Characteristics of Science and Engineering Instrumentation in Academic Settings: 1993

CHAPTER I

Aggregate Stock of Scientific and Engineering Instrumentation in Research Usage in Academic Settings

Data Considerations for this Chapter[13]

This chapter presents an overall picture of research instrumentation at approximately 1,500 departments and facilities located in the 318 academic institutions represented by the Instrumentation Survey. It provides inventory information—such as the estimated total numbers of instruments and their total aggregate cost—on all in-scope instruments costing $20,000 or more at these institutions, for supersystem and non-supersystems.[14] This inventory-based information corresponds to the data on instrumentation expenditures and future instrumentation needs at these same departments and facilities; the latter data were collected via the survey's Department/Facility Questionnaire.[15]

Context of this Chapter

In 1993, there were approximately 61,700 items comprising the national stock of S&E research instruments costing $20,000 or more located at colleges, universities, and medical schools that performed over $3 million in R&D. The monetary value of the aggregate cost[16] of these items was $6.3 billion. Figure 1 shows the distribution of the research instruments by type. For this report, they were tabulated into five categories, as follows.

Figure 1

Five Major Categories of Instruments[17]

Computers and Data Handling Equipment (Computers)

In terms of number of pieces, computers comprised 19 percent of all in-scope research instruments (12,000). The aggregate cost of research computers in 1993 was $1.9 billion: 30 percent of the aggregate cost of all instrumentation. Graphics and CAD equipment are included in this computer category, as are peripheral (not stand-alone) equipment such as: storage devices; networking equipment such as modems, switches, ports, and gateways; and image analyzers/processors and visualization facilities.

Chromatographs and Spectrometers

This category comprised 22 percent (13,800) of the total number of in-scope instruments in 1993 and accounted for 21 percent ($1.3 billion) of the total aggregate cost. Included in this category are NMR/EPR spectrometers, chromatographs and elemental analyzers, and UV/visible/infrared spectrophotometers.

Microscopy Instruments

Items in this category accounted for 9 percent (5,600) of the total number of in-scope instruments as well as 9 percent ($.5 billion ) of the total aggregate cost of all research instruments. This category includes electron microscopes, which accounted for $.3 billion of the bulk of the microscopy investment.

Bioanalytical Instrument

The number of individual items in this category accounted for a much larger percentage of the stock of in-scope instruments (10,200: 17 percent of the total number of instruments) than their share of total aggregate cost ($.5 billion: 7 percent of the total investment in instruments). Included in this category are instruments such as centrifuges and accessories, DNA/protein synthesizers/sequencers/analyzers, and scintillation/gamma radiation/counters/detectors.

"Other" Instruments

This category was the largest in terms of number of pieces (20,100, accounting for 33 percent of the total number of in-scope instruments) and in terms of aggregate cost ($2.1 billion, representing 34 percent of the total aggregate purchase price of all instruments).

The title "other instruments" indicates a miscellaneous grouping of instruments, none of which was large enough to constitute a separate major category; however, there were large concentrations of homogeneous groupings of instruments in this category. For example, the combination of electronics instruments and lasers totaled approximately 7,000 instruments with an aggregate purchase price of $.4 billion; this represents 11 percent of the total number of instruments and 7 percent of total dollar cost. The combination of research vessels, telescopes, and other major instruments (including nuclear reactors, wind/wave/water/shock tunnels, and major prototype systems) comprised almost 1,300 instruments totaling $.6 billion (6 percent of total instruments and 10 percent of total cost). Also included in this category were temperature/pressure control/measurement instruments, robots, and all items not included in any other category.

Numbers of Instruments by Field

Whether measured by numbers of instruments or by total aggregate purchase price, the distribution of the aggregate stock of academic instrumentation varies widely across fields.

Where to find the data: Figure 2 shows the distribution of the approximately 61,700 instruments by field of science and engineering. Table 1 presents the number and percentage distribution of the major types of instruments among the various disciplines. For readers wishing for more research detail, Appendix Table A-2 offers information by type of instrument and the primary field in which the instrument was used; the table also shows the breakdown of instruments costing between $20,000 and $1 million and those costing $1 million or more.

Figure 2

Selected Findings

In ranking the distribution of total numbers of instruments, the biological sciences had the highest total (approximately 21,000 or 34 percent of the total number of instruments). (See Figure 2, Table 1.) Engineering had the second highest number of instruments (approximately 18,050 or 29 percent of the total number of instruments inventoried in all fields). Both disciplines typically had many separate departments at these institutions; perhaps this explains the presence of more instruments in the aggregate. Agriculture had the smallest number of research instruments costing at least $20,000 (600). None of these had an original purchase price of over $1 million (see Table A-2).

The relative importance of different types of instruments varied within most disciplines; however, the distribution of computers was widespread. Most research computers were not in the computer science units; only 15 percent of the total number of computers (1,700) were inventoried to computer science units. Over one-third of all computers costing at least $20,000 were inventoried to engineering (4,100 of the approximately 12,000 computers), and biology ran a distant second with 2,100 computers (18 percent of the total number). The environmental sciences and physics/astronomy each had 12 percent of the total number of research computers.

Given the decreasing prices of high-powered computers, the widespread ownership of research computers would have been even more apparent if the minimum price criterion for inclusion in the survey had been lower than $20,000. Although the computer science fields no longer own the bulk of computers by total numbers, they do have the bulk of expensive supercomputers and mainframes, especially those costing over $1 million. (See cost discussion below.)

The two fields with the largest numbers of chromatographs and spectrometers were chemistry (4,000) and biology (3,800). As would be expected, biology had 90 percent of the bioanalytical instruments (9,200) and 65 percent of the microscopy instruments (3,600). Within the microscopy category, biology had 45 percent of all the electron microscopes (800) but 73 percent of all the other microscopy instruments (2,800). Other fields with a considerable number of electron microscopes were engineering (400) and environmental sciences (250) (see Table A-2). Approximately 50 percent of the "other" instrument category was inventoried to the engineering units. (This category included items such as temperature and measurement instruments, electronics instruments, and robots and manufacturing machines.) An additional 15 percent of "other" instruments was inventoried to physics/astronomy, which had a large number of the lasers and optical instruments and all of the telescopes (see Table A-2).

Distribution by Aggregate Cost of the Instruments

Where to find the data: See Figures 1 and 2 for a depiction of the distribution of instruments by aggregate cost. Table 2 presents the distribution by total cost and by percentage distribution among fields. Those who wish to study the distribution of instruments at a more disaggegated level can look at Tables A-3 and A-4, companion tables that present the relative importance of each type of instrument in detail. Table A-3 is analogous to the categories in Table 2, but instead calculates the percent of total cost attributed to each type of instrument. Table A-4 presents the most detail: total cost of each type of instrument inventoried from the various disciplines, grouped by detailed type of instrument and broken down by those costing between $20,000 and $1 million, and those costing $1 million or more.

Selected Findings

The aggregate cost of computers was the highest for any single type of instrument, totaling $1.9 billion, or 30 percent of the total cost of all instruments costing at least $20,000. Computers represented at least 10 percent of the total investment in instruments in every field, although the range of importance varied widely among the fields (see Table 2).

As would be expected, computers were the most important type of instrument in the computer science field, comprising 98 percent of the cost of instruments inventoried to computer science units. Sixty percent of the total investment in computers was inventoried to the computer science units, largely because 90 percent of the approximately 130 computers costing over $1 million were inventoried to computer science.[18] This is consistent with the advent of less expensive yet very powerful smaller computers, which have the capability to network with large computers and are housed within the laboratories of the researchers; the large mainframes and supercomputers are still mainly inventoried to computer science units.

Chromatography and spectrometers comprised the next highest individual category in terms of total cost ($1.3 billion, or 21 percent of the aggregate cost of all instruments). Chromatographs and spectrometers were proportionately most important to the chemistry units: 70 percent of the total cost of inventoried instruments in chemistry was in this category, with 27 percent of total cost in NMR/EPR spectrometers alone (see Table A-3). The agriculture units were also heavily invested in this category, with chromatographs and spectrometers comprising 42 percent of the total cost of their inventory.

Microscopy instruments totaled $.5 billion, which represented 9 percent of the total cost of the aggregate stock of instruments. This category of instruments was proportionately most important to the biological sciences (23 percent of the total cost of their inventory) and environmental sciences (17 percent of the total cost of their instruments, with the bulk of that investment in electron microscopes).

Bioanalytical instruments totaled $.5 billion, which was 7 percent of the total cost of instruments in all fields. However, bioanalytical instruments were proportionately very important to certain fields. They represented 36 percent of the total dollar value of instruments in the biological sciences and 20 percent of the total cost in the agricultural science units.

"Other" instruments, the diverse grouping of miscellaneous instruments, totaled $2.1 billion, which was 34 percent of the total cost of the aggregate stock of instruments in 1993. The cost of instruments in this category was especially large in three fields: physics/astronomy (65 percent of their total inventory was in this category, with 21 percent of the total in nuclear reactors/instruments alone); engineering (61 percent of the total cost of their inventory, which was rather evenly spread among the other instrument types); and other multidisciplinary sciences (52 percent of their total inventory was in this category, with 33 percent of the total in lasers and optical instruments alone) (see Table A-3).

Change in the Stock of Instruments between 1988-89 and 1993[19]

Where to find the data: Table 3 provides a detailed presentation of the changes in the two surveys in the numbers of instruments and their aggregate purchase costs. It also presents the median cost of in-scope instruments in each survey year.

Selected Findings

In the four to five years between the two surveys that covered 1988-89 and 1993,[20] there was a 30 percent increase in the numbers of instruments costing at least $20,000. This change corresponded to a 43 percent increase in the aggregate purchase price of the instruments, from $4.5 billion to $6.3 billion (see Table 3).

Every category of instrument increased in both numbers and dollar amounts between 1988-89 and 1993. The median price of all in-stock instruments decreased between the two surveys, however, from approximately $70,000 in 1988-89 to $65,000 in 1993.

The total inventory of microscopes increased 34 percent in numbers and 102 percent in total cost between the two surveys. (Chapter III's discussion of the current age of instruments presents additional data to corroborate the increase in these instruments between the two survey years. For example, Table 10 shows that 27 percent of electron microscopes and 47 percent of other microscopes were less than four years old in 1993.)

Chromatographs and spectrometers increased 27 percent in numbers and 71 percent in total cost between the two surveys. Electron/auger/ion scattering instruments in particular increased sizably in percentages, although they represent a relatively small proportion of the total inventory in these institutions: in numbers these instruments increased by 63 percent between the two surveys (from approximately 200 to over 325), and they rose 250 percent in dollars, from approximately 20 million to over $75 million. On average, these are expensive instruments, with a median price of $185,000 in 1993.

The entire category of bioanalytical instruments costing $20,000 or more increased only 6 percent in numbers between the two surveys (from 9,700 to 10,200), but they increased 17 percent in dollar terms, from approximately $.4 billion to approximately $.5 billion. These tend to be relatively inexpensive instruments; they are the least expensive major category of instrument, with a median price of $40,000.

The largest percentage increase in numbers occurred in the other instruments category (the compilation of miscellaneous instruments, none of which were large enough to be classified as a major category for tabulation in this report). This category increased 53 percent in numbers (from 13,100 to 20,100) between the two survey years; this corresponded to a 93 percent increase in aggregate cost, from $1.1 billion in 1988-89 to $2.1 billion in 1993. Of all instruments in this group, only research vessels/planes/helicopters decreased in total aggregate value (a slight decrease from approximately $140 million in 1988-89 to $130 million in 1993). (See Table 3 for additional data by detailed type of instrument.)

Finally, the number of computers costing at least $20,000 increased by 26 percent between the two surveys—from 9,500 to 12,000—while the aggregate cost of the stock of these computers remained virtually the same (almost $1.9 billion). Because computers are increasingly essential to the performance of research in every discipline, an additional analysis was made of the changes in the inventory of computers between the two surveys.

Change in the Stock of Computers between 1988-89 and 1993

Data Considerations for this Analysis

The true nature of the change in the number and importance of research computers may not be completely reflected in the survey data. The recent experience of phenomenal growth in the power of computers, at the same time that their prices were decreasing, unfortunately coincided with the planned change in the survey's minimum purchase price criterion from $10,000 in 1988-89 to $20,000 in 1993. Therefore, the survey's new $20,000 threshold eliminated as out of scope many of the less-expensive but powerful research computers that were purchased since 1988-89.

The survey questionnaires were not specifically designed to produce qualitative comparisons of the instruments that were extant in each survey year; nevertheless, normalizing the database of computers between the two surveys is particularly difficult. More than perhaps for any other instrument, a cost comparison is a poor indicator of change. For example, a computer purchased for $25,000 in 1993 would be vastly superior in power and capacity to a computer purchased for $25,000 in 1988. Given that caveat, it can be said that the phenomenon of a large increase in the numbers of computers between the two surveys but a level amount of total cost was due primarily to the interaction of two events:

a) There was a 35 percent decrease in the number of expensive computers costing at least $1 million that were used for research.

The drop in the number of more expensive computers was due in large part to a change in the academic mission of many of the mainframe computers housed in central computer facilities at the institutions. When the number of these special computer facilities reported to the 1993 Instrumentation Survey was lower than the number reported in the 1988-89 survey, follow-up calls were made to ascertain the reasons for the decline. Respondents reported that many of the mainframes utilized as research instruments in the 1988-89 survey were being used primarily for administrative purposes in 1993. As a result, these mainframes were no longer considered research instruments and consequently were out of scope for the 1993 survey.

Because there were fewer research computers costing more than $1 million remaining in the database in 1993, there was a 12 percent corresponding drop in their aggregate cost since 1988-89 (from $1.2 billion to $1.1 billion). The median cost of these large computers, however, increased from $4.5 million to $6.4 million. Thus, while the numbers of expensive mainframes and supercomputers used for research decreased, those used for research in 1993 were very expensive and were widely utilized for research across the entire range of science and engineering fields. (See the discussion of usage of instruments in Chapter III.)

b) There was a 47 percent increase in the number of computers costing between $20,000 and $50,000.

Computers in this price range increased to approximately 7,400 in 1993, up from about 5,000 in 1988-89 (see Table 3). At the time of the 1993 survey, 55 percent of the computers costing between $20,000 and $50,000 were less than four years old. (See also the discussion of the average age of instruments in Chapter III and Table A-5.)

The recent phenomenal growth in the computational capabilities of personal computers and workstations coupled with their continued decrease in price has changed the purchase pattern for computers used for research in all disciplines. Since the last survey in 1988-89, the emphasis on many campuses is increasingly on the purchase of smaller, very powerful computers that are located in the laboratories and offices of the researchers themselves. Many of these computers have networking capability, allowing users to conduct research and collaborate using off-site mainframes and supercomputers.


Footnotes

[13] For the reader's convenience, the text discussion in this chapter generally rounds dollars to the nearest million and numbers to the nearest hundred. The accompanying tables, however, contain unrounded data.

[14] See the Technical Notes for an explanation of the sampling scheme for the survey and the sampling errors for all instruments as well as a detailed explanation of supersystems and non-supersystems.

[15] The companion report based on that questionnaire (Academic Research Instrumentation : Expenditures 1993, Needs 1994 [NSF 96-324]) indicates that in 1993, these same departments and facilities spent $1.2 billion on the acquisition of new instrumentation. For the reader's convenience, Appendix Table A-1 reproduces a table from that report, presenting trend data since 1982 for annual expenditures at these in-scope institutions. The full report is available on the Web at http://www.nsf.gov/statistics/nsf96324/.

[16] In this survey, the cost of an instrument includes its price at the time of purchase plus the cost of all its dedicated accessories, if any, without any adjustments for depreciation or inflation.

[17] For a more complete listing of instruments included in each of these categories, see the Technical Notes.

[18] It is understood that large computers and supercomputers are used by researchers from a variety of disciplines, and many are heavily used for big multidisciplinary research projects. However, in this report all computers which fit the "supersystem" definition were automatically tabulated in the computer science discipline. All other instruments, including other non-supersystem computers costing over $1 million, were tabulated according to the primary field of usage.

[19] In attempting to reduce the respondent burden of the survey, the minimum purchase price criterion for an in-scope instrument was changed from $10,000 to $20,000, beginning with the Cycle IV survey. To develop a comparable database for this report, the instruments in the 1988-89 database were standardized using the same minimum price criterion of $20,000 in constant 1993 dollars, using the GDP implicit price deflator. See the technical notes for more information on changes in methodology in Cycle IV of the survey.

[20] Data for survey Cycles I-III were all collected over a two-year period, with half of the in-scope fields surveyed each year. For Cycle III, engineering, chemistry, physics/astronomy, and computer science instrumentation data were collected as of 1988; data for the agriculture, biology, environmental sciences, and multidisciplinary fields were collected as of 1989. In Cycle IV, all the fields were surveyed in the same year, so that all instrument data cover 1993.


top toc help home