You have data on your screen and two ways to look at it. One shows every respondent or customer on a row of their own. The other collapses those rows into a grid of totals. Both are useful, and they answer different questions.
This article shows the same data both ways, then sets out when each one is the right tool. It is part of our guide to crosstabs explained: one-way and two-way tables, banners, and crosstab vs. pivot tables.
Key takeaways
- A list shows one row per respondent or customer and preserves detail, while a crosstab totals records by two variables at once to compare groups and reveal patterns.
- Use a list for verbatims, quality checks, small samples, look-ups and investigating surprises; use a crosstab for comparing segments, large samples, significance tests and reporting.
- In the example, Southwest buys 68% of items but contributes 4% of the amount, while Northeast buys 14% of items and contributes 93%.
- A crosstab can hide what a list reveals: 93% of the amount rests on two customers, and an empty cell may mean no data rather than no behavior.
What is a list?
A list, sometimes called a record list or respondent-level listing, has one row per entry and one column per field. Each row describes a single customer, respondent or transaction. Nothing has been combined.
Here is a small customer list. The numbers are illustrative.
| Name | Region | Items bought | Amount |
|---|---|---|---|
| John Adams | Northwest | 4 | $42 |
| Jane Bergen | Northeast | 2 | $900 |
| Jill Dougherty | Southwest | 15 | $66 |
| Jeff Jones | Northeast | 1 | $640 |
A list is the shape most data arrives in. A survey export is a list: one row per respondent, one column per question. You can add as many columns as you have (address, age, job title, a verbatim comment) and every one stays attached to the person who gave it.
What is a crosstab?
A crosstab, short for cross-tabulation, counts or totals records by two variables at once. One variable defines the rows and the other defines the columns. Each cell holds the result for one combination. If the term is new, what crosstab software does is a good next read.
Suppose you care less about who the customers are and more about where the sales come from. Cross region against sales and you get this:
| Region | Customers | Items sold | % of items | Amount | % of amount |
|---|---|---|---|---|---|
| Northeast | 2 | 3 | 14% | $1,540 | 93% |
| Northwest | 1 | 4 | 18% | $42 | 3% |
| Southeast | 0 | 0 | 0% | $0 | 0% |
| Southwest | 1 | 15 | 68% | $66 | 4% |
| Total | 4 | 22 | 100% | $1,648 | 100% |
Here is the arithmetic. Northeast items are 2 + 1 = 3 and its amount is $900 + $640 = $1,540. Percent of items is each region’s items divided by 22 (Northeast 3 / 22 = 13.6%, rounded to 14%). Percent of amount is each region’s amount divided by $1,648 (Northeast $1,540 / $1,648 = 93.4%, rounded to 93%).
What does the crosstab show that the list hides?
Read the list again and the pattern is hard to see. Read the crosstab and it is immediate.
- Southwest buys the most items (68%) but contributes 4% of the amount. That is $66 / 15 items, or $4.40 per item.
- Northeast buys few items (14%) but contributes 93% of the amount. That is $1,540 / 3 items, about $513 per item.
- Northwest sits between the two, at $42 / 4 items, or $10.50 per item.
The comparison is the point. A crosstab lets you put one group beside another and see the gap, which a list cannot do without you doing the sums by hand.
The crosstab also shows its limits. The Southeast row is zero, but the list has no Southeast customers at all. The zero means “no one in this data”, not “customers who buy nothing”, so it would be a mistake to conclude that Southeast customers are not buying.
And 93% of the amount rests on two customers, one of whom (Jane Bergen, $900) accounts for 55% of the total on her own. The list is how you find that out.
What are the pros and cons of a list vs. a crosstab?
| List | Crosstab | |
|---|---|---|
| Unit shown | One row per respondent or customer | One cell per group combination |
| Best at | Inspecting individual records | Comparing groups and spotting patterns |
| Shows base sizes | Implicitly (you can count rows) | Yes, if you include counts or a base row |
| Handles open-ended text | Yes, verbatims stay intact | Only as coded counts |
| Scales to large samples | Poorly; thousands of rows cannot be read | Well; size barely changes the table |
| Supports significance testing | No, it has no groups to compare | Yes, with a test such as chi-square or a column test |
| Shows outliers | Yes, you can see the single odd row | Not on its own; averages absorb them |
| Can be rebuilt into the other | Yes, into any crosstab | No, you cannot recover the rows |
In short, a list preserves detail and a crosstab preserves meaning. Each costs you what the other gives.
When should you use a list?
Use a list when the question is about individual records, or when the sample is small enough to read.
- Verbatims. Open-ended comments need reading, ideally beside the respondent’s score or segment.
- Quality checks. Straight-liners, speeders, duplicate IDs and impossible values hide inside averages and show up in a list. Our guide on how to analyze survey data covers cleaning before analysis.
- Small samples. With a handful of interviews, such as 10 or 12 in-depth conversations, percentages imply a precision you do not have. Read the rows.
- Look-up. When you need all the information about one customer or one respondent in one place.
- Investigating a surprise. When a crosstab cell looks odd, go to the records behind it.
When should you use a crosstab?
Use a crosstab when the question is about groups, such as whether purchase intent differs by age band, or whether the new concept scores higher than the old one.
- Comparing segments. Customers versus non-customers, regions, waves of a tracker.
- Finding patterns in large samples. With 1,000 respondents, only a summary is readable.
- Testing differences. A crosstab is the natural place for significance testing, because the groups are already defined. Be aware that survey results come from a sample, so a gap between groups may be noise. Our explainer on one-way and two-way tables covers what each table shows before you test.
- Reporting. Stakeholders usually want the summary, with the records available on request.
If you are weighing a crosstab against a spreadsheet pivot table, see the difference between crosstabs and pivot tables.
What are common mistakes when choosing between them?
- Reading a list for patterns. Scanning 400 rows for a trend is slow and unreliable. Summarize first, then look at records.
- Trusting a crosstab without checking the base. A percentage of 3 people is not a finding. Always show the count behind each cell.
- Treating an empty cell as a zero result. As the Southeast row shows, zero may mean no data, not no behavior.
- Throwing away the list. Keep the respondent-level file. You will want it the first time a number looks wrong.
- Sharing only the crosstab when one record drives it. A single large customer or an outlier can produce a striking cell. Check the rows behind any result you plan to act on.
How do you do this in practice?
Most analysis moves between the two views. You start with the list to check that the data is clean, build crosstabs to compare groups, then return to the records behind any cell that surprises you. Halo Reports produces crosstabs and banner tables with weighting and significance testing from respondent-level data, so the summary and the source stay connected. You can also ask questions in plain language and get answers from your own research, with the source shown.
Next time you open a data file, do three things in order: scan the list for obvious problems, cross your key question by one or two groups, and open the rows behind the biggest difference before you report it. For wide, complex studies and global trackers, insights teams can keep both views on one governed platform.