Merseyside UFO Bulletin, volume 1, number 2, March/April 1968
Over forty years ago, ufologists were beginning to look at ways of using computers to aid their research. However, back then computers were the size of a small garage, processing time was scarce and expensive, and programs and languages were hand-crafted. In this article from the second ever Merseyside UFO Bulletin, Bernard Wignall, from the Cambridge University Group for the Investigation of UFOs (CUGIUFO) describes a scheme for processing UFO data which they inaugurated in 1968. The young scientists associated with CUGIUFO were the people memorably branded as “white-coated godlings of the laboratory”, by BUFORA big cheese John Cleary Baker. – JR
Over the last 20 years two main approaches to the serious study of UFOs have emerged: the investigation of individual reports in detail, with the eventual aim of deciding the probable cause of the report, and the large scale study of the whole body of UFO reports. This article is concerned with the latter method and deals with not only the current state of the art but also with the prospects for the immediate future.
The essential tool in this aspect of the subject is statistics and most of the results obtained are of a statistical nature, a fact which must always be remembered when interpreting any of these results. Thus the fact that two things are correlated is of little use until the statistical significance of the correlation is considered. Essentially the problem is to find patterns that are somehow characteristic of the phenomenon and which can be used for predicting the results of future investigations. In this way we can begin to apply the full force of scientific method to a problem which seems so far only to have reached the stage of conflicting opinions with little immediate prospect of objectively deciding between them.
There are two main ways of finding patterns: first, the deduction of a pattern from some hypothesis about the phenomenon and second, the observation of some pattern in the existing body of data. In the first case the hypothesis can be tested by looking for the predicted pattern, and in the second case the pattern can be used to suggest a possible hypothesis, or to check existing ideas. In this way a high level of objectivity can be maintained and accuracy is not forfeit. Statistics is a precise science when properly applied.
We now come to the question of how we deal with large numbers of reports and just what sort of information is inherent in them. There are tons of thousands of UFO reports from the last twenty years in the files of various UFO groups, as well as any reports in government files and many uncollected press articles from all over the world. Even this number comes from only a small sample of the people who have seen something unusual enough for it to be termed a UFO. A recent Gallup poll in the U.S.A. estimated that over five million Americans had seen a UFO. Most of the reports available have very little information in them; the date, place and a brief description being average, and almost all have a mundane explanation in terms of common natural or man-made objects. Even so, such reports can be useful in helping us understand the processes by which we receive reports and providing a standard by which to calibrate things. An important point is that this sort of report will not usually influence results as it will be eliminated in the statistical procedures.
The obvious method of coping with such large numbers of sightings is to use electronic computers and the associated data handling techniques. The first person to do this in any detail was Vallée (1) who eventually built up a catalogue of over 3,000 reports and published several results on the frequency of reports and a possible correlation with the proximity of Mars (2). The largest computer file of sightings is probably the Condon Committee’s catalogue, which numbers over 7,000. These are drawn from about twenty sources, the largest contributor being Vallée. The only other computing project of which the author is aware is the joint Cambridge University and Imperial College scheme the details of which will occupy most of the rest of the article.
Although some programmes relevant to UFOs had been run by the Cambridge Group previously, the joint project began last summer, when the possibility of getting computer time on the Imperial College I.B.M.7094 [like the one in the picture] was mentioned by a senior member of Imperial College. From there, after negotiations with I.B.M., the Imperial College Group was given a job number and free computing time by the company. The main condition is that some of the time be used for a computing service so that everyone could benefit from this scheme. Computing time on a large machine can cost as much as £300 per hour end this would normally be beyond the scope of the largest UFO organisations.
The main difficulty is accumulating the raw data and processing it into a machine readable form. The first method tried was a multi-stage process which involved reducing the original report to a standard format which would later form the main reference file at Cambridge, coding the report according to the code used by the Condon Committee and punching the code on standard cards ready to be fed into the computer. Unfortunately this proved a lot slower than was originally intended, and a backlog of unprocessed sightings quickly mounted up. To get round this difficulty, the computing section of CUGIUFO (Cambridge University Group for the Investigation of Unidentified Flying Objects) decided to write their own very simple computer language to help process the large numbers of reports. Using this language, which should be in routine use in the next month or so, whoever is processing reports merely writes the details of the sightings on a special form, which is then sent away to be punched on cards. These are then put into the machine, which codes the report and also prints out a standard format for files. In this way it is hoped to build up a comprehensive file of world wide reports for use in research.
At this point a word about the type of problem that can benefit from the use of a computer will be of interest to those who may want to use the computing service. Anything that requires a large number of arithmetic operations can usually be done much more easily on a computer. This is especially true if the process is to be repeated for several sets of data, as once the programme is written very little effort is required to run it again and again with different data. The other applications use a basic body of data such as the list of reports described above. The machine can be used to data sort, i.e, prepare selected lists of reports which have some common factor, either geographical or qualitative. Thus one could ask for all reports of red spheres seen over Liverpool in 1966, or the proportion of reports where the principal witness was a woman over 40 years old.
When the data has been sorted in this way a statistical correlation can be performed, either with external data (e.g. the proximity of Mars) or more internal data (e.g, thundery weather). A special case of the latter is the analysis of the overall frequency of reports, which looks at any regularities or patterns in the number of reports per week or month. An example of this kind of analysis can be found in (3).
These are just scale of the ways in which it is possible to build up a picture of the overall behaviour of the phenomenon, a picture that is independent of the observer and subject to scientific methodology. The construction of the main bank of data is just beginning. When it has reached a reasonable size it will be a very powerful resource tool and, thanks to the generosity of I.B.M., it will be available to everyone.
Vallée, Jacques. ‘How to classify and codify UFO sightings’, Flying Saucer Review, vol. 9, 5. Also Vallée, Jacques, and Vallée, Janine, Challenge to Science, appendix 4, Analysis of UFO Activity.
Vallée, Jacques. ‘The study of the periodicity of the UFO phenomenon in its correlation with the oppositions of Mars’. Flying Saucer Review, vol. 8, 5.
Vallée, Jacques and Janine. Challenge to science, chapter 8.