Topic desk
Media bias is usually hiding in the frame.
Measurement Press treats bias as a measurable editorial pattern, not a team sport. The work is to identify where a reader is pushed toward an interpretation before they see the evidence.
Core patterns
| Pattern | What to inspect | Reader harm |
|---|---|---|
| Base-rate omission | Does the story give numerator without denominator? | Makes rare events feel representative. |
| Denominator substitution | Does the story switch between population, opportunity, reporting, respondent, or time-window bases? | Makes a true number feel larger, smaller, or more representative than the evidence supports. |
| Composition shift | Does the story compare averages, rates, or shares while the mix of people, cases, products, geographies, or respondents changed? | Makes an aggregate movement look like broad change when the blend may have changed. |
| Public-record compression | Does the story explain what the record includes, what it excludes, and whether the count needs a rate base? | Makes a partial administrative record feel like the full universe. |
| Trend-window framing | Does the chart start at a convenient peak or trough, hide seasonality, or imply cause from timing alone? | Makes an ordinary or ambiguous movement feel like proof. |
| Event-order compression | Does the story turn sequence, proximity, or reaction timing into causal language without a mechanism or comparison? | Makes a true chronology feel like a settled explanation. |
| Causal verb upgrade | Does the story use caused, drove, reduced, changed, or prevented before a counterfactual is visible? | Makes descriptive evidence sound like proof of impact. |
| Asymmetric sourcing | Are accusations and rebuttals sourced at the same standard? | Creates false confidence on one side. |
| Source-role flattening | Does the story treat access, expertise, sponsorship, and direct experience as if they provide the same evidence? | Makes an interested or narrow source carry more confidence than its role can support. |
| Same-source looping | Do several citations trace back to the same release, dashboard, quote, or report? | Makes repetition look like independent confirmation. |
| Anecdote substitution | Does one vivid case, quote, or example stand in for prevalence, trend, representativeness, or cause? | Makes an illustrative detail feel like a measured pattern. |
| Survey method stripping | Does a poll headline hide sample source, wording, field dates, weighting, or uncertainty? | Makes a limited respondent claim feel like public consensus. |
| Sponsored-report compression | Does a report headline hide sponsor role, sample selection, denominator, comparison, or limitations? | Makes a useful but bounded research asset feel like independent proof. |
| Verb loading | Do verbs imply intent where evidence only supports action? | Smuggles motive into reporting. |
| Comparison class drift | Does the article switch from peer comparison to absolute outrage? | Makes ordinary variation look exceptional. |
| Context laundering | Is a secondary summary standing in for the original record? | Lets a weak interpretation look sourced. |
| Quote-weight imbalance | Do quote length, placement, verbs, and response standards match the strength of the evidence? | Makes one position feel stronger before readers inspect the record. |
| Disconfirmation gap | Does the story name the fact, denominator, comparison, or counter-source that would weaken the frame? | Makes a claim feel settled before readers know what evidence it has survived. |
Choose by reader job
Start with the frame you need to inspect, then follow the checklist that makes the missing denominator, source trail, or comparison visible.
Audit the article frame
Use when a headline, lede, quote mix, or causal verb makes a public claim feel more certain than the source trail supports.
- Media claim audit worksheet
- Worked media claim example
- Source triangulation checklist
- Disconfirming evidence checklist
- Headline and source-mix checklist
- Anecdote and exemplar checklist
- Case-study generalizability checklist
- Timeline and event-order framing checklist
- Causal claim review protocol
- Evidence-to-claim language matrix
- Quote weight and response standard checklist
- Subtle media bias field guide
Test the base of the number
Use when a count, rate, percentage, incident total, or chart window needs the right population, time period, eligibility rule, or comparison class.
Check respondent evidence
Use when public opinion, customer research, brand lift, or sponsored research is compressed into a broader claim than the disclosed method supports.
Choose the review workflow
Use when the next step is deciding whether the evidence is strong enough to cite, share, challenge, or turn into a decision record.
Contextual sponsor fit
Research and source quality
Pages about source trails, public records, survey disclosure, denominator checks, and citation quality fit research tools, media-literacy education, and professional training categories.
Decision support
Pages about claim confidence, evaluation worksheets, and evidence review fit governance, knowledge-management, insight, and analyst-enablement categories.
Complete reference map
Begin with the source and vendor evaluation desk if you need to choose a workflow. Use the claim confidence rubric when a story needs language that matches the evidence, the causal claim review protocol when a headline, chart title, or summary sentence says something caused, drove, reduced, or changed an outcome, and the evidence-to-claim language matrix when the final sentence needs to be rewritten with the right denominator, comparison, and verb. Use the source library when a critique needs official references for source handling, survey disclosure, corrections, public records, data citation, or search-quality language. Use the topic taxonomy when you need to map a reader job to the right desk, guide, or case study. Use the worked media claim example when a reader needs to see one claim move through source triangulation, denominator choice, and final wording. Use the field guide for recurring patterns, the headline and source-mix checklist for the first impression, the anecdote and exemplar checklist when one vivid case may be standing in for scale, trend, representativeness, or cause, the case-study generalizability checklist when one example or case result is being used as a broader rule, the source triangulation checklist when multiple citations may be repeating the same origin, the source role and incentive map when access, expertise, sponsorship, or direct experience may be doing more work than the evidence supports, the disconfirming evidence checklist when a frame needs its strongest weakener named, the quote weight and response standard checklist when a story feels balanced but may apply uneven evidence treatment, the denominator framing examples when a count, percentage, rate, or survey share needs the right base, the composition mix and average checklist when a blended average, rate, share, or survey result may have moved because the underlying mix changed, the public records and denominator checklist when official data or incident counts drive the frame, the before-and-after chart and trend claim checklist when a chart implies cause from timing, slope, or a highlighted event, the timeline and event-order framing checklist when chronology, proximity, or lag windows make cause feel obvious, the survey and poll claim checklist when public opinion or customer research drives the frame, the sponsored research and vendor report checklist when a white paper, benchmark, or commissioned study carries the claim, the audit worksheet for a fast read on the frame, the source and vendor evaluation worksheet for meeting notes, and the scorecard to grade whether the source trail is strong enough for the article's conclusion. The habit is simple: identify the claim, find the denominator, check the comparison class, and look for the strongest disconfirming context.