Computer-Assisted Reporting p.40

By: BRUCE GARRISON NEW HIGH-TECH tools have set in motion an evolution in investigative and daily news-gathering strategies.
Innovative story ideas and approaches to established subjects appear regularly in U.S. daily newspapers. Central to these new approaches is use of the desktop computer.
Computing has been used for investigative and other reporting projects for several decades ? dating back to the mainframe systems used to cover the presidential election in 1952. However, computer use has not been widespread until recently.
""When you deal with mass human behavior, statistics are the way to do so. Most reporters deal with anecdotal information. This goes beyond the anecdotal to describe trends,"" said Steve Doig, Arizona State University Knight Chair professor and former Miami Herald research editor.
""There are different ways of being a reporter,"" he said. ""Using a computer to analyze information and produce a story about human behavior is only one way. It is a new way of doing something we have always done as reporters. But before we had computers, the work took much more time to do ? using index cards.
""What 'precision journalism' does is turn anecdotal evidence into statistical evidence. Instead of saying, 'We found a home that's all beat up,' and then talking about it, which is the approach you always had to use before, now we can say, 'Here's the horrible case of a drunk driver and there are 172 cases like it.' You couldn't do that kind of thing before.""
Because of the new ""big picture"" strategy made possible by computer-assisted reporting (CAR), the name commonly given to this type of journalism, new approaches to old subjects have developed. Furthermore, a widening range of public records has been digitized and some are now also available online.

Turning point
Perhaps the turning point was the development of powerful statistical and relational database software programs for use on desktop computers capable of doing much the same sort of analysis as was previously done on mainframe systems.
Movement toward high-tech journalism can be attributed to easier use, lowered economic barriers, and few regulatory hurdles, recent research by Steven Ross and Dan Middleberg has shown. Larger hard drives, increased memory, and, of course, faster processors also influenced this transition.
How have these recent technological changes affected the nature of news reporting?
In the beginning, projects were the type that needed extra budgetary support and caught editors' eyes. But by the mid-1990s, it had become routine for daily newspapers to make frequent online checks for background, to check other newspapers, to find sources, to identify experts, and to verify information. It was also more common for news organizations to use databases for in-house research as well as for story generation.
Rosemary Armao, a reporter for the Baltimore Sun, is former Investigative Reporters and Editors' executive director. Brant Houston is the current IRE executive director and former National Institute for Computer-Assisted Reporting managing director. In their work at IRE and NICAR, they have observed considerable idea innovation and subject diversity of projects using CAR.
""They [reporters] sense something went awry,"" Armao and Houston wrote in a recent IRE book. ""They saw an irregularity in a system. They possess an imagination for stories. They are indignant over public trust betrayed.""
Ross and Middleberg found that about 87% of journalists in a 1996 national survey said either they or their staff members have Internet access and that 85% go online at least monthly. Both figures increased from those in a similar study a year earlier.
Ross and Middleberg also found a wide range of uses in reporting, including article research and reference, e-mail, downloading data, reading other publications, finding new sources and experts, reading news groups, and reading press releases.
For further understanding of how technology is changing journalism, an ongoing national research project focusing on the development and use of CAR has been underway at the University of Miami since 1993. In the 1995 and 1996 studies, portions of the research focused on the types of topics of CAR daily stories and special projects.
For the 1995 and 1996 studies, data collection began with a mailing to 510 Sunday and daily newspapers with Sunday circulation of at least 20,000. Two follow-up mailings were sent in February and March each year.
One open-ended question sought information about the top three recent stories and projects used in general CAR. A second open-ended question focused on the top three recent stories and projects that used online CAR tools.
In 1995, 287 responses were received, a rate of 56%. In 1996, 233 responses were received, a rate of 45%. The 1995 mean newspaper circulation was 113,735, and respondents represented all regions of the nation (18% East, 33% South, 27% Midwest, and 22% West). The 1996 mean circulation was 105,240, and respondents represented all regions of the nation (22% East, 34% South, 26% Midwest, and 19% West).

Most common CAR subjects
The open-ended question about general CAR subjects sought a response of three recent examples. Responses were analyzed at two levels, the ""first-listed"" subjects and the ""top three"" listed subjects.
Data reveal that there was remarkable range in responses for both years for both measures. No single category was listed by a majority of respondents. In fact, the wide diversity in responses may be the key finding in this study. Some similarity and common subject matter spanned the two years also, but there were notable differences.
Education was the most popular subject in 1996 (see Table 1). Analysis of public school test scores and general administration of school systems (6.9%) combined to make it the most-used topic.
Stories and projects involving local or statewide public education reflected two approaches. One technique involved analysis of standardized student test scores, such as the SAT and PSAT or IQ-type tests. The other approach was school system administrative performance analyses, such as personnel, salaries and pensions, and budget matters that included construction, contracts, and other spending issues.
Education was listed as the first response by 3.8% of newspapers in 1995. While that figure may seem low, it underlines the overall diversity of responses to the question.
Elections, political campaign contributions, and other political subjects ranked second (6.4%) in 1996. This reflects only minor change from a year earlier, but it should be remembered that local and regional election reporting often fluctuates on a seasonal basis.
Elections and campaign contribution stories ranked first a year earlier (5.1%).
Coverage of local and/or regional general crime statistics dropped during the year. It was the second most popular subject in 1995, named by 4.5%, but by just 1.3% in 1996. Another crime statistics story involved specific types of crimes, such as murder or sexual assault, in a particular jurisdiction and this approach was taken by an additional 2.1% of newspapers in both 1995 and 1996. Even added, crime statistics stories account for only 6.6% of the first-ranked responses in 1995 and 3.4% in 1996.
Local spending policy stories, such as those analyzing city budgets or other finance or tax issues, increased from 1995 to 1996. Public spending subjects were listed by 3.5% in 1995 and by 3.9% in 1996. Stories and projects involving local or state courts and court sentencing were listed by 3.5% in 1995, but dropped to 2.6% a year later.
Other leading subjects listed first by respondents included:
u Motor vehicle registration and driver's license data analyses.
u Traffic tickets data analyses.
u Elections data and other types of political stories, local campaign contributors and local candidate contribution analyses.
u Home mortgage-loan patterns data analyses.
u Public salaries and public pensions data analyses.
u Local or regional census data analyses.
Table 2 shows the change in subject rankings when any top three mention was considered. Among first, second, or third rank subjects, public education became the most popular subject for CAR projects in 1996 and the second most popular in 1995. Education stories or projects about public schools' performance, administration, or teachers were listed in 5.2% of all subjects in 1996 and 3.4% in 1995.
Stories involving elections, campaign contributors, or other political issues ranked second at 6.4% in 1996 and first at 5.1% in 1995. General crime statistics stories were second at 3.1%.
Other leading topics included ? in order of popularity:
u Courts and sentencing.
u Local spending policies.
u Census data.
u Traffic tickets or motor vehicle and driver data-oriented stories.
u Stories about gun permit holders.
u Mortgage loan policies and patterns.
u Salaries of public employees.
u Murder or sexual assault crime statistics and truck and car accidents.

Most frequent topics sought online
Respondents listed up to three online research uses or story-project topics.
It is clear from the responses that online research resources were most in demand for gathering background information for a story. There was a dramatic increase in this category from 1995 to 1996. A total of 8.4% of the first-ranked mentions listed use of online research for background, Table 3 shows. This proportion grew to 19.7% of first mentions in 1996.
Perhaps the most surprising change was the second-most mentioned subject in 1996, accessing government information. This use category was listed by 15.5%, but had been listed by only 1.7% a year earlier. This no doubt reflects increased access to online government information at all levels, particularly through newly developed and easier-to-use Web sites.
Locating people and identifying potential sources, as well as specific expert sources, was another widespread use. In 1996, the level of searching for people jumped to 5.2% from 3.5% a year earlier. There was, however, a decline in top-priority use of online tools for finding experts. The figure slipped from 4.2% to 2.6% in a year.
Combining the two different source and people-oriented uses accounted for 7.7% of responses in 1995. That total increased only slightly in 1996 to 7.8%.
The most-often listed subject for online research was 4.9% for business or financial stories in 1995. Another 4.9% mentioned using online resources for stories about the growing public use of the Internet or, specifically, the Web, during 1995.
Crime and courts stories were also frequently named (3.5%) subjects for online research. These categories were not measured in 1996.
Other popular applications of online services included:
u Checking to learn what other publications had written about a subject.
u Accessing online databases.
u Accessing government information.
u Locating information about candidates or about elections.
u Writing school or education stories.
u Accessing financial and business information from government databases.
u Downloading files of varying types from a remote server.
Table 4 shows subtle changes in these rankings that occurred when any of the three responses, regardless of order used by the respondent, were counted in 1995 and 1996.
Story backgrounding remained the leading reason to use online services and it experienced a large increase in one year (from 6.6% in 1995 to 15.6%).
Accessing government information at local and national levels grew considerably in the one year between the surveys. In 1995, it was cited on 5.1% of responses and increased to 9.4% in 1996. Searching for people or for sources ranked third in 1996 at an increase from 2.7% a year earlier. Checking other newspapers' stories on a subject was mentioned by 5.6% in 1996, an increase from 2.8%.
Company and business research grew in use also, from 2.7% to 4.7%. Accessing online databases grew from 2.6% to 3.6%, but finding experts dropped from 2.9% in 1995 to 2.0% in 1996.

Diversity of subjects
Perhaps the most significant finding in this two-year analysis was the diversity of reporting subjects.
There was no single subject that dominated CAR daily stories and special projects in 1995 or in 1996. Instead, a wide range of subjects was reflected in responses to both questions.
Subject diversity also reflects the considerable variety of accessible public data available for CAR. It further shows the independence in approaches to public issues.
It is clear that there are remarkable changes occurring in journalism related to computer-based information gathering. Journalists continue to use traditional techniques, such as hard-copy document checking and interviewing, but CAR tools have certainly expanded journalists' reach and the approaches. In addition to the variety, the widened range of information gathering and analysis is reflected.
One of the major problems using CAR is the inconsistent level of access to digital public information in online database form or in other database forms, such as nine-track tape, Jaz and Zip drive disks, compact discs, and even standard 3.5-inch diskettes. Assuming the resources are available and the skills are in place in newsrooms, differences exist in how journalists can access data. These differences not only exist from state to state, but also from county to county and even at the municipal level.
Clearly, stories using state or local data that were being done in one part of the country were not necessarily being done in other places because of data access issues. Only in the case of data that were available from the federal government ? such as campaign contributions or the census ? were CAR stories or projects involving these subjects possible at any newspaper that had the chance to participate in the two surveys.
In cases where local data were used for stories or projects ? such as those involving educational systems or traffic tickets ? the prospects of approaching these topics is dependent upon availability of data in electronic form or the willingness of journalists to build their own databases from hard copy documents.
The inconsistency in digital data availability may account, at least in part, for the wide variety found.
There is considerable idea sharing for daily reporting and special projects employing CAR, just as there is in other forms of news-gathering. Stories that work in one news market are quickly redone elsewhere, whether they were serious investigative topics involving election contributions or more light-hearted ones, such as dog or cat names or breeds or marriages. Use of electronic mail, distribution lists, faxes, and the Web has facilitated the spread of story ideas.
Data sharing, especially within newspaper ownership groups or through reporting organizations such as NICAR, is also evident in the findings. These factors explain some of the repetition that is evident, although not as much of it seems to be occurring as might have been expected, given the ease in checking what other news organizations are doing.
It also appears that there may be shifts in the degree of skill used as well, if the subjects of general CAR stories and projects and the uses and topics of online stories and projects are indicators.
For example, it is apparent that journalists were not only using online databases with remote connections, but they were downloading files, searching for information, and reanalyzing that information at some newspapers.
Daily stories and projects tended to focus on public issues using public information. While access to private information is not readily achieved, there is potential for stories and projects based on private sector subjects such as private education, business, and private medical care. These topics remained relatively undeveloped in 1995 and 1996.
There were also a considerable number of uses of online resources and stories that were produced using online resources. At the time of these two surveys, reporters, editors, and news researchers were only beginning to appreciate the benefits of online research for their work on stories and special projects.
It is clear that several functions of online research exist in newsrooms, however. Backgrounding, checking in-house clips, checking competing publications, finding experts, locating people, and identifying potential story sources all take advantage of the thoroughness, speed, and systematic approach of both online and off-line database searching now available to newsrooms.
News story and project sources most often using online research were those of a public nature, such as general crime, courts, the Web/e-mail/Internet, candidates and elections, health, and education.
Because organizations in the public sector were among the first to have digitized their data and make them available to the public, they were more often in use and more often used for investigative stories and projects. Other subject areas, such as local crime data, may not be as readily available and fewer stories were being produced.
It is evident in these data that the availability of CAR has created new approaches to stories that have been traditional subjects of investigative journalism, such as crime and law enforcement, the judicial system, politics, and public spending.
Furthermore, it has led to new areas of investigation and analysis, such as performance of state and local school systems, the Web, campaign contributions, traffic tickets, and the census.


?(Garrison is a journalism professor at the School of Communications at the University of Miami. He is
director of the school's four-year national computer-assisted reporting research project. He is also author of ""Successful Strategies for Computer-Assisted Reporting"" (1996) and ""Computer-Assisted Reporting"" (1995).
Additional information about the research discussed in this article is located at http://www.miami.edu/com/car/ index.htm.) [Caption]

?(Table 1
First-ranked recent general
CAR subjects
Table 2
Top 3 recent general
CAR subjects
6.9
5.2
School test scores,
administration analysis
School test scores,
administration analysis
3.8
3.4
19951996
19951996
6.4
5.0
5.1
4.4
Elections, contributions,
other politics
Elections, contributions,
other politics
19951996
19951996
3.9
3.5
Local spending policies
Local spending policies
3.1
2.2
19951996
19951996
Other subjects:
Local economy
General crime statistics analysis
Traffic tickets and DMV data analysis
Murder or rape statistics analysis
Truck and car accidents analysis
Hospitals, nursing homes
Weather, disasters
Courts, sentencing
Public salaries
Census data
Gun permits
HMDA, mortgage-loan data analysis
Other subjects:
Courts, sentencing
Murder or rape statistics analysis
Traffic tickets and DMV data analysis
Census data
General crime statistics analysis
Public salaries
HMDA, mortgage-loan data analysis
Source: University of Miami, computer-assisted reporting research projects.) [Caption]

?(Table 3
First-ranked recent online
CAR subjects-uses
Table 4
Top 3 recent online
CAR subjects-uses
19.7
15.6
Background information
Background information
8.4
6.6
19951996
19951996
15.5
Access government
information
Access government
information
9.4
1.7
5.1
19951996
19951996
Search for people or sources for stories
Search for people or sources for stories
6.4
8.3
3.5
2.7
19951996
19951996
Other subjects:
Company or business profiles
Check other newspapers' stories
Finding experts
Access online databases
Web/Internet-related stories
Crime or courts stories
Find information about candidates, elections
Schools, education stories
Access EDGAR, other SEC information
Download files from remote server
Other subjects:
Check other newspapers' stories
Company or business profiles
Access online databases
Finding experts
Download files from remote server
Crime or courts stories
Web/Internet-related stories
Find information about candidates, election
Health issues
Schools, education issues
Environmental issues

Access EDGAR, other SEC information
Source: University of Miami, computer-assisted reporting research projects.) [Caption]
?( E&P Web Site: http://www.mediainfo.com)
?(copyaright: Editor & Publisher June 21, 1997)