Presenting Data in Tables
Two consecutive slides carry the whole of the week's instruction on tables. This page reproduces all seven rules and the typology, and marks the one figure the source states twice in two incompatible registers.
Where the table rules sit, and how much of them there is
The whole of the week's instruction on tables is two consecutive slides. For this material they are unusually specific: they tell you how much precision to print, how to order rows and columns, where to put the comparison you want a reader to make, what to do about grid lines and labels, and how to classify a table by the number of variables in it.
That specificity is worth using, and it is worth knowing exactly where it stops. Several of the decisions a table forces on you are not settled anywhere in the supplied material — nor on the companion slides covered at Presenting Data in Graphs.
THE SEVEN TABLE RULES, AS THE SOURCE STATES THEM
| Rule, quoted | Slide | What it settles | What it leaves open |
|---|---|---|---|
| "Use the appropriate number of effective digits (often 2-3)" | First | How much precision to print | What makes a change important; and the figure was stated two lines earlier as a requirement, not a tendency |
| "Provide averages for rows and columns where appropriate" | First | That averages belong in the margins of a table | Which average. The week defines three and does not choose between them |
| "Where possible order rows and columns by size" | First | The default ordering of a table | What to do when a natural order — time, phase, category — competes with size |
| "Present the main comparison in a column as numbers are easier to read vertically rather than horizontally" | First | Which axis carries the comparison | The reading claim is asserted with no source, evidence or exception |
| "Avoid overly intricate grid patterns; use regular spacing of rows and columns etc." | Second | That heavy ruling and irregular spacing are faults | The list closes with "etc."; no other formatting rule is given |
| "Use labels where appropriate to clarify the information in the table" | Second | That labels are required | What a label must contain. Units, sample size and data source are named nowhere |
| "Provide a brief summary to focus the reader's attention on the main features of the table" | Second | That a table needs prose beside it | How long it should be, where it sits, what it must cover |
The count of seven is derived by reading the two slides; the source states no count on either. Four rules appear on the first slide and three on the second.
Effective digits: a requirement and a frequency, two lines apart
The prose sets the figure as an obligation — "The basic requirement is to keep the number of effective digits … to two or three by appropriate rounding". The rule that follows one line later restates the same figure as a tendency — "often 2-3". A requirement and an observation about frequency are different kinds of statement, and the source does not say which of the two governs.
Reproduce both when you cite this. If you round to three digits and someone asks on what authority, the honest answer is that the supplied material states it as a basic requirement in one sentence and as common practice in the next.
Ordering, averages and the direction of reading
Three of the four rules on the first slide are about arrangement rather than content, and they are the ones most likely to change how a finding lands. Ordering rows and columns by size puts the extremes at the ends where a reader can find them. Presenting the main comparison down a column exploits the claim that numbers are easier to read vertically. Providing marginal averages gives the reader a reference point for every cell.
The second of those rules is the one that will stop you. It requires an average and does not name one, and the week's own vocabulary table defines three — see Measures of Central Tendency. The choice is not cosmetic: the deck's own worked example of an extreme value shows a survey average of $10,000 collapsing to a maximum of $1,000 once one respondent is removed, which is a demonstration that the choice of average changes the answer. That example is offered by the source as an illustration and is discussed on Handling Outliers; the deck never connects it to the table rule.
The typology: three table types by number of variables
THE TABLE TYPOLOGY AS THE SOURCE GIVES IT
| Type | Number of variables | Construction guidance supplied |
|---|---|---|
| univariable | One variable | None beyond the seven rules |
| bivariable | Two variables | None |
| multivariable | "more than two variable", in the source's wording | None |
The source says the type you select depends on "the number of variables you select" and gives no guidance on selecting them. For the two-variable case it supplies no layout, no statement of what goes in rows versus columns beyond the vertical-reading rule, and no method for computing row or column percentages. The word cross-tabulation does not appear anywhere in the week.
Percentages, accumulated percentages and ratios
The slide names three ways of turning counts into something interpretable: straight percentages, accumulated percentages — glossed in parentheses as "Pareto" — and ratios. The observation behind it is worth acting on: a table of raw counts makes a reader do arithmetic, and a table of proportions does that arithmetic for them. Which of the three suits your data depends on how the variable was measured, and the week never joins those two ideas; see Levels of Measurement.
That is as far as the source goes. It names three treatments, defines none of them, and gives no example of any. No percentage figure appears anywhere in the week's presentation slides, so there is not even an illustration of what a correctly rounded percentage looks like under the effective-digits rule.
What the two slides settle and what you must bring
Settled by the source
- Print two or three effective digits, by rounding
- Put averages in the row and column margins where appropriate
- Order rows and columns by size where possible
- Run the main comparison down a column, not across a row
- Keep the grid plain and the spacing regular
- Label the table, and put a short summary beside it
- Name your table by the number of variables it holds
Not settled anywhere in the week
- Which average a margin should carry
- Whether the digit rule counts significant figures or decimal places
- How to lay out a two-variable table, or compute its percentages
- What a label must state — units, sample size and source are named nowhere
- What a Pareto treatment or a ratio actually is
- Any measure of spread to report beside an average
- What a finished table looks like — none is shown
Working the rules into a research report
The source does not prescribe the following; it is a way of using what the source does say without over-claiming. Each row below turns one of the seven rules into a decision you have to record somewhere in your method or your table notes.
Turning the rules into decisions you can defend
A pre-flight check on any table you are about to publish
- Every column is rounded to a consistent number of effective digits, and the caption or notes say which convention you used
- Row and column averages, where present, are named — not just "average"
- The comparison you want read runs down a column
- Rows and columns are ordered by size, or the departure from that order is explained
- The grid is plain; nothing is ruled that does not need to be
- Every row, column and unit is labelled, and the number of cases behind the table is stated somewhere — the source never asks for it
- A short summary sits beside the table and states the main feature you want noticed
- Anything beyond the seven rules — a percentage base, a measure of spread, a two-variable layout — is attributed to the source you took it from
What to carry forward
- Seven rules, one typology and three percentage treatments — that is the whole of the week's table guidance, and it is the most concrete instruction in it.
- Reproduce the two-or-three-digit figure in both of its forms. The source states it as a requirement and, one line later, as a tendency, and does not resolve the difference.
- The rule requiring marginal averages is unusable as stated until you decide which average, because the week defines three and chooses none.
- Pareto, ratios and the two-variable table are labels without content in this material. Take them from a cited source and attribute them there.
- Nothing here tells you how to show variation, and no table anywhere in the week demonstrates any of the rules.
Frequently asked questions
How many digits should a table in my report carry?
The supplied material states that the basic requirement is to keep the number of effective digits to two or three by appropriate rounding, and then restates the same figure as what is "often" done. It does not say which formulation governs, and it does not say whether it means significant figures or decimal places. Apply one reading consistently and say in your text which you applied.
What are effective digits?
The source defines them as "those that represent important changes in the data". It gives no test for importance and no worked example, so the definition tells you the principle and not the procedure. In practice you decide what size of difference matters for your research question and round to the precision that shows it.
Which average should I put in the row and column margins?
The material does not say. The table rule requires averages "where appropriate" and the week's vocabulary table defines the mean, the median and the mode without ranking them for this purpose. Name whichever you use in the label, and be aware that the week's own extreme-value example shows the choice changing the answer.
Does the source explain how to build a cross-tabulation?
No. It names a bivariable table as one holding two variables and supplies nothing else — no layout, no statement of what belongs in rows against columns beyond the vertical-reading rule, and no method for row or column percentages. The word cross-tabulation does not appear anywhere in the week.
What is meant by accumulated percentages (Pareto)?
The source names the treatment and glosses it in parentheses as Pareto, and stops there. No chart, principle or example of that name appears anywhere in the week, so the term is a label with nothing behind it in this material. If you use it, define it from a cited source.
Why does this page keep saying what the source does not contain?
Because the week is represented by a slide deck alone — its study notes were listed in the upload manifest and were not supplied — and in every other week the notes carry the definitions. Recording an absence tells you accurately where a second source is needed, which is more use than a plausible definition attributed to material that does not hold one.
References and source attribution
- Veal, A. J. 2005, Business Research Methods: A Managerial Approach, 2nd ed., Longman. — the sole entry on the week's reference slide. It is cited for the normal-distribution slide and accounts for none of the table material.
- A publisher's companion website for a fourth-edition social research text, cited on the week's terminology slide as a bare address with no author, year or title stated. It is the source of the three definitions of an average that the table rules rely on.
- The prescribed text, Chapter 16, pp. 353–374 — the week's only set reading, on statistical software. No author, year, title or publisher for it is given anywhere in the week's slides beyond a single incidental surname.
- The supplied teaching source: the week 11 slide deck on analysing data and the presentation of quantitative data, whose two tables slides are reproduced in full on this page. No study notes for this week were supplied.
Suggested questions for Ask KEVOS
- Check whether my results table follows the seven table rules in this material.
- Which average should I report in a table margin, and what does the supplied material say about that choice?
- Help me write the brief summary that sits beside a results table.
- What does this teaching material leave out about building a two-variable table?
- Rewrite my table caption so it states units, sample size and rounding convention.