Common arts institution reporting problems

What to take away

Most institutional stories fail on category errors, not on hidden accounting. The numbers are real. They belong to another entity, another period, or another population.

1. The brand and the legal entity are treated as one

A venue, a foundation, a resident company and an operating nonprofit share a name, so their revenue and decisions get blended.

Draw the entity map first. Attach each filing, contract, grant and program to the organization that signed it.

2. Publication year replaces fiscal year

A report published in 2027 covers a year that ended in 2026. A calendar-year attendance total covers twelve months to December.

Label the covered period in the notes, the tables, the charts and the prose. Flag partial-year and seasonal data.

3. The largest figure wins

A campaign goal, a grant ceiling, an annual budget and an endowment market value are four different measures. Articles treat them as one measure of size.

Define the figure and state its restriction, period and stage. Say why it answers your question.

4. Visits become unique people

Repeat admissions, multi-event tickets, staff, online devices and estimated passersby all become "audience members."

Keep the raw unit. The Census Bureau's statistical analysis standard requires sample analysis to account for design and estimation method, and expects conclusions from sample data to carry measures of uncertainty. An institution's survey is not a Census product. The principle still holds: method sits beside the conclusion.

5. A voluntary survey represents everyone

Respondents with spare time, English, a device or a strong opinion become the voice of all visitors and nonvisitors.

Publish recruitment, questions, field dates, sample, response rate and weighting. Name the groups missing from the sample. Write "respondents" when respondents are what you measured.

6. Mission is reported as outcome

Words like inclusive, local, innovative or educational get copied from institutional copy into the article with no evidence attached.

Turn each word into an observable question about programming, budget, staffing, ticket price and who holds decision power.

7. Access equals a ramp or a discount

One feature stands in for several things. These include routes, communication, restrooms, and sensory conditions. They also include digital systems, schedules, and transport.

Report each feature separately and date the check. One visitor's account is one account, not a finding.

8. Labor disappears behind programs

The article covers artistic ambition and omits pay, staffing losses, contractors, volunteer dependence, schedules and technical work.

Name who delivered the program and under what reported conditions. Protect vulnerable sources and put the allegations to management.

9. Institutional data is assumed to be standardized

Two museums use "attendance" differently, or a field changes definition between reporting years.

Ask for the data dictionary, the survey instrument and the notes. The Institute of Museum and Library Services museum data collection report treats systematic collection as a policy goal. Standardization is a research aim, not something matching labels give you.

10. The chart removes qualification

A narrow axis exaggerates a change. Nominal dollars hide inflation. A preliminary point sits level with audited years.

Label the chart itself with these:

Common questions

Can a deficit be newsworthy?

Yes, once you explain its scale, whether it recurs, and the cash position behind it. Restricted funds, debt and management response all belong in the same paragraph. A deficit without those is a number, not a story.

Does more attendance mean deeper participation?

Not automatically. Frequency, duration, making, learning and satisfaction are separate measures. An institution can sell more tickets and still lose the relationship it claims to build. Ask which measure the claim actually rests on.

Should a newsroom calculate its own ratios?

Yes, when the underlying definitions are compatible and you disclose the calculation. Show the inputs and have someone check the arithmetic. A ratio built from two incompatible "attendance" figures is worse than no ratio.

Can an institution correct a preliminary figure without an editor's note?

If the article published the old number, readers need a transparent correction or update that explains the change. A silent swap leaves the published record wrong. Refer to the publication's corrections policy and name the primary source.