Newsroom Tools for Computer-Assisted Reporting in 1995

Bruce Garrison

School of Communication, University of Miami

P.O. Box 248127, Coral Gables, FL 33124-2030

305-284-2846 (v) and 305-284-3648 (f) bruce@miami.edu

A paper presented to the Communication Technology and Policy Division of AEJMC at Anaheim, Calif, August 10-13, 1996.

Contents

Newsroom Tools for Computer-Assisted Reporting in 1995

ABSTRACT

This paper focuses on hardware and software tools used in newspapers for computer-assisted reporting in 1994 and 1995. The findings are based on two national mail surveys of daily newspapers. The study found hardware and software differences were significant in terms of the size of the newspaper and its available resources. Among the most popular resources in 1995 were 486-type processors, DOS/Windows, 14.4-Kbps modems, Procomm Plus, XyWrite, Excel, and Paradox and FoxPro. There was slight growth in the limited-use of analytical mapping, statistical packages, programming tools, and personal information / text database managers.

Newsroom Tools for Computer-Assisted Reporting in 1995

INTRODUCTION AND LITERATURE

For several decades, the only computers available when needed for newsgathering were the large systems known as mainframe computers. These are the highest-level computing systems that can handle the most difficult and largest of processing tasks. Traditionally, mainframe systems were used by larger institutions such as corporations, educational facilities, government agencies, the military, and research centers. These costly systems are designed for multiple users through time sharing. Mainframes are also most often associated with nine-track tapes, the storage tapes used for data generated and analyzed on mainframe system reel-to-reel magnetic tape drives.

For newspapers and other mass media, mainframe systems were primarily used in the 1950s through the 1970s for purposes mostly outside the newsroom. Companies large enough to have their own mainframe systems, such as metropolitan dailies, wire services, and networks, often used the mainframes for accounting, billing, customer record keeping, and other business purposes. In the 1970s, some news companies began to use them for production-oriented activities such as typesetting in “cold-type” systems (Garrison, 1983). But these mainframe and mini computers, as powerful as they can be, were rarely used as newsgathering tools. At some enlightened newspapers in the 1960s and 1970s, mainframes were used for occasional data processing and analysis, but mainly for special projects such as analysis of political poll data (Meyer, 1979; Reavy, 1996). A few database-oriented projects produced from government data copied onto nine-track tape appeared during this era, but these were extremely rare (Garrison, 1995; Reavy, 1996).

Desktop computing began to appear in the early 1980s. As these more affordable computers became more powerful, applications that had been previously reserved for mainframes began to appear. Gradually, into the 1990s, PCs, especially networked PCs, began to replace mainframes. And slowly, use of desktop computers in newsrooms increased since PCs were easier to use and affordable for all levels of newsgathering.

The result has been a decline in mainframe use around the world, especially in government data collection that previously depended on mainframes or mini computers. At the same time, there has even been a decline in use of mainframes by journalists for news stories. In the mid 1990s, only the largest of databases still required mainframes. A recent massive traffic analysis of 16.2 million Texas driver records and 3.7 million driving citations completed by the Houston Chronicle is an example of a project that still cannot be handled by desktop systems (Hunt, 1995).

At large daily newspapers, large news magazines, the television networks, and those other news media with national or large regional coverage areas, projects involve databases and analysis of millions of records, hundreds of variables or fields, and multiple data tables. USA Today, the Wall Street Journal, and U.S. News & World Report, for example, regularly use computers for certain projects but must produce stories with a national perspective. Their CAR specialists routinely use extraordinarily large databases (Garrison, 1996; Garrison, 1995; Loeb, 1995).

The incredible growth of desktop systems since about 1980 is widely discussed and known (Garrison, 1995; Reavy, 1996). In the past decade, there has also been a rapid growth in the use of portable PCs. In fact, most market studies show portables--- with their increasing power and speed--- taking a larger share of the total number of PCs sold each year in this decade. PCs operating MS-DOS and Windows dominated the world's computer industry sales each year. Yet, despite that dominance, some journalists prefer other computers such as Macintoshes for their newsroom work. Smaller news organizations, for instance, often have entire newsrooms based around Macintosh computers. At news organizations using MS-DOS based PCs, it is common in some graphics or advertising departments to find a room full of Macintosh computers. Some small publishing companies that produce newsletters and other forms of printed mass communication also prefer Macintosh computers.

Use of computers in basic newsgathering – commonly known as computer-assisted reporting – has been growing in this decade. Journalists use computers to connect to others through online resources and they use computers for database analysis in ways in which mainframe computers were once associated. But there may be important differences in the ability to use computers in reporting at some newspapers. Earlier research on uses of computers in newsgathering has determined that variation in computer use exists because not all newspapers use computers in the same way or even use them at all. One of the major distinctions exists between large and small newspapers. “[S]mall newspapers lag far behind large and medium-sized ones in newsroom computerization,” Brooks and Yang (1993, p. 16) concluded.

The present study focused on determining what computer tools were in use in 1995 in daily newspaper newsrooms in the United States. A second area of focused on whether newspaper size, measured by Sunday circulation, was a factor in use of CAR and its tools. Generally, the study sought to answer these research questions:

1. What are the characteristics of personal computer hardware in use for computer-assisted reporting?
2. What are the types of personal computer software in use for computer-assisted reporting?
3. What differences in large and small daily newspapers' use of hardware and software exist?
4. What changes in use of CAR tools have occurred from early 1994 to early 1995?

THE STUDY METHOD

An on-going national study of the development and use of computer-assisted reporting has been underway at the University of Miami in Coral Gables, Fla., since 1993. To begin the project, a national database of newspapers, managing editors, and CAR supervisors was built. From the database, an initial mailing of 514 cover letters, questionnaires, and stamped, self-addressed envelopes was sent to the nation's largest Sunday and daily newspapers in late December 1993. One follow-up mailing was sent in late January 1994. Circulation of minimum 20,000 on Sundays was used as the cutoff point to control research costs.

In late December 1994, the second data collection wave began with a mailing to 510 Sunday and daily newspapers. Again, a circulation minimum of 20,000 on Sundays was the minimum for inclusion in the population. Two follow-up mailings were sent. The first follow-up was mailed in early February 1995 and the second was sent in mid March 1995. In 1994, a total of 208 responses were received, a response rate of 41 percent. In 1995, a total of 287 responses were received, a response rate of 56 percent. Since the two waves of this study involved surveying an entire population, significance tests are not reported. Circulation figures were obtained from the latest available edition of the Editor & Publisher International Year Book (Anderson, 1994; Anderson, 1995). To simplify data presented below, “missing” and “none” responses to questions were combined. In many cases, it was apparent from the response patterns of some respondents that no response to a question was equivalent to no use.

One possible explanation for the jump in participation level is general involvement in at least one of the two major forms of CAR. In 1994, perhaps, more newspapers were not involved in any type of CAR and did not have anything to report. Despite pleas to respond to the survey even if there was no CAR at a newspaper, many respondents chose not to reply. The increase in participation in 1995 reflected a spurt in growth in involvement, even at the most basic level.

Editors were asked either to complete the questionnaire themselves or to forward it to the person in charge of online news research and computer-assisted reporting. In some cases, as many as two or three persons completed various portions of the questionnaire related to their newsroom specializations. Questionnaires were developed from discussions and interviews during the Investigative Reporters and Editors and National Institute for Computer-Assisted Reporting conferences on Computer-Assisted Reporting at Raleigh, N.C., in October 1993 and in San Jose, Calif., in October 1994. The instruments consisted of four sets of questions including institutional and personal information, computer-assisted reporting, online news research, and field reporting use of computers.

The median daily circulation of the 287 newspapers responding in 1995 was 52,800 copies. Therefore, small newspapers were categorized as those with circulations under 52,800 and large newspapers were categorized as those over 52,800. In most cases, variables in the 1995 study phase are reported in contingency tables by circulation size. Data from the two surveys were processed using programs from the Statistical Package for the Social Sciences for Windows, Version 6.1.3 (Norusis, 1995).

HARDWARE AND OPERATING SYSTEMS FINDINGS

With the decline of mainframe systems, which PC desktop systems were most used for CAR? This study measured the usage of processors, operating systems, and other critical hardware. In Table 1, the most commonly used processor for CAR stories or projects was a 80486-type processor that has been on the market since 1993. A total of 43% of all newspapers responding to the question used “486s” in 1995. In 1995, only a handful of newsrooms (3%) used the “Pentium” processor. A little more than one in nine newspapers used any type of Macintosh computer for CAR (12%) and another one in ten still used older 80386 processors for CAR. The differences in small and large newspapers were striking when processors were considered. Larger newspapers, for example, rarely used Macintoshes for CAR but were widely known to use them for production of graphics. Almost one-quarter of smaller newspapers used them for their CAR work. In fact, it was the second-largest category of processor for the smaller dailies and only by a small margin. This suggests that most small newspapers used their hardware for more than one purpose.

Table 2 reveals changes in processors used for CAR in 1994 and 1995. As shown, there was some evidence of upgrading for CAR, moving slowly toward high-end processors such as 80486s and 80586 Pentiums. Perhaps an important observation is that there continued to be a wide range of processors in use, with no apparent shifts in any particular direction between the 1994 and 1995 studies.

Operating systems

For CAR applications, the most-used operating system in 1995 was a combined DOS and Windows set up. More than half of the newspapers responding to the question (52%) reported using DOS with Windows, as shown in Table 3. The combination referred to either basic Windows Ver. 3.x or Windows for Work Groups Ver. 3.11, or any other version of Windows. Since the data in the 1995 survey were collected early in the year, prior to the debut of the large Windows 95 beta program or the final release of the product, the newest form of Windows was not considered in the study. Beyond DOS and Windows, there were other OSs in use, but none in any notable proportions. About 7% of newspapers responding reported using DOS only, interestingly, an artifact of slower and older hardware, it appeared. There were meaningful differences in small and large newspapers also, most likely due to the greater proportion of small newspapers using Macintosh systems.

Table 4 shows the shifts in use of operating systems / platforms between 1994 and 1995. The change that was most noticeable was movement from DOS only to DOS / Windows. Use of DOS / Windows rose from 22% to 42% in a single year, while DOS-only users dropped almost 9% to 6%. These numbers reflected the national shift in use of DOS-only software for data analysis to Windows-based programs that had been introduced about two years earlier.

Compact disc drives

One of the hardware peripherals that has rapidly grown in use because of its ability to store large amounts of data is the compact-disc drive. Beginning in 1995, most new desktop systems were sold with them installed internally in the desktop box and, in 1996, most new portables offered CD-ROM, as well. This was due in part to the large storage capacity and use of CDs for multi-media systems. The most common type is CD-ROM, which is a read-only storage drive. In 1995, CD-ROM drives were becoming more common in newsrooms. In the 1995 survey, Table 5 shows more than half of the newspapers reported using one or more CD-ROM drives. A total of 38% reported using one CD-ROM drive and another 33% reported using two or more CD-ROM drives, or more than 71% used a CD-ROM drive in some CAR capacity. As with other hardware examined in the study in 1995, there were important differences in large and small newspapers. Large dailies typically have more than one CD-ROM drive while smaller dailies were more likely to have only one drive--- and that one was often shared, several responses indicated--- and a much larger proportion had none.

Table 6 shows that 39% of reporting newspapers used one or more CD-ROM drives in newsrooms in 1994 and 58% in 1995. This reflects growth from 4% to 27% of newspapers with two or more CD-ROM drives and a decline from 36% to 31% in the number of newsrooms with CD-ROMs in use for CAR.

Other critical CAR hardware

Hard drive capacities were growing in 1995. Modem speeds and capabilities to fax were increasing. More storage was required and different media for storage were being used. These were the trends in hardware. Additional CAR hardware findings include:

• Hard drive manufacturers in 1995 and 1996 were making larger and faster drives and, at the same time, lowering prices. This meant that newsrooms using personal computers for CAR had greater capacity available to them. And some were taking advantage of the newer drives. The mean hard drive size is over a gigabyte (1,055 MB) in 1995. The sizes reported ranged from a 20 MB drive still in use to a 10 GB drive on a server. The most frequent reported sizes in early 1995 were 1 GB (10%), 500 MB drives (5%), and 2 GB (4%). A year earlier, the mean hard drive size was 853 MB for reporting newspapers. The sizes ranged from a 20 MB drive in use to a 6.7 GB server in use for CAR.
• There has been a shift in modems required by online users in newsrooms. The data transmission needs of World Wide Web, other Internet services, and commercial online service users have forced news organizations to upgrade modems. In late 1994 and early 1995, the industry standard was still 14.4-Kbps, but it was changing to 28.8 Kbps at the time of the 1995 survey, and slower newsroom modems were rapidly being replaced. Table 7 displays modem speeds in 1995. The most popular modem speed for those responding to the question in 1995 was 14.4-Kbps (49%) and 9.6-Kbps (22%), but 11% had 28.8-Kbps and about 2% used 57.6-Kbps or faster connections. Perhaps differences in large and small newspapers' resources are evident in the speeds most often used. There were interesting differences by circulation size and, as Table 7 shows, the proportions were one or more modem generations of faster speeds for larger dailies.
• Much CAR that involves databases requires use of storage media beyond hard drives and floppy diskettes. While this might be acceptable for some databases and some projects, larger databases or limited capacities of hard drives require users to find alternatives for storage. Table 8 shows digital tape (31%) to be the most popular form of storage among those responding, but it should be noted that an even larger percentage, 38%, did not use additional storage media. Optical disks were also in use by about 9% of the responding newspapers and another 16% reported other forms of storage such as nine-track tape, Zip drives, Bernoulli drives, and various forms of CD-ROM drives. Again, there were noticeable differences in large and small newspapers and how each handled data storage problems. A much larger proportion of smaller newspapers had no additional storage media in use than larger dailies, although digital tape was the storage of choice when it is used. Regardless of what is in use, it seemed clear that if new storage technology develops, the database people in newsrooms check it out.

CAR SOFTWARE FINDINGS

Word processing

There are a variety of word management tools, including word processors. As shown in Table 9, XyWrite (27%) remained the most-used word processor and text editor in newsrooms responding, but Microsoft Word (20%) and Novell WordPerfect (14%) were becoming more common as newsrooms moved to PC-based systems and adopted the leading office software suites. Diversity in word processing seemed to be the growing trend, however. There were notable usage differences that occurred when looking at preferences by newspaper size. XyWrite, Word, and WordPerfect were the top three, in order, for all newspapers, but smaller newspapers tended to use other word processors such as those used exclusively on Macintosh computers.

Table 10 shows changes in word processor use from 1994 to 1995. Perhaps the most meaningful changes were brought about by shifts from “missing” to that of a particular word processor. This could have been caused by respondents who did not know the type of proprietary software in use. More respondents knew what was in use in 1995 and reported it. Clearly, there was a shift toward PC-based software such as Microsoft Word, but almost all individual categories increased slightly.

Spreadsheets

Many journalists venturing into the quantitative side of CAR for the first time do so with a spreadsheet program. Spreadsheets have considerable database building and computational power and provide a solid base for fast learning. Because users can learn the basics of database construction and data processing with a spreadsheet program, and because spreadsheets also offer a variety of database manipulation features such as formulas, sorting, advanced bivariate and multivariate statistics, they are ideal for beginners.

Table 11 shows that Microsoft's Excel had become the dominant spreadsheet (31% of those who reported using spreadsheets), but many journalists also used 1-2-3 (20%) and Quattro Pro (17%). However, noticeable proportions did not use spreadsheets or did not respond. The single-greatest difference in large and small daily newspaper use of spreadsheets was not so much the products selected as the use of spreadsheets at all. A total of 33% of smaller dailies did not use spreadsheets while only 8% of larger dailies did not use them.

Table 12 displays the growth of Excel as the favorite spreadsheet. Excel's use grew from 17% to 26% in a single year, while Novell's Quattro Pro and Lotus 1-2-3 grew very little. The fact that there was a 16% drop in “missing / none” responses in the single year indicated more awareness of products in use as well as more use of spreadsheets that began in the past year. The growth of “none” in use was a reflection of more specific answers provided. It is assumed that many of the respondents that would have reported none in 1994 simply did not answer the question.

Relational database managers

For news organizations already using spreadsheets and those needing to go beyond analysis of data in single tables, more advanced tools exist. These include relational database management systems, also called relational database programs. These are popular for what is commonly called “database journalism.” Table 13 reveals that in 1995, Borland Paradox was the preferred (24%) software among those responding, but Microsoft's FoxPro (21%) was also widely used. Another Microsoft relational database program, Access, was growing in popularity as part of the Microsoft Office suite's “Professional” edition, was used by 5% in 1995. There were strong differences in database software use by size of newspaper, but as was the case with spreadsheets, the most glaring difference was in use. Almost half of the small dailies, 47%, did not use the tool. Only 11% of the larger dailies in 1995 did not use a relational database program. Table 14 shows most product use categories increased slightly, but FoxPro experienced the most growth, from 12% to 18% in a year.

Analytical mapping

Analytical mapping tools were high on the wish lists of many CAR supervisors. Many newspapers, so far in the mid 1990s, have added numerous software tools to their CAR resources, but mapping programs had not yet gained widespread use. One of the major reasons, until recently, had been the high cost of the software. Some producers lowered their prices in 1994 and 1995 from original levels that often exceeded $1,000 per product. Most newspapers (71% of those responding to the question) were not yet using computer mapping products, Table 15 demonstrates. But the decrease in this figure shows signs of some growth in use. Of those newsrooms using geographic information systems (GIS) software, MapInfo (15%) and Atlas GIS (9%) were the most widely used. There were also differences in use by size. Larger newspapers (43%) used the product far more than smaller ones (10%). Table 16 shows changes in mapping software use between 1994 and 1995. MapInfo gained in use, from about 5% to 12%, while Atlas GIS did not change in proportion of use.

Statistical packages

Another growth area is use of packages designed for advanced statistical analysis. As analytical demands for CAR projects grow, so does the sophistication of software. However, not many newspapers (just 16% of those responding to the question) used statistical software in 1995, as shown in Table 17. The Statistical Package for the Social Sciences (SPSS) was the most popular CAR statistics tool with 9% use. The Statistical Analysis System (SAS), 3%, and others, 4%, were used by only a few daily newspapers. The differences in large and small newspapers were again characterized by the nature of CAR projects and the need for the most powerful analytical tools, as well as the abilities of staff members to use these tools. Smaller newspapers (3%) did not use statistical packages often. By comparison, 27% of larger newspapers used them. Table 18 shows virtually no change in use of statistical software during 1994-95.

Text database and personal information managers

Specialized database programs called personal information managers (PIMs) are perhaps best for handling an individual's data needs. There are also a handful of text database programs that can be used to mimic many of the features of PIMs and complete other important tasks. While information management tools make sense for reporters and editors to use, they had not found their way into newsrooms in 1995 in any substantial levels, Table 19 indicates. But there was some growth, as Table 20 shows. In 1994, 93% did not use PIMs or text managers, compared to 79% a year later. askSam was used by 2% of respondents in that year. Use of askSam grew about 7%, but use of other products was minimal.

Software development and programming

Perhaps the highest level of computer literacy achieved in newsrooms is competent programming to customize computing needs. There are numerous software development and programming tools and they vary in complexity and computing power. Regardless, most journalists involved in CAR do not program. Some of the most sophisticated users did some original programming, but these were a very small group in newspaper newsrooms, data in Table 21 show. Twenty percent of newsrooms in 1995 used original programming software. Microsoft Basic (5% of those responding) and Microsoft's newer Visual Basic (4%) were gaining some use. As with most of the other software products discussed, there were differences in use according to the size of the newspaper. Programming tools, the limited use of them noted already, were almost exclusively the province of larger newspapers. While 28% of larger newspapers used some form of programming software, only 10% of smaller dailies used them.

Table 22 shows minimal growth in use of development and programming tools from 1994 to 1995. There was a small amount of growth in using Basic and Visual Basic. But many CAR journalists who did programming typically use the macros of their spreadsheet or database programs and these uses were not measured.

Communications software

For CAR that involves online research, E-mail, accessing tools available on a remote computer system, or even filing a story, a communications package is essential. As shown in Table 23, Procomm Plus dominated newsroom preferences. In 1995, about 45% of newsrooms that answered the question use Procomm Plus. All other programs -- such as Crosstalk, Windows Terminal, Smartcom, and others -- constitute only about 39% of the respondents. However, one in six newspapers, 16%, did not use any communications software. As with other software, use of communications packages was a function of size. The primary difference, again, was not as much in the product used as it was whether the category of product was used. While only 6% of larger dailies did not use a communications package, 30% of the smaller dailies did not use communication software. While Procomm was the favorite product of larger dailies, other products, such as those designed for Macintosh computers, were more often favored by smaller dailies. Table 24 shows there was not much change in terms of communications software used. With no breakthrough products introduced, there were some minor shifts.

What tools do CAR journalists want?

In 1995, what sort of hardware tools did news organizations hope to acquire? Table 25 shows CAR hardware and software wish lists in 1995. The most common preference was for new personal computers. It was often the complaint that there were either not enough or no computers devoted to CAR. In some cases, individuals wanted both new computers and upgraded existing systems.

Of the 172 newspapers reporting wish lists in 1995, new computers were at the top of the list for almost 8%. Another 6% sought upgrades for existing desktop computers, pointing to a considerable hardware stress problem at almost one quarter of the newspapers. Third on the list, also about 6%, was analytic mapping software. Internet access or improved Internet access (e.g., World Wide Web) was fourth at about 5% of the newspapers. Another hardware upgrade, adding a CD-ROM drive, was fifth with about 4% of respondents seeking one or more drives.

A wide range of other hardware items followed. These included such items as nine-track drives, local area networks for both CAR computers and the entire newsroom, additional RAM, optical disk drives, and new work stations. Software preferences included relational database programs, programming tools for relational database programs, and statistical packages. Yet some news organizations simply sought basics such as spreadsheets, online services, and databases.

The top general category software was high-end analytical software. The second overall choice was one such tool -- analytical mapping software such as MapInfo or Atlas GIS. It was a first preference by 6%. Improved access to the Internet, which included updating World Wide Web browsers and access connections, was fourth. Database packages constituted a variety of preferences also, including such things as “any database package” or specific relational database programs such as FoxPro by name, Access, or Paradox. Visual Basic also rated high. Some news organizations sought spreadsheet software upgrades, acquisition of statistical packages, text database managers, scanner software to accompany scanners -- presumably for data entry -- and data visualization software.

CONCLUSIONS

Journalism is in a period of technological transition in this decade. Information gathering, an important part of the job of any journalist, is being re-invented. Use of computer tools in the past decade for news reporting has improved as availability of less-expensive tools has increased. The result is a new approach to newsgathering that has emerged in the middle of this decade. The spread of CAR is apparent in the data presented from this study. There was growth in the proportion of newspapers using CAR. There was also evidence of the increased sophistication of CAR among newspapers already using CAR.

Perhaps one of the most significant observations is that journalism is not controlling its own destiny with computer technology. There is no doubt that the computer industry determines the direction of business and home computing in the nation. News organizations, like other businesses, are at the mercy of the computer industry since news companies rarely develop their own hardware or substantial software used in CAR. There are occasions when news businesses have developed computer tools for other purposes, such as pagination and archiving. If any CAR programming is done, it is front-end programming for databases. Original software, such as Elliot Jaspin's Nine Track Express, is an exception instead of the rule. Rarely do news companies fund development of original computer hardware.

The introduction of Windows 95 last August caused numerous discussions in newsrooms and in newspaper information systems departments about possible use. If the decision was made to upgrade, then decisions related to upgrading hardware followed. It could be argued that some of the Fortune 500 corporations may have had significant influence on Microsoft's programmers as they developed Windows 95 and its various features. But it is also safe to assume that most news companies, the wire services, the television networks, and news magazines, were not significantly involved. There is little software and hardware in the PC market that is developed exclusively for the news business simply because of the high research and development costs involved, and few news companies see the investment worthwhile. It seems likely that use of the Internet's World Wide Web for distribution of the news product will change this to a large degree.

Some journalists recognize the influences of the computer industry on newsgathering, but most do not. Some feel that this impact is not so noteworthy. It is certainly important to note that larger news companies have developed their own software products, or had custom development done for them, for daily uses outside of CAR such as pagination and archiving. It is very difficult, if not impossible, to identify a computer-based product that was developed for news use that eventually wound up in use by the general corporate world or the public in general.

Journalists adapt these CAR tools to the best of their abilities. In recent years, software companies have developed upgraded products that offer more and more customizability. This helps journalists with specific applications of these products in mind and provides a mechanism for individualizing these applications for particular newsroom uses, such as ease in data entry or speed in data retrieval or report printing. Despite this, perception of the impact of the computer industry on CAR is mixed.

A handful of respondents in both 1994 and 1995 remarked on open-ended questions that computers and CAR approaches offer a “leveling” device among the reporting resources of large and small newspapers. Perhaps this is so, once small newspapers obtain the hardware, software, data, and expertise to become involved in CAR on a regular basis. But the data presented here demonstrate that there are still considerable differences in CAR resources used by large and small newspapers participating in this study. A small newspaper with the right CAR tools and people on staff to use them can, of course, produce strong daily stories and special projects. But CAR can be expensive, especially for online services, for a news organization on a tight budget. And the effect of “leveling” cannot occur at a newspaper that has not invested in the various resources needed CAR on either a special project or routine reporting basis.

One of the shortcomings of this study was that it did not include the two-thirds of U.S. daily newspapers that have circulations of fewer than 20,000 on Sundays or did not publish a Sunday edition. It would seem to be important to investigate use levels and reasons for lack of use of CAR at these newspapers. Such a project would give a much more complete and appropriate look at all newspapers, especially since CAR usage was beginning to expand in smaller news organizations in 1995. Ideally, and it is within realistic expectations, such a study of computer use and hardware/software configurations for newsgathering would also include broadcast news outlets, news and other major magazines, and news-oriented subscription newsletters.

TABLE 1: Processors, 1995

n = 287, missing observations = 57, * Circulation over 52,800. ** Circulation under 52,800.

TABLE 2: Processors, 1994-95

TABLE 3: Operating Systems, 1995

n = 287, missing observations = 54, * Circulation over 52,800. ** Circulation under 52,800.

TABLE 4: Operating Systems, 1994-95

TABLE 5: CD-ROM Drive, 1995

n = 287, missing observations = 53, * Circulation over 52,800. ** Circulation under 52,800.

TABLE 6: CD-ROM Drive Use, 1994-95

TABLE 7: Modem Speeds in Use, 1995

n = 287, missing observations = 81, * Circulation over 52,800. ** Circulation under 52,800.

TABLE 8: Data Storage Media, 1995

n = 287, missing observations = 64, * Circulation over 52,800. ** Circulation under 52,800.

TABLE 9: Word Processors, 1995

n = 287, missing observations = 41, Circulation over 52,800. ** Circulation under 52,800.

TABLE 10: Word Processors, 1994-95

TABLE 11: Spreadsheets, 1995

n = 287, missing observations = 45, * Circulation over 52,800. ** Circulation under 52,800.

TABLE 12: Spreadsheets, 1994-95

TABLE 13: Database Management Systems, 1995

n = 287, missing observations = 48, * Circulation over 52,800. ** Circulation under 52,800.

TABLE 14: Database Management Systems, 1994-95

TABLE 15: Analytical Mapping, 1995

n = 287, missing observations = 47, * Circulation over 52,800. ** Circulation under 52,800.

TABLE 16: Analytical Mapping, 1994-95

TABLE 17: Statistical Packages, 1995

n = 287, missing observations = 47, * Circulation over 52,800. ** Circulation under 52,800.

TABLE 18: Statistical Packages, 1994-95

TABLE 19: Text Databases / Personal Information Managers, 1995

n = 287, missing observations = 51, * Circulation over 52,800. ** Circulation under 52,800.

TABLE 20: Text Databases / PIMs, 1994-95

TABLE 21: Development Tools, 1995

n = 287, missing observations = 60, * Circulation over 52,800. ** Circulation under 52,800.

TABLE 22: Development Tools, 1994-95

TABLE 23: Communications Packages, 1995

n = 287, missing observations = 63, * Circulation over 52,800. ** Circulation under 52,800.

TABLE 24: Communications Packages, 1994-95

TABLE 25: First Choice for Additional CAR Tools, 1995

REFERENCES

Anderson, I. E. (1994), Editor & Publisher international year book 1994, New York: Editor & Publisher.

Anderson, I. E. (1995), Editor & Publisher international year book 1995, New York: Editor & Publisher.

Brooks, B.S. & Yang, T. (1993, August). Patterns of computer use in newspaper newsrooms: A national study of U.S. dailies. Unpublished paper presented to the Association for Education in Journalism and Mass Communication, Kansas City, Mo.

Garrison, B. (1983, Spring). Impact of Computers on the Total Newspaper. Newspaper Research Journal, 4(3), pp. 41-63.

Garrison, B. (1995). Computer-assisted reporting. Hillsdale, N.J.: Lawrence Erlbaum Associates.

Garrison, B. (1996, in press). Successful strategies for computer-assisted reporting. Mahwah, N.J.: Lawrence Erlbaum Associates.

Hunt, D. (1995, May 14). Ticket to trouble: Wheels of injustice? Houston Chronicle, 2 Star ed., pp. A-1, A-20.

Loeb, P. (1995, October 13). Computer-assisted reporting success stories, unpublished presentation at the Society of Professional Journalists national convention, St. Paul, Minn.

Meyer, P. (1979). Precision journalism, 2nd edition, Bloomington, Ind.: Indiana University Press.

Norusis, M.J. (1995). SPSS for Windows base system user's guide, release 6.1. Chicago: SPSS, Inc.

Overberg, P. (1995, September 19). Personal communication to author.

Reavy, M.M. (1996, March). How the media learned computer-assisted reporting, unpublished paper presented to the Newspaper Division, AEJMC Southeast Colloquium, Roanoke, Va.