ARC logo

Scientific Association Records Programs: A Beginner's Guide

Records Problems and Problem Records

In this chapter we address troublesome records and issues that cut across several functions of scientific societies.

E-mail (Electronic Mail)

Archives and records managers are still experimenting with how to appraise and schedule electronic mail. Rancorous public arguments over White House e-mail had distorted the evolution of policies on the issue. Much email in scientific associations is akin to telephone or internal memo traffic: it is often of housekeeping or transitory nature (such as setting up dates and times when a group can meet, or publicizing the forthcoming staff blood drive). Thus, it is documentary overkill to save tapes of all email traffic in the organization, as has seriously been proposed for some major political bodies. But what of the small fraction of email messages in scientific associations that is significant?

E-mail consists of correspondence and bulletin board notices, and the strategy for preserving them may be different. At AAAS, most staff members print out important electronically sent and received correspondence and file it with their paper records. In the Science and Policy office, the director, Albert Teich, has asked his staff to place a copy of valuable email messages in the office chron file [chronological file, i.e., log of daily activities] to keep everyone updated, as well as archiving crucial messages on line (most e-mail software provides for this and it is handy for short term retention needs) or in paper subject files. He reports that compliance is not perfect, but that a notable number of documents do get shared and saved this way.

Bulletin board notices are another matter. Some boards, especially those maintained on the World Wide Web, may record an important public face of the association or have intrinsic value for historical or journalistic research. FYI (American Institute of Physics news bulletin board) and Robert Park's justly renowned electronic comments on the Washington science scene are an outstanding example; these are influential far beyond the physics community. Other boards may have considerable at-the-time interest but far less durable value: at AAAS, the "non-AAAS" staff board is replete with notices of apartments sought or for rent, cars for sale, kittens for adoption, and the like. While this board gives a homey feel to the organization, the archives does not preserve its everyday traffic. Every once in a while, when a significant debate flares on this board (such as over the scientific merit of the movie and book Jurassic Park), the archives will try to print off the arguments, which run to about ten pages before petering out. However, this creates an "artificial" collection, filtered through the eyes of the archivist, that researchers must use cautiously.

Regional, Topical, and Affiliate Group Records

These groups may perform many of the same functions as the parent society for a smaller constituency--run meetings, publish newsletters, serve as a forum for graduate student work, and so on. The number and complexity of these bodies varies from one society to another. The Geological Society of America's [GSA] six regional sections, given the nature of geological field work, are very lively and active parts of the association, and there is significant documentation on them at the national headquarters. AAAS has four regional divisions that serve the far-flung members of the association (Pacific, Arctic, Caribbean, and Southwestern and Rocky Moutain) who find it a problem getting to national meetings in the East, South, and Midwest. Both GSA's and AAAS's geographically aligned groups concentrate on science relevant for their regions, and both are prized by graduate students as the best place to debut one's first research papers.

The collections policy for the association needs to address the issue of who saves the records of regional groups. If the national headquarters interacts closely with them, there may be enough materials saved to document their activities in central records. Serious consideration ought to be given, in such a case, to having the other records of the regional group placed in a depository in the region rather than with the archives of the parent body. This may prove convenient not only for the officers of the regional branch, but also for scholars who are studying scientific topics germane to the geographic area.

Topical, disciplinary, or scientific specialty branches of the association are almost always run by elected officers who keep the records of their work with their personal papers. In doing a records survey, archivists and records managers need to watch for what material on these groups is saved by the central organization. In the Geological Society of America, for example, newsletters of the twelve specialized divisions (geophysics, sedimentary, Quaternary, and so on) are written by division officers but distributed with help from the main office in Denver, giving an opportunity for the headquarters to document their operations. At AAAS, the discipline sections contribute significantly to the content of the annual national meetings, and their work shows up in the review process conducted by the central meetings office and in governance records in the executive office. If the survey shows that such documentation is adequate for the current purposes and history of the association, having the rest of the records deposited as personal papers of the scientists in libraries and other archives is probably preferable to trying to collect them centrally.

An association may maintain formal connections with other societies that share common goals, activities, and concerns. The Geological Society of America's fifteen affiliated organizations hold business meetings and sponsor short courses and symposia at the GSA annual meeting. AAAS's 238 affiliates appoint representatives to its topical sections and may work collaboratively on projects in the human rights or international areas. It is unlikely that an association would manage or collect the records of affiliates; rather, each will probably handle its own documentation. The exception might be an association whose history center might accession the records of a small affiliated organization if there were no other depository that could take them.

Project Records/Grant and Contract Records

These kinds of records were mentioned above as being generated by many functions of scientific societies. They consist of files on the administration of the project--who took part in it, finances, and how it was funded, operational records such as minutes and internal memos--drafts and materials related to review and production of reports, and data. It is the third class of information that causes records managers and archivists the most heartache because of the size of the files, the arcane nature of the evidence, and the unpredictable future uses for the material. Data files must be retained for the minimum amount of time required by the funding agency or if none is specified, for at least long enough for the scientific community to review and assess the published results of the project (about three years). As mentioned above, data that might be needed in longitudinal studies should be retained until that application is completed. But what happens after these minimums?

A group of archivists and researchers (see Haas, et al., 1985, in "Further Readings" list of this report) addressed this problem for scientific records in general and arrived at the following useful guideline. It is not necessary to keep data for long that can be assembled again from libraries, laboratories, or surveys. An example is the agency budget books that form the basis of the popular AAAS report series on research and development expenditures by the federal government. However, if the data cannot be reassembled, such as evidence from questionnaires sent out by the project or interviews done by staff or committee members, researchers may have to come to your archives to use the evidence because it exists nowhere else. These data may still take too much expensive space to save in their entirety. This brings us to the question of sampling.


Scholars from the humanities are often uncomfortable with sampled records. They are convinced that whatever they need will be in the discarded portion, and some may not be trained in how to use and interpret samples. The sciences, however, use sampling techniques frequently and users of scientific association records are apt to understand the need for and the limitations of sampled material.

Gerald Ham recently appraised the slim literature on archival sampling. He points out that the selection an organization makes of what to save of its records is itself a form of deliberative sampling, and that for extremely large record series (larger than those generated by scientific associations--for example, FBI case files) carefully planned sampling may be the only basis for appraisal. More common, but still underutilized, in Ham's opinion, is sampling to reduce a large collection of files to a size that is economically justifiable to save.

Ham notes that statistical, systematic, and judgmental sampling techniques are used by archivists. The first well know to scientists; it assumes a homogenous [for instance, numeric vs. alphabetic] collection of data (files in this case) from which a mathematically valid selection can be made using a random number table. However, the evidence in project files is often already a sample of a population; in that case, one would be sampling a sample, and few statistical procedures could be performed with validity. For this reason, and because random sampling is more time consuming than the next procedure, it is rarely used.

In the second sampling technique, the archivist chooses files [often records in scientific parlance when used in reference to data sets] at set intervals, such as the first of every twenty, as AAAS did with its files on the Guide to Scientific Instruments, or a year in every five of a long chronological run. This method does not meet the usual criteria for random sampling as the term is used in the sciences, although some archivists lump it with random sampling because it is based on arithmetic.

The third sampling technique is deliberately to choose files whose information content is likely to be valuable for future researchers or to choose files only to illustrate an operation, the most controversial method used. One variant that Ham mentions is the "fat file" technique where the largest files in a series are chosen because they are presumed to document significant activity. Ham reports that this has been applied successfully in sampling voluminous court records, but a moment's reflection on the distortion it might cause in, say, member correspondence might give an archivist or records manager pause. Another example is to save all the files of "significant" or "historically important" scientific figures in a series of files on persons, but this reflects the values, often unarticulated, of the selector in his or her culture (thirty years ago, few women's names would have been chosen). Because of the limitations of purposeful sampling, it is often used in conjunction with one of the probability techniques to give a better idea of the original documentation.

Ham includes a caveat from hard-won experience: "the sampling process must be meticulously documented [italics added][with] precise and detailed information, mathematical and otherwise, on the sample design and how it was drawn, and a detailed description of the original universe of records, including what was not retained." Only then can researchers draw useful albeit cautious inferences from the data.

<< Previous | Table of Contents | Next >>