Last updated October 2026
What we hold Mitzpe’s analysis to — forensic sourcing, calibrated language about uncertainty, a hard line between donors and the analysis, and corrections you can see — and how to tell us when we miss.
For the separate standard a book must meet to carry the Mitzpe Press imprint, see Mitzpe Press’s Editorial Standard.
Sourcing
Every factual claim in a Mitzpe publication can be checked. We go to primary sources before aggregators. When sources disagree, we say so and show where. When a claim is our own inference, we show how we got there.
We publish in English. We read the Hebrew, Arabic, and English-language press, and regional reporting in other languages when it matters. Translation is part of our production work, and when a translation choice changes how a passage reads, we flag it.
When an assessment turns on exactly what someone said and how they said it, we check the translation against the original. Quotes that only illustrate a point rest on the pipeline’s translation.
We don’t run other outlets’ reporting as our own. When an assessment depends on someone else’s reporting, we name the outlet in the body of the piece, where the reader will see it.
Calibrated probability
Mitzpe writes about futures nobody can be sure of. So we use calibrated language — “likely,” “probable,” “low confidence,” numerical ranges where they help — and stay away from the hedges commentary leans on: “could,” “may,” “some say.” A calibrated statement can be tracked, challenged, and updated. A hedge can’t.
When we commit to a probability, we commit in writing. We don’t quietly rewrite a past call when events go another way. The call stays where we published it, and a new assessment links back to the one it updates.
Anonymous sources
We use anonymous sources sparingly, and only when naming them would expose them to retaliation, professional harm, or physical danger. We check every anonymous claim against a second source or named context before we run it. We say why the source isn’t named and describe them as closely as we safely can (“a senior official briefed on the deliberations,” “an analyst at a partner institution who declined to be named”). We don’t use pseudonyms.
AI in production
Mitzpe uses large language models throughout its production pipeline.
Corrections
When a Mitzpe publication gets a fact wrong, we correct it where readers can see it. The correction goes at the top or bottom of the piece, dated, and says what was wrong and what’s right. If the error was in the body of the analysis, the original text stays up in strikethrough. If it was a number, a name, or a date, the correction line records the change.
We don’t make silent corrections. Someone who read the original should be able to see what changed, and someone arriving later should see that the piece was corrected, and how.
We correct factual errors: a misstated number, a misattributed quote, a wrong date, the wrong actor. When the facts under an assessment shift, we deal with it in the next assessment, with a link back to the earlier call.
To request a correction, email [email protected] with the link to the piece, the claim you’re questioning, and your source for the correct information. We answer correction requests within five business days.
The donor–editorial firewall
No donor shapes, edits, vetoes, or pre-clears anything Mitzpe publishes. The full firewall — what a donor can and can’t do, and how we disclose conflicts of interest — is on the Disclosures page.
Bylines and accountability
Every Mitzpe publication carries a byline. On Dossiers, Vantage, and Watchwords it names the analyst who owns the call, and when several analysts worked on a piece, it names all of them. The Israel Brief, published Monday to Thursday, has carried the institute’s byline since October 2026.
Voice on uncertainty
Our readers use analysis to make decisions. So we say what we know, say what we don’t, and don’t cover the gap with more confidence than the evidence supports. Overstating certainty is an editorial failure here, and so is understating it. A signal at low confidence is still a signal. A signal at high confidence carries more weight, and we tell you which one you’re looking at.
Contact
- To request a correction
- [email protected]
- For other editorial questions
- [email protected]