LOZANOV'S   RESEARCH

Lozanov's research in the 1960s and 1970s was conducted predominantly at the Institute of Suggestology at Sofia, Bulgaria. Most of his studies were carried out within Suggestopedia, that is to say that he looked for effects that the method may have on memory and on physiological and psychological measures. He also conducted a small number of comparative studies in order to determine the effectiveness of Suggestopedia when compared to other teaching methods. Although Lozanov is not a linguist, the majority of studies were conducted with language students, since Lozanov (1978) believed that results in language learning were more easily measurable than in other fields of learning. Language teachers were extensively trained in the use of the method, and teaching was carried out in the favourable environment described above. Some experimentation, however, was carried out in natural teaching environments, most notably a two-year experiment in primary schools.

Findings of Lozanov's research are reported in Suggestology and Outlines of Suggestopedy (1978), the only major publication on Lozanov's empirical work which is available in English in the West. (He also published a scientific magazine entitled Suggestologija which is not readily available in the West and has not officially been translated). On the basis of his findings, Lozanov made a series of claims for the effectiveness of Suggestopedia. These, however, need to be examined critically since Lozanov's research procedures and the style in which experiments are reported have been the subject of severe criticism by a number of notable linguists (Scovel 1979, Mans 1981, Schiffler 1986a). The credibility of Lozanov's research is further threatened by the fact that claims of a highly dramatic nature have been falsely attributed to Lozanov or to Suggestopedia, and have been generalised and promulgated by the popular press and by some commercial language teaching enterprises.

The objective of this section is to present the claims that Lozanov himself makes for Suggestopedia, to examine their validity in the light of the evidence that he provides, to examine the soundness of his research procedures, and to point out distortions of his claims by other sources.

Five broad categories of claims for Suggestopedia can be identified in Lozanov (1978):

1. Exceptionally large amounts of materials are assimilated by the students.
2. Functional use of these materials is exceptionally high.
3. Retention of these materials is exceptionally high over long periods of time.
4. Students' physiological and psychological state is influenced positively.
5. Higher achievement can be produced in suggestopedic courses as compared to traditionally taught courses.

We will now look at each of these claims in detail.

1. Volume of material. One of the most dramatic claims which can be attributed to Lozanov (1978:322) is that an average of four times more new words can be given and assimilated in suggestopedic instruction than in instruction by other methods of language teaching such as audio-visual, audio-lingual and conventional (presumably grammar-translation) methods. In a basic suggestopedic course approximately 2000 items of vocabulary are taught in 96 lessons of instruction. This means that an average of 20.8 words are given per lesson. According to Lozanov (1978:322) the corresponding figures for traditionally taught courses are 7.0 words per lesson for audio-lingual courses, 5.55 for audio-visual courses, and 5.35 for conventional courses.

This data is simply given in a table (p.322) without further explanation, except that the figures were obtained from "official data". It appears therefore that this claim is not based on sound experimental research, but on observational data from an unknown source. There is also no further description of the nature of the courses used in the comparison. Lozanov's teaching was conducted under such favourable conditions that a comparison with traditionally taught courses can hardly be valid. Not only did the teaching take place in small groups, in pleasant surroundings and for several hours at a time, but students were also exceptionally motivated. According to Schiffler (1986a) suggestopedically taught students were selected from extensive waiting lists. Lozanov's own (1988) description of selection procedures suggests that students were also chosen according to psychological characteristics, in order to make groups as homogeneous as possible.

Lozanov provides no basis on which a valid comparison can be made about the volume of material assimilated by students taught at the Institute of Suggestology and that of students taught elsewhere. The term "assimilate", used by Lozanov to describe the learning process and possibly the learning outcome, is also difficult to interpret. Does this mean that students are able to use these materials in a meaningful way, or do they simply recognise or recall them? From the way that Lozanov describes the various tests given either the day after a suggestopedic session or at the end of the course (p.166, 203, 210), it is clear that these were translation tests, predominantly from the foreign language into the mother tongue. This form of testing gives information on students' recall only. Both Mans (1981) and Baur (1982) interpret this as a serious limitation of Lozanov's research.

Lozanov refers to hundreds of suggestopedic sessions in which between 100 and 1000 lexical items were presented to the students, and after which students were able to recall an average of 90% and more (p.166). However, we are only given sporadic information about the nature of the courses, the number and background of students involved or the length of the individual sessions. During the decade of experimentation at the Institute, Lozanov clearly had access to a vast pool of data regarding all aspects of the suggestopedic teaching, and it is unfortunate that he reports this in such as haphazard and unsatisfactory manner. Scovel (1979:261) is quite justified when he points out Lozanov's inability to substantiate his speculations with empirical proof. On the basis of the evidence which Lozanov provides in his 1978 publication, therefore, the above claim would be more soundly based were it rephrased in these terms: Highly motivated students, taught suggestopedically in small classes and in a pleasant environment, are able to recall exceptionally large amounts of materials.

2. Functional use of materials. Although language tests as described by Lozanov (1978:166, 203, 210) appear to be chiefly related to memory skills and passive knowledge of the language, he makes the following claims concerning the students' ability to handle the lexical items with which they have been presented in a suggestopedic course (1978:321-322):

(a) Students assimilate on average 90% of the 2000 lexical units presented.

(b) More than 60% of the vocabulary can be used actively and fluently in everyday conversation; the rest of the vocabulary is known at translation level.

(c) Students speak within the framework of the whole essential grammar.

(d) Any text can be read.

(e) Students can write with some mistakes.

(f) Students make some mistakes in speaking but this does not hinder communication.

(g) Pronunciation is satisfactory.

(h) Students are not afraid to talk to native speakers.

(i) Students are eager to continue studying the same language, if possible in the same way.

As discussed above, only the first and the second half of the second claim have been supported, at least on a limited basis, by Lozanov's experimental research. The other claims can only have been arrived at by means of the assessment of naturalistic data. With the exception of the third and fourth claim, these claims are not really dramatic or sensational from an applied linguist's point of view, considering again that highly motivated students in small groups had almost 100 hours of intensive teaching with the addition of music and suggestion, which have been shown to be instrumental in improved learning. However, Lozanov provides insufficient background detail to allow a satisfactory evaluation of these claims. It may be that such claims are indeed valid, but on the basis of Lozanov's (1978) reports, they must be treated with caution. They can at best be considered as items of anecdotal evidence.

3. Retention of materials. One of Lozanov's major interests was to test the retention rate of materials "assimilated" by the students over various periods of time. As a results of extensive tests, he claims that forgetting is minimal in Suggestopedia, and that retention is still exceptionally high after as long as 2 years after the original learning. Again, however, most results are simply listed in tables with no precise information on how tests were conducted. The initial assessment appears to have been based on the results of written translations of lexical items presented at random the day after the suggestopedic sessions (p. 203), while the delayed assessment was taken at various intervals after an entire course had finished (p.213).There is no precise information, however, on which basis students were selected for the delayed tests or on the nature and conditions of this testing.

Results are provided in two formats. Either individual students are referred to, or the results of a group of students are given. Lozanov usually states the students' initial recall rate, their delayed recall rate, the time elapsed between the two tests and whether or not the students had reviewed the materials in the meantime. The tables do not give information about how many words had been taught or tested in each instance. Lozanov lists results, which tend to be inconsistent, at random, and, without providing sufficient evidence of standard statistical analyses, makes claims regarding the statistical reliability of results. An example of this can be found on pages 213-215.

Table 21 (p.213) shows the "Percentage of Forgetting in Suggestopedic Memorization". The results of 21 subjects are referred to. The data for 12 students, however, is incomplete. Of the rest, five students recalled 76.3% initially, 67.2% after 12 months without reading the materials and 79.6% after having read it again. The other four students recalled 93.5% initially, 57.0% after 22 months without reading and 81.0% with reading. The selection of data in this table is surprising, considering that Lozanov had access to the initial recall of 416 subjects (p.205) whose average recall was given as 93.2% (p.204). A more interesting selection would have been to re-test as many subjects as possible from that sample at random after several intervals of time each.

Lozanov, however, refers to individual results instead, such as that of B.A. who took part in an experiment in which students were presented 1000 words in a single session (p.213). B.A. recalled 98% initially, 53.3% after 20 months without reading and 73% with reading. Without providing any evidence of relevant statistical analyses, Lozanov then goes on to say: "The large number of words on which the experiment with B.A. was based, as well as the great differences between the percentages in the comparisons we made, ensures quite high statistical reliability" (p.214). How this claim is to be interpreted cannot be ascertained from the data provided.

Lozanov then gives a graphic representation of the "Reproduction of suggestopedically memorised material" (p.214), which shows that initial recall is around 90% and delayed recall after 24 months around 57%. The source for this data, however, is not discussed. Lozanov also provides examples of 44 subjects who obtained an average of 85% on delayed tests up to 16 months after the initial learning (p.215), and goes on to say that "the tendency was always towards a delayed deterioration in the retention of material" (p.215). Immediately following this statement, however, he provides a table with the results of 10 students whose average initial recall was 99.4% and whose delayed recall after 3 years with either little or no review of materials was 99.2%. This last figure appears to be substantially inflated when compared to all the other figures Lozanov reports for the delayed recall tests.

As a result of his experimentation, Lozanov observed a radical deviation from the classical curve of forgetting provided by Ebbinghaus (p.214), in which one hour after the memorisation only 48% of material is recalled which deteriorates to 28% after 48 hours. Since Ebbinghaus experimented with nonsense syllables, Lozanov carried out an experiment with 133 subjects learning nonsense syllables in order to provide a more valid comparison. The experiment was conducted in three conditions, each varying according to lists learnt, testing procedures and subject numbers. The study is again poorly described (pp.216-217) and results are given in Figure 30 (p.217). The average recall of the first condition (15 subjects) is given as 90% immediately after memorisation, 80% after 4 hours, 85% after 24 hours and 85% after 48 hours. The corresponding figures for the second condition (40 subjects) were 55%, 45%, 45% and 45%, and for the third (3 groups of 26 subjects, not tested after 48 hours), 58%, 42% and 58%. It is interesting to note that the difference in results between conditions 1 and 2 is more dramatic than the difference between condition 2 and the Ebbinghaus results. Lozanov does not refer to these differences but concludes rather obscurely:

In the three variants of this experiment, a tendency to form a reminiscent type reproductiveness curve was noticeable, i.e., in subsequent checks students tended to reproduce more and more of the material presented, and delayed reproduction approached reproduction in the immediate check. (p.218)

Although it is impossible to ascertain the validity of Lozanov's results on the basis of the data provided here, the radical deviation from the Ebbinghaus curve is often referred to in advertising for commercial language courses.

Again, it is unfortunate that Lozanov made such poor use of large amounts of data regarding delayed recall. Although there is some indication in his reports that recall may be exceptionally high, both 24 hours after presentation of materials and in some cases even after 2 to 3 years, it is impossible to make definite claims about the extent of students' recall ability over time on the basis of the data provided. The most consistently supported finding appears to be that the mean recall rate of suggestopedically taught students after 24 hours is indeed around 90% which may decline to around 57% after a period of two years. Again, this finding is not as dramatic as it may appear at first when we consider that highly motivated adult students were required to translate language items predominantly from the foreign language into their mother tongue. There is also no precise information about what use students made of the language in the period between testing.

4. Physiological and psychological benefits. In order to test whether the high results reported in Suggestopedia were obtained at the expense of students' health, Lozanov carried out some investigations on the physiological and psychological effects of the method. These led to the following claims:

It can be claimed with certainty that suggestopedic instruction has no unfavourable effect on the health of students.

In a comparatively small percentage of students (17.4%), suggestopedic instruction had a favourable effect on some functional disorders.

A number of complaints of a neurotic nature disappeared during instruction, giving grounds for the elaboration of methods for group psychotherapy for neurotic patients, by means of suggestopedic instruction. (p.223)

It was established that in the suggestopedic schools neurotic disorders in children have decreased by half compared with those of the control schools. (p.226)

These claims are based on the data of 396 questionnaires in the case of the adults and on the reports of 12 psychotherapists and 4 university professors in the case of the children. We are not given any precise information on how questionnaires were evaluated, and no further descriptions of the nature of the two-year examinations and reports in the schools are provided. Lozanov goes on to say that "the psychotherapeutic, psychohygienic and psychoprophylactic sides of suggestopedy were experimentally studied and corroborated by I.Z. Velovski (1971,1975) and by other authors too" (p.226). These authors, however, are not identified, and Velovski does not appear in the English bibliography. Instead, Lozanov provides excerpts of letters received from students (p.224) which support his claims on a naturalistic basis. The only claims which appear reasonably well supported by the data from the questionnaires (p.223) are that suggestopedic teaching has no negative effects on the health of adult students, and that in some cases positive effects on functional disorders are observed by the students. In the absence of experimental data, the other claims can be considered on a naturalistic basis only.

5. Achievement as compared to other methods. Lozanov (1978) reports two large comparative studies in which the results of experimental classes receiving suggestopedic teaching were compared to those of control classes receiving conventional teaching. The first experiment was a three-week study with 75 adults, assigned to 3 experimental and 3 control classes, being taught English and French. The second was a two-year experiment in two primary schools. One school was assigned to the experimental condition while the other served as a control condition. In both studies achievement was found to be around 20% higher in the experimental classes. In the primary school experiment, however, Lozanov makes further more dramatic claims regarding achievement, which need to be discussed since they may have been the basis for exaggerated claims about the effectiveness of Suggestopedia by other sources. A misinterpreted result of the adult experiment by the research committee working on the project also needs to be discussed, since it resulted in a highly exaggerated claim being falsely attributed to Lozanov.

Since the serious flaws of the experimental procedures of the adult study have already been discussed at length by Scovel (1979), we will refer to them only briefly. Lozanov's experimental procedures are poorly described, especially with regard to the assignments of groups, to the method used in the control groups, and to the tests given. Experimental data is poorly presented and sometimes even inaccurately calculated: the result of the 20.5% higher achievement of the experimental groups, for instance, is given as 21.5% (p.17).

A more serious misinterpretation appears in the claim (p.27) that results in the experimental groups were 25 times higher than in the controls (Lozanov's book, however, is so poorly organised, and with such large gaps in information, that it may be possible that a different experiment is being referred to). Scovel (1979:256) attributes this claim to Lozanov:

The strong claims made about the potential benefits of suggestopedy do not come solely from his publishers or disciples, however, they emanate, in fact, from Lozanov himself. In Chapter 2 of the book under review, the claim is made that "As seen from the results obtained in experimental groups, memorisation in learning by the suggestopedic method is accelerated 25 times over that in learning by conventional methods."

This attribution is surprising, not merely because this highly dramatic claim is so far removed from all the others that Lozanov makes, but because it is clear from the text that it is not Lozanov himself who makes the claim but the research committee working on the project.

Naturally the author of a book must be responsible for its contents; in Lozanov's case, however, it is possible that he did not have a chance to proofread the book before its publication in the United States. If Bancroft's (1976) observations are accurate, then it is possible that sections of the book were cut and rearranged without Lozanov's knowledge or approval. This would explain the poor organisation of the book and the missing information. An error such as the above may also have been produced in the translation. We are here not trying to make excuses for the unscientific nature of Lozanov's presentation which cannot be denied, we are simply trying to establish which claims have been made about the effects of Suggestopedia and to attribute these to the proper sources.

Further dramatic claims in the literature may be based on another quite uncharacteristic finding which Lozanov himself reports in relation to the primary school experiment. Two schools in two different villages were chosen for this experiment. Some information about the children's reading abilities was given, but no information about the teaching method used in the control school. One school was designated as the experimental school while the other became the control school. One serious flaw in the research design was that the experimental children were taught in homogeneous groups while the control children remained in heterogeneous groups. The experiment was conducted over two years.

For the first year achievement was generally around 20% better (p.325) in the experimental school. The most dramatic finding was reported at the beginning of the second year when children in the experimental group solved 77.39% of problems presented while the control children solved only 5.28% (p.328). This means that results in the experimental group were 14 times higher than those of the control group. However, this result related to the testing of the second year material in mathematics which had already been covered in the experimental school in the first year, but which could only just have begun to be taught in the control school.

This constitutes an unfair comparison, and it would have been more valid to conclude that materials were covered in half the time and compare results at the end of the second year.

Instead Lozanov reports the results of two other schools, not hitherto mentioned, which appear to refer to similar tests under similar conditions. Here results are 63 times higher for the experimental students. Table 47 (p.330) shows 65.83% for the experimental group and 1.04% for the control group. These findings are so far removed from all others in the book, that the research procedures, especially the basis on which the testing was conducted, must be seriously questioned.

Lozanov (p.327) claims that the above results were corroborated by a large scale experiment which followed, including a total of 1500 pupils and 146 researchers. No further information on the design, procedure or subject matter for testing is provided for this experiment. The results were that the experimental children who had been taught in a five day week with no homework, assimilated 80.3% of the materials for the first grade and 81% of the materials for the second grade. The control children who had been taught in normal teaching time (presumably one day more per week with the addition of homework) assimilated 63.3% and 66.4% respectively. We do not know at what stage the second year materials were tested, but it seems more likely this time that tests were given to the control group at the end of the second year. These results hardly support the dramatic result quoted above. Instead they appear to corroborate the consistent findings of the experimental group performing about 20% better than the control group throughout the comparative experiments reported in Lozanov (1978).

Yet the report of the dramatic results together with the sensational claim made by the research committee above, may have been the basis of Ostrander and Schroeder's (1979:22) claim that "learning can be speeded up by five to fifty times" as an example of what can be achieved by Suggestopedia. Another experiment may have contributed to this claim. This was an experiment (Lozanov 1978:30) in which 1000 unknown words were presented to a group of highly educated professionals and academics in a one day suggestopedic session. Sources quoting this experiment fail to mention that it included 10 days of elaboration on the words and was a one-off experiment, even though Schuster (1978) points this out in his review of Lozanov's 1978 publication. The claim that can therefore be made for this experiment is that 1000 words were learnt in 11 days of intensive teaching. The American publicity release (Scovel 1979:256) for Lozanov's book, however, claimed that 1000 words could be learnt "daily", and according to (Ostrander & Schroeder 1979:43) this achievement could even be improved:

With the Bulgarian approach, 500 words a day was just 'Mach1'. By 1966, a group learned 1000 words in a day, and by 1974, a rate of 1800 words was charted. In 1977, Lozanov reported, some tests showed people capable of absorbing even 3000 words per day.

No experiments with more than 1000 words are known to this author and at no stage does Lozanov claim that even 500 words were being learnt 'per day' which implies that 3500 words could be learnt in a week. Gross distortions of this nature did not enhance the credibility of Suggestopedia as a viable teaching method. Since the book Superlearning was more readily available to the general public than Lozanov's (1978) publication, this claim became falsely associated with Lozanov and with suggestopedic language teaching. Both Gassner-Roberts (1987) and Schiffler (1987) quote commercial language enterprises which still advertise their courses on the basis of this and similarly exaggerated claims.

This practice has become so widely spread that some language teaching enterprises believe that they have to dissociate themselves from such claims. Hinkelmann (1988:1) writes:

Leider werden über die Superlearning-Methode immer wieder unsinnige Behauptungen aufgestellt und damit Vorurteile geweckt. So wird manchmal behauptet, man könne damit 50 mal schneller oder 1000 Vokabeln pro Tag lernen. Für derartige Behauptungen gibt es jedoch keine wirklichen Beweise.

[Unfortunately there are always nonsensical claims being made about the Superlearning method which evoke prejudices. Sometimes it is claimed that one can learn 50 times faster or 1000 words per day. However, there is no real proof to support such claims.]

Conclusions - Lozanov's research. Lozanov carried out a great deal of research over a long period of time with a large number of subjects looking at many aspects of suggestopedic instruction. Unfortunately his data is so poorly reported that it is difficult to check the validity of many findings. In general, Lozanov's own claims about the effects of Suggestopedia are not highly dramatic, especially if we take into consideration the favourable conditions in which his experimentation took place. He does, however report isolated and highly uncharacteristic findings in a school experiment which can be interpreted as achievement having been 63 times higher in the experimental group. The only other dramatic claim, that results were 25 times higher in the experimental groups, appears to have been falsely attributed to Lozanov. Neither finding is corroborated anywhere else in Lozanov's research or by other sources. Yet claims of a similarly dramatic nature have appeared in the popular press and in the advertising of some commercial language courses (see Gassner-Roberts 1987, Schiffler 1987).

Having examined Lozanov's research in detail, it can be said with certainty that there is no support whatsoever for claims that learning can be improved by 5 to 50 times or that 1000 words can be learnt daily. There is some indication that achievement may be improved by about 20%, that large amounts of materials may be given, that retention rates and functional use of materials are high, that materials may be learnt in half the normal time in primary schools and that there may be positive effects on the students' psychological and physiological state. These indications are interesting enough to merit further investigation.

Extensive research has already been carried out in the West following Lozanov's (1978) and Ostrander and Schroeder's (1979) publications. We will now examine these studies in detail. In the light of the limitations of Lozanov's research resulting from unsatisfactory research procedures and poorly reported data, an effort will be made to describe studies in as much detail as possible.