Public Accountability Studies: the method
Version 1, draft, November 11, 2026.
The five tests were written for businesses, and a public program fails the first one by design. You can't leave your taxes. The money was never optional, so what it bought should be easy to see.
A Public Accountability Study follows one public expenditure from the law that funded it to what it delivered. It says what the records show, what the people responsible said when asked, and which questions are still open. It does not say who to vote for, and it does not guess at anyone's motives.
What a study covers
One expenditure: a line in a budget, one contract or one grant. Pick one whose purpose was stated in public, so there is something to measure it against, and one small enough that a single person can finish the study. If a question leads to a second program, that is a second study.
The five questions
- What was promised? The law or budget line that funded it, the purpose it states, and any target it sets. If no target was set, that is a finding.
- Who was paid, and how much? Every vendor, grantee or agency that received the money, with amounts and dates, from contracts and payment records.
- What was delivered? What the contract or grant required, and what the records say was actually provided.
- Who checked? Who measured the result, and whether that was someone other than the party being paid. This is the fourth test, accountability is independent, applied to public money.
- Did it leave people more capable? The question at the center of the constitution, asked of the outcome. It is answered only from the record. If the record can't answer it, the study says so.
What counts as evidence
Every statement in a study carries one of three labels.
- Verified. A public record we link to or obtained, quoted word for word, with the date we got it.
- Reported. What an agency, vendor or official says, quoted and attributed, and not treated as verified until a record shows it.
- Unanswered. A question we asked in writing, with the date we asked and who we asked. If nobody answers, the study says so.
Anonymous claims, our own estimates and anyone's guess about motive are not evidence, and they don't appear.
How we ask
- Start with what is already published: the bill and the budget, the agency's own reports, and the state's spending records.
- Then ask for records. In Arizona, A.R.S. § 39-121 says public records "shall be open to inspection by any person." Under § 39-121.01, a custodian who fails to respond promptly has denied the request, and an agency must list any records it withholds and why, when asked. Every request asks for that list.
- Then send the office responsible the study's questions in writing, before anything is published, and give it 30 days to answer. Its answer is published in full beside the study. No answer is published as no answer, with the dates.
The rules
- No motives. A study says what happened and what the record shows. It never says why someone did it.
- No sides. Studies don't endorse candidates, parties or ballot measures. The same five questions go to every program, whoever started it.
- Everyone named is asked first. Nobody reads about themselves in a study before they had the questions.
- Studies stay published as they were written. A correction or a later answer is added beside the study with its date, never in place of it.
- Studies are free and never sold. Records requested for a study are requested for a noncommercial purpose, and that has to stay true.
- Paying us never changes what a study says.
- We say where we stand. Every study names who wrote it and any stake they have in the program.
What a finished study includes
- the expenditure, named precisely, with the law or budget line that funded it
- the answer to each of the five questions, every statement labeled verified, reported or unanswered
- every record relied on, linked or attached, with the date obtained
- the questions sent, the dates sent, and every answer in full
- the questions still unanswered
- the author and any stake they have
- the date it was published, and the dates of any additions since
Do one yourself
Anyone can use this method. Pick one expenditure where you live, ask the five questions, and publish what you find, including what nobody would tell you. If you do, write to me. I'd like to read it.