Apple, Google news algorithms suppressed negative stories on Michigan Senate hopeful El-Sayed, study says

framing by emphasis: The paragraph frames the absence of negative headlines as a suppression issue without immediately presenting counter-evidence or platform responses.

“Apple News and Google News displayed no negatively framed headlines about progressive Michigan Democratic Senate candidate Abdul El-Sayed”

vague attribution: The study is cited without initial detail on methodology, leaving readers to accept its findings at face value.

“according to a Media Research Center study published Friday”

fear appeal: Bozell's quote implies a risk to democratic decision-making by suggesting voters are being misled, amplifying concern.

“Most Americans don't search dozens of news websites every day”

appeal to authority: Bozell is presented as an authoritative voice without critical examination of his group's ideological stance.

“MRC President David Bozell warned”

selective coverage: Focuses only on stories from right-leaning outlets, omitting whether left-leaning stories were also excluded or whether the omitted stories met general newsworthiness thresholds.

“at least 104 stories from right-leaning outlets about El-Sayed's record, statements and policies that did not appear in the feeds”

source asymmetry: Highlights omission of right-leaning stories without assessing balance or inclusion of left-leaning ones.

“104 stories from right-leaning outlets”

misleading context: Presents methodology without acknowledging its limitations, such as keyword specificity or lack of personalization controls.

“Researchers searched for "Michigan," "Abdul" and "El-Sayed," then used sentiment analysis”

cherry picking: Acknowledges methodological flaws but does not critically assess their impact on the study’s conclusions.

“It did not specify the time or location of the reviews, whether researchers were signed in or how personalization was controlled”

vague attribution: Refers to inconsistencies in the report without naming the source or clarifying who identified them.

“The report also cited 13 stories from left-leaning and other outlets in one section but placed the total at 16 later in the report”

outrage appeal: Mentions a rape accusation without contextualizing its relevance to algorithmic bias, potentially evoking moral outrage.

“Platner was forced to exit the race after a rape accusation from a former girlfriend”

false equivalence: Implies equivalence between El-Sayed and Platner despite vastly different circumstances, potentially to tarnish both.

“MRC President David Bozell connected the latest findings to the group’s previous investigation”

cherry picking: Cites disproportionate representation of left-leaning outlets without discussing whether content quality or newsworthiness was assessed.

“69% of the outlets featured in Google News’ top news selections were left-leaning, compared with just 3% that were right-leaning”

uncritical authority quotation: Presents Bozell’s claim about 250 missing stories without independent verification or platform response at this point.

“MRC’s investigation found that Google News did not feature more than 250 stories”

editorializing: Bozell’s repeated questioning frames doubt as conclusion, implying suppression without proof.

“Our findings raise serious questions about whether Google News is providing users with the complete picture”

fear appeal: Suggests users are being deceived, appealing to distrust in tech platforms.

“consumers should not have to wonder whether major, newsworthy reporting... is being left out”

loaded labels: Labels Hasan Piker as 'far-left streamer,' introducing ideological bias.

“far-left streamer Hasan Piker”

moral framing: Presents policy positions (e.g., abolishing ICE) without neutral context, framing them as controversial by association.

“his support for Medicare for All, abolishing ICE and shifting funding away from police departments”

episodic framing: Presents a single controversial quote without broader context of El-Sayed’s full position or CNN’s framing.

“"Yes, killing tens of thousands of people makes you pretty damn evil," El-Sayed said.”

selective coverage: Repeats the claim about Platner without incorporating Google’s rebuttal, which was mentioned later.

“failed to promote at least 112 stories scrutinizing Platner”

vague attribution: Presents Google’s rebuttal without naming the spokesperson or providing a direct quote until later.

“Google disputed that analysis”

missing historical context: Presents Google’s rebuttal late in the article, after the initial claims have been established in the reader’s mind.

“These claims are totally false and based on a completely flawed methodology”

vague attribution: Cites Apple’s statement without a direct quote or named spokesperson.

“Apple said”