An opposition research example is a documented case, not a vibe about politics being negative. The useful question is narrow. What did a named source say was collected, and what did that source say was done with it? A page that stacks ancient anecdotes, unsourced lists, and later moral verdicts answers a different question. This guide stays with four modern U.S. cases that a reader can open again, then shows how to write the next example so the source survives the retelling.

The wider method is opposition research. How a public statement is stored, dated, and separated from a paraphrase is the public statements guide. A line is ready to leave the notes only after the check in how to verify political claims before they reach a campaign memo. The worksheet that orders the file is the opposition research checklist.

What an example is allowed to prove

An example proves only the facts the cited report contains. If National Public Radio quotes a journalist calling 1988 a turning point, the example can say that NPR reported that judgment. It cannot say that every campaign before 1988 lacked a research file. If a research starter says an advertisement was produced by a political action committee, the example can say that. It cannot add a private meeting, a budget, or a motive the starter does not print.

That limit is the same one the methodology uses on a current profile. A published claim cites a public record. Posture words such as may, could, and appears consistent with mark an open thread. A historical case deserves the same posture. "The ad ran" is a claim when a source says it ran. "The ad decided the election" is a claim about cause. Most of the sources below do not establish cause. They establish what was said, who described the operation, and where a reader can hear or read that description.

Examples also fail when they migrate onto the wrong person. Governor, U.S. Senate, and U.S. House general-election nominees and certified write-ins for 2026 are collected on the certified nominees page from official state lists. A state is absent from that file when no list is on file. Nothing in a 1988, 1992, 2012, or 2014 case identifies anyone on that list. The case stays in its year. The current file starts from the current filing.

How to read a case before you reuse it

Reuse is the danger. A campaign desk hears "do the Horton ad" or "stand up a rapid-response desk" and treats a nickname as a method. The nickname is not the method. The method is the stack of sentences a second person can open. Before a case moves into a training memo, write three lines.

First, the publication, the date, and the URL or the page. Second, the names the publication actually uses, including the speaker of any quotation. Third, the sentence you will not write because the publication does not support it. That third line is what keeps a famous case from becoming a template for unsourced attacks. Campaign propaganda is the separate question of how a message selects, distorts, or repeats a claim. An example of research is not a license to apply a propaganda label. The label needs its own source.

Step 1. Name the source before the story

Start the note with the source, not with the conclusion. "NPR, February 6, 2007" is a beginning. "Oppo won the 1988 election" is a conclusion the report may not contain. Read the report once for names and dates, and a second time for quotations. A quotation stays inside quotation marks and names the speaker. A paraphrase stays labeled as the publication's summary.

If two publications disagree, keep both. Disagreement is part of the example. Collapsing it into a single production credit is how a later memo cites a document that does not exist.

Step 2. Separate the record from the later retelling

Some cases have a record and a retelling. The record might be a convention speech, a television advertisement, or a recording of a local broadcast. The retelling is a journalist, a participant, or a reference publisher describing that record years later. Both can be cited. They are not the same document.

When the note says "Bush said," it should point at the speech the source quotes, and it should say which source supplied the quotation. When the note says "the ad showed," it should point at the description of the pictures and the words, and it should name the describer. A still image remembered from a documentary is not a substitute for the description you actually opened.

Step 3. Write what the source does not say

Every example on this page has a missing piece. NPR does not print the media buy. The Associated Press sentence on the Akin recording does not print the full transcript. Politico's 2014 scene-setter names stories. It does not attach the underlying court file, thesis, or finance report. The honest example includes that gap in the same paragraph as the fact. A gap is not a failure of the guide. It is the boundary of the source.

The 1988 Horton advertisement, as two sources describe it

On February 6, 2007, NPR published "Opposition Research: Know Thine Enemies." Journalist John Harris, author of The Way to Win, told NPR that the turning point for opposition research was the 1988 presidential race between Vice President George H.W. Bush and Massachusetts Governor Michael Dukakis. NPR reported that by the fall of 1988 Dukakis had a 17-point lead in the polls, tied in that report to his claim of a Massachusetts economic recovery. NPR reported that the Republican National Committee, led by Lee Atwater, created an aggressive research operation and that dozens of volunteers staffed three shifts to feed a 24-hour news cycle.

NPR quotes Bush at the 1988 Republican National Convention in New Orleans: "Should public school teachers be required to lead our children in the pledge of allegiance? My opponent says no, and I say yes." The same report quotes conservative activist David Bossie describing the Willie Horton advertisement. Bossie said Dukakis let Horton out on a weekend furlough program Dukakis supported, that Horton then committed further crimes, and that the episode defined Dukakis and damaged his candidacy. Those sentences are Bossie's, inside an NPR report. They are not a court opinion, and they are not the advertisement's own script.

A separate account, the EBSCO research starter "Willie Horton incident," says the advertisement was produced by the National Security Political Action Committee, a committee that supported Bush, and that it was not officially produced by the Bush campaign. The starter says the thirty-second ad ran for twenty-eight days on cable television. It describes the ad's words, including a claim that Horton murdered a boy in a robbery, received weekend passes despite a life sentence, fled, and attacked a couple, and a closing line, "Weekend prison passes. Dukakis on crime." The starter also records the criticism that the ad used Horton's image in a racialized way, and it records that Bush denied the advertisement was racist.

A research memo that cites this case should carry both links. One link supports Harris's judgment, the poll figure as NPR stated it, the Atwater operation as NPR described it, the convention quotation, and Bossie's characterization. The other link supports the committee credit, the cable run, the ad's described wording, and the dispute about race. Neither link is a full archive of the spot. If the memo needs the pictures, it needs the spot or a source that reproduces them, not a sentence that only names Horton.

The 1992 rapid-response desk, as NPR described it

The same NPR report treats 1992 as the next documented shift. It says Democrats applied the lesson, and that Bill Clinton's campaign built a rapid-response operation around faxes. Clinton communications director Jeff Eller told NPR that the fax let the campaign respond inside the same news cycle and get its side into the hands of the traveling press. Harris added that the campaign used cell phones and early text pagers to suggest questions to reporters covering Bush.

What this source supports is a description of tools and of an intent to answer inside one news cycle. It does not itemize the research files, the staff chart, or which charges were true. A desk that wants a 1992 "example of a sourced hit" will not find that hit in this report. It will find an example of distribution. The research question, "which public record was the fax about," stays open until a different document answers it.

That distinction matters for current work. A fast reply is not a sourced line. The verification step still sits between the note and the sentence someone will repeat. Speed was the point Eller described. Speed does not replace the citation.

The 2012 Akin recording, as the Associated Press described it

An Associated Press explainer, "How do politicians collect opposition research?", gives a concrete recording example. It says that in August 2012 an American Bridge tracker assigned to Missouri's Republican Senate candidate, Todd Akin, heard him talk about "legitimate rape" during a local morning show. The group sent the clip to the media. The AP says the campaign did not recover.

The sentence supports a narrow set of facts: a tracker, a local morning show, a quoted phrase, a group that distributed the clip, and the AP's judgment that the campaign did not recover. It does not, in the line this guide relies on, reproduce the full interview, name the station, or print a transcript. A memo that needs those details opens the broadcast or a source that carries them. The phrase stays in quotation marks because the AP put it in quotation marks. The effect on the campaign stays attributed to the AP.

This case is also a boundary. Hearing a candidate on a local morning show is not the same act as intercepting a private call. The legal limits on recording and on computer access are a different guide. The example here is the one the AP printed: a public appearance, a recording the tracker heard, and a clip moved to reporters.

The 2014 midterms, as Politico framed them

In September 2014, Politico published "How oppo took over the midterms." The piece argues that memorable stories in that cycle, including Alison Lundergan Grimes's bus, John Walsh's thesis, Bruce Braley's chickens, and Pat Roberts's recliner, carried the hallmarks of opposition research. It describes the work as combing public records and tracking candidates for a vote, a court filing, a financial transaction, or a quote. It is a reported essay about volume and tone, not a docket.

Use it for what it is. Politico named those stories and described the method in general terms. The essay does not attach the thesis, the court filing, or the finance report. A later researcher who wants Walsh's thesis cites the thesis or a story that quotes it, not this scene-setter alone. The scene-setter is evidence of how a national newsroom talked about research in 2014. It is not the underlying record.

What the four sources do and do not establish

CaseSource opened for this guideWhat that source supportsWhat it does not establish
1988 Bush and DukakisNPR, February 6, 2007, and the EBSCO Horton research starterHarris's "turning point" judgment, NPR's 17-point figure, the Atwater operation as described, a Bush convention quotation, Bossie's account of the ad, and EBSCO's account that NSPAC produced the cable adThe media buy, the full video, or a causal proof that the ad decided the election
1992 Clinton campaignNPR, February 6, 2007Eller's description of same-cycle fax replies, and Harris's description of phones and pagers used to prompt reportersWhich charges were true, or the contents of any research file
2012 Todd AkinAssociated Press explainerA tracker, a local morning show, the quoted phrase, distribution of the clip, and the AP's report that the campaign did not recoverThe full transcript, the station, or a recording archive
2014 midtermsPolitico, September 2014The names of several stories and a general description of records and trackingThe thesis, court file, or finance report underneath any one story

What these cases do not authorize

None of these sources authorizes a shortcut around identity. A surname in a 2014 story is not a match to a 2026 filing. None of them authorizes a sentence about a private motive. Bossie offered a conclusion about Dukakis. A memo can quote Bossie. It cannot pretend the quotation is the advertisement, or that the quotation is Dukakis's own words.

None of them is a how-to for acquiring nonpublic material. The AP's tracker example is a public appearance. NPR's fax example is distribution. Politico's list is a newsroom's framing. A desk that wants the public-record shelves, rather than a highlight reel, should use opposition research sources and keep each claim on the document that contains it.

A checklist before an example enters a training memo

Use this list on the next case, including the four above.

1. Write the publication, the date, and the URL before any adjective. 2. Quote the speaker, not "the campaign," unless the source itself uses that subject. 3. If a second source disagrees about who produced an ad or who said a line, cite both and say they disagree. 4. Mark any sentence about electoral effect as the publication's judgment, unless the source is an official return. 5. Leave a blank where the underlying record, the spot, the transcript, the thesis, or the filing, was not opened. 6. Refuse to move the case onto a current candidate. Check the office and the year. For 2026 governor, Senate, and House nominees, start from the certified nominees file and the candidate's own filing, not from a historical nickname.

The method for the campaign's own candidate, before a launch, is how to do self-opposition research before launch. These four cases are not that method.

References

NPR, "Opposition Research: Know Thine Enemies," February 6, 2007, npr.org/2007/02/06/7226716.

EBSCO Research Starters, "Willie Horton incident," ebsco.com.

Associated Press, "AP Explains: How do politicians collect opposition research?", apnews.com.

Politico, "How oppo took over the midterms," September 2014, politico.com.

The citation rule for a current file is the methodology. The place a statement is stored is public statements.

FAQs

What counts as an opposition research example?

A usable example names the campaign or cycle, the record or recording that was used, and the publication that reported it. A retelling that drops the source is a story about research, not an example a later desk can check. The field manual is opposition research.

Was the Willie Horton ad produced by the Bush campaign?

NPR's 2007 report quotes a description of the ad as a GOP message that defined Michael Dukakis. A separate EBSCO research starter says the advertisement was produced by the National Security Political Action Committee, not as an official Bush campaign ad, and that it ran on cable. Those are two sources. A memo cites both and does not collapse them into one production credit.

Is a tracker recording opposition research?

The Associated Press described a 2012 case in which an American Bridge tracker heard Todd Akin discuss rape on a local morning show and sent the clip to the media. That is a documented recording used in a campaign. A recording made by breaking a wiretap statute or by entering a closed system is a different fact, and this page does not treat it as ordinary research.

Can a historical case be copied onto a 2026 candidate?

No. A 1988 advertisement does not describe a person on the 2026 ballot. Current general-election nominees for governor, Senate, and House are listed from official state lists on the certified nominees page, and a state is absent there when no list is on file. The historical case stays in its own year.