Making testers out of teachers: the work of a Swedish State Research Institute 1946–1956

ABSTRACT Between 1946 and 1956, the Swedish Psychological and Pedagogical Institute (SPPI) organised several summer courses for the purpose of training teachers in intelligence testing. The aim of the courses was to make these teachers the first gatekeepers who would meet and direct the youngest pupils into ordinary classes or into special classes. This paper investigates the course leaders and the participants in these courses, as well as the content taught. It is argued that these testing courses are examples of a shift in assessment in education from trusting teachers’ ‘judgements’ of pupils’ skills and abilities to externally standardising the ‘measure’ of these merits. It is also argued that researchers and teachers were part of a larger change in the politics of IQ. SPPI’s role within this process was that of a new and modern institution serving society: making IQ testing a public familiarity and the stratification of children that often followed from it.

Sweden also has a long history of external tests aimed at 'standardising' grades, ie more or less deciding them. 3 So, why have Swedish teachers accepted that external tests, in a way, 'override' their own judgement concerning the students' skills and knowledge? One way to understand this is to look at how psychological testing (mental testing) gained increasing influence in the educational field during the twentieth century. Other scholars have claimed that this was a result of the effective politics of IQ; 4 the legitimisation of merit; 5 the successful match, or marketisation, between applied psychology (group testing) and education; 6 and/or the higher status of male testers compared with female teachers. 7 While I believe that all these apply for Sweden, 8 I also argue that Swedish teachers were, and still are, trained to accept external tests as at least equally valid as, and more reliable than, their own assessments.
It is clear from the history of psychology that part of the reason the way of looking at individuals as measurable subjects became disciplinarily successful is because of its administrative relevance to bureaucracy and politics. Experts on individual differences were attractive for sectors that administered social programmes, such as the military or the educational system: insofar as it could assist in the administrative processspecifically by providing a culturally acceptable rationale for the treatment of individuals by categories that bureaucratic structures demanded. 9 In education, mental testing of individual differences became increasingly fashionable as education became compulsory and gradually demanded more and more effort and time from the pupils. In Sweden during the 1940s, children could be placed in at least 6-7 different special needs classes, and their general school readiness was thoroughly tested, 10 all with the intention of adapting teaching methods in an optimal way. How various kinds of mental testing merged into the educational discourse and organisation due to its administrative usefulness in Sweden during the twentieth century is well recorded. 11 Thinking surrounding individual mental differences became an important tool for larger educational reforms in Sweden. However, how the techniques of mental testing and group testing reached teachers and presumably also affected their own assessments is not so well recorded. We know that teacher union magazines helped introduce the psychological perspective and language, 12 but what about training?
In this article, I will look more closely at one institution that initially played a major role in training teachers to think like testers: the State Psychological and Pedagogical Institute (SPPI), formed in 1944. The SPPI was prepared through great lobbying work by many of Sweden's most progressive educators who, in 1935, had formed the 'Society for a Psychological and Pedagogical Institute'. The progressive movement in Sweden in the 1930s and 1940s consisted of a discursive alliance representing a kind of progressivism as a programme for teacher professionalism. One important ingredient in this professionalism was closeness to psychological and pedagogical research. 13 This was also exactly what SPPI was supposed to promote. This ambition was strongly influenced by similar institutes in Germany, the United States and Scotland, 14 according to the regulations of SPPI: 15 §2 The institute shall promote and conduct research activities on the psychologicalpedagogical field with a direct focus on the needs of school, and seek collaboration with the psychological-pedagogical and medical research at universities and colleges. The institute also, where appropriate, can provide individual researchers with an opportunity for independent research within the institute's field of activity. Cooperation should be organised between the institute and other institutions and associations in this area. 16 Further, it was regulated that the institute should, when appropriate: §4 . . . in cooperation with universities colleges and other institutions organise training courses of various kinds, both for future teachers and partly for educated teachers. In addition, as far as possible, the institution should arrange training courses for nurses in kindergartens. 17 testing. Representatives from SCRE and from the Society for a Psychological Pedagogical Institute cooperated in the 1930s in the so-called International Examinations Inquiry (IEE) with, among others, representatives from Teachers College in New York. IEE was an almost 10-year-long project funded by the Carnegie Corporation, aiming at standardising examinations in the western world. The project placed great trust in the future of testing. 19 According to Martin Lawn, research institutes such as these were flexible ways of managing specific tasks and influencing policy: 'they appear to conform to disciplinary procedures and yet they are free to inquire and suggest policy directions'. 20 They did not, like the universities, have a specialised staff of researchers but worked cooperatively, as knowledge networks.
When SPPI was founded, there were four professors in psychology and education in Sweden, one each at Gothenburg, Lund, Uppsala and Stockholm universities. According to the regulations of the institute, the professors were evident members of the board: §7 The institute is a board consisting of twelve members. Self-written members are the head of the higher education institute, the institute's director, the professors of psychology, and in pedagogy and pedagogic psychology at Uppsala University, professors in psychology and pedagogy at Lund University, as well as Stockholm and Gothenburg colleges, also the head school doctor. Other members, of whom at least one should represent the primary schools, at least one the grammar schools and one the vocational education and one the interests of the public, particularly the parents, and alternates for them are appointed by the Ministry for not more than three years at a time. 21 As I will show, the institute promoted a test-based grading system based on three learning strategies. First, the institute built up a knowledge base on teacher-assigned grades (and their problems) and on group testing in education. This was accomplished with a research project financed by the institute and also by developing its tests. The SPPI initially developed the so-called 'school readiness test' with the purpose of calibrating the course plans according to what children at a certain age could 'normally' learn. The institute then began to develop a national standardised test. Second, the institute gave a 'further development course' in psychology and pedagogy, which presented theories on intelligence and intelligence testing. Third, the institute gave 'crash courses' to teachers on actual intelligence testing. These three learning strategies taken together helped teachers become accustomed to testing. Even if only a few teachers attended the actual courses and were 'made into testers', SPPI developed national tests for all teachers to use, accompanied by a strong research-based testing discourse. 22 In two main sections, I will advance the argument that teachers' 'internal' assessments were gradually replaced with, or at least complemented by, increasingly external testing, in a process whereby teachers were 're-professionalised' to accept this development through training. The first section describes the tradition of teachers' internal assessments and illustrates how it was challenged. The second section describes the work of SPPI establishing testing as something 'ordinary' in education and, at the same time, educational research as something helpful to teachers through the courses of SPPI. In the concluding remarks, Idiscuss some long-term consequences of this period in Swedish educational history and the repetitiveness of state-funded institutes that arise over time with the purpose of basing the teaching profession in scientific knowledge and routines.

From assessing to testing merits
During the nineteenth century, grading and assessments were not particularly regulated in Swedish primary schools. A standardised grading scale was introduced in 1897, but how to assign the grades was generally left to the teacher. When grading, a teacher relied heavily on his or her observation of the pupil's performance, and not least on how the pupil responded to homework inquiries and to the end-of-the-school-year questioning held by the school inspector. 23 Some teachers kept diaries in which they took notes on their pupils' progress; for example: Karin Lindblad. Conduct well. Attention good. Generally good diligence. Her attitude has also got remarks this year; However, there is a clear improvement. Progress is due to her ability to follow the lessons so well, especially she seems to have [found it] easier to follow the maths this year. The writing is still bad. 24 the learning processteachers, students, parentswas severely limited. As the scale of the education system became larger and more complex, the limitations of these early forms of grading became more acute. 26 When grades were to be communicated beyond the school site, marking systems had to be made more reliable, more universal and more standardised. In Sweden, some early studies conducted by a lecturer, Frits Wigforss , at the Teachers College in Kalmar, indicated that teachers' grading lacked reliability. 27 Wigforss suggested that standardised tests should be deployed to help teachers calibrate their grading in relation to a national norm (Figures 1 and 2). 28 The bell curve/normal distribution gradually became canonised as the main principle that should be obeyed in grading. 29 The overarching aim was to replace the admission test at the grammar school with the simpler and faster procedure of using the final grades from primary school for the Figure 1. The national standard tests were supposed to adjust the grading scale but not the teacher's own ranking. 30  selection procedure. 32 Before standardised grades were used, teachers had to prepare and train the few pupils who wanted to go to grammar school so that they could succeed in the admission examination. The consequence of this was that the rest of the pupils received an education that was far too advanced for them. More than 75% of the student population in 1950 did not go to grammar school. 33 To preserve the elementary teachers' right to teach at a more basic level, grades were seen as a possible substitute for the old admission test on classical subjects. The grammar teachers did not like this idea, so it had to be legitimised, and this was possible with the help of standardised testing. By standardising the grades with help from external testing, it also became possible to accept primary school as the foundation of the educational system. Sweden had a tracked school system with parallel classes in years 4 to 6. Between 1949 and 1962, all the school districts in Sweden changed that into a unified, comprehensive school system. Many of the ideas behind the standardised test came from the United States. As in the United States, the Swedish test developers were trained during the Second World War to construct and carry out group tests. After the war, many of the psychometricians who had been working for the military with group testing moved over to the field of education, and especially then to SPPI. The ongoing school reform, which aimed for a comprehensive school system, provided standardised testing with an ideal playground is used as a metaphor for a place perfectly suited for correlations and deviations . The new field offered this psychological brand of educational science a shelter for its own development. Not being very coherent in theory, skills in methodology gave scientific status to this brand. 34 Abstract and numerical calculations functioned as a way to gain legitimacy. The whole procedure of student admission could now be conceptualised as an objective matter. Initially, the teachers did not really like this, 35 as their judgement had been replaced. Porter wrote about a similar development in the United States, where the teachers were mostly unmarried women who did not really have the professional self-confidence to stand up against these 'men of standardisation'. 36 In Sweden, however, the feminisation of the educational field came later. Self-confidence probably played a bigger part than gender. However, I would also argue that they learned to accept these new norms, partly through the efforts made by SPPI.

SPPI and the promotion of standardised testing
One of SPPI's major purposes was to 'promote and conduct research activities on the psychological-pedagogical field with a direct focus on the needs of school', as stated above. In relation to the 'comprehensive school reform', the introduction of further-reaching standardised testing become logical in several ways, not least in relation to admission, and thus a way of legitimising the comprehensive school reform as such.
As mentioned, SPPI promoted standardised testing in three ways: (1) by itself developing tests and test theories 37 and by supporting external research projects to do this; (2) by delivering introductory courses in psychology and pedagogy, in which theories about IQ and IQ testing were big parts; and (3) providing courses that actually trained teachers in the craft of testing. 38 In this article, I will focus on the second and third strategies but will also add something briefly about the first.
In a board meeting protocol from 11 December 1944, we can see that the board calculated that these three areas would cost approximately 55,000 SEK. The board also wanted to continue investigations on 'admittance to higher education and problems related to that'. This was an international engagement in what was called the International Examinations Committee (IEA). 39 In total, the board anticipated that it would spend 75,000 SEK 40 on investigations and courses, mostly within the testing area. 41 At this time, most educational research in Sweden was financed through the four professors. In comparison, the average professor's salary 1945 was 12,000 SEK, 42  clear that SPPI meant a substantial positive change in the funding of Swedish educational research. The conditions were also far better for SPPI compared with the SCRE in Scotland, which on a yearly basis was funded with only £3000 until the mid-1950s. 43 Supporting research on testing I have not been able to find how it was possible to apply for funding from SPPI. However, a large portion of research funds went to internal initiatives. According to a protocol from 17 September 1945, the four professors in psychology and education in Sweden, all of whom were appointed to SPPI, were given a share of 24,000 SEK to develop 'psychological tests'. 44 These tests were supposed to measure the 'standard' knowledge, or ability, within certain cohorts, in order to better calibrate the curriculum to what a 'normal' schoolchild could be anticipated to achieve. There was a division of labour here between the four universities and their professors: With a slightly increasing amount of money and applications, SPPI continued to fund projects, largely but not exclusively related to abilities and testing, until 1953. It is also worth noting that approximately a quarter of the total applications were submitted by teachers, principals and school doctors, and many approved applications concerned practical and didactic issues.
Parallel to the development of mental tests and to the approved applications, Frits Wigforss had a standing grant to develop standardised testing for assessing school content knowledge in maths, writing and reading.
In conjunction with the new curriculum for the primary schools, called U55, responsibility for development of standardised tests from SPPI and Skolöverstyrelsen (the National Board of Education) passed to the newly formed Teachers College in Stockholm. This happened in 1953, the same year Frits Wigforss died. The new head of testing development was the newly admitted professor of practical pedagogy at the Teachers College, Torsten Husén

Courses in psychology and pedagogy
Evidently, SPPI was an important actor in funding research on testing of various kinds. It also arranged training courses in testing. The archives include 53 documented course programmes between 1946 and 1956 for 11 different types of courses (Table 1). 56 Some of these also show how many people applied and, in some cases, attendance lists are attached. Most of the institute's courses were summer activities (the course programmes often encouraged participants to bring bikes and swimwear) so that as many teachers as possible would be able to attend. 57 One problem, however, was that it was often difficult to find children for the practical activities during summer especially then concerning the testing course. Consequently, Professor Katz proposed arranging these courses close to Barnens Ö ('The Children's Island'a famous summer camp in the archipelago of Stockholm for socially disadvantaged children) and similar summer camps for children elsewhere. 58 Most courses lasted for about one week, according to the course programmes. However, the test courses were sometimes longer. In 1953 a three-week course was organised, and in 1955 a four-week test course was organised, in parallel at Karlstad, Arvika and Filipstad. At least 53 (51) 59 courses were initiated over the 10 years that SPPI organised course activities. Unfortunately, there are no annual reports in the institute's archive, so it is not possible to say whether additional courses were offered during these years. However, the institute's diary for incoming and outgoing mail does indicate that a few courses in testing had already been held in 1945. 60 Of the 11 different courses that SPPI organised, the educational training course for vocational teachers was the most commonly run, followed by the test course and courses in psychology and pedagogy (Table 1).
The course programmes show that 'the Institute's pedagogical continuing education course' mainly focused on vocational teachers. Twenty of the documented courses were Participants. Based on attendance lists, at least 514 people participated in a course organised by SPPI. The archive shows that each course had approximately 20 participants. 63 Assuming that there were at least 20 participants in each course, even in courses for which attendance lists are missing, SPPI trained and educated more than 1000 people through its courses. 64 Of these, 400 would have been vocational teachers. The application pressure on the courses was high, where such information is available. For example, a testing course in June 1951 had 220 applicants, and one starting in June 1955 had 250. 65 The participants in the Psychological Measurements and Educational Statistics courses were exclusively academics, college teachers and principals. In other words, the universities, teacher seminars and grammar schools were probably the main target groups for these courses. This would also apply to the courses for psychology and philosophy teachers. Other attendance lists testify that the remaining courses were aimed at upper secondary school teachers, primary school teachers and preschool teachers. 66 Course leaders. These courses involved a large number of course leaders: teachers in psychology and pedagogy, primary school inspectors, directors of education, lecturers, principals and others. Jon Naeslund, later the author of the well-known book Allmän undervisningsmetodik (General Education Methodology), held most of the courses for vocational teachers. Wilhelm Sjöstrand, later a well-known professor of education at Uppsala University, and Jan Gästrin held the first course in psychology and pedagogy. At the time, both were associate professors, and they worked mainly on intelligence testing. Sjöstrand was also the lead teacher for this course in 1947, together with, among others, Stellan Arvidsson and Alf Nyman. Arvidsson was in particular responsible for a part of the course that concerned individual education (a form of internal differentiation). At this time, Arvidsson was also deeply involved in the ongoing comprehensive school reform. Gästrin also returns for several courses. Torsten Husén, another wellknown figure, participated in many courses. Frits Wigforss, Elisabeth Dahr and Rosa Katz (wife of David Katz) also appeared as lecturers. Representation from the Education Department at Stockholm University came from the teachers Gösta Ekman and Arne Trankell, both skilled test researchers (the latter an expert on the challenges faced by left-handed people). Also appearing on the lecture lists were Sten Henrysson, later a driving force behind the Swedish SAT (Högskoleprovet), and Sven-Erik Henricson, a few decades later head of the testing department at the National Board of Education; both were doctoral students of Torsten Husén. Gothenburg is represented via Elmgren's PhD student Karl-Gustaf Stukát. 67 The collection also includes a large number of lesserknown people but, in sum, many of those who made a name for themselves in Swedish education in the decades to come served as teachers on the courses run by SPPI. 68 The representatives from the discipline of psychology and pedagogy came from the younger generation. However, from the institute's diary for incoming and outgoing mail, it appears that the older generation was also involved, at least in the testing courses. There are notes on correspondence between John Elmgren, John Landquist, Rudolf Anderberg and David Katz (as well as Sjöstrand and Siegvald from the younger generation). 69 In other words, a large number of course leaders appeared at SPPI, a system that could be summarised more systematically with the finding that representatives of the 1940 school committee, the 1946 school commission (that is, the comprehensive school reform), psychological-pedagogical research, and teacher education participated in this form of teacher education work.
Course schedule and literature. The course programmes in the archive often have a schedule and a bibliography list. Our interest is primarily in the courses on psychological and measurement terminology. 70 The courses in psychology and pedagogy had some recurring themes and consisted of lectures in both the morning and afternoon. Sjöstrand and Gästrin, for example, lectured in their first course on topics such as knowledge development, intelligence, testing, fatigue and overstrain, aptitude testing, the psychological structure of the curriculum, and admission processes during the transition from primary school to higher school. Psychology and school organisation was the very essence of the course. However, there were some variants of this course; for example, the course week 9-14 August 1948 had some additional educational elements such as the 'mother tongue's position in the overall teaching' (Lecturer Ester Hermansson), and 'teaching in English at the lower stage'. The course was also given in 1949 with a special focus on art teachers, who were taught about gestational psychology, behaviourism, characterology, psychoanalysis, test psychology, left-handedness, and more obviously artistic subjects such as 'from children's painting to youth painting'.
Many of the courses in psychology and pedagogy mixed (differential) psychology with subject education. However, one of the courses that maintained the worst balance was the continuing education course for kindergarten teachers. As early as the third lesson on the first day (25 June 1951) there was a demonstration of individual testing (Terman), according to the course schedule. After lunch there was yet another lesson on the Terman test, followed by two lessons on the theme of 'heritage and environment'. The following morning began with a 'double lecture' on heritage and environment, followed by testing according to Büler-Hetzer. Teaching followed throughout the week on aptitude tests, psychological tests, mental hygiene tests, school readiness and left-handedness. Clearly, the study of school readiness seems to be the main subject of this test orientation. These teachers would be the first outpost of differential thinking when the school received new children.
Therefore, it is clear that it was not only those courses called 'test courses' that taught the traits of testing. However, the test courses were further refined with practical exercises in testing.

The test courses
The three courses called 'psychological measurements and pedagogical statistics' were led by Husén and Ekman and by two newly appointed professors, licentiate of arts Henrysson and Principal Dahr. With some minor variations, the courses were structured so that, after the lectures by Ekman on day 1, day 2 had seminars in pedagogical statistics and demonstrations in individual and group tests. On day 3, Henrysson lectured on methods of measurement in schools, after which the participants were allowed time for an excursion. Henrysson continued his lecture on day 4, when Dahr also taught about admission to higher schools. Day 5 was then devoted to personality tests as well as group tests for 15-year-olds and older. According to the programme, a debate was held on the last day, in which Husén and Henrysson were appointed as respondents. The participants were expected to attend the course week equipped with Ekman's book, Intelligensmätning (Intelligence testing), Husén's book, Testresultatens prognosvärde (The forecast value of test results), and Wigforss' book, Matematisk statistik för pedagoger (Mathematical statistics for educators).
However, this course was not very test intensive compared with the 11 (nine?) test courses. These courses would make testers of the primary teachers so that the teachers themselves could sort children into the 'special needs classes' (hjälpklasser). From 1950 to 1952, SPPI issued a letter to all school inspectors in Sweden in which they examined the needs of test-trained teachers, partly on the basis of proposals for suitable teachers to train. There are 51 answers in the archive, the vast majority of which suggest that such a need existed. For example, Äppelviken's primary school inspector, Helge Haage, responded as follows: Testers are common in the larger school districts but are missing for the smaller ones. It is therefore desirable to train those who can live somewhat centrally in an inspection area and are willing to assist in testing in the different school districts. There is a need for such testers, for example, at Djurö, Ed, Ekerö, Färentuna, Hilleshög, Lovö, Munsö, Möja, Skå, Sånga and Värmdö. 71 The courses took off, and Helge Haage and the other colleges of school inspectors returned with lists of teachers who reported interest in the courses from their area, and which the school inspectors ranked in first, second and third place. Judging from the number of applicantswhich was, as mentioned, quite highthe courses appear to have been very attractive. SPPI was also careful to offer them in many parts of Sweden, not just in the larger cities. 72 The estimated 200+ teachers participating in the testing courses came from all over Sweden and were supposed to fulfil the nation's need for testers. The initial idea was that each of the four universities should arrange one testing course per year, supervised by the professor of psychology and pedagogy. 73 The research committee had described the basic structure of these courses as follows: 4-5 lectures for orientation; 15 seminars; 80 hours' practical work (testing, discussion, etc.) distributed so that each participant gets 10 hours' individual supervision. Every participant shall also independently test at least 10 pupils.
The theoretical training through lectures and seminars is held by a professor or an associate professor, the practical part by an assistant. 74 For an example of what the participants could expect from these courses, we can look at what the teachers who came to Ljungskile Folk High School for three weeks in the summer of 1953 had to do. Even before they arrived, they were requested to read books by Anderberg, Wigforss and Terman-Merrill. Until 1952 there had been no pre-course reading requirements, but it turned out that the course leaders did not feel the participants' knowledge of psychology was good enough to be able to follow the course. 75 During the first week, the 18 teachers conducted two tests on Monday, 18 on Tuesday, 21 on Wednesday, 21 on Thursday and another eight tests on Friday, or a total of 70 tests. The tests lasted for a maximum of 1.5 hours. Teachers also had seminars and lectures in the evenings. For the first week's tests they performed these on each other, and they had access to children in the second week. Five groups of five children from Ljungskile primary schools and from Örebro school district experienced how it felt to be tested with Terman-Merrill's or Ohlander-Wessman's test. According to an early Working Committee protocol, the 'subject of the experiment'that is, the childrenshould be paid 1 SEK each (about 2 euros at today's value). 76 In the third and final week, the teachers were instructed to test intellectually handicapped children in special needs classes. This was called 'doing field tests'. The course ended with a written examination of the course participants. Children who passed received a certificate. The test course in January 1955 went on in a similar manner: the first week had lectures throughout, and the second week comprised test exercises under the supervision of the course management. In the third week teachers were allowed to do five of their own tests as well as some related exercises. In the fourth and final week, they would conduct two test tests in the field. If they passed their exams it meant they had carried out nine tests each and passed a written assignment. Now they were (believed to be) ready to meet the children for 'advocacy' at, for example, Värmdö.
In 1957, most of the work carried out at SPPI was moved to the newly funded Teachers' College in Stockholm . It was assumed that the college would continue offering these courses. In a referendum to the proposition called 'Den första lärarhögskolan' (The first teachers' college), 77 the board of SPPI clearly stated that it approved of the idea of letting the teachers' college take over these courses. 78 If this is in fact the case, I have not yet been able to verify it. They definitely continued to develop the standardised test and the heritage from Frits Wigforss, thus moving away from mental testing towards knowledge assessments. 79 Overall, SPPI's course activities illustrate a way in which specific knowledge was disseminated to school teachers in the 1940s and 1950s. It should be quite clear that SPPI was an important actor in the dissemination of both testing techniques and psychological concepts. SPPI can also be seen as a shared platform for some progressive scholars and school reformers, on which to base a new ambition for Swedish education.

Conclusion and final remarks
There is substantial evidence that, in just a few decades after the Second World War, Swedish teachers went from assessing students' knowledge to trying to measure it. 80 How was this possible? Taken together, the activities of SPPI aimed, somewhat successfully, to train teachers to understand, conduct and accept (mental) testing as a natural part of everyday education and thus as playing an important role in the 'discursive machinery', promoting more testing in education.
Even if only a small proportion of Sweden's teachers participated in the testing courses arranged by SPPI, the courses themselves can be seen as a more general expression of a movement that eventually encompassed almost all Swedish teachers. While the testing courses focused heavily on mental testing or school-readiness testing, the work that SPPI itself carried out and financed in the testing area during these years gradually turned from ability testing towards knowledge assessments and the development of standardised tests. In 1957, most of what the Institute worked on was moved into the newly funded teachers' college in Stockholm, where Torsten Husén carried on with the development of the Swedish national test. In one way, it is possible to think that SPPI normalised testing, and when this enterprise moved to the teachers' college the process of institutionalising these kinds of external tests in Swedish education was completed. At the same time, these tests were no longer fully external, because they were developed at the heart of Swedish teacher education. In the years that followed, the national tests became even more calibrated to the curriculum and the everyday teaching in schools, in an effort to measure more complex skills. In this endeavour, the Swedish progressive movement, which aimed to professionalise teachers through scientific training, was relatively sensitive to teacher practice, even though it clearly proclaimed that tests were more objective and reliable than teachers' own assessments. The standardisation of grades, using external testing, gave the grades increased legitimacy, but whether it contributed to teachers' professionalisation regarding assessment is highly uncertain. According to much sociology of professions, there is a need for symbolic and social work to produce boundaries or domains of a profession. 81 Post-Second World War Swedish education appears to have been another tendency, moving in the opposite direction rather than unifying divergent professions in and around education. 82 SPPI was a place that organised the discourse on testing in that it tied together representatives from the four universities' psychological and pedagogical departments and also connected them with representatives of comprehensive school reform and with teachers. SPPI became a hub for 'professional alliances'. 83 It financed their meetings and their research on policy-relevant matters (mainly testing) and helped disseminate a testing discourse to participants through courses. Most of the individuals mentioned in this paper were visible for several decades in the discursive and material construction of the comprehensive school system. SPPI gathered international experiences, offering opportunities to disseminate written material and meet teachers and, perhaps not least, the opportunity for the researchers involved and the reformers to meet each other. SPPI helped 'market' the science of psychology and education, allowing it to show its usefulness. 84 This kind of state-relevant research became criticised in the 1970s and 1980s, however, and the sciences of psychology and of pedagogy (in Sweden they were separated as two different disciplines in 1949) distanced themselves from policy as well as teacher practice. Therefore, it is interesting to see that a new agency, similar to SPPI, was formed in 2014 -Skolforskningsinstitutet (the Swedish Institute for Educational Research), to 'compile the best available knowledge of educational working methods and practices' and to distribute funding for practice-centred research. 85 There again appears to be a need for an institute to make teaching more research based, teachers more informed and research more useful for educational practice.
One way to understand why SPPI (and similar research institutes) emerged is in relation to the need for a new kind of central control. SPPI promoted a new type of control instrument based on the production of psychometric information. SPPI is, in the broadest sense, an example of how the organisation of a modern society begins to require targeted knowledge in certain sectors. With such an interpretation, it is possible to consider SPPI as a complement, at the time, to the universities and to the National Board of Education (Skolöverstyrelsen). The Scottish equivalent of the State Psychological Educational Institute, SCRE, has been described as an institution that took care of a type of knowledge production that the universities considered themselves too politically neutral to provide, but also as an institution that took care of 'modern' issues for which the school bureaucracy was not developed. 86 SPPI was created because existing school authorities did not adequately force onwards the modernisation of the school.