• E-mail
• Print
• Comment
• Font Size
• Digg
• del.icio.us
• Discuss article

The British Journal of Psychology Centenary: A Preliminary Content Survey and Its Problems

Posted on: Sunday, 12 December 2004, 03:00 CST

The preliminary findings are reported of a review of the contents of the British Journal of Psychology 1904-2003 undertaken to mark the journal's centenary. This identifies (a) the top 11 categories of paper published and their patterns of distribution, (b) an apparent change in the nature of papers published after 1983, (c) the overall pattern of category-introduction, showing a dramatic change in 1965 and (d) a provisional typology of the principal patterns of distribution over time. Major conceptual and practical problems arising in the categorization of papers, and their implications, are discussed. Some reflections are offered on the possible value of research of this kind and on the nature of the journal itself.

In order to mark the centenary of the British Journal of Psychology, Britain's first specialist Psychology1 journal, the author has undertaken a comprehensive survey of its contents since 1904. This paper offers a preliminary report of its findings and a summary of the methodological and conceptual difficulties which arose in implementing what was initially expected to be a relatively simple descriptive research task. While historians of science (including those of Psychology such as the author) have long viewed 'celebratory' approaches with some disdain, it must be conceded that celebration is part of the historian's social, if not academic, function. In the present context, however, it was felt that the celebratory aspect was not incompatible with a more serious agenda. Given current moves towards giving historical and conceptual issues more prominence on undergraduate courses, the journal's centenary provides an opportunity for providing at least a glimpse of the potential of this long-marginalized field for making a serious contribution to the discipline in its own right. While readers may understandably be turning to this paper for light nostalgic relief, the hope is that they may find its content more substantial than they expected. The thought had indeed occured of sub-titling it 'theoretical and methodological issues arising in the serial classification of a temporally content-fluid text sequence'.

Following a brief general synopsis of the founding and history of the journal, an orthodox 'Method, Results, Discussion' format is adopted, deviating from this only in that the 'Discussion' is more extensive than normal, is itself subdivided, and will introduce additional illustrative data. The full data is too extensive and complex for a full report at the present time.

The history of the British Journal of Psychology

The founding of BJP has been previously sketched by Clark (1979) in his Editorial for the 75th Anniversary volume (see also Hearnshaw, 1964). By 1904 British psychologists were becoming uncomfortably aware that seven Psychology journals had already been founded abroad and that Britian lacked any comparable serial publications. Mind (founded by Alexander Bain in 1876) had never been exclusively Psychological in character and from the later 1880s had become increasingly philosophical (as it remains), despite being under the editorship of G.F. Stout (Neary, 2001). The recent (1901) creation of the British Psychological Society (then still the 'Psychological Society') had heightened the felt need for a British journal and in 1904 under the joint editorship of James Ward and W.H.R. Rivers, BJP was launched. The first nine volumes appeared somewhat irregularly (see Table 1).

Table I. Publication dates of Volumes 1-9

Volumes appear annually thereafter, but with an overlap of calendar years until Vol. 41, 1950, in which year the last issue of Vol. 40 was also published. Usually a volume comprises four issues, but occasionally there have been double issues, and Volumes 34-37 (Sept. 1943-May 1947) each comprised only three issues. From 1911- 1962, BJP also issued a series of 'Monograph Supplements', amounting to 35 in all. These have been disregarded in the current survey.

While we now take it for granted that the journal is a BPS publication, this was not originally the case with Ward being its owner. From as early as 1906 the BPS Minutes indicate that discussions had begun about the Society's relationship with the BJP and, on 13th March 1909 we read that 'A proposal was made informally. . . by W. McDougall to the editors suggesting that the Society should bear half the loss (if any) on the Journal, conditionally upon the Society being represented on the Editorial Board'. It came to nought as the Journal editors (still Ward and Rivers) were reluctant to yield editorial control over acceptance and rejection of submitted papers. The matter is ruminated upon annually, and only resolved in 1914. From its founding until the founding of the Bulletin of the British Psychological Society in 1948, BPS notices were nevertheless generally included in the BJP.

One initial factor in the Society's attitude towards the BJP was that despite Rivers's presence, Ward himself remained biased against experimental, psychometric and physiological research. This ceased in 1910 when C. S. Myers took over as Rivers's co-editor, becoming sole editor in 1912 (see Table 2 for editorships). F. C. Bartlett's 23-year editorship saw the journal through from the late 1920s to the post-Second World War period and played a major role in creating its current ethos.

Table 2. Editorship of the British Journal of Psychology

One might assume that both the size and number of papers published per volume have steadily increased. If broadly true, this has not been a smooth expansion and requires some qualification (see Table 4.) Although an expansion in size officially took place with Vol. 55, 1964, the jump from the 300 + -400 + pages (typical of the period after 1938) to 512 was in some respects a return to an earlier norm - Vol. 25, 1934-5, was also 512 pages, and paginations in the high 400s were not uncommon. The 557 of Vol. 8, 1915-1917, was not exceeded until Vol. 60, 1969. Since the massive 1996-2001 Volumes (peaking at 721 and 704 pages for 1997 and 1998 respectively), the pagination has returned in the mid-500s typical since Vol. 60, 1969. Regarding numbers of papers, the highest so far was 56 in Vol. 78, 1987, otherwise the number from the mid-30s to high-40s since Vol. 67, 1976, and first exceeded 40 in Vol. 58, 1967. This indicates that number of papers and volume size are less than perfectly correlated (time constraints have prevented this being calculated as yet), implying that mean paper-length can vary and also requires tracking. This would all be gratuitous detail were it not that it has a potential bearing on how trends in content are evaluated.

A further important reminder should be made at this stage. If initially monopolising the field, the BJP was soon joined by other, more specialist, journals. Disregarding those which have been founded in the last two decades, and non-UK serials, the most important for present purposes are those in Table 3.

Table 3. Founding dates of some other major British Psychology journals*

Clearly the expanding presence of other journals vitiates any simple reading of content-trends as representative of the discipline's interests as a whole. (We return to this point later.) Expansion rates in numbers of scientific journals on any topic, and the existence of threshold points at which more specialist journals come into existence, have long been known to follow a regular pattern from the first Transactions of the Royal Society of London in the late 17th century onwards, Price (1961) including a now classic account of this. Psychology in Britain is no exception.

The BJP has clearly established itself, under a succession of eminent editors, as the most visible British Psychology journal on the national and international scene. During its 100 record it has played a major role in the discipline's development, publishing papers by not only British but also many eminent foreign psychologists (some are noted in the Discussion, other early authors included C.G. Jung, Mary W. Calkins, J. Piaget. and E. Rignano). However, our concern in the present paper is somewhat different- what has the BJP actually been about ?

Method

A content analysis based primarily on journal paper titles was undertaken. In order to avoid anachronism it was felt that the classification categories used should be those emerging 'naturally' from the material. One deliberate exception to this was the use of 'Learning Difficulties' where a succession of terms such as 'idiocy', 'imbecility', 'mental defectives', 'mental retardation' and 'backwardness' have been deployed in the past (this still begs the question of how far 'learning difficulties' and e.g. 'idiocy' should be treated as equivalent, especially as the latter and its alternatives included 'moral imbecility' into the 1920s). Another, more serious, was the somewhat complacent decision to use the received familiar category 'Motivation and Emotion' throughout. This is obviously an extremely broad 'catch-all' category (see Danziger, 1997) and hardly denotes a unitary research topic. In the event it was invariably amplified by an additional classification (from the outset it was clear that many papers had to be multiply classified in order to encode their content adequately). Following the policy \successfully adopted in analysing 'race'-related publications prior to 1940 in Richards (1997) the data were aggregated into 2-volume blocks for ascertaining category ranking. Other similar exercises (e.g. Clark op.cit) have usually sampled single volumes at 5 or 10 year intervals. While this may give a broad impression of historical trends it is too crude to enable any more sensitive analysis and can be misleading if the volume in question is atypical. More focused tracking nonetheless sometimes required a return to the annual data. This preliminary analysis will be concerned with the following:

(1) temporal changes in frequency of use of major content- categories;

(2) temporal change in character of papers as indicated by number of content-category classifications required per paper;

3) identification of timing of radical changes in agenda as indicated by rates of addition of new content-category classifications;

(4) preliminary identification of types of change in frequency of use of different content-category classifications.

Three major points should be stressed. Firstly, the data refer to frequency of use of content categories. Since papers may be multiply classified this means that figures for two categories should not be taken as disjunctive, i.e. the fact that in a given period 10 papers are classified as 'visual perception' and 6 as 'developmental/ child' does not imply a total of 16 different papers, since some of the latter may concern children's perception and be included under both. Some implications of this are elaborated on in the 'Discussion' section. Secondly, since this has been a single-handed research exercise the reliability of the classification category allocations is unknown. There is unlikely to be any systematic bias since no specific experimental hypothesis was being tested as this was intended as a purely descriptive exercise. Nonetheless a subjective element may well have entered in at two levels: (a) slight covert shifts in category-usage decisions over what was an extended research time period, and (b) the author's differing levels of specialist understanding regarding the nature of some research topics, especially those reported in more recent papers. Thirdly, the question may be posed as to why Psychological Abstracts categories were not used. The major drawback in this is that Psychological Abstracts (from 1927, and before it Psychological Review, 1905-1935) classify each publication under a single heading. This aside, the author felt it would be advisable to start from scratch since the purpose of this research clearly differed from that of the editors of Psychological Abstracts and Psychological Review, each of which use an extremely elaborated and continuously evolving synoptic and multiply nested mapping of the discipline's fields. The possibility also had to be held open that the categories emerging in historical retrospect might in some way diverge from those used by abstractors at the time of publication.

The implementation of this method, straightforward as it appeared, raised a number of problems which will be addressed in the Discussion.

Results

Basic data

Temporal tracking of numbers of papers and journal size is provided in Table 4.

* Total number of papers Volumes 1 - 94 = 3222

This excludes occasional brief editorials (two were included: that of Volume 1 and Clark (op cif) as being historical). Critical Reviews were included, but not ordinary book reviews. One paper remained unclassified as the title was too opaque to interpret and the volume in question missing from Senate House Library, only Contents pages being available on the computer catalogue.

* Total number of content categories = 345

These are of all levels of generality, from 'General', 'Theory' and 'Perception' to 'Debt', 'Boredom', 'Respect' and 'Stalking'. This figure is slightly under the number actually employed as in a few cases (e.g. 'Developmental (adolescence)' and 'Adolescence') it became apparent on inspection that the same category had been worded differently on different occasions. (A further slight reduction was made in the course of tracking category-introductions.) Few of these will be reported or discussed in the present paper.

Major categories

The 11 categories most frequently used are given in Table 5. A 'top 10' would have been neater, but each of these clearly needed mentioning and required some comment. Their frequencies are tabulated in 2-volume blocks in Table 6. Frequencies for Volume 1 are however given separately as it was thought this might be of independent interest, necessarily therefore frequencies for Vol. 94 are also given separately. Peak frequencies are in bold.

Table 4. Total papers published and volume size in 2-volume blocks

Table 5. 11 most frequent categories

Perception

Accounting, alone or jointly, for 18% of the published papers this includes those classified as Perception (General and Visual), Perception (Auditory), Perception (Other) (i.e. olfactory, taste and tactile) and Perception of Causation (in the Michotte sense). Social Perception and similar categories were excluded. Face Perception/ Recognition was also excluded as being considered too distinctive a research tradition (this has accounted for 53 papers since first appearing in Vol. 66, 1975). Perception has been a persistently important category, with no clear pattern of historical change but does oscillate widely between 2 entries (Vols. 36-37, 38-39) and 35 (Vois. 58-59), exhibiting a notable surge during the late 1960s and 1970s (156, 27.5%, appearing from 1964-1979). Since Vol. 68, 1977, it has ranked below Memory on all but two occasions (Vois 76-77, 89- 89). It was ranked first throughout the first 11 volumes, dropping to third behind Theory and Intelligence in Vols 12-13(1921-1923,9 entries) and second behind Intelligence in Vols 14-15 (1923-1925, 8 entries), before sinking to only six entries in the subsequent three volumes. In 1928-1929 (Vol. 19) it returned to first place, remaining there until Vol. 33 (1942 -1943), after which there is a second decline until the Vol. 48-49 block (1957-1958) when it returns to first place, thereafter remaining first or second until Vol. 69 (1978), retrieving its fortunes only fitfully thereafter (e.g. Vols. 88-89, 1997-1998, when it comes first again). The factors contributing to these oscillations are not transparently obvious.

Memory

The Memory story is more dramatic. Only 57 (16%) precede Vol. 55, 1964. After that it consistently remains in the top six and from Vol. 68, 1977, is only occasionally below second. This pattern presumably reflects, first, the rise to prominence of cognitivist reseach on the topic and secondly a later broadening both of types of memory studied and of theoretical approaches. This creates some some conceptual difficulties with the category, reserved for the Discussion.

Table 6. Frequencies of occurence of top 11 category- classifications by 2-volume block. Maxima indicated in bold

Table 6. Frequencies of occurence of top 11 category- classifications by 2-volume block. Maxima indicated in bold

Cognitive/thinking

This exhibits a somewhat similar pattern to Memory, however the divide is earlier, only 48 (14%) of the papers in this category appearing pre-1955 (Vol. 46). Only in this volume does it first receive a high ranking (3rd), and never subsequently drops below 6th except on one occasion (Vols 62-63, 1971-1972), this is deceptive however, as the 11 entries would usually have guaranteed it around 4th or 5th place, but a fortuitous clustering of papers on other major topics raised the barrier as it were. The peak for Cognitive papers is in 1981-82 (Vols. 72-73) with 26 entries. There are serious conceptual problems with the category, which will be explored in the Discussion. These indicate there there is considerable lee-way for interpretation and it would not be difficult to boost it into 2nd place overall. One should note in particular that overlap with Memory is obviously high during the heyday of D. Broadbent's and A. Baddeley's research and that a number of occasions on which the two categories were used disjunctively might well be open to challenge or doubt.

Developmental/child

The distribution of these papers differs from the previous ones in displaying a broad peak between Vol. 60, 1969 and Vol. 77, 1986 during which 139 (54%) are published. While prominent in early volumes (22 papers up to Vol. 19, 1928- 1929) and ranking 3rd in the Vol. 2-3 block (1906-1910) and 1st in Vol. 4 (1911) the genre declined dramatically from the late 1920s into the early 1950s, although retaining a persistent low-level presence. An untypically higher representation in the Vol. 72-73 block, 1981 -1982, with 36 entries, should be noted, this being nearly double that achieved elsewhere, 16 for Vols. 62 - 63 a decade earlier being its largest presence elsewhere. One difficulty with this category which should be noted is that it is not always evident from a paper's title that the subjects or participants were children. Since this was almost certainly the case in a number of papers categorized by title alone the total is likely to be an underestimate.

Personality and individual differences

Following a relatively constant but low-level presence since the beginning, this expands very rapidly from 1955 (Vol. 46) and remains very prominent until 1981 (Vol. 72). During these years 135 (58%) of the papers so-classified were published. The popularity of scores on the various versions of Eysenck's Extraversion/Neuroticism inventory as independent variables and, to a lesser degree, the appearance of Personal Construct Theory (and the Repertory Grid technique) were undoubtedly a factor in this, but probably not the whole story, especially after 1970. The highest number of entries was for Vols. 62-63, 1971-1972, with 17. The category title, though customary, does of course combine work in a variety of often opposed theoret\ical traditions.

Language

The assymetry of distribution here is very marked. Of the 227 papers so categorized, 177 (78%) appeared after 1966 (Vol. 57), Volume 58 showing a distinct jump, with 9 papers, as much as the total for the preceding six volumes. This is perhaps the most problematic of all the classification-categories however for reasons we will address in the Discussion. Vols 1972-1973 marked its peak, with 18 entries.

Learning

As the archetypal 'experimental' topic one might have expected this to be more prominent, but the distribution bears out the general historical consensus that British psychologists only began to address it at all widely in the 1950s and did not exhibit the enthusiasm for the topic evident during the inter-war years in the United States when the various versions of behaviourism were at their height. It did, nonetheless, rank 3rd in Vols 22-23, 1931- 1933 and 2rd in Vols. 26-27, 1935-1937, signifying that some notice was being taken of developments in America. 123 (54%) of the Learning papers appear between Vol. 46, 1955 and Vol. 65, 1974, the highest being 22 in Vols 56-57, 1965 -1966. The peak is clearly in the years 1962 -1966 during which it maintains 2nd rank after Perception. After 1974 it retains a moderate presence but generally disappears from the higher ranks. In comparison with the categories discussed so far this is relatively straightforward, although not entirely so.

Intelligence

Again this is a relatively unproblematic category. The distribution clearly reflects the early 20th century excitement about the topic, with 108 (54%) of the papers appearing in the first 19 volumes, although the highest single level of representation is 14 in Vols. 24-25, 1933-1935, placing it in the 2nd rank at the height of the debate over Spearman's g (General Intelligence) theory and the methodological statistical wrangles associated with it. It retains a strong, if not constant, presence until around 1940, also being ranked 2nd in the Vol. 30-31 block (1939-41), but 7 of the 8 papers in question appear in Vol. 30. From 1970-1974 (Vois 61-65) Intelligence apparently disappears altogether, before enjoying a brief revival in the late 1980s.

Methodological

This excluded purely statistical papers which were classified separately. The distribution of Methodological papers shows them appearing in periodic waves, although never disappearing. The years in which it receives top 6 rankings are: 1921-1923, 1939-1941, 1955- 1956, 1977-1978, 1991-1998. Its highest presence was as recently as Vols 88-89, 1997-1998, with 12. Between these peaks it occasionally disappears but typically sustains a low level presence.

Motivation and emotion

Aside from vanishing completely from Volumes 30-38 (1939-1948), this broad category maintains a fairly stable presence from the outset. It occasionally reaches the higher rankings (3rd in Vols. 16- 17, 1925-1927; 1st in Vols. 38-39, 1947-1949; 2nd in Vols. 40-41, 1949-1950; a surge between 1957-1962 - especially Vols. 49, 51-53, and so on). Given the somewhat Protean nature of this category it is not easy to interpret this pattern, or lack of it.

Theoretical/conceptual

The count for this category included a few papers initially categorized as Philosophy. This too has sustained a fairly even presence, but with a collapse during the years 1928-1939 (Vols. 19- 29) during which six relevant papers were identified. Its two peaks, when it reached double figures, were Vols. 12 -13, 1921 -1923, and Vols. 88-89, 1997-1998, at 13 and 11 respectively. Otherwise the only point of real interest is perhaps the generally strong showing of Theory papers over the period 1993 - 1999 (Vols. 84-90) during which 17% (26) of all the theoretical and/or conceptual papers were published. (If this level had been maintained throughout the journal's history, this category would have been in 4th place with 349 papers.)

Hypotheses regarding the explanations of these various distributions patterns will undoubtedly readily occur to many readers. At this stage however it would be premature to speculate too freely regarding what they signify.

Change in character of papers

A number of later volumes were classified early on in the research and the author's impression was that these seemed to require higher levels of multiple categorization than early ones. The complete data enabled us to verify this, showing a quite rapid and dramatic shift. For each volume we have totals of both number of papers and number of categories used. In earlier volumes the former almost always exceeds the latter, in the later ones the reverse is the case - more categories are used than papers published. The turning point can be located very precisely to Vol. 74, 1983, when the numbers are equal (43), thereafter papers exceed categories on only one occasion (Vol. 90, 1999 when five fewer categories were used than papers published, 39:44), prior to this date we have to back to Vols. 38, 39 (1947 - 1949) to find an earlier zero difference and to Vol. 11, 1920-1921 to find an earlier, and unique, excess of categories over papers, 31:29. The size of the differences during the entire periods either side of Vol. 74 oscillates quite widely, but the switch itself appears to be a point in a trend. (See Table 7.)

Table 7. Papers:Categories difference Vol. 68, 1977-Vol. 79, 1988

A complementary analysis of the category-disappearance distribution has yet to be undertaken.

From Vols. 70-78, a sustained swing in the data is quite evident. This suggests a major shift in the character of the papers being published, but what precisely the nature of this shift is remains, as yet, open to interpretation. This finding does however suggest that further scrutiny of the fluctuations in the N Papers: N Categories difference might identify subtler movements in the nature of papers being published and locate historical moments warranting deeper scrutiny.

It should also be noted in this context that a comparison of the relationship between journal size and number of papers indicates that paper-lengths themselves have varied. Very crudely it would appear that mean paper length exceeds 10pp until Vol. 35, 1944- 1945, after which there is a fairly close oscillation around 10pp until Vol. 70, 1979, after which it increases up to 16 in some volumes and never drops below 10 again. (In Vol. 5 it had briefly reached c.34 pages per paper.) Quite how this 3-phase trajectory should be interpreted is unclear, but it does not correspond to editorial changes.

Radical agenda changes

On the assumption that a change in the agendas of the journal's contributors will be reflected in the introduction of new content- classification categories, we can use the rates of category- addition to identify when these changes occur as an albeit crude quantification of their magnitude. As one would expect the first Volumes serve primarily to introduce all the major categories. Table 8 presents the situation somewhat simplistically (a more fine- grained analysis will be undertaken in due course).

Table 8. Category introduction levels

This appears to indicate that by Volume 45, 1954, the fields covered by journal contributors had been exhaustively mapped. Appropriately perhaps the single 1954 addition was 'Ageing'. From then until 1964 no further additions were required. The next additions are for Vol. 56, 1965: 'Decision making', 'Perceptual defence' and 'Subliminal perception'. The period since then has really seen no such settling down as characterized by the previous phase, continuing to display high levels of category introduction and 'turnover'. While we may all share an underlying impression that the 1960s saw a radical change in Psychology as everywhere else, it is surely remarkable to see this registering quite so dramatically in data on a mainstream journal's contents, (and one hardly thinks it can be entirely laid at Arthur Summerfield's door as the new editor in 1965).

Further analysis of the data in shorter time-blocks would undoubtedly reveal some interesting, if less extreme, episodes on agenda change. In particular the relationship between this finding and the previous one needs further exploration.

Type of change

The earlier discussion of patterns of change in the 11 most frequently used categories suggests that it may be possible to create some kind of typology for the various kinds of pattern which emerge. Very tentatively the following appear to be discernible from the data as so far analysed or examined.

(1) Eruptive. After a prolonged low-level presence, a topic dramatically expands. This would be the case with 'Cognition/ Thinking' and 'Memory'.

(2) Collapse. A topic hitherto used frequently suddenly collapses. This is the case with Statistics in the BJP following the establishment of the British Journal of Psychology (Statistical Section). Of the 82 'Statistics' entries only 8 have been published since 1947 (Vol. 38), when that journal was founded. Something like this also occurs, as we saw, with 'Intelligence'.

(3) Oscillatory. The presence of the topic rises and falls with greater or lesser regularity but it neither disappears nor becomes dominant except very transiently. Both 'Methods' and 'Theory/ Conceptual' categories fit this description.

(4) Moment of Fame. A topic rises into prominence, flourishes, and then declines again. This would to a large extent apply to 'Learning' coverage in the BJP.

(5) Evaporation. A category comes to refer to an increasingly wide range of actual phenomena and loses its coherence. This is less easy to demonstrate with frequency data but does seem discernible in the late 20th century with regard to both 'Perception' and 'Language'. 'Personality and Individual Differences' is another possible candidate.

(6) Adaptive Mutation. This might be said to apply to a category like 'Motivation and Emotion' (and possibly 'Personality & Individual Differences'), serving as a convenient unifying label for a var\iety of different research topics and phenomena, which are in continual flux. Again, this is difficult to demonstrate from the frequency data, although a fairly constant and stable level of representation is one of its hall-marks.

(7) Constant. The category persists throughout at a fairly constant level, although may exhibit fluctuations. 'Perception', which erratically oscillates between 2 and 35, cannot really be considered to be of this type. This pattern is more common among relatively low frequency categories, e.g. 'Music', which appears in both in Volumes 1 and 93- The 39 papers so-classified are evenly, though intermittently, spread across the journal's lifetime, although the pattern is vaguely U-shaped due to its disappearance from Volumes 34-58.

When the less frequently used categories are more fully analysed further light may be shed on this. One might finally remark that 10 (of 20) categories used in coding Vol. 1 were required in coding Vol. 93 (which required 47).

Such, in outline, are the results which we are currently in a position to report. The pivotal turning points identified with regard to the Papers:Categories ratio in 1983 and the 'agenda revolution' after 1965 are, I believe, novel findings providing empirical grounding for what were previously at best no more than personal impressions of events. While preliminary in nature, these results give at least some hint of the potential value of data of this kind. It is now however time for a pre-emptive cold shower, rendering explicit a number of problems which underline the need for caution in using and interpreting such data.

Discussion

This discussion will focus on three issues. The first is the methodological and conceptual difficulties arising in the use of content-classification categories for a long historical textual sequence of the kind represented by the BJP. The second is to present some explicit conclusions regarding what questions can and cannot be legitimately addressed with this data. The third is to offer some brief personal conclusions and reflections on this research and its significance.

Difficulties in the use of content-classification categories

The underlying problem is that it appears to be impossible to establish rules for how categories are to be used which are applicable for the entire century-long text sequence. This does not occur in orthodox content (or discourse) analysis where functionally unambiguous rules can be established for categorizing the presence of particular themes and motifs (usually decided beforehand) in a text. The classic example would be those used for coding the story- responses to TAT and n.Ach projective tests. Generally speaking one should ensure that categories are either disjunctive or hierarchical, which would, in the present case, mean that each paper could receive only one classification, or a superdordinate + subordinate one, faciltating an easy correspondence between category- counts and paper-counts. It was impossible to adopt such a strategy here. Leaving aside the need for multiple classification in order to code papers dealing with more than one topic (thus excluding any simple disjunction), a number of other factors may be noted.

(i) Changes in terminology for refering to what appears to be the same topic (e.g. the various terms for 'learning difficulties' cited earlier). This constancy of referent may however be partly illusory since changed terminology usually reflects a shift in how a phenomenon is being conceptualized and defined. Consequently its boundaries of applicability may well have shifted from those of its apparently synonymous predecessor (as we noted, 'idiocy' included 'moral imbecility', a notion long vanished -'sociopathy' would be the nearest current term but has quite different connotations).

(ii) A term's meaning can alter over time in various ways. Since 'cognition' and 'cognitive' exemplify several of these, as well as figuring so prominently in the foregoing analysis, we may consider this case in a little more detail. The term 'cognition' was widely used by 19th century philosophers and proto-psychologists (e.g. Jardine, 1874) and early modern psychologists like Spearman (e.g. Spearman, 1923). Its use as a technical synonym for 'thinking' prior to the 1950s is fairly unproblematic. The emergence of Cognitive Psychology seriously complicates things thereafter. 'Cognitive' can, for a start, now refer either to theoretical character or subject- matter. But these cannot be easily differentiated since the whole message of the theory is that numerous phenomena hitherto clearly distinct from 'cognition/thinking' can be treated as aspects of it (e.g. memory). Yet we cannot henceforth simply subsume all such papers hierarchically under 'cognitive' or 'cognition' because not all 'memory' papers are theoretically cognitivist. Moreover, a clutch of more specific research topics such as, for example, 'reasoning', 'problem-solving', 'concept attainment', 'cognitive style' and 'probability judgment' begin to differentiate out from the 1960s onwards. It then becomes difficult, without time- consuming scrutiny of each paper, to determine whether these should be jointly coded as 'cognitive', a task further exacerbated by the near impossibility of formulating clear criteria for the use of 'Cognitive' in the theoretical sense. 'Cognitive dissonance', for example, arguably has its closest affinities with Freudian 'rationalization' strategies, and actually emerged within Social Psychology, yet did so in the 'cognitivist' climate with which its name identifies it. Any such criteria would, moreover, be in a sense arbitrary 'constructions' by those formulating them, not precise delimitations of a 'natural' category. In the case of 'cognition'/ 'cognitive' a combination of splitting of referent (between theory and subject matter), theoretical reconceptualization of the topic and redefinitions of subject-matter has occurred which, taken in conjunction with the persistence of traditional usage, creates huge difficulties with regard to establishing consistent rules for its use.

(iii) While remaining constant in meaning the kinds of research topic to which broad category terms refer can become increasingly diverse. Both 'Perception' and 'Language' are cases in point. Even while sub-categories of perception-research genre may be introduced this does not fully resolve the difficulty since boundaries again become fuzzy - 'social perception', 'face recognition', 'person perception', 'reading' and use of the Stroop technique etc. are sufficiently far from 'perception' as a traditional research genre to suggest that specific categories should be introduced for each and that they be excluded from the 'perception' count. 'Language' is perhaps the most difficult of all categories, since it pervades later 20th century research in a variety of guises. Much cognitive, perceptual and memory research uses linguistic stimulus material or focuses on aspects of language (such as naming, word-recognition, and language-use errors) as behavioural indices of cognitive, memory and brain processes and structures. Should sight-reading of musical scores, voice recognition and verbal memory all be coded as 'language' in addition to psycholinguistics, discourse analysis, writing and social psychological research on verbal communication? Many of the specific topics addressed in both perception and language are, additionally, historically fairly transient. Again, it becomes difficult to establish any general rules for quite when to use these categories. Memory and Psychophysics present similar features, although have moved in different directions. While Memory has flourished and its identified varieties multiplied, the category 'Psychophysics' has ceased to have any clearcut denotation. The classic psychophysical phenomena such as reaction times and sensory thresholds are now rarely studied on their own, commonly featuring instead in more complex experimental research. Visual perception long ago assumed a separate life of its own and whether research involving a phenomenon like auditory acuity or effects of noise is appropriately classified as 'Psychophysics' depends very much on the details of the individual case. Here it is not perhaps so much an instance of the term denoting an increasing diversity of phenomena, as of increasing diversity of usage and roles of those phenomena in experimental research, and reduction of interest in 'pure' psychophysical research as once understood.

(iv) The phenomenon of category-use expansion identified in the previous section is most evidently related to the increasingly complex nature of much post-1970s research in which the number of variables being explored has proliferated. We encounter titles such as 'Stress and blood donation: the effects of music and previous donation experience' (Ferguson et al., 1992) and 'Staff ratings of children's behaviour in hospital: comparability of factor structures' (Clough, 1978). Even classifying the relatively innocuous sounding 'Expedition stress and personality change' (Watts et al., 1992) proved, on further inspection, to require five different categories: Personality, Stress, Sex Differences, Social Psychology, Motivation and Emotion. As the titles just cited indicate, papers have, over time, become both more multi-faceted and more narrowly focused. Their titles also become more technical. The upshot is that examination of the text itself becomes increasingly necessary for classification purposes. Methodologically, however, this means that papers are being classified by two different procedures which could, in principle, yield different results. As was previously mentioned, this is likely to have occurred in relation to the 'Developmental/Child' category when textual examination reveals that the subjects or participants were children, though this is not mentioned in the title. One might note, incidental\ly, that all this places the widely perceived trend towards greater specialization and compartmentalization in Psychology (Wood, 2004) in need of some reformulation, since research, at least in this journal, increasingly incorporates variables traditionally isolated within separate research genres. Even the present paper might feasibly require quadruple coding as 'History of Psychology', 'Classification', 'Social Constructionism' and 'Methodology'. This hardly corresponds to specialization in the usual sense of attending to an ever narrowing range of phenomena.

(v) Further problems worth noting are (a) that some categories are inherently somewhat ambiguous (e.g. the category 'Ageing' includes 'Age-differences' and 'Life-span Development', while the terms 'theoretical', 'philosophical' and 'conceptual' cannot be easily disentangled - a single 'Theoretical' category eventually being settled on to cover all three); (b) even apparently distinct categories can prove to have blurred boundaries, the overlap between 'Handedness' and 'Brain Lateralisation' being a case in point; (c) in recent years a number of papers have appeared in which a reflexive approach is taken towards traditional topics, for example papers by Blackmore (1997) and Wiseman et al. (2003) which look not directly at 'paranormal' phenomena but people's beliefs in them. Other papers have similarly looked at beliefs about other psychological phenomena rather than directly at the phenomenon in question. In the case mentioned, should this still be classified as 'Parapsychology'? (d) Developments discussed under (iv) above have also led to a proliferation of very specific categories, some used only once, which are nevertheless clearly required if the paper's character is to be captured. 'Marriage' is used only for Russell and Wells (1994) 'Personality and quality of marriage', 'Celebrity worship' was uniquely required for McCutcheon et al (2002) 'Conceptualization and measurement of celebrity worship', and 'Boredom' was introduced for two consecutive papers by Hill and Perkins (Perkins & Hill, 1985, Hill & Perkins, 1985), with only one subsequent use in 1994. 'Absent-mindedness', 'Debt', 'Shyness' and 'Respect' are among other single-use categories. (e.) Unexamined so far is the pattern of category disappearance. One can however assume that this can arise for several reasons including loss of interest in the topic, subsumption or merging of the category into another, cessation of the research programme which generated the papers and marginalization of the topic as a result of negative evaluation of its value among psychologists. The last would be the case with 'Psychoanalysis', 32 of its 36 entries being in Volumes 5-15 (1912- 1925) and the remainder scattered singly across the early 1930s. This clearly reflects the 'psychoanalysis craze' of the late 1910s- early 1920s (Richards, 2000). (f.) Finally pragmatic 'presentist' considerations may come into play, thus one might be readier to introduce a new category (e.g. 'Sport') in coding earlier papers than their character strictly warrants simply because this now a topic of great interest.

This review of the major difficulties encountered in categorizing the journal's past papers ought - or need - not be read too negatively. While they certainly imply that the findings should be treated with some caution and that any quantification clearly has some margin of uncertainty, they do lead us to ask some deeper questions. They should not be viewed as mere methodological problems to be overcome but as a genuine reflection of the dynamic instability of the material itself, as will be argued in the final section below. I believe that the idea that had the research been a group, rather than individual, project, with scrupulous attention to inter-coder reliabiality etc., the categorizations could have somehow become more 'accurate' and 'objective' is mistaken - hence the use just now of the phrase 'margin of uncertainty' rather than 'of error'. Although this may have removed a few idiosyncrasies the outcome would have remained a product of collective pragmatic decision-making: the categories would not have more nearly approached some ideal representation of reality than those decided by an individual. Indeed a coding system which actually eliminated all the uncertainties and ambiguities just identified would be less accurate, since these are rooted in the uncertainties and ambiguities of the categories deployed in the material itself.

Turning to the second issue, which kinds of questions can and cannot be addressed by using this data, the most obvious observation is that it cannot be interpreted as a direct reflection of the discipline's interests and character in Britain as a whole, except for the first dozen or so volumes. The growing presence of more specialist journals, founded precisely because the topics they address have become widely popular in the discipline, will clearly result in these being underrepresented in the BJP. The collapse of statistics coverage has already been mentioned. 'Clinical/Abnormal' is (excepting pre-1939 papers on Psychoanalysis) clearly underrepresented throughout the run, totalling only 86. But the British Journal of Psychology (Medical Section) had largely assumed responsibility for this area after 1920 (along with Psychology and Psychotherapy, also founded in the early 1920s), being superceded in 1962 by the British Journal of Social and Clinical Psychology, which later split into the separate Social and Clinical journals currently published. The Journal of Mental Science, later British Journal of Psychiatry, was another outlet. 'Clinical' enjoys a brief spell of prominence over Vols. 55-63, 1964 - 1972, ranking 3, 3, 4, 3 in the four 2-year blocks, but in Vols 64-67, 1973 - 1976, it disappears completely. Several recently popular fields and topics are notably almost absent: little on Psychology of Women, sex and gender issues, sport, Forensic Psychology, counselling or, perhaps surprisingly, Michael Apter's 'Reversal Theory'. Apart from the last these are all catered for by other journals. Nor does any continued wider cultural and disciplinary interest in Psychoanalysis and Jungian Analytical Psychology after 1945 find an echo in the BJP. Like Clinical, Social Psychology is also relatively underrepresented (111 entries), figuring most prominently during the period 1940-1954, during which 30 (27%) of papers so-classified appeared.

There are however a number of more specific questions on which the data may shed further light, some of which we have already begun to address. If the respective weightings of different categories cannot be generalized without qualification, the same is not necessarily true of the changing nature of the categories themselves. We have identified one major turning point already (1964), but the data would permit a finer-grained analysis of both the tempo of change and the areas in which it was taking place. Such an analysis would require more sophisticated techniques e.g. for tracking category-associations in multiple classifications and year- by-year monitoring of category introduction, plus the formulation of some explicit hypotheses. It might also involve a systematic reappraisal of the initial classification with the benefit of hindsight. Such an exercise would surely have an important and direct bearing on current debates about the extent to which Psychological categories are 'natural' or 'socially constructed". More specifically, further analysis of the data could reveal more about the changing nature of the BJP itself and how (and whether) this can be related to different editorial regimes. (It is noteworthy, for example, that from 1984-1996 under the D.E. Blackman and A.J. Chapman regimes its size, with one exceptional volume of 526 pages in 1980, is virtually constant at 558 + /- 4, and then expands on the advent of Vicki Bruce to 690, 721, 704, 596, 696 respectively for the volumes published during her editorship, thereafter returning to its previous proportions. This history is clearly too recent to address, but must mean something!) While we are rightly celebrating its centenary, we should not ignore the fact that the BJP's changing content might also be read as reflecting various editorial policies aimed at securing its status and defining (or redefining) its role at a succession of historical junctures. More trivially we might also look at the rise of multiple authorship (on which the BJP may be considered a representative academic journal) and the proportion of foreign authors. These would require a further, but straightforward, survey. (The author's impression on the first of these is that the real turning point is as recent as the late 1980s, the single authored papers dropping from 20 in 1985 to 10 in 1992, 6 in 2000 and 4 in 2002, although 2003 contained 11. By contrast, Volumes 9-12, 1917-1921 contained only 5 co-authored papers.)

In sum, the precise nature of the relationships (there have surely been several at different periods) between the BJP contents and the situation in British Psychology as a whole remains unclear. The research reported here, uniquely comprehensive though it is, can only be considered as an exploratory foray. Nevertheless it hopefully serves to illustrate the potential of this type of historical research, which is unenvisaged by Chase (1995) in one of the very few discussions of the use of historical analysis as a Psychological research method. It also quite effectively blurs the line betweeen History of Psychology and Historical Psychology.

To close with some personal reflections. This paper was undertaken in the spirit of providing the BJP with a fairly simple, primarily celebratory, review of its first 100 years. I imagined much space would be devoted to 'great papers' and 'great authors' it had published. The second two issues of Volume 14, which appeared in 1924, would hav\e been an obvious highlight, featuring as they did E. Claparde, W Khler, K. Koffka, O. Lippmann, G. Rvsz and LL. Thurstone (twice). Another would have been Vol. 11, Part 1 of 1920, which contains a 5-paper symposium on 'Is Thinking merely the Action of Speech Mechanisms?', featuring EC. Bartlett, G.H. Thomson, TH. Pear and J.B. Watson himself. A closer look at the early 1930s exchanges on statistics and intelligence between C. Spearman, G.H. Thomson, R.B. Cattell, WK. Stephenson and others would also have been interesting. It was not to be. What began as a preliminary orienting operation rapidly took over. Whatever my 'social constructionist' leanings as an historian of Psychology I had thought to set them aside on this occasion, but should perhaps have taken my convictions more seriously. While I anticipated that the data might lend themselves to a constructionist interpretation what, naively as I now feel, I did not expect was that this issue would emerge in the very process of content-classification itself. It struck me increasingly that what I was seeing was a surface manifestation of an ongoing process of socially defining, dissolving and redefining the very subject matters of Psychology itself. To use a geological analogy, the data bear the same relationship to this underlying psychological process as the surface topography of a landscape does to subterranean rock formation. It was, moreover, particularly gratifying to discover that even the somewhat elementary kind of analysis, which was all I have so far been in a position to undertake, could yield at least two quite unforeseen and clearcut results: a switch in the character of the papers being published in 1983 and a virtual revolution in the discipline's own category system (at least as represented in the BJP) in 1965. There also seems to be much more to be gleaned from the various historical trajectories we can trace for different categories of paper, the interpretation and implications of which are often less straightforward than might be assumed.

As the major topics of Psychological research are constantly hived off into their own specialist journals, the British Journal of Psychology is perennially faced with the need to remake itself and reconsider its own function. One role it has always played is to provide a forum for new ideas and research genres, and for important work which fits uneasily into the existing scheme of things - one thinks of face-recognition and face-perception (which first appear in 1975, Vol. 66), handedness - which was pioneered by Chris McManus in particular during the early 1990s, and Susanna Millar's almost single-handed studies of psychological issues related to blindness through the 1980s. In that respect History of Psychology, presently in its own state of radical self-recreation, might be considered a worthy candidate for further appearances in the future (see Bunn, 2001; Richards, 2002). I only hope the present paper has not irrevocably scuppered its chances!

1 In this paper I adopt my usual policy of capitalizing Psychology/Psychological when refering to the discipline and using lower-case psychology/psychological for its subject matter. This distinction is now being adopted by some other disciplinary historians as a way of facilitating discussion of the relationship between the two.

References

Blackmore, S. J. (1997). Probability misjudgment and belief in the paranormal: A newspaper survey. British Journal of Psychology, 88, 683-690.

Bunn, G. C. (2001). Introduction. In G. C. Bunn, A. D. Lovie & G. D. Richards (Eds), Psychology in Britain. Historical essays and personal reflections (pp. 1-29). Leicester and London: BPS Books, The Science Museum.

Chase, J. (1995). Historical analysis in psychological research. In G. M. Breakwell, S. Hammond & C. Fife-Shaw (Eds), Research methods in psychology (pp. 314-326). London: Sage.

Clark, A. D. B. (1979). Editorial - Seventy-five years of the British Journal of Psychology 1904-1979. British Journal of Psychology, 70, 1-6.

Clough, F. (1978). Staff ratings of children's behaviour in hospital: Comparability of factor structures. British Journal of Psychology, 69, 59-68.

Danziger, K. (1997). Naming the mind. How psychology discovered its language. London: Sage.

Ferguson, E., Singh, A. P. & Cunningham-Snell, (1992). Stress and blood donation: the effects of music and previous experience. British Journal of Psychology, 88, 277-294.

Hearnshaw, L. (1964). A short history of British psychology 1840- 1940. London: Methuen.

Hill, A. B., & Perkins, R. E. (1985). Towards a model of boredom. British Journal of Psychology, 76, 235-240.

Jardine, R. (1874). The elements of the psychology of cognition. London: Macmillan.

McCutcheon, L. E., Lange, R., & Hooran, J. (2002). Conceptualization and measurement of celebrity worship. British Journal of Psychology, 93, 67-88.

Neary, F. (2001). A question of 'peculiar importance' George Croom Robertson, Mind, and the changing relationship between British psychology and philosophy. In G. C. Bunn, A. D. Lovie & G. D. Richards (Eds), Psychology in Britain. Historical essays and personal reflections (pp. 54-71). Leicester and London: BPS Books, The Science Museum.

Perkins, R. E., & Hill, A. B. (1985). Cognitive and affective aspects of boredom. British Journal of Psychology, 76, 221-234.

Price, D. J. de Solla (1961). Science since Babylon. New Haven CT: Yale University Press.

Richards, G. (1997). 'Race', racism and psychology: Towards a reflexive history. London: Routledge.

Richards, G. (2000). Britain on the couch: the popularization of psychoanalysis in Britain 1918 - 1940. Science in Context, 13, 183- 230.

Richards, G. (2002). Putting psychology in its place: A critical historical overview (2nd ed). London: Psychology Press/Routledge.

Russell, R. J. H., & Wells, P. A. (1994). Personality and quality of marriage. British Journal of Psychology, 25, 161 - 168.

Spearman, C. (1923). The nature of 'intelligence' and the principles of cognition. London: Methuen.

Watts, F. N., Webster, S. M., Morley, C. J., & Cohen, J. (1992). Expedition stress and personality change. British Journal of Psychology, 83, 337-342.

Wiseman, R., Greening, E., & Smith, M. (2003). Belief in the paranormal and suggestion in the sance room. British Journal of Psychology, 94, 285-298.

Wood, J. (2004). Integration not polemics. The Psychologist, 17, 433.

Graham Richards*

BPS History of Psychology Centre, London, UK

* Correspondence should be addressed to Dr Graham Richards, BPS History of Psychology Centre, 33 John Street, London WCIN 2AT; UK (e- mail: graric@bps.org.uk).

Copyright British Psychological Society Nov 2004

Source: British Journal of Psychology

More News in this Category