4. Results

This page is part of a graduate thesis. Learn more.

The first section of this chapter will describe some commonalities, or themes, among the arguments in each of the 28 stories analyzed. These themes specifically address RQ2, regarding reasons and evidence; RQ4, regarding important descriptive assumptions; and RQ5, regarding fallacies, particularly Damer’s fallacy of Irrelevant Authority.

The description of each theme will focus more on weaknesses than strengths in the arguments from the reporters; throughout the stories, weaknesses that impeded readers’ ability to accept arguments were more prominent than strengths. But the first section of the chapter will close by looking for a pattern in the stronger arguments encountered during the analysis.

The second section of this chapter will answer the research questions more thoroughly by providing a full analysis of the quality of arguments in four stories from the sample, two each from the Public Service and Investigative series.[1] One conclusion from each of the four stories will also be visualized with a Toulmin model.

The four stories selected for a full analysis in this section aim to be representative of this thesis’s findings regarding each of the research questions. For readers wanting additional evidence, the author’s full analyses of each story from the Public Service and Investigative series are included in the appendix.[2]

For the sake of reducing monotony, in this chapter the six research questions are usually addressed by their content, not number. Numbers in parentheses refer to paragraphs in the story under discussion, and references to Damer all indicate (2005).

Commonalities among the arguments

Insufficient evidence

As stated in chapter 2, this thesis set out with a particular focus on fallacies in journalists’ arguments. But fallacies were not found to be a common source of weakness. Instead, far more stories than expected showed difficulty marshaling enough evidence to support their claims.[3] Some provided little or no evidence at all.

The kinds of conclusions for which insufficient evidence was presented were not limited to broad or narrow conclusions, or long or short stories.

Sometimes, where more than one claim was presented, the arguably more serious claim was the one left wanting of support, stranding it as not much more than an assertion.

On September 6, the Los Angeles Times argued that the city of Bell “went on an aggressive push to increase municipal revenue by impounding cars in the city, police officers say” (1). The police officers quoted by the reporter[4] presented plenty of evidence to support the claim that the city “aggressively pushed” to impound cars (4, 16, 18, 19, 21). The city, too, confirmed the impound effort. On the other hand, none of the quotes or paraphrases of the police or of city officials spoke to the motivation behind the increase.

So although the reporter claimed that the impounds were to increase municipal revenue, which is an important claim, nothing she reported police officers as saying confirmed it. The most direct evidence supporting the reporter’s conclusion about the motivation behind the impounds was a reference to another police officer, James Corcoran, who, the reporter said, told “city leaders” in 2009 that the impounds were meant to increase revenue, not seize cars endangering the city. But the reporter did not tell readers anything about the conversation itself, such as what the city said in return, or how Corcoran “complained” to officials. Was it in a meeting? By letter? Readers weren’t even told the source of the reporter’s knowledge about the complaint or its content. It had no attribution.

Meanwhile, the reporter faced an additional problem in not adequately responding to a counterargument offered by Bell Police Capt. Anthony Miranda, “who has helped lead the department since Chief Randy Adams resigned in July.” Miranda “said the goal [of the increase] was to make the city an undesirable place for gang members by cracking down on traffic enforcement.” The reporter did not present any testimony or evidence to counter Miranda’s claim, and even provided some support for it: In transitioning to a second issue in the story she began, “The enforcement policy may have been aimed at gang members and undesirables, but it put a heavy burden on others too” (10). At the least, this was tantamount to acknowledging that multiple justifications existed for the increase.

So the reporter in this story easily showed that Bell aggressively impounded cars, but offered little to show that the purpose of the impounds was, as she stated, to “increase municipal revenue.” But although the reporter in this case made some attempt to provide evidence for their claims a surprising number of conclusions in stories for both the Times and the Herald-Tribune were offered with no attempt at supporting them.

Some of these conclusions were fairly serious ones. For instance, the Times wrote in their lede on November 2 that “officials in Bell arbitrarily required some businesses to make payments to the city totaling tens of thousands of dollars annually, in at least one case threatening a business owner with closure if he failed to comply, according to interviews and records reviewed by The Times” (1). The claim that “in at least one case” a business owner was threatened was never revisited or substantiated in the story. A less serious instance of conclusion offered without evidence was found in the Times’s story of September 1, which claimed it “highly unusual” that some loans given by the city were made without certain documentation, a claim the reporters never returned to. St. John also committed these omissions, the most prominent will be discussed later in this chapter.

Many other conclusions rested on generic references to “documents” or to a single “expert” of questionable authority. These claims lay somewhere between somewhat-supported and wholly unsupported states. Such claims were found in, for example, the Times’ stories of July 15 (about the national ranking of Bell officials’ salaries), August 8 (about the “norm” of compensation packages for city managers) and September 1 (about some characteristics of loans given by the city of Bell)

“Insufficient evidence” is a relative claim, and some readers might find sufficient what this study labels otherwise. But the evidence described above provides a good sense of what is meant by “insufficient” here. Although no attempt at formal quantification was made, evidence problems similar to the above sketches affected roughly half of the conclusions examined for this study. Bearing in mind that single stories usually contained multiple conclusions, 18 of the 25 Investigative and Public Service stories contained at least one such evidence problem.

Problematic descriptive assumptions

RQ4 asked, What important descriptive assumptions do the stories contain? Although much media criticism focuses on assumptions in stories relating to some sort of political bias, the pattern of note in this study focuses on those descriptive assumptions that are less polarizing, but still capable of instilling doubt in the strength of reporters’ conclusions.

The frequency of key descriptive assumptions in stories was less than the frequency of conclusions with poor evidence, or even instances of appeals to irrelevant authority, discussed below. Instead, what was noticeable, particularly in the Los Angeles Times, were key descriptive assumptions embedded in conclusions likely to draw the most outrage from readers.

For example, in the first story of the Times’ series, the reporters questioned whether the city manager and other top officials in Bell deserved their salaries, each in the hundreds of thousands of dollars.[5] Paragraph 2 attempted to justify their concern by noting that the salary of the Bell police chief, Randy Adams, was higher than those of the Los Angeles police chief, Los Angeles County sheriff, and New York City police commissioner.

If readers accepted the descriptive assumption that, usually, pay for police chiefs (or any city official) should be positively correlated with a city’s population or square mileage, then the comparison gave them reason to think that Adams’s salary was overkill. But readers were given no basis to accept that descriptive assumption.

Even were some basis for it given, the assumption could be easily challenged as oversimplistic, considering other information in the story. The story offered several bases on which city officials’ pay might be structured: the job responsibilities, work product, or what residents feel comfortable paying. Readers could also probably think of other ways to calculate that salary, such as past successes.

The discussion of the salary of Robert Rizzo, the city administrator and the Times’s highest-value target, met a similar fate. His salary was compared to administrators in “a far wealthier city with about 7,000 fewer people,” a city “with a population close to 500,000,” and the Los Angeles County Chief Executive. Each comparison was packed with a descriptive assumption — that pay for city administrators should correlate with population, citizen wealth, or Los Angeles County. While any of these assumptions might be justified, the reporters did not provide a reason to think so or a reason against other possible bases for comparison.

One basis on which the salary might be structured that was given in the story, however, is the law. If nothing else, it probably would have been justifiable for the reporters to assume that an answer to “What should Rizzo and Adams be paid?” is that the pay should be legal. But the reporters quoted a member of the Los Angeles County District Attorney’s office saying that the pay was legal, so that approach wouldn’t help justify their conclusion.

In another story, the Times reporters asked, “How did the Bell City Council get to be exempt from state limits on their salaries?” and answered, “by placing a city charter on the ballot [that passed] in a little-noticed special election that attracted fewer than 400 voters” (1).

A simple reference to differences in the law would have sufficed here as evidence — the law says X under one type of city but Y under charter cities. But the story said that the charter passed in the special election limited council members to “the same salaries as their counterparts in general law cities of the same size. In fact, [Bell City Council members] receive $150 for serving on the council” (17). The story continued: “The City Charter bypasses the limits that state law would impose on pay for boards and commissions,” and, indeed, members of Bell’s boards and commissions earned thousands of dollars monthly (18).

The evidence offered by the reporters contained a key descriptive assumption, then: that all members of the city council were also members of a board or commission when the new charter was approved. Any council members at the time the charter passed in 2005 who were not also members of a board or commission would not have been in line for a pay raise made possible by avoiding state salary limits, which contradicts the reporters’ claim in the lede referencing the council as a whole. And although it might seem obvious that all the council members would have sat on a board, the reporters never said so, and they later noted that only in 2008 did the council mandated that “commissioners must be council members” (13). So there remains some reasonable doubt as to the strength of the reporters’ evidence in powering their conclusion.

All stories must make assumptions, but the descriptive assumptions in these stories are not ones that can be taken lightly. Moreover, they are found in arguments supporting blockbuster conclusions: That officials in Bell took disgustingly high salaries, and that some of the officials, namely those on the city council, did so by gaming the democratic system.

Two other Times stories similarly included descriptive assumptions that posed roadblocks to the acceptance of important conclusions. In attempting to show that Bell citizens had a legitimate gripe against their tax rates, on July 30, the Times left unstated and unjustified a descriptive assumption that property tax rates should be positively related to citizen wealth. In the same story, the reporters provided alternate bases for property taxes, such as improvements to the city or fulfilling previous commitments, both of which, as reported, would qualify as explanations in Bell. The Times’s November 2 story about arbitrary taxation also included a descriptive assumption about the basis on which a particular type of fee was levied that was required for the reporters to show it was levied arbitrarily. That assumption was not only not justified in the story, but was rebutted by a source, whose rebuttal was not countered in turn.

St. John, in the Herald-Tribune, also relied on a questionable descriptive assumption in her November 15 story. She attempted to show that hurricane models were “flawed” in part because the models themselves must make assumptions, while leaving unjustified her own assumption that assumptions are not useful. Generally, though, descriptive assumptions did not play as strong a role in her series.

Irrelevant Authority

Irrelevant Authority, another prominent fallacy, appeared in the two forms described in the previous chapter. One form of Irrelevant Authority consisted of reporters relying on authorities or experts whose weight was questionable. The other form consisted of reporters relying on their own authority to ram a conclusion through to readers. This second form of appealing to authority provided the main justification for nearly the entire series by the Milwaukee Journal Sentinel that won it the Pulitzer Prize for Explanatory Reporting.

Although there were plenty of appeals to authority in the Investigative and Public Service series, there was no noticeable pattern of when questionable appeals would appear. In her October 24 story, for example, St. John generically referred to “industry watchers” (81) to support her claim that insurers’ financial reserves were declining. But she named none of the watchers or described why they would have expertise about the situation.

The same goes for Dave Mora, the “West Coast regional director of the International City/County Management Association, and a retired city manager,” who was called upon, along with “experts in city government,” by the Times in their initial story in the series on July 15 (12). The experts were said by the reporters to have been “amazed at the salaries” Bell paid to its officials. None of these “experts” were named or described, no reason was given to think that the “West Coast regional director” of the management association would be a position with expertise or training in paying city officials, and no sense of whether “expertise” in such things could exist was given. Additionally, Mora’s experience as a city manager was questionable in that the reporters did not say where or when Mora worked as a manager or why readers should believe his experience remained relevant today.

There were instances of similarly questionable appeals to authority throughout the Public Service and Investigative categories, but they were not overwhelming. For that matter, it was not clear that the questionable appeals even outnumbered the substantiated ones.

Although the Journal Sentinel often cited scientific experts in their stories, it appealed most of all to the authority of its reporters on the subject at hand.

In their three stories and 11,000 words about the effort to save the life of Nicholas Volker, the reporters, Mark Johnson and Kathleen Gallagher, rarely named the source of their information within the text of the story itself. The information lacking such explicit sourcing was both broad and minute, including powerful claims about the importance of medical techniques and details about what people said or thought on a particular day. Within just the first story of the series, on December 19, the reporters included information or made claims about:

  1. The contents of email sent between two doctors, and the state of mind at the time of both those doctors.

  2. The frequency of hereditary diseases in America, the implications of reading Nicholas’s genes (which was proposed in the email) for treating hereditary diseases, and how the “sequencing question” presented an unusual “urgency.”

  3. The “new era in medicine” that would begin if DNA sequencing began in 2014 at the hospital where one of the doctors works.

  4. Nicholas’s biography and medical history, including results and details of his surgeries.

  5. Conversations among Nicholas’s family members about his future.

  6. The composition, opinions, and actions of Nicholas’s medical team.

  7. Routines and conversations among Nicholas, his family, and the doctors who at one point performed surgery on him daily.

  8. The dreams (of the sleeping kind) of one of Nicholas’s nurses.

  9. Conversations had as part of conflicts between Nicholas’s family and doctors.

The reporters presented most details in their stories as omniscient narrators. The words “according to” or anything like them rarely appeared, despite the inclusion of facts dating to Nicholas’s birth in 2004 that the reporters could not have directly accessed.

The reporters, then, must have expected readers to accept claims in the story on their authority. How could readers determine the extent to which the reporters’ authority was acceptable? By studying the introduction to the series, reprinted in full here, where the reporters outlined their methodology:

This series began with a tip in January, two months before the case became public. Medical College of Wisconsin experts had sequenced the DNA of a young boy at Children’s Hospital of Wisconsin. In early 2010 only one other case of doctors sequencing a patient’s genes to obtain a diagnosis had been reported in medical literature, and it had received little attention in the popular press.

Reporters Mark Johnson and Kathleen Gallagher received permission from Nicholas Volker’s family, the hospital and the Medical College to tell the story of how Nicholas’ genes were sequenced and how doctors used the information to treat him. The reporters interviewed more than 60 people, including members of the Volker family, the doctors and nurses who treated Nicholas, the scientists who handled, sequenced and analyzed his DNA sample, and experts in genomic medicine.

They also accompanied Nicholas to doctor appointments. Reporters visited a lab at the Medical College to watch scientists sequence DNA.

A number of written sources were used. The reporters read Amylynne Volker’s 500-page online journal chronicling more than three years of her son’s treatment. In addition, they read James Watson’s “DNA: The Secret of Life,” and more than three dozen medical and scientific papers.

Some scenes described in the series were witnessed by the reporters. Other scenes that took place before 2010 were reconstructed based on extensive interviews with the participants and accounts in Amylynne Volker’s journal.

The reporters’ argument from authority, then, rested on separate appeals to different authorities: to the Volkers, scientists, research papers, and so on.

It seems clear that the reporters’ methodology gave them the authority for some of their claims. The thoughts of Nicholas’s mother, for example, probably derived from her journal (which is one of the few sources the reporters explicitly referenced in the articles) or from conversations with the Volker family. The conversations and journal, along with interviews of “the doctors and nurses,” probably made up the sourcing of details about Nicholas’s medical history.

But many of the claims by the reporters remained on shaky ground even accounting for their methodology — in particular, the reporters’ scientific claims, such as those regarding hereditary diseases and the potential “new era” in medicine. These claims regarded developments in medicine and science that would affect the well-being of large numbers of people. Hence, they were likely to influence democratic discussions about, for example, ethical regulations (a topic that appears in the Journal Sentinel series — for example, on December 22, paras. 27–34) or public funding of research.

In their methodology, the reporters cited only one book or journal article by name. The rest were unidentified in the methodology and the stories themselves. But more justification from the reporters was needed to show that readers should consider the reporters such authorities on these matters that they need not cite names. As it stands, it’s difficult to see how the reporters do not fall into Damer’s category of irrelevant authorities. Damer wrote:

an unidentified authority is questionable because there is no way for us to determine whether the unnamed authority is in fact a qualified one. … The testimony of an unidentified authority may have some bearing on the truth of the conclusion, but because it is unidentified, there is no way to know whether it does so. It must therefore be treated as if it is coming from an irrelevant or questionable authority. (70, emphasis added)

Given that the reporters did not claim to have preexisting scientific expertise, readers must assume that their knowledge largely came from the sources listed in their methodology. Yet only one of those sources was named.

The rest were labeled generically: “experts in genomic medicine”; “scientists”; “more than three dozen medical and scientific papers”; “participants.” Readers did not know which “experts in genomic medicine” were interviewed, what they said, or whether the experts disagreed among themselves. Readers did not know how many of the “more than 60” interviews were of “experts” and “scientists.” Readers did not know which “medical and scientific papers” the reporters read, how many of the papers turned out to be relevant to the series, whether the reporters understood the papers, or how they resolved their confusion if they did not.

Perhaps most importantly, the methodology betrayed no obvious reason why readers should accept that the interviews and journal articles actually supported the things the reporters claimed. For example, what did the experts think about the reporters’ conclusion about the “new era in medicine” sequencing can bring? The missing warrant between evidence and conclusion affected the scientific claims, but also more specific claims such as those regarding the state of mind of doctors at a particular time.

Ultimately, as said, this thesis attempts to provide only a metric for readers to judge the acceptability of news articles, not an answer key with which readers can mark articles as passing or failing. Whether the methodology provided by the reporters was enough to set a reader at ease would be up to the reader. For some readers, the Journal Sentinel articles might have been important for their description of the Volkers’ struggle and hope, a subject about which the reporters provided decent reasons to think they have some authority. These readers might have comfortably ignored the shakier, broader claims in the series. But readers who found themselves captivated by all of the issues and ideas the reporters presented would have encountered several obstacles between them and a comfortable acceptance of the reporters’ claims.

Stronger arguments

Much has been said about what went wrong in many of the arguments under review, but, of course, some things went well. There was not a clear pattern, however, of when arguments tended to be strong.

Often, as might be expected, the stronger arguments were straightforward. The main arguments in the Los Angeles Times stories of September 17 and September 22, for example, were essentially descriptions of new activity by officials, which did not require much to report convincingly. Other times, reporters drew on documents to support conclusions, but would go beyond simply citing “documents.” They would provide details about the documents, such as their source or direct quotes from them. St. John’s story of April 19, detailing the crash of an unorthodox insurance company, was an example of a story that used documents in this way to its advantage.

Encouragingly, there were also some instances of reporters who arrived at their own conclusions through study and observation, then reported on them with good evidence. One of these stories, from the Times, will be discussed at length below. In another, in the Times’s story of September 17, the reporters took the hard news of the day, a new investigation into Bell’s finances by state officials, and convincingly demonstrated that the investigation “rais[ed] questions” about the city’s solvency (1).

Similarly, the most interesting part of St. John’s look at the use and abuse of hurricane modeling, on November 15, was its section on the expansion of modeling to include calculations for “costs for political meddling, government ineptness and even human greed” (13). An array of documents and quotations supported her conclusion; a reader would have needed to engage in some serious contortions to quarrel with it.

That said, instances of strong argumentation were not found frequently. Again, although no formal quantification of the stories was made, the review conducted for this study found only two articles in which all of the arguments were unproblematic (still subject to challenge, perhaps, but no so easily rejected).

Individual story analyses

The remainder of this chapter presents in-depth analyses of four stories from the sample, two each from the Public Service and Investigative categories. Each story will also include a Toulmin model of one of its conclusions.

The stories analyzed below in this section were chosen for two reasons. First, they exemplified the general findings regarding the six research questions. Second, they exemplified the range of the quality of argument found in the whole sample: generally problematic, occasionally successful.

‘336 voters opened Bell’s wallet’ (LAT, July 23)

This article raised two issues. The first was, how did the Bell City Council get to be exempt from state limits on their salaries? The conclusion offered by the reporter was found in the lede:

The highly paid members of the Bell City Council were able to exempt themselves from state salary limits by placing a city charter on the ballot in a little-noticed special election that attracted fewer than 400 voters (1).

There was some awkward wording in this conclusion. The council probably did not look for special elections that were “little-noticed” and then place the charter in one of them, but instead placed the charter in a special election that happened to be little-noticed. But in either case, the reporter concluded that the council exempted itself from salary requirements by passing a city charter on a special election ballot.

What reasons and evidence were offered to support the conclusion that the charter exempted council members from salary requirements? As outlined earlier in this chapter, the evidence presented could be accepted only by also accepting an important descriptive assumption.

If the conclusion to “How did the council exempt themselves from state limits” was “by becoming a charter city,” the second issue raised by the article was: Why did Bell seek to become a charter city? Much of the article was spent addressing this second issue.

The reporter did not explicitly provide a conclusion on the issue, but it could be inferred that his argument was that the council members passed the charter to increase their salaries.[6] This conclusion could be inferred starting with the lede, which directly connected council members “exempt[ing] themselves from state salary limits” and “placing a city charter on the ballot.” The second paragraph said that “since passage of the measure, salaries for council members … have jumped more than 50%.” Later, the reporter also claimed that although some council members “insisted that the ballot measure was not motivated by a desire to increase salaries,” they “did not cite any other ways the charter changed how Bell did business” (9).

What basis did the reporter offer for his conclusion? The main reason was just mentioned: that “some city council members” could not name any other way that Bell “changed how [it] did business” as a result of becoming a charter city. But the use of council members’ absent minds was a clear example of the fallacy of Argument from Ignorance. Damer defined this fallacy as:

arguing for the truth (or falsity) of a claim, because there is no evidence or proof to the contrary or because of the inability or refusal of an opponent to present convincing evidence to the contrary (p. 146, emphasis added).

As Damer noted, the argument from ignorance shifts the burden of proof away from the person — in this case, the reporter — attempting to convince the reader of something. However, it was particularly important in this story for the burden of proof to remain with the reporter because of the ambiguity contained in the same sentence: what did the reporter mean by the way Bell “did business”?

The reporter actually cited multiple council members offering reasons in support of becoming a charter city. The reporter said that “the move was billed as one that would give the city more local control” (4); he was told by a then-council member that “the way I understood it, we would have better control of governing ourselves” (12). The reporter also paraphrased as the council members’ “argument in favor of [the measure] that the state had taken $30 million in tax revenues from Bell over the previous 15 years” (24), although it isn’t clear when that argument was made.

The reporter’s claim that council members could not state ways in which Bell changed how it “did business” is now in trouble. The phrase was ambiguous, but it seems fair to say that how Bell “did business” could be related to its “independence” or “control” relative to the state, or that it could be related to the amount of tax revenues it didn’t have to send to the state. If the phrase meant either of these things, however, then the reporter’s evidence has been challenged, for council members could indeed cite changes in how Bell “did business” resulting from the adoption of the city charter. Also challenged would be the reporter’s appeal to expert authority in the form of David Demerjian, of the district attorney’s Public Integrity Division, who was quoted as asking, “What explanation is there for why the city becomes a charter city[?]” (11). According to the reporter’s own work, the council in fact had at least two possible explanations — local control and less taxes — which the reporter did not attempt to rebut.

Because these alternative explanations remained in play, it is difficult to allow the reporter to shift the burden of proof to the council to show that salaries did not motivate the switch to charter-city status. The reporter has more work to do first, namely, defeating the counterarguments from the council members. Inside a Toulmin model (Figure 1), these counterarguments would serve as the Rebuttal to the Data that a new charter was passed that allowed city council members to increase their salaries.

Toulmin model for July 23 LAT story

Toulmin model for July 23 LAT story

But the reporter stated only that “some city council members” could not cite such changes. It could be that those council members were just not the ones he quoted. But this would not assist the reporter. At the very least, it exposes him to charges of being uncharitable to the council members’ argument by not focusing on the strongest version of it. The response could also expose the reporter to the fallacy Damer calls Denying the Counterevidence (p. 172). A more charitable approach would have been to ignore the council members who could not cite changes in Bell and instead try to counter those who could.

Another way for the reporter to have approached the story in general might have been to push harder on his second paragraph, which noted that since the ballot measure passed, council member salaries “have jumped more than 50%, from $61,992 a year to at least to $96,996.” The obvious issue this statistic raised is, “Why have council member salaries increased?” and the likely conclusion is “for no reason.” Associating the increase in salary with the time since the charter passed would be partially shielded from the Post Hoc fallacy (Damer, p. 160) because the reporter demonstrated that the increase in salaries could not have happened without Bell becoming a charter city. Additionally, the reporter would have been shielded from the Argument from Ignorance fallacy because he provided evidence to suggest that the boards and commissions responsible for most of the council members’ salaries served little purpose: “Board meetings in Bell are supposed to take place during council meetings, although their names seldom appear in council minutes. When the boards held separate meetings, they sometimes lasted a minute” (18–19).[7]

In this story, then, the reporter tried to explain the source of the change that allowed council members’ salaries to escalate, and the council’s motivation for that change. Each point would be of crucial importance to a citizen in a democracy, and that the reporter alerted readers to questions surrounding the charter switch was undoubtedly valuable. But because of a needless descriptive assumption, along with an important ambiguity and two alarming fallacies, the reporter’s own answers to those questions were underwhelming.

‘Bell’s codes a cash cow’ (LAT, December 16)

The issue tackled in this story, the 15th of 16 submitted to the Pulitzer committee, was probably so familiar to reporters and readers that the reporters felt comfortable waiting until paragraph 12 to mention it:

The program is a prime example of the aggressive efforts that officials in one of Los Angeles County’s poorest cities were making to raise revenue, some of which was used to help pay the large salaries of top officials.

So the issue was, “What did Bell do to try to increase revenues?” The “program” referred to was one that the reporters tried to show existed through a descriptive argument:

The city of Bell extracted tens of thousands of dollars from plumbers, carpet cleaners, even people scavenging for bottles and cans, by seizing vehicles for alleged code violations and then pressuring the owners to pay arbitrary fines.

In hundreds of cases, city officials created documents that looked like official court papers declaring individuals were making a payment to the city as part of a “civil compromise.” Normally, such cases would be reviewed by a judge to ensure that they had been settled fairly. But the vast majority of these cases do not appear to have been presented to a court. (1–2)

The reporters were unusually adventurous in their conclusions: that the city seized vehicles for “alleged code violations,” created questionable documents that were then signed by citizens, and, as a result, arbitrarily “extracted tens of thousands of dollars.” The reporters didn’t even qualify their claims with an “according to” statement. So what evidence did they offer?

The evidence was labeled as “documents” (16) and “records” (2). However, the reporters described the documents and their brethren more clearly than had been seen in previous stories, providing firmer ground on which their claim could stand. For example:

Normally, such cases would be reviewed by a judge to ensure that they had been settled fairly. But the vast majority of these cases do not appear to have been presented to a court. Times reporters reviewed 164 cases, roughly one-third of those located in the city’s records, and found only three that were filed. All three were dismissed. (2)

The reporters specified not only the number of cases they reviewed but also provided the additional context of how many cases total they believed the city had. Thirty-three percent of the cases seems like a fair number on which to base a generalization, should the reporters or readers want to say that most of the cases in the city’s files exhibited a similar pattern.

The reporters also provided a good amount of detail about the reports:

In each, the person cited for violations signed documents titled “Superior Court of The Southeast/North Judicial District,” referring to the courthouse in Huntington Park. Although most of the individuals had been cited for code infractions, the documents stated that they were agreeing to pay a fine to settle a “criminal action.”

And while the justification for impounding property was that it was needed as evidence for an investigation, many of the cases reviewed by The Times were settled within hours of the citation, and many of the people cited admitted their violations at the scene. There were no reports showing that any evidence was gathered from the impounded cars. In addition to the fines, the city charged a $400 release fee and received additional money from the tow yard depending on how long a car had remained there. (17–18)

Additionally, and crucially, the reporters put a copy of one of the settlements online and linked to it in the story. The file showed the fine paid by the citizen, the supposed code violation, and pictures of the van that was impounded.

True, readers would have had to take on the reporters’ authority that the remainder of the documents reviewed showed something similar to the online version. Only one “arbitrary” set of fines was listed in the story — two strawberry sellers, one fined $75 and the other $200 (8) — and Figure 2 outlines in a Toulmin model how these Data would demonstrate the Claim. The reporters’ process of reviewing the cases also wasn’t clear. But the reporters’ attempt, anyway, to contextualize and detail these documents gave readers a much stronger reason than had previous stories for accepting that authority.

Toulmin model for December 16 LAT story

Toulmin model for December 16 LAT story

A second source of evidence came from anecdotes with Bell citizens who said they were victims of the city’s code enforcement (9–10, 19–24, 27–31). The reporters did not specify why the three individuals were chosen or how the conversations took place, but they paraphrased each as saying that their vehicles were impounded and that they were fined, in line with the reporters’ initial claims. By themselves, of course, the anecdotes were not enough to justify the claims made by the reporters in the story. But fortunately for the reporters the anecdotes combined with the strong documentation to present a fairly convincing story. It also helped the reporters’ case that they provided three anecdotes, more than most stories in the series offered.

The reporters on this story, then, presented a generally well-researched and fairly clear argument in support of their claim. But it’s worth noting an important ambiguity in the story. Although the statement of the issue referred to a “program” of impounds and fines, and other paragraphs to a “practice” (3, 5), nothing in the statement of the conclusion in the first two paragraphs suggested anything systematic or programmatic took place.

The conclusion could be reasonably read to say simply that Bell “extracted tens of thousands of dollars” in total from citizens at various times. But perhaps readers are supposed to infer that the “program” meant not just the fines but also the creation of “documents that looked like official court papers.” This second interpretation seems plausible given the reporters’ reference to their review of “more than 450 code-enforcement cases involving civil compromise,” suggesting the fines and suspect documents were meant to be considered jointly. But even that reference seems to suggest something more of a “practice,” something that happened to be done repeatedly, than a “program,” something more formal.

In any case, the harm caused by this ambiguity was not that the reporters couldn’t demonstrate what they said about the fines, strictly speaking; rather, the harm was that context of what they described was obscured.

‘Weak insurers put millions of Floridians at risk’ (H-T, February 28)

This first story in St. John’s series asked a broad question: “Are private insurance companies in Florida equipped to protect homeowners in case of a hurricane?”

She opened with what appeared to be a clear statement of the conclusions she planned to demonstrate:

The Herald-Tribune spent more than a year examining Florida’s property insurers, tracing the ownership of more than 70 companies through shell corporations and reviewing the financial filings of each. It found:

  • One in three privately insured Florida homeowners relies on insurers that exhibit one or more signs of financial risk.

  • More than 100,000 homeowners relied on companies barely capable of paying for house fires, let alone hurricanes. These insurers’ reserves come so close to the state’s $4 million minimum requirement that they operate with only a few hundred thousand dollars of their own to pay claims.

  • During the 2009 hurricane season, at least 38,000 Florida homes were insured by companies state regulators knew would fail. Homeowners were not told until after hurricane season, when one company was shut down and the other had to sell.

  • Lawmakers and regulators have ignored warnings and encouraged private companies to stretch their limited cash further. They have pushed companies to insure more and more homes without increasing the money set aside to pay claims, a practice that put state residents farther out on a limb.

  • Larger dangers loom. Despite rising property values, one in three Florida carriers has decreased the cash set aside for storms. (6–11)

St. John provided other conclusions in the story besides the above. For example, her lede stated that “millions of Floridians now bet their homes on property insurers that teeter on the edge of financial failure.” The second section, under the heading “Why upstart insurers dominate in Florida,” was a short argument about that topic. The clarity of the story was harmed, in fact, by the lack of a clear issue statement, leaving readers unsure of how or whether these non-bulleted conclusions were meant to mesh with the neatly listed ones.

This analysis will focus on only the claims in the list, primarily because of the language used to introduce them. It would be disingenuous for St. John to present the list, prefaced by describing her yearlong examination of insurers, and not expect readers to take it seriously. The list screams for a critical examination.

So, starting from the top, what evidence did St. John present to demonstrate that “one in three privately insured Florida homeowners relies on insurers that exhibit one or more signs of financial risk”?

She was fairly clear in how she defined whether a company was at risk and why she chose that definition. She consulted with “a half-dozen experts” from “the industry” who gave “several important indicators of financial weakness and they provided benchmarks for each” (60–62). She turned to people in insurance, she said, because the state “will not name the companies [‘on the verge of collapse’] or say how many are in trouble” (59). However, she did not specify what records she examined, where she obtained them, or her methodology in examining them.[8]

Reporting the result of her examination, she wrote:

The Herald-Tribune found that about 30 companies out of more than 70 reviewed appear fiscally sound. Forty-two failed at least one of the benchmarks.

That means one in three privately insured homes in Florida – some 2 million families – relies upon an at-risk insurer for hurricane protection. (63–64)

Her reasoning, outlined in a Toulmin model in Figure 3, contained two weaknesses that make it difficult for readers to accept.

The first weakness was an unsteady use of terms. Her statement of the conclusion in bullet-point form referred to insurers “that exhibit one or more signs of financial risk.” However, she later said her data “means” that one in three privately insured homes “relies upon an at-risk insurer.”

The latter statement was stronger. It said not that the companies showed a sign of financial risk but that they actually were at risk. It also conflicted with St. John’s statement three paragraphs earlier:

A half-dozen experts consulted by the Herald-Tribune cautioned that no single measure told the strength of an insurer. (61)

Her experts cautioned that “no single measure told the strength of an insurer.” St. John reported that 42 companies failed at least one benchmark, but that only 14 failed two or more (65). So 28 firms of the 42 failed only one benchmark, which, by her previous statement, was not enough to determine “the strength of an insurer” and certainly not enough to show that the insurer was “at-risk.”

So were readers to accept either the stronger or weaker form of St. John’s conclusion, they would be better off accepting the weaker one — which required showing only what St. John does, that the companies showed signs of financial risk — but they must parse St. John’s flip-flopping prose to get there.

Toulmin model for February 28 H-T story

Toulmin model for February 28 H-T story

The second weakness concerned St. John’s conclusion that “one in three privately insured homes” relied on (take your pick) an at-risk insurer or one with a sign of failure. There simply were no data offered to indicate the source of the population figure. That 42 of “more than 70 reviewed” companies showed warning flags doesn’t indicate anything about the number of people those companies insured. Presumably she didn’t mean that one-third of insurers showed signs of risk, because even 42 divided by 80 is more than 50%, not one in three, and had she reviewed “more than 80 companies” she probably would have said so. The reader simply must take on her authority that the number of people affected by the “at-risk” companies was what she said.

Moving on to St. John’s second conclusion:

  • More than 100,000 homeowners relied on companies barely capable of paying for house fires, let alone hurricanes. These insurers’ reserves come so close to the state’s $4 million minimum requirement that they operate with only a few hundred thousand dollars of their own to pay claims. (8)

Not enough evidence was presented to support this conclusion, although it came close. Under the heading of “Not enough money to pay off house fires,” St. John discussed the quick rise and fall of an insurance company called Northern Capital Select.

Financial statements and reinsurance contracts show that in 2009 it was operating with barely a $300,000 cushion above what it needed to meet state solvency requirements – not even enough to cover a handful of house fires. (45)

Assuming that St. John was correct to say that $300,000 is insufficient to cover house fires, she left out any detail about how she knew that Northern Capital Select insured more than 100,000 homeowners — a fairly easy fact to include and one that was necessary to demonstrate her claim.

Further in the story St. John discussed the even quicker rise and fall of another insurance company, Magnolia. She did mention that Magnolia was “responsible for the financial security of 100,000 homeowners” by the end of 2008 (98). But throughout her discussion of the financial collapse of the company she never said anything about whether the company’s coffers were falling toward state solvency requirements, as she did with Northern Capital Select, and which she said would put them in house-fires-only territory.

It is certainly possible, even likely, that Northern Capital Select insured more than 100,000 homes or that Magnolia could have covered only “a handful of house fires.” But it was left for readers to assume so. St. John instead omitted simple statements that would have provided the final push in presenting an easily acceptable argument.[9]

Bullet points 3, 4, and 5 can be addressed swiftly: St. John never returned to them. Not a word was said in support of her accusation that “lawmakers and other regulators have ignored warnings,” that “one in three Florida carriers has decreased the cash set aside for storms,” or that “larger dangers loom.”

A person reading her whole series would eventually find that some of these conclusions were addressed more fully in future stories. But no reader of only the first story could know that, especially when the second story in the series, analyzed below, appeared two weeks later.

Possibly the most glaring weakness in St. John’s story, then, was that it was unclear to readers what she wanted to argue. The problem goes beyond not explicitly stating an issue or conclusion, which happened occasionally in the Los Angeles Times stories. Instead, there were so many of what seem to be conclusions that when they weren’t addressed again, readers could be confused as to whether it was they or St. John who missed something.

‘How insurers make millions on the side’ (H-T, March 14)

St. John again wrote an extensive story that covered quite a bit of ground. It addressed the issues “How do Florida insurers use affiliated companies to generate profits?” and “Does the arrangement help or hurt consumers?”

The existence and purpose of affiliated companies were not in question in the story. Affiliated companies are companies that run day-to-day business for insurance companies, though they can be found in other sectors, and they are usually owned by the same people who own the insurers. Premiums sent to the insurance firms are often sent on to the affiliates in return for the affiliates’ “services.” There, the premiums emerge as profit because, as the affiliates aren’t formally insurance companies, they’re exempt from financial regulations the state of Florida places on the insurance industry.

According to St. John, both regulators and the insurance companies see affiliates as a means of quickly generating profits for insurers that, as startups, have limited revenues. Affiliates thus serve the important function of helping keep a steady amount of insurance available for Floridians when it is not always easy to find after the departure of major insurance firms earlier in the decade.

If the interpretation of the two issues addressed in this story seems broadly drawn, it is. It is broad because this story, like the previous offering, struggled with an opacity of purpose. St. John offered many potential conclusions in her introduction, some of which were eventually addressed and some of which weren’t. As with the first article, she sprinkled new conclusions throughout that had a tenuous or unclear relationship to where the article began.

So, for example, the lede contained a descriptive conclusion:

Today, nearly half of Florida’s home insurance is provided by companies whose primary profit comes not from insuring homes but from diverting premiums into a host of side ventures.

But no data, or even an explanation of how she got the “half” figure, were presented to back this claim.

Shortly thereafter, she introduced her investigation again:

The Herald-Tribune spent more than a year investigating the Florida insurance industry, including reviewing the financial filings of more than 70 Florida-only companies that now provide nearly three-quarters of the private property insurance in the state.

It found that: (8–9)

Followed by another bulleted list of claims.

Those bulleted claims will be the focus of the examination here, primarily because of, once again, their presentation as a list transitioning into the next section of the article. What’s more, the bullet points differed from those in the first article, suggesting that each story meant to cover different conclusions from the investigation.

St. John’s first finding read:

  • Overhead costs – expenses not related to hurricanes or other disasters – are 50 percent higher in Florida than the national average. The higher overhead cost Florida homeowners an added $900 million in 2009 alone. (10)

There was evidence presented for this claim. But the evidence suggested that the claim was very poorly written, in the best light, or close to dishonest in the worst. It arrived in the next section:

A yet-unpublished analysis by the Insurance Consumer Advocate, an independent state position created by the Legislature, found that Floridians pay 50 percent more for overhead costs than the national average. (21)

Unless the newspaper conducted an identical study, which St. John did not cite, then clearly the newspaper did not “find” the statistic regarding overhead costs. The state Insurance Consumer Advocate did.

Perhaps all St. John meant to say in her finding was that the investigation found that the Advocate found that overhead costs were high. But that potential meaning did not come across in the actual article. Instead, readers were led to believe the newspaper uncovered the fact by its own elbow grease, as visible most clearly when the three paragraphs are shown together:

The Herald-Tribune spent more than a year investigating the Florida insurance industry, including reviewing the financial filings of more than 70 Florida-only companies that now provide nearly three-quarters of the private property insurance in the state.

It found that:

  • Overhead costs – expenses not related to hurricanes or other disasters – are 50 percent higher in Florida than the national average. (8–10)

In any case, the conclusion regarding the state’s high overhead costs was demonstrated, albeit by an argument from authority (details of the report or its methodology were not revealed). The sketchy wording appeared to be a one-time slip, or perhaps an instance of miscommunication among reporter, copy editors, and news editors. But it was troubling nonetheless. Moreover, the second part of St. John’s first bullet point, regarding the additional $900 million that Florida homeowners supposedly paid, was not substantiated in the story or even cited to the Advocate’s report. Readers were not able to tell where the statistic came from, making it difficult to accept.

The second conclusion concerned whether companies “inflated” their overhead costs, where overhead costs referred to the fees that insurers sent to their affiliate companies for managing the insurers’ daily affairs:

  • In cases where the Herald-Tribune could see both sides of the ledger, the overhead charges were inflated. Of the $72 million in management fees that Southern Oak paid its affiliate over five years, nearly half – $35 million – was profit, insurance regulators now say. Three other carriers paid themselves an average 44 percent profit. (11)

There were some data missing that would have helped readers understand this claim. First, “inflated” was undefined, although the surrounding statements seemed to indicate it meant that the insurer simply paid more than what the services cost. Second, readers weren’t told how many cases there were in which St. John could “see both sides of the ledger,” or exactly what that meant. So readers don’t know whether the four carriers mentioned in the paragraph were meant to comprise the only examples of being able to see the ledgers.

St. John discussed three examples of “inflated” overhead charges in greater detail. Her first target was Homeowners Choice, which “paid Homeowners Choice Managers $24 million (and $2 million more to others) in 2009 for management services that cost $15.4 million” (34). The data appeared to be credited to “financial reports [filed] with the Securities and Exchange Commission,” although it was not clear which documents were used or how they revealed those figures. Her second target was First Home, which was mentioned in one paragraph, and which the state Office of Insurance Regulation had ordered to “return $1.3 million in management fees” (86). Her third target, in the next paragraph, was Southern Oak, which had recently been

ordered to show why it should not return $10 million in “excessive profit,” a portion of the $35 million in profit regulators said the [affiliate company] made off the insurer since its inception in 2004. (87)

These three cases went some way towards demonstrating that some companies inflated their payments to affiliates beyond what the services actually cost. That said, the stronger case would have been to attempt to falsify the conclusion rather than to present examples demonstrating in the end only a generalization. Such a case would have showed that St. John searched for and failed to find a company that did not inflate its costs.

St. John’s third conclusion, that “Some insurers devote so much of their premium to reinsurance and paying related companies they have little left for claims” (12) was not addressed again.

The fourth conclusion received some attention. The conclusion was:

  • Insurers have contracted so much of their work to unregulated sister companies that some are essentially shell operations with few employees. Homeowners Choice, for instance, pays one affiliate to negotiate reinsurance contracts and another to manage policies, and buys catastrophe protection from a third. (13)

The second sentence of the paragraph, with data about the roles of Homeowners Choice affiliates but not about the insurer itself, was insufficient to demonstrate the conclusion, even regarding just Homeowners Choice, let alone “some” companies. Later, St. John reported the result of trying to contact an insurance firm called Hillcrest:

The Herald-Tribune also attempted to reach Smith and his family through their insurance company, without success. There is no Hillcrest office to contact. The company pays the Tower Hill insurance group to run its business.

“We’re what you call a ‘virtual operation,’” said Hillcrest chief finance officer William Thompson, who earns his $172,000 salary working from Tallahassee. (63–64)

It is unclear how much recounting these efforts helped show her conclusion. Clearly, Hillcrest did contract its work to other companies. But was it a “shell operation[] with few employees”? It was difficult for readers to know — no data on that question were provided. Readers knew it had at least one employee, the CFO. To have him call the operation “virtual” was as ambiguous as it is dispositive: Did Thompson mean that the employees, however many there were, work online or that the company didn’t have much of an existence? The role of this Rebuttal is seen in the Toulmin model for this conclusion in Figure 4. St. John’s conclusion, then, was partially demonstrated, but readers probably would want more data about before accepting it.

Toulmin model for March 14 H-T story

Toulmin model for March 14 H-T story

The fifth conclusion also received attention, but, again, was plagued by changes in terms.

According to St. John, “Lax state rules encourage executives to pay sister companies as much as possible. The Legislature barred regulators from requiring insurance affiliates to report their finances” (14). There is an unaddressed assumption here about an incentive to move money to where it’s subject to the fewest regulations, but that can be accepted for now. What was more troubling were the reasons St. John used to demonstrate her claim:

Like most states in the mid- 1990s, Florida adopted model laws aimed at regulating how insurers use managing companies called MGAs.

But in Florida, the Legislature added words excluding the most common kind of managing agent in the state, those controlled by the insurance company’s owners.

So there are laws that require managing agents to charge a fair rate and allow regulators to audit their books, and laws that impose penalties for violators.

But those laws do not apply if the insurance company owners form their own MGA and charge themselves for the services. (79–82)

This proved only a weaker version of St. John’s initial claim, which was that regulators were “barred from requiring insurance affiliates to report their finances.” According to her later reporting, only a set of affiliates, albeit the “most common” set, faced “lax state rules.“ This dissonance between reasons, evidence, and conclusion could easily have been remedied through an additional word (”some sister companies"), but as it stands, readers can’t be sure whether to assume it was a slip-up and accept the weaker conclusion, or to not be able to accept yet the stronger, originally stated conclusion.

Finally, the sixth conclusion:

Even while complaining of losses, Florida insurers from 2006 through 2008 paid $38 million in bonuses and $32 million in other perks to 180 of their officers. (15)

was not addressed again. There was no source given for the bonuses data and only one further mention of bonuses in the story, that of “$525,000” given to one executive “for the past two years” (39). There were no quotations of insurers “complaining of losses.”


This chapter has reviewed the results of an analysis of the reasons, evidence, and conclusions in a total of 28 articles from three Pulitzer Prize-winning series‘. The analysis found three common causes of weak argument among the stories: Appeals to irrelevant authority, particularly in the work of the Milwaukee Journal Sentinel; important descriptive assumptions that instill doubt in the strength of a conclusion without any explicit justification; and, perhaps most importantly, a surprising lack of evidence to back many of reporters’ claims.

This chapter also provided a full analysis of four stories from the series. These stories highlighted the themes described in the first half of the chapter while also hinting at some other problems found in the stories, such as unclear (or missing) statements of reasons and conclusions. The story-by-story analysis also presented the December 16 Los Angeles Times story as an example of an argument excelling.

The next and final chapter, then, presents some additional commentary about the results and concludes with suggestions for further research.

  1. The Explanatory series, which relied heavily on arguments from authority, will be analyzed in the thematic discussion of that type of argument below.  ↩

  2. They will also be available on the author’s website, https://argumentinjournalism.wordpress.com.  ↩

  3. “Claim” and “conclusion” are used interchangeably in this chapter.  ↩

  4. Although this thesis continually refers to “the reporter” and “reporters” as the progenitors of the arguments under analysis, it acknowledges that such characterizations are oversimplifications. The news reports seen by the public are the product of complex interactions not only within newsrooms but among journalists, sources, market forces, and the “public” itself (Gans, 1980). The next chapter addresses this contextual gap in its suggestions for further research.  ↩

  5. The reporters did not explicitly say that “top officials in Bell are overpaid.” But much of their story consisted of providing evidence that answered the question “Do top officials in Bell deserve their salaries?” in the negative. Kovach and Rosenstiel confronted a similar scenario in analyzing The New York Times’s reporting during the 2008 presidential race that Sen. John McCain might have had an affair with a lobbyist. Drawing on semiotics — “the study of understanding language and symbol” (p. 113) — they considered it reasonable to say that the story’s conclusion was that McCain had the affair even though the report didn’t say so outright. The connotation and implication of the words used in the story, though, conveyed it (2010, p. 112).  ↩

  6. This was the second instance in as many stories of having to try to find an unstated conclusion to make sense out of the many paragraphs related to, but not part of, the initial claim of a story. More instances followed as more stories were analyzed. The entire exercise evoked comparisons to May’s (1988) argument that reporters largely argue by invited inference. Readers were invited to infer something based on these extra paragraphs that contain important information. It seemed the paragraphs must serve some purpose, rather than loiter as isolated nuggets of knowledge, but readers are rarely told what that purpose was. An inference must be made, which made readers’ task of analyzing the journalism that much more difficult.  ↩

  7. The lede of the story also hints at a third issue: Why did the charter pass? It’s possible to say the story raises this issue because of the headline’s and lede’s reference to low voter turnout. However, it isn’t clear that the reference to low turnout is meant to serve as a reason for any conclusion about why the charter passed — that is, that the charter passed because the election was “little-noticed.” If the reporters did mean to link low turnout to the charter’s passage, they would also have had to justify their implication that a higher turnout could have led to a different result in the election, and no such justification is given.  ↩

  8. To St. John’s immense credit, however, she and the Herald-Tribune compiled an online database of insurers and information about them such as “capital” and “South Florida Risk.” Readers unconvinced or curious about her conclusions can use the database to supplement the information in the stories. As said, however, the potential availability of evidence that would influence the strength of St. John’s written arguments, while important in a larger sense, are not considered in this thesis’s analysis.  ↩

  9. Neither Northern Capital Select nor Magnolia were included in the Herald-Tribune’s online insurance company database.  ↩


Leave a comment

Fill in your details below or click an icon to log in:

WordPress.com Logo

You are commenting using your WordPress.com account. Log Out /  Change )

Google photo

You are commenting using your Google account. Log Out /  Change )

Twitter picture

You are commenting using your Twitter account. Log Out /  Change )

Facebook photo

You are commenting using your Facebook account. Log Out /  Change )

Connecting to %s