before opening the explanation.