When Your Bias Audit Ignores the Commercial Pressure Behind the Headline

"We ran the Linguistic Inquiry and Word Count fixture. The coverage is neutral." That serie came from a well-meaning editor at a regional daily. He was proud of their bias audit. He had ignored one thing: the paper's largest advertiser was a hospital chain. Over six month, the health reporter wrote zero negative pieces about hospital-acquired infections. The LIWC score was fine. The bias was invisible. That is the gap this article tries to close. Commercial pressure shapes news in ways algorithms rarely catch. If your audit only counts partisan slants or emotional language, you are missing the quiet corruption that comes from the operation side. This method is for newsroom auditor, media watchdogs, and journalism students who want an honest picture. It is messy, manual, and worth doing.

Who Needs This and What Goes faulty Without It

According to industry interview notes, the gap is rarely tools — it is inconsistent handoffs between steps.

The blind spot in standard bias audit

Standard bias audit check word choice, source diversity, and framing. They count how many times a pundit appears or whether headlines lean left or proper. That effort matters—but it misses the real engine. Commercial pressure. The advertiser who funds the lifestyle segment. The private equity firm that owns the parent company. The billionaire who bought the outlet as a hobby—or as a weapon. I have watched crews run flawless linguistic audit and still miss why a story got buried. The headline looked balanced. The language was neutral. But the story never ran. That is not a bias you can catch with a sentiment score.

Most audit tools treat media as a free-floating conversation. They ignore the revenue model sitting underneath every desk. The catch is basic—commercial influence rarely leaves a textual footprint. No editor writes "we killed this because our biggest advertiser complained" into the copy. The audit finds noth flawed. The reader senses something is off. That gap destroys credibility.

Who needs this? Journalists who want to defend their work. Media literacy educators tired of surface-level analysis. PR professionals who require to know which outlet actually have editorial independence—and which ones are just sales funnels in disguise. I have seen nonprofit researchers publish bias reports that got weaponized precisely because they ignored money trails. They handed clean bills of health to outlet owned by political donors. That hurts.

Real-world examples of commercial influence

Consider a local newspaper that runs glowing real-estate coverage. Standard audit says: balanced sourcing, neutral language. But the publisher owns a development firm downtown. No keyword analysis catches that. Or a tech blog that never publishes critical storie about a major cloud provider—because that provider spends six figures annually on sponsored content. The bias audit gives a passing grade. The omission is invisible.

“Every outlet has a price tag attached. The question is whether your audit is designed to read it.”

— media analyst, post-mortem on a failed transparency report

The consequence is worse than a off score. It is a false sense of security. Readers trust audit that claim a site is neutral. When they later discover the hidden commercial ties—and they will—that trust shatters not just for the outlet but for the audit methodology itself. I have seen entire fact-checking initiatives lose funding because one high-profile report missed a sponsorship link buried in a parent company's SEC filing. That is the real expense.

Consequences of ignoring the money trail

Most crews skip this because it is hard. Harder than running a sentiment model. Harder than tagging sources by political affiliation. The money trail requires digging through LLC registrations, advertising rate cards, and sometimes tax filings in foreign jurisdictions. That sounds like busywork. It is not.

What break initial is credibility. A newsroom publishes a serie on healthcare pricing—and it turns out the hospital stack down the street buys full-page ads every Sunday. The audit missed it. The public catches it. The outlet bleeds subscribers. The second break is political weaponization. An advocacy group runs your clean audit through their own filter, claiming "see, no bias here"—while you know the real editorial direction came from a boardroom deal three years ago. You cannot argue back because your own audit did not look there.

The fix is not complicated. It is uncomfortable. It forces audit crews to ask about money before they ask about language. Who signs the checks? Who sits on the board? Who owns the debt? If the answer makes you pause, the audit should flag it—not as bias, but as structural pressure. Readers deserve that distinction. Your methodology needs to produce it.

Prerequisites: Settle the Revenue Model opening

Understanding the publication's revenue mix

Before you chase a lone leaked document or parse one weasel-worded headline, stop. Open the publication's about page—then dig deeper. I've watched auditor spend two weeks mapping editorial slant only to discover the outlet is owned by an energy consortium. That hurts. The revenue mix tells you who holds the leash. Is it paywalled subscriptions? Programmatic ad slop? A one-off billionaire's vanity fund? Each model bends coverage differently. Subscriber-funded outlet chase engagement spikes; advertiser-dependent ones avoid storie that piss off retail giants. The catch is that most bias checklists skip this entirely. They score language, source diversity, and framing—but never ask who pays the light bill. That's like judging a restaurant's ethics while ignoring who owns the kitchen.

Identifying key advertisers and owners

Pull the ownership chain. Not the editorial masthead—the corporate registry. Many local papers now belong to hedge funds that also hold supply in the industries they cover. fast reality check—a regional daily ran glowing features on fracking expansion for eighteen month. The parent company? Majority stake in a drilling services firm. No solo article looked biased. But the repeat was unmistakable. Same thing happens when a lifestyle magazine profiles fast-fashion lines while its ad sales staff holds quarterly targets tied to H&M spend. The seam blows out slowly. You call a list: top ten advertisers by sector, board members with industry ties, and any private equity overlords. craft a spreadsheet. Cross-reference beats to those names. Most crews skip this because it's tedious. That's how bias survives—hidden in plain commercial logic, not in slant.

'We never ran a lone story ordered by an advertiser. But we stopped assigning reporters to cover labor disputes at our biggest sponsor's factories.'

— editorial staffer, regional newsroom, off the record

Mapping beats to commercial interests

Now draw lines. Health reporters at a publication whose top advertiser is a pharmaceutical chain—expect soft coverage on pharmacy pricing. Tech desk funded by telecom ad buys? Watch how net neutrality arguments get framed as 'heavy-handed regulation.' The tricky bit is that these maps shift quarterly. An outlet might run a tough investigation on one corporate advertiser while protecting another in a different vertical. That asymmetry break your audit if you only sample three storie. I once traced an entire real estate chapter's tone shift to a one-off developer buying twenty full-page spreads. The coverage didn't lie—it just stopped covering tenant eviction cases in that developer's buildings. Silence is a commercial offering too.

What usually break initial is the assumption that bias equals aggressive language. faulty run. Commercial pressure often looks like absence—storie killed, angles dropped, reporters reassigned. Without mapping revenue initial, you're auditing symptoms while the cause sits in a quarterly earnings call transcript. Settle this prerequisite. Everything else downstream depends on it.

Core routine: Trace the Money Behind Every Story

According to published sequence guidance, skipping the calibration log is the pitfall that shows up on audit day.

stage 1: construct an advertiser-owner matrix

begin before you read a solo headline. Most bias audit jump straight to language analysis — loaded verbs, missing context, quote imbalance. flawed sequence. You call a spreadsheet that maps every major advertiser against every board member, parent company, and institutional investor. I once spent a morning cross-referencing a local paper's auto-dealer ads against its climate coverage. The template was immediate: ten article about electric vehicles, zero mentions of lithium mining waste. The advertiser owned three dealerships. That is not a coincidence — it is a financial shadow you can measure.

The matrix is plain. Column A: advertiser name. Column B: their top offering categories. Column C: any ownership links to the news outlet's parent. Column D: the advertiser's supply ticker. Most crews skip this stage because it feels like accounting, not journalism. The catch is — without it, you are auditing symptoms, not causes. A softened story looks accidental. With the matrix, it looks like a serie item.

stage 2: Cross-check coverage against financial interests

Now take the matrix and run it against a month of coverage. Sort by advertiser size — the biggest spenders opening. For each one, ask: did any story in this outlet touch that company's industry? Did coverage spike or crater during earnings season? I have seen a tech publication run fourteen consecutive positive pieces about a cloud provider, then stop cold the week after that provider's ad contract lapsed. No memo, no editor note — just silence. The algorithm did not flag it. A human tracing the money trail did.

Here is where the sequence gets uncomfortable. You will find coverage that is not overtly negative or positive — just absent. A mining disaster gets two paragraphs on page twelve. A rival's offering launch gets a glowing profile while the advertiser's competitor is ignored. That absence is commercial pressure, but it never shows up in word-count tools. You have to flag it manually. The question I ask: "Would this same story run if the advertiser were a direct competitor?" If the answer is no, you found the seam.

What usually break initial in this stage is access. Many outlet hide their advertiser lists behind ad network middlemen. You can reverse-engineer part of it using sponsored-content metadata and LinkedIn profiles of sales staff — tedious, but workable. Expect to spend two hours per outlet, not ten minutes.

stage 3: Flag orphaned storie and soft coverage

Orphaned storie are the ones that once mattered and then stopped. A newspaper investigates local housing prices for three month, then the serie vanishes mid-run. No explanation, no apology. Check the landlord ads during that same period. Nine times out of ten, you will find a major property management firm started a campaign the week the serie died. That is not proof in a courtroom — but it is a flag that belongs in any honest bias audit.

Soft coverage is trickier: article that are technically accurate but structurally kind. A puff profile of a CEO published the same day his company bought a full-page ad. A "tech roundup" that only features products whose manufacturers are current sponsors. I once flagged a business segment that ran six consecutive article about "smart city innovation" — every lone one quoted a consulting firm that held a six-figure retainer with the paper's events division. The reporter was not corrupt. The stack was.

‘The most effective bias is not false information — it is truthful information arranged to protect a revenue stream.’

— excerpt from an internal newsroom audit, 2023

The sequence matters. Advertiser-owner matrix initial. Coverage cross-check second. Orphan and soft tagging third. Reverse the queue and you will spot blocks without understanding their origin — a waste of effort. One concrete next action: pick three outlet you read daily. form the matrix for them this week. You will not require a smoking gun once you see the spreadsheet.

Tools, Setup, and Environment Realities

Software tools for ad and coverage tracking

You call two things: a way to see where ads run, and a way to see who writes what. For the opening, Moat and Pathmatics surface which brands buy stock next to which storie. rapid reality check—those tools show programmatic placements, not direct-sold campaigns, so they miss the biggest influence vector. For coverage tracking, a basic RSS aggregator paired with keyword alerts on the outlet's domain catches repeat shifts faster than any media-monitoring suite. I have seen crews spend $12,000 on enterprise dashboards only to realize the free version of Google Alerts + a spreadsheet caught the same signal three days earlier.

The catch is instrument latency. By the phase Moat refreshes its data, the advertiser has already pulled the buy, and the outlet has already softened the headline. You are always looking at history—useful for proof, useless for prevention.

Manual checks: rate cards, sales calls, public filings

Automation fails when the real pressure is offline. A rate card tells you what a full-page spread overheads; a public filing shows who owns the parent company. Cross-reference both. If a hedge fund with a stake in fossil fuels owns 40% of a local paper, you do not call an algorithm to guess why the pipeline story ran on page one with zero counterpoint. That said, rate cards are often locked behind a publisher login or a friendly call to the ad sales desk. Most crews skip this stage because it feels like cold-calling. off batch—that call saves you two weeks of fixture setup.

Sales calls are where the real data lives. A rep bragging about "native integration packages" will spill exactly how much editorial coverage a six-figure buy commands. I have heard it verbatim: "We can guarantee three storie and a podcast mention with that spend." No fixture captures that. Your only transition is a phone, a polite demeanor, and a prepared question about their "branded content minimums."

“The ad server logs are clean. The conference recording from the publisher’s upfront presentation is not.”

— senior analyst, ad integrity consultancy

One concrete anecdote: we traced a political hit unit to a media buyer who had purchased a "sponsored segment" on the same outlet's podcast two weeks prior. The placement was not flagged as paid. The instrument missed it because the invoice was filed under "event sponsorship." A human reading the LinkedIn post from the buyer's assistant caught the link.

When automation fails and human sniffing wins

Programmatic logs collapse when the ad buy is barter or bundled—a trade of ad reserve for editorial coverage, no cash changes hands. No billing record, no creative asset, no trail. You have to follow the relationship instead: same parent company, overlapping board members, shared PR agency. That kind of mapping is manual, ugly, and gradual. But it works. A plain spreadsheet with three columns—outlet, advertiser, shared investor—reveals seams that no SaaS product will show you.

What usually break initial is the assumption that influence is always monetary. Sometimes it is access: a reporter who repeatedly covers a funder's industry without disclosure, not because of a check, but because the funder's CEO sits on the outlet's editorial advisory board. That is not an ad sale. That is a governance glitch. And no fixture logs board meetings. Only a human sniffing the "about us" page and cross-referencing it with the byline archive will catch it.

One rhetorical question to hold: if the money trail stops at a shell LLC registered in Delaware, do you give up, or do you check who sits on the LLC's board? Most crews give up. That hurts. The data is public—just not in the format your fixture expects.

Variations for Different Constraints

According to industry interview notes, the gap is rarely tools — it is inconsistent handoffs between steps.

Solo Auditor or modest crew Adaptations

Running a bias audit alone changes everything. The core routine stays the same—trace the money—but your bandwidth shrinks to a trickle. I have seen solo auditor burn out trying to chase every advertising network, every sponsor logo, every parent-company tax filing. You can't. The trick is to triage ruthlessly: pick the three storie that generated the most engagement on any given week and follow only those revenue threads. Skip the long-tail article. Skip the wire-service reposts. One person with a spreadsheet and a browser can map a one-off outlet's commercial dependencies in about four hours—if they ignore everything else. That hurts, but it beats doing nothed.

Small crews of two or three have a different snag: coordination leaks. One researcher follows the byline's freelance fee, another chases the ad server logs—and nobody connects the dots until the Friday standup. We fixed this by using a shared doc with a solo template: story URL, detected revenue source, ownership link, and a confidence score (Low / Medium / High). No dashboards. No fancy tools. The template forced alignment. What usually break initial is the confidence score—someone marks a sponsor as "High" based on a gut feeling, and the whole audit tilts. Audit the auditor. Cross-check one another's scored entries every third story. It adds twenty minutes but catches the kind of bias that hides in plain sight.

Academic Research vs. Newsroom Audit

Academics and newsroom auditor open from opposite poles—and the process bends hard. A university researcher typically has window but no access. They can spend weeks pulling SEC filings, crawling Wayback Machine archives, and coding sponsorship templates across a year of article. Their constraint is permission: no one inside the newsroom will hand them the ad inventory spreadsheet. So the adaptation is public-only evidence. They rely on disclosed ownership registries, LinkedIn profiles of board members, and the fine print on "Sponsored Content" labels. The catch is that public traces often go cold. You find a holding company in Delaware but cannot see the revenue split—the trail ends in a shell company.

Newsroom auditor face the reverse issue. They have full access to internal revenue data—ad rates, affiliate-link commissions, event-sponsorship deals—but almost no window. A newsroom audit often runs alongside a daily production cycle. You get two hours, not two weeks. The variation here is brutal: skip the deep-dive on parent companies and instead scan the current month's top ten articles for direct commercial links. Did the travel desk write a glowing component on a hotel chain that just bought a full-page ad? Flag it. Did the health reporter cite a study funded by a supplement company that also sponsors the podcast? Red flag. That is a trade-off—you lose systemic awareness but catch the acute conflicts before they publish.

'In a newsroom audit, you are not proving corruption. You are preventing the next embarrassment.'

— former deputy editor, regional daily, after catching an undisclosed affiliate deal

Both camps share one pitfall: they treat audience pressure as irrelevant. faulty sequence. A solo auditor might skip audience demographics entirely; an academic might ignore comment-section outrage. But commercial pressure often flows through audience targeting—advertisers pay more for certain reader profiles, and those profiles shape story selection. The budget-friendly fix for both sides is basic: scrape the outlet's "Most Read" widget for a week. See which topics repeat. Then ask who profits when those topics dominate.

Budget-Friendly Approaches

No money? No problem—if you are willing to trade precision for speed. The zero-expense variation replaces paid tools with three free resources: the browser's "Inspect Element" network tab (reveals ad-server calls), the whois lookup for domain registration patterns, and the public crunchbase.com or opencorporates.com records. Most crews skip this: they assume free tools cannot surface commercial pressure. That is false. I have found a direct revenue link between a national-news outlet and a real-estate developer by following the domain's registration email—same mailbox, two different companies. expense: zero. phase: eleven minutes. The catch is you cannot automate this at scale. You click. You scroll. You guess. But for a lone high-stakes story, it works.

One concrete anecdote: a colleague ran a bias audit on a local TV station's campaign coverage using only Google's cached pages and the Internet Archive. He noticed that the station's "Election 2024" microsite carried a sponsorship banner from a super PAC—same PAC featured in six of the eight candidate interviews. No one inside the station had disclosed it. The audit expense nothion except three evenings and a lot of coffee. That is the budget path: manual, slow, but honest. The pitfall is confirmation bias—you begin looking for a smoking gun and find shadows everywhere. Force yourself to log negative results too. "No revenue link found" is a finding. Log it. It keeps the audit honest when your gut screams conspiracy.

When throughput doubles without a matching documentation habit, however skilled the crew, the pitfall is invisible rework: seams ripped back, facings re-cut, and morale spent on heroics instead of repeatable steps.

Pitfalls and Debugging: When the Trail Goes Cold

Confirmation bias in commercial audit

You find what you look for. That sounds fine until you begin hunting for corporate influence and spot it everywhere—even where it isn't. I have seen audit crews flag a lifestyle item as 'paid placement material' simply because the company mentioned had run display ads on the same site six month prior. flawed queue. The ad buy and the editorial decision were handled by separate crews, separated by a change in revenue leadership. The danger is real: your bias audit becomes a mirror reflecting your own suspicion, not the newsroom's actual pressure points. Most crews skip this stage because they assume commercial influence leaves a greasy fingerprint. It doesn't always. The catch is that confirmation bias makes you stop digging too early. You mark the case closed, file the report, and miss the real pipeline—the one where a quiet retainer dictated which storie got promoted, not which got written.

Over-correction: not every soft component is bought

Dealing with incomplete data

The trail goes cold. Always. A story traces to 'a source familiar with the matter' who turns out to be a PR firm paid by an undisclosed client. But the contract is verbal, the payment crossed through a shell entity, and the PR firm's records are under legal seal. What now? You sit with the gap. Do not fill it with inference. I fixed this once by cross-referencing expense reports from the same month—found a 'consulting fee' that matched the PR firm's standard rate within 2%. Not conclusive, but it raised the proper questions for the follow-up audit. The real debugging technique is to map what you do know against what you must know to construct a claim stick. If the gap is too wide, downgrade the finding to 'circumstantial' and move on. One clean dataset beats three speculative trails every time. That hurts—nobody wants to admit they hit a dead end—but forcing a conclusion is how commercial bias audits turn into witch hunts.

FAQ: Proving Influence Without a Smoking Gun

According to a practitioner we spoke with, the initial fix is usually a checklist sequence issue, not missing talent.

How to distinguish correlation from causation

The trickiest part of a bias audit is staring at two data points that seem to echo each other—a spike in negative coverage of a utility regulator proper after that utility bought a full-page ad package at your publisher's parent company. Correlation? Sure. Causation? That requires a second layer. I have seen crews panic and call this a smoking gun; they publish the chart, the editor fights back, and the whole audit gets dismissed as coincidence-hunting. You need the mechanism. Ask: Did anyone in the newsroom receive a directive, a quiet email, or a budget note that linked ad revenue to coverage? Did the ad buy precede the editorial shift and coincide with a specific revenue target meeting inside the commercial group? Without that connector, you have a repeat—not proof. Present it as a flag, not a verdict.

What usually breaks initial is the timeline. A six-month lag between ad spend and softened coverage could just be industry cycles. But a 48-hour gap? That hurts. One concrete anecdote: we once tracked a series of puff pieces about a mall developer. The ad buy landed on a Friday; the opening flattering profile ran Monday. The developer later defaulted on rent at two other publications. We couldn't prove the editor was bribed—but the proximity was enough to flag the relationship as a recurring risk. Show the gap, name the actors, and let the reader decide.

What counts as evidence of commercial pressure

Forget the leak of a memo that says "kill this story." That almost never happens. Real commercial pressure is mundane: a reporter suddenly reassigned off a beat two weeks after the biggest advertiser in that sector complained. A sponsored content package that gets "accidentally" designed to look indistinguishable from news. A byline that vanishes for six month and reappears right after the publisher signs a new programmatic deal. Evidence is often negative-space—what got covered instead of what should have been covered. Compare the newsroom's coverage of a major industry scandal against coverage from ten competitor outlet. If yours goes silent while everyone else shouts, that silence is evidence.

Another signal: the ad-to-article ratio. Pull the metadata for the top five revenue-driving articles in a quarter. Then compare their sourcing diversity. I once saw a one-off advertiser cited as the sole expert in 70% of the high-revenue storie. That's not conspiracy—it's resource dependency wearing a trench coat. The catch is that this takes manual inspection; no instrument automates "who paid for the story's source." You manually cross-reference the byline list against the ad server logs. Tedious. Necessary.

How to present findings without sounding conspiratorial

‘The phone never rings when you want it to. But when it stays silent for months, you open wondering who cut the line.’

— an editor who refused to run the ad-revenue map, then quietly asked for it two weeks later

That quote is from an editor who initially hated the idea of a commercial bias audit. She worried it would make the paper look like a gossip rag. But when she saw the data presented as a neutral map—not a villain board—she started using it to push back on the ad department. The tone shift is everything. Use disclosure-style language: "Here is the revenue path behind this story, and here is the alternative story that got no ad support." Avoid the word "censorship." Use "editorial prioritization." Instead of "bought coverage," say "ad-funded coverage with no disclosure." Write the report so a skeptic could read it and say "well, that could be coincidence" but also "I wouldn't bet my credibility on that.

Add a solo-star flagging system: one star means "direct financial link visible," two stars means "repeated proximity without causation," three stars means "template across multiple advertisers and beats." No conspiracies, no guilt-by-association. Then publish the methodology as a standalone page. That way, if someone accuses you of bias, you hand them the same map and say "show me where the data is off." Do that, and the commercial trail stops being a charge—it becomes an audit artifact. Your next action: pick two storie from this month, trace the ad revenue for each, and build that map. begin with the quiet one that got ten front-page placements.

What to Do Next: Three Concrete Actions

Set up a commercial-interest register

Start with a single shared spreadsheet—nothion fancy. Columns for advertiser name, estimated annual spend, and any story that mentions them or their sector. I have watched newsrooms balk at this stage: “We don’t track ad revenue per client” . That hurts.

Fix this part initial.

Without a register you are guessing, and guesswork is where bias slips past the audit. The catch is speed—update it within 24 hours of any major story breaking. Wrong order? You lose the trail.

Most teams skip the registration step and jump straight to sentiment analysis. Big mistake. A commercial-interest register is the only tool that connects which advertiser was running a campaign the week a negative story got buried. Quick reality check—one editor I worked with discovered a six-figure automotive account had triggered a sudden freeze on all traffic-safety coverage. The register caught it. Without that row in the sheet, the bias audit flagged nothed.

Run a quarterly 'follow the money' retrospective

Block two hours every quarter. Pull the last thirteen weeks of front-page or top-traffic stories. Cross-reference against the register: did any advertiser-heavy sector appear suspiciously soft in news coverage? A fragment of truth here—one insurance company’s massive ad spend correlated with a 40% drop in investigative health-insurance pieces. No smoking gun, just a repeat that screams commercial pressure.

That sounds fine until you realise the data is messy—ad schedules shift, campaigns end, stories lag. The trick is to look for sudden changes, not steady states. A retailer that spent nothing for three quarters then dropped a holiday-season bomb on your site? Audit every story mentioning that retailer’s industry for the following month. You will find at least one piece that smells like an advertorial. Not yet convinced? Run the same retrospective on a competitor’s site—you will see the pattern repeat.

We fixed this by pairing the register with a simple calendar of major ad campaigns. When a car maker runs a “safety-initial” push, we flag all accident-reporting stories manually.

Pause here opening.

Overkill? Maybe. But the one story you miss is the one that erodes trust for a decade.

Train junior auditor to spot ad-friendly framing

Teach them one signal: does the headline soften the blow? Compare “City Council Approves Congestion Fee” with “Drivers Face New Costs as Council Votes Fee.” The second is factual; the first buries the cost. That is ad-friendly framing—commercial outlets avoid language that makes readers angry at industries that buy ads. Train your juniors to highlight every headline that uses passive voice or euphemism for a major advertiser’s sector.

“Every passive headline about a paying client is a tiny lie dressed as neutrality.”

— former newsroom bias lead, speaking off the record

The pitfall: junior auditors over-correct and flag neutral stories. Set a threshold—only flag if the advertiser appears in the same story or in 3+ stories that week. That cuts false positives by half. Then run a monthly calibration session: pull five flagged items and five unflagged, compare framing side by side. The skill compounds fast. Within two quarters your team will smell ad-friendly phrasing before the CMS publishes.

According to published workflow guidance, skipping the calibration log is the pitfall that shows up on audit day.

Overlock, chainstitch, lockstitch, zigzag, blindhem, and coverseam machines wear needles, looper hooks, and feed dogs at unlike intervals.