Dropout in adult education as a phenomenon of fit – an integrative model proposal for the genesis of dropout in adult education based on dropout experiences

In adult training practice, dropout marks the transition from participation to non-participation. There are only a few theoretical models, especially from the second half of the 20 th century, that address this phenomenon. With special consideration to the congruence model of Boshier (1973) and the integration model of Tinto (1975, 1993) the study focuses first on the theoretical discussion of empirically grounded models for the explanation of dropout in the field of adult education. Against the background of analyses of 40 problem-centered interviews with dropouts from adult education, the two models are examined as to their explanatory contributions. Based on these empirical and theoretical explorations, a newly developed typology of as well as a model for dropout are proposed which topicalize dropout in adult education as a phenomenon of fit.


Introduction
The phenomenon of premature termination of participation in adult education 1 , also known as dropout, is well-known in adult education practice. In this phenomenon, individuals discontinue their participation in continuing education before the regular end of the education program, even if they were registered for the education program and participated in it up to a certain point. This phenomenon is known from all areas of training and occurs regardless of the length of the training (Hoffmann, Thalhammer, von Hippel & Schmidt-Hertha, 2020, 2021. Today the field of international research on dropout is characterized by a broad scope of analytical perspectives which include diverse groups of participants or research settings as well as special course formats (e.g. Bariso, 2008;Park & Choi, 2009). A lot of research has been done on higher education dropouts, including some work on the retention of 'older' students (usually defined as older than 25), as is shown by the summary given by McGivney (2004). The qualitative study by Jacot, Frenay and Cazan (2010) points to the relevance of competing areas of life, in particular work or family commitments, which may be responsible for dropout. The study emphasizes the importance of acknowledging the relevance of what is being learned for mature students. An earlier quantitative survey by Shank and McCracken (1993) on dropouts in adult vocational training additionally underlines issues of financial support as well as social and academic integration while participating in training. Furthermore, processes in learning events are being investigated, e.g., by Ten Dam (1995), who relates the participants' individual self-perception to the respective external ascriptions by the teacher. A bundle of studies can also be found on attrition in adult literacy programs or adult basic education (e.g., Kambouri & Francis, 1994;Perin & Greenberg, 1994;Ziegler, Bain, Bell, McCallum & Brian, 2006). In this area of education, dropping out can mean quite different things for different learners. As many programs last longer than a year, the likelihood of significant changes occurring in other areas of life during that time span is rather high and personal learning goals may have been reached long before a course ends (Pickard, 2013). Similarly, research has been carried out in the field of adult language learning with regard to attrition and to the retention of participants (e.g., Evans & Tragant, 2020). A challenge for future research is to look not only at highly intensive educational programs lasting several years, but to also take into consideration that the majority of adult education programs is not that long-termed and is not necessarily of the same biographical relevance as a university study program. What is therefore also needed is an integrative theoretical model for analyzing drop-out in different adult education contexts. Evans and Tragant (2020) suggest looking not only at motivation as a positive factor allowing to keep learners on track, but also at demotivation as the negative counterpart. Garrison (1988) identifies five variables (intrinsic and extrinsic motivation, intrinsic and extrinsic constraints, skills) as determining factors for either change or stability in the behavior of adult learners. This approach can be linked to the differentiation made by Darkenwald and Gavin (1987), who consider a discrepancy between expectations and experiences to be decisive for dropout. In this, motives for participation or orientations in participation are not only multiple, but also mostly neither conscious, nor rational. This also complicates the empirical research into educational dropouts regarding the recording of dropout reasons.
In addition, research trends can be identified with regard to participation in adult education: characteristics of educational participants are investigated quantitatively, through quantitative surveys on sociological characteristics, or qualitatively, by accessing educational biographies. Against this background Jacot et al. (2010) point out that dropout research and theory building should take greater account to the linking of the individual with the institution (Jacot et al., 2010).
First of all, it becomes clear that there is a multitude of findings that need to be considered in theoretical models for dropout. In this article the guiding question is 'which theoretical models are currently available in dropout research in adult education': First, it will be shown that there currently seem to be only two theoretical models that appear to be relevant to adult education research: This is (1) the integration model by Tinto (1975Tinto ( , 1993 and (2) the congruence model by Boshier (1973). Both models will be presented and their significance for research on dropout will be outlined (chapter 2). Subsequently a current empirical study, based on 40 problem-centered interviews with dropouts from adult education in Germany will then be presented (chapter 3). The explanatory value of the two models for the discourse on adult education will be discussed on the basis of current empirical findings from the study to be presented (Chapters 4.1 and 4.2). In an extended continuation an empirical based dropout typology of non-fit will be presented (chapter 4.3) and, based on this typology, an integrative model for the genesis of dropout in adult education will be proposed (chapter 5). Both empirically developed theoretical proposals were generated using the grounded theory methodology (Strauss & Corbin, 1998) and address dropout experiences of adult education as a phenomenon of fit.

Theoretical background and state of research
Garrison stated as early as 1987 that a theoretical research framework was required for the field of adult education in order to be able to understand the dropout phenomenon. He argued that there were too few attempts to develop a systematic strategy and too many comparative studies that did not take into consideration theory building (Garrison, 1987). To this day, that doesn't seem to have changed. Both then and now in adult education research, only one model is known to have been explicitly developed to explain dropout specifically in the field of adult education: This is the congruence model of Boshier (1973), which emerged about 50 years ago. However, this model has hardly been used since then, critical reflection on the circumstance is lacking, and alternative or complementary explanatory models for the dropout phenomenon has not been developed in adult education research. Against this background, Boshier´s model is first presented in more detail (chapter 2.1).
From today´s perspective, however, empirical and theoretical debates that explicitly deal with the phenomenon of dropout in adult education seem to have become a rarity in adult education research (Hoffmann et al, 2020(Hoffmann et al, , 2021. This is particularly striking when comparing dropout research in adult education with other areas of educational research. For example, in the field of university research 2 , there is at present an increase in scientific analyses of dropout (among others in Germany), as is also revealed by diverse recent government research funding. Of principal interest are questions as to the causes, the extent, and the consequences of dropout in higher education as well as questions as to the effectiveness of existing approaches to and methods of securing academic success. In this specific pedagogical context, the integration model developed by Tinto (1975Tinto ( , 1993 has proven helpful in explaining dropout and has become established. Even if higher education is comparatively close to adult education due to the target group, the question arises whether the gained insights can be transferred to the field of adult education. After introducing the Tinto model (chapter 2.2), this will be examined in more detail in chapter 4, based on actual data analysis.
But first of all, the congruence model by Boshier (1973) and the integration model by Tinto (1975Tinto ( , 1993 will be presented and their significance for research on dropout will be outlined (chapter 2.3).

The congruence model by Boshier
The theoretical model developed by Boshier (1971a;1972b) to explain participation and drop-out in adult education is based on the fundamental assumption that participation in further education is dependent on the interaction and fit between psychological factors (learner) and external factors perceived by the subject (learning environment). Building on his former work on theories of motivational psychology and on models from personality psychology from the 1950s and 1960s (Boshier, 1971b), on the one hand, and on his empirical work on dropouts from New Zealand adult education programs (Boshier, 1971a;1973), on the other, he started developing a theoretical model for dropout among adult learners. Boshier`s congruence model is inspired by the work of Houle (1961), Maslow (1967), and Rogers (1963). His main differentiation is between deficiencyoriented and growth-oriented learners. 'The growth-motivated person is inner-directed, autonomous, open to new experience, willing to be spontaneous, creative, and free from deterministic attitudes' (Boshier, 1973, p. 256), and has satisfied their basic needs in the sense of Maslow's (1967) hierarchical model of needs. Deficiency-oriented people, worry about disappointments, 'are afraid of the environment' (Boshier, 1973, p. 256), and feel driven by outer circumstances. This is what Boshier also calls intra-self incongruence with reference to Rogers' (1963) terminology. He presupposes 'self/institution incongruence' as a consequence of intra-self incongruence for participation, whereas dropout evolves from 'intra-self and self/other incongruence' (Boshier, 1973, p. 257). If there is no congruence between individual needs, expectations, or values and the institutional arrangements in the form of the perceived attitudes, values, or convictions of the persons linked to the institution, then this results in dropout. However, his model should not be understood as deterministic or dichotomous. Just like growth motivation/intra-self congruence does not exclude the possibility of dropping out, deficiency motivation/intra-self incongruence does not automatically lead to dropout. Rather, both motivational points of origin have to be seen as extreme ends of a continuous unilinear scale, with individuals usually showing a mix of deficiency and growth motivation (Boshier, 1977). Finally, Boshier, in an experimental study (1972a), tried to show that growth or deficiency motivation can be influenced by the learning environment; in this, however, he was not entirely successful.

The integration model by Tinto
During the 1970s, the US-American sociologist and education theorist Vincent Tinto addressed the phenomenon of dropout. In his theoretical writings "Dropout from Higher Education" (1975) and "Leaving College" (1993) he aims at explaining the processual nature of dropout cases, not a description of dropout. These works build on the initial study by Tinto and Cullen (1973) as well as on the study by Spady (1970Spady ( , 1971. In order to adequately grasp the issue of dropout, Tinto creates a synthesis of the research on dropout in the field of higher education (Tinto & Cullen, 1973;Tinto 1975Tinto , 1993. On this basis, he develops a longitudinal and interactional model for dropout (1975, p. 95; 1993, p. 113), which, from an institution-oriented perspective, focuses on the development of dropout and on influencing factors specific to different phases of this process. His theoretical model 'can be viewed as a longitudinal process of interactions between the individual and the academic and social systems of the college during which a person's experiences in those systems […] continually modify his goal and institutional commitments in ways which lead to persistence and/or to varying forms of dropout' (Tinto, 1975, p. 94). Tinto (1975) assumes that students enter the institution with their very own family backgrounds (e.g., social status attributes), individual attributes (e.g., sex or race) and precollege experiences (e.g., academic and social attainments) and that these, in turn, influence their goal and institutional commitments. With regard to the question of either graduating or dropping out, the highest prominence is given to academic and social integration, which Tinto theoretically grounds on Durkheim's suicide theory (1951) and on his emphasis on the relevance of social and intellectual integration with regard to one's social life and society: the higher the degree of integration in the academic system (expressed, e.g., in academic achievements, grades, intellectual developed, or the individual identification with norms and values of the academic system) or in the social system (expressed, e.g., in relations and interactions with fellow students and faculty members such as lecturers, professors, or university staff), the stronger the commitment to the institution and to the goal of achieving graduation. Ultimately, this commitment will decide whether a student will dropout or finish their course of studies as scheduled. This institutionally-centered perspective is broadened about 20 years later, when Tinto (1993), in his model, gives greater prominence to 'external communities with their own set of values and behavioral requirements' (Tinto, 1993, p. 115), so that college becomes but one of several commitments which have to be balanced. Accordingly, these extrainstitutional commitments, too, determine departure decisions.

Interim conclusion and deduction of the study's guiding question
Looking at these two models it can be shown that adult education research does not have an integrative theoretical model for analyzing drop-out in different contexts of adult education. Both dropout models presented above focus differently on the interaction of individuals and institutions. Tinto's model emphasizes the processual nature of dropout and the special significance of academic and social integration of learners for the context of higher education. Boshier's model puts greater emphasis on the significance of additional factors which, independent of the individual, influence a learner's motivation for participation or their decision to either stay on or drop out.
The only existing theoretical dropout model emerging from and pertaining to the field of adult education is rarely taken up. Since Boshier's model seems to be the only one that directly addresses dropout in adult education, it is somewhat surprising that, so far, it has hardly been referred to in relevant studies. One reason for this lack of attention might be a certain inconsistency in Boshier's psychological point of origin, in that he refers to a two-fold model of personality, referring to Maslow and Rogers on the one hand, and, on the other, he advocates taking into account the interaction of individual dispositions and environment. It remains unclear, to what extend the environment really can influence idiosyncratic drop-out tendencies and how far drop-out can be clearly predicted by individuals' relation to the learning contents. Thus, the question remains to what extent dropout can be predicted by personality (deficiency motivation) and in how far the environment is of relevance to the further course of attendance. To get one step ahead with that question, a more open and explorative research design is needed that also enables a further development of theoretical concepts. The question arises to what extent the model of Boshier should be taken up again more strongly in dropout research and what weakness and strength it has (Chapter 4.1).
In contrast, the model of Tinto is paid much more attention. Since this model has been developed, numerous studies have built on it and have continued to advance it, partly with regard to specific contexts (most recently Kerby, 2015;Chrysikos, Ahmed & Ward, 2017;Li, 2017;Klein, 2019;Neugebauer, Heublein & Daniel, 2019), and finally, 'such model has been cited in a considerable number of research works […] and has contributed to the proposal of hybrid models that share its goal' (Carmo Nicoletty, 2019, p. 61). Limitations can be seen in 'that the process [of modelling the dropout process] itself is highly dependent of several intertwined factors, such as geographic […], social […], cultural […], temporal and economical […], besides those related with the course […] and with the institution; as a consequence, it is a process hardly prone to generalization' (Carmo Nicoletty, 2019, pp. 62f). The field of higher education is comparatively close to adult education due to the target group. Nevertheless, the question arises whether the gained insights can be transferred to the field of adult education (e.g. with regard to the different framework conditions and the relevance of exams and qualifications).
Against this background and with reference to the results from a recent study on dropout in the field of German adult education, it will, in the following, be examined 'whether and to what extent the models by Tinto and by Boshier contribute to the explanation of dropout from adult education' (chapter 4). The desiderata shown are taken into account in development of a new, empirically founded theoretical typology of nonfit (chapter 4.3) and a integrative model for the genesis of dropout in the field of adult education (chapter 4.4).

Methodology and data analyses
To answer the question 'whether and to what extent the models by Tinto and by Boshier contribute to the explanation of dropout from adult education' (chapter 4.1 and 4.2) empirical data from a current study carried out in Germany are used. The presented study is a qualitative analysis of empirical data to the dropout phenomenon in adult education 3 . The aim of the study is to explore processes of dropout in the field of adult education as a first step that should contribute to refine existing theoretical models or provide impulses for new theoretical concepts. Therefore, we follow the Grounded Theory Approach (Strauss & Corbin, 1998) to generate new insights into the genesis of dropout in the adult education sector and to contribute to a deeper understanding of individual dropout decisions. As a main database for this approach we collected interviews and analyzed this data following the concepts of Strauss and Corbin (1998).

Data collection
The study is based on 40 qualitative problem-centered interviews (Witzel, 2000) with individuals who have dropped out of adult education courses in Germany. In various adult education institutions and open public places (e.g. libraries, supermarkets, hair salon, advisory services) information was posted about the study and people with dropout experiences were searched for voluntary interview participation. Selection of the interview partners was partially carried out according to the theoretical sampling (Strauss & Corbin, 1998), but also -and in deviation from the strict interpretation of theoretical sampling -on the basis of plausible presuppositions (e.g. age, gender, the type of adult education, duration and funding of training) (e.g. Reich-Claassen, 2010; Schmidt, 2011).
In this way, a sample that was as rich in contrast as possible was generated: The age of the participants varied between 24 and 68 years. A total of 29 were female and 11 male. Using the highest school-leaving qualifications, care was taken to ensure that the sample does not only include people with a high level of education: 35 people have an advanced school leaving certificate, three have a secondary school certificate, one person has a lower secondary school certificate and one dropped out of school. The interview participants have experience of dropping out of adult education in very different training courses: There are distinct variances in the duration (between two weeks and several years), but also with regard to the content-related topics, the group sizes and the funding of the training. Despite the great variance, all interview participants have one thing in common: they have had at least one or more significant dropout experience in adult education. In this way, a total of 48 individual dropout experiences could be reconstructed by analyzing this data corpus. The analyses only include dropout experiences (the termination of participation before to the regular end of a course) from adult education programs which suffice the criteria of the definition of continuing education formulated by the German Education Council (1970). The spectrum of adult education comprises vocational and general continuing education programs, the expansion of basic education, as well as political adult education measures.

Data analysis
In analyzing the interviews as individual cases, methodical techniques based on the premises of Grounded Theory (cf. Strauss and Corbin, 1998) were used. For example, each empirically determined abortion experience was systematized and visualized in its own complexity. In this way, among other things, more in-depth approaches to the respondents' motives for action and interpretation schemes were explored, which go well beyond the explicitly stated reasons for dropout.
In addition to the analysis of individual cases cross-case comparative analyses (Miles & Huberman, 1994) were carried out to develop types on the basis of deductively and inductively gained categories. The objectives of our typology are both descriptivestructuring and theoretical-heuristic; the aim is to reveal content-related semantic connections (Kluge, 2000). Empirically grounded typologies are situated between empiricism and theory and are thus an 'intermediate step in theory building' (Kluge, 2000, p. 51). The construction of types has the potential to serve as a basis for the deduction of action-relevant knowledge to be applied in practice (Schmidt-Hertha & Tippelt, 2011, p. 23). Our typology represents an interim step on the way to generalizable statements, e.g., with regard to the development of a model (cf. chapters 4 and 5).
Having developed other concepts of empirically grounded categorizations of ideal types, Kluge (2000) advances a systematic and intersubjectively traceable stage model. Our empirically grounded typology is a multi-dimensional typology, based on combinations of features (Kluge, 2000). It is not a typology of persons but a typology of non-fit among dropouts from adult education. This entails that a person with diverse dropout experiences can be assigned to various types (see also Kluge, 2000).
The development of such typologies requires comparative dimensions which allow for a positioning of the empirical cases. These comparative dimensions were at first inductively elaborated and then through a supplementary deductive procedure related to theoretical models and further differentiated. Thus, for instance, the integration model by Tinto was used because it is considered a validated model with regard to the explanation of dropout from study courses in the context of higher education. Likewise, social and academic integration (according to Tinto, but adapted to adult education programs) were explored as comparative dimensions (among many others). In the course of the case analyses and case comparisons, however, the situational fit/non-fit crystallized as the core category 4 . Our two central comparative categories for the actors and contexts involved in dropout are both empirically significant and theoretically compatible; in their combination, they describe different kinds of non-fit. In a multiple field table (see figure  1 and chapter 4) we have depicted all theoretically possible combinations of characteristic values, which we have then transferred into a proposition of a model for the genesis of dropout in adult education (see figure 2 and chapter 4).

Results
The different models by Tinto and Boshier presented above each focus on the dropout phenomenon, but still, in each case a different perspective is assumed and different facets of dropout are focused upon. Therefore, we will explore which empirical phenomena can be explained either well, less satisfactorily or not at all with the respective model. This allows us to show that the individual models contribute in different ways to the explanation of dropout and to reveal their limits.

The explanatory contribution of Boshier's congruence model
In the strongly psychologically grounded approach Boshier centers on the social relations. The focal point of his observations is thus the maintaining of inner harmony both with oneself and with the environment. With regard to the environment, the institutionally framed relations between learners and teachers (e.g. lecturers or trainers) and between learners and other students (Boshier, 1973, p. 259) are above all taken into consideration.
Ideally, the learner's experience of self matches up with the ideal self-concept (congruence). The degree of 'congruence' or balance between the students' concepts of self and the two most important aspects of the adult education environment (the other students and the lecturer) are associated with dropout behavior (1972b, p. 88). If the equilibrium in the adult education situation becomes unbalanced, the participant will feel 'uncomfortable in a group', 'bored', or 'generally at loggerheads with the situation he found himself in' (Boshier, 1972b, p. 260). It is assumed 'that so-called ‚non-courserelated' reasons for drop-out are often used as a rationalization for dropping-out for ʹcourse-relatedʹ reasons' (Boshier, 1972b, p. 97). Boshier 'suggested that dropout, particularly occurring for course-related reasons, can be understood as a function of the cumulative effects of self/other incongruence developing in the adult education situation, but which initially stems from ʹdeficiencyʹ motivation, an intra-self (self-ideal) incongruence' (Boshier, 1973, p. 167f). In pointing to the causes for the motivations for dropping out, Boshier refers to the psychological basic model, thus linking the reasons for dropping out solely to the individual dispositions of the learner.
It has to be kept in mind that Boshier's theoretical model is based on empirical data that is related exclusively to non-credit liberal adult education, a sector which is open to each and every one who wishes to study. Since we applied a broader concept of adult education in our study, we also considered cases of dropout beyond this limited framework of adult education programs (e.g. obligatory participation in further training).
When referring to examples of adult education which are roughly similar to those examined by Boshier, then the congruence model is useful for explaining cases of dropout. The limitations of Boshier's congruence model with regard to the explanation of dropout become evident once we look at adult education programs featuring final exams.
Due to the different requirements and framework conditions, such offers of adult education are only comparable to a limited extent to the programs investigated by Boshier. Even if the participants show pronounced 'growth motivation', the influence of third bodies on the dropout behavior is not to be underestimated. Thus, the analyses of empirical data carried out in our study reveal the decision-making power of the employer or the supervisor in cases of participation in vocational further education programs, in particular. This can above all be observed when participation in the training course is financed not by the participant themselves and/or when, for example, arrangements have been agreed upon with regard to the possibility to include working hours in the training program. In these cases, the analyses of the interviews show that the congruence model by Boshier is of only limited explanatory value: other persons or third bodies (e.g., employers or the state) which, in addition to the learners, the other participants, and the lecturers, are at least indirectly involved in the process of adult education are not (or hardly) taken into consideration by Boshier.
It can thus be stated that the congruence model by Boshier is of only limited use with regard to the explanation of dropout from adult education if third bodies are involved in the decisions relating to the training and/or if participation is linked to certain obligations (e.g., exams).

The explanatory contribution of Tinto's integration model
In contrast to Boshier, in his approach Tinto tries to describe a social phenomenon as being (multi-)causally determined, contextually embedded, and as processual in nature. Therefore, he focuses mainly on categories of social and academic integration. He presupposes a network of meaning with regard to the relations of the actors found in the academic field, and he adopts a mainly institutional perspective: the individual's behavior is seemingly of interest mainly in relation to the institution 'college' and only within the lifeworld context of the individuals.
In the course of the project, the dimensions of social and academic integration repeatedly served as heuristics for empirical observations. The great importance given to social relations pertaining to the dimension of social integration is also reflected in our material. These relations do not refer only to the limited group of persons linked to the institution (such as counselors, lecturers, other learners), but also to the private and professional environment of the participants, sometimes even to persons who decide on a premature termination without involving the person participating in a program (e.g., employers or an agency). The dimension of academic integration, too, can be found in our material because instances of content-related overload/underload are also an issue in adult education programs. Usually, however, performance expectations on the part of the institution are less important in adult education; often there are no process-accompanying tests or certification-related exams. In those cases, however, in which examination formats are integrated, these may well be a reason for dropping out.
In relation to the context of adult education, however, Tinto's model is of greater informative value with regard to dropout from the university as an institution of the tertiary sector. His claims have to be considered contextual in two ways: (1) They can be regarded as primarily institution-related because he specifically focuses the college context with all its framings, conditions, and dynamics. (2) They can be considered linked to the tertiary educational sector because individuals spend most of their time during training periods (both while teaching and while studying) in these institutions and even outside of these timeframes they are dealing with the related contents. During that time, the foundation is laid for the subsequent professional phase.
On both points, differences can be identified in institutions of adult education compared with school education and initial vocational training institutions: places of adult education are subject to different dynamics, are differently equipped, have different systems of application, enrollment or mentoring, conclude with (or without) diverse certificates, etc. Institutions of adult education differ in many ways from the constitution and the rules of the institution 'college'. Differences cannot only be identified as to the constitution of an institution, but also in the question of how these institutions and the educational offers they create are linked to the lifeworld of the individuals. In this context, comparisons show that participants do not usually attend adult education programs fulltime for a timespan of several years (unless, e.g., they are enrolled in a full-time retraining), which is generally the case with regard to enrollment at a college. Therefore, the factors of social and academic integration in Tinto's model decrease with regard to the field of adult education, because a large part of adult education measures requires participants to spend but a few hours per month in the institution, and other factors outside the institution of adult education gain in importance. Compared to college students, participants in adult education are in a different phase of life; they often have a time-consuming job, are in the family-phase of life and sometimes in a phase of caring for relatives. They find themselves perhaps in the rush hour of life or in their final years of professional activity and thus move in different contexts, of which adult education is but one. Models explaining dropout from adult education thus need to include extra-institutional contexts. Tinto himself came to that conclusion when further developing his model for the college sector and therefore added to his model of 1993 the category 'external environment' with its values and attitudinal conditions. Although external environment is not considered a dominant explanatory force in Tinto's model, it at least considers that students have a life outside of college, despite the fact that they spend most of their time in college. This category -included, but still ranked secondary -proves to be a crucial category for the explanation of dropout from adult education in our research in the context of adult education. It allows to grasp reasons for dropping out that lie outside of the relation between individual and institution -and which may perhaps be located in the individual and their changing lifeworld or which may be triggered entirely by external bodies. This is substantiated repeatedly by our empirical material.
The model thus offers some potential for the transferability of its explanatory force from the college sector to that of adult education, but it is not in all respects suited for explaining dropout in the field of adult education. While Tinto proposes a causalanalytical model for the university sector which fans out commitments relevant to college students' decisions to either drop out of or remain on a course, our typology of fit (see below) offers a hermeneutic-understanding approach to reasons for dropping out from adult education programs, made possible by the interviews and the reconstructions of reasons for dropping out from subjective accounts of experiences.

Empirically grounded expansion proposals: a typology of non-fit
In reference to Tietgens (1992), dropout is regarded as a possible result of search movements: while teachers in adult education, in their planning processes, try to anticipate the needs and expectations of the target groups and to identify as well as arouse interest, the addressees, for their part, search for appropriate programs. Both sides (supply and demand) are thus 'searching' and adult education only happens if there is a concordance of expectations (Tietgens, 1992, p. 67; see also Hippel & Tippelt, 2010). The concept 'fit' used in this context implies the tensions between the individual and the (surrounding) field (such as the field of adult education or, more specifically, an institution of adult education), in which the individual moves temporarily (see next chapter for further elaboration). A dropout thus represents one possible reaction to these tensions.
Against this background and based on the reconstructions of the reasons for dropping out given by the dropouts interviewed for this study, a typology of non-fit has been developed to explain dropout (sc. Figure 1). Altogether seven different types could be worked out in the analysis of the empirical material: 1. The type intra-individual non-fit points to a non-fit within a subject with regard to lasting or temporary/situational forms of dealing with the world and themselves. In those cases, dropout ensues when lasting features, such as, e.g., structure aversion in several sectors of life or constant subordination across the lifespan, also arise in the adult education situation and ultimately make it impossible to continue to attend the training. These stable patterns in dealing with oneself and the world thus conflict with a successful participation and a dropout accordingly provides an opportunity to resolve the non-fit and to temporarily (re-)create a fit within the subject; finally, these (intra-individually dealt with) situations of conflict are resolved. 2. The type life context-related non-fit refers to changed living conditions, such as, e.g., changing career aspirations, which cause an education course to become irrelevant. 3. While these two types still focus exclusively on the individual and their private and professional lifeworld, the type inter-personal non-fit now specifically looks at the adult education sector. In an adult education situation at least two systems of persons encounter one another within a context of adult education as well as other contexts, which, due, e.g., to different expectations, may result in interpersonal non-fit which may manifest itself in heated (open) but also cold (covert) conflicts. This does, in this case, not include conflicts with institutionally affiliated persons such as lecturers, but rather conflicts between, e.g., the individual and other participants or between the individual and their partner. 4. The type individual-institutional non-fit refers to a discrepancy between individuals and institutions, such as the institution of adult education in itself or other institutions such as, e.g., a government body: the individual's expectations, stipulated requirements for the lecturers, institutions, assessments of fit, interests do either not fit from the beginning or cease to fit in the course of the training activity. Dissatisfactions on the part of the individual with the institution are perhaps brought up and discussed, but cannot be resolved, so that a dropout ensues, which is usually initiated by the individual. 5. This discrepancy is also taken up by the type institutional-individual non-fit. The dissatisfaction and the dropout originate from the institution, for instance when the funding institution (e.g., the employment agency/job center) or the institution of adult education are dissatisfied with the participants' absenteeism and therefore deny adult participation. 6. In contrast to the types mentioned so far, the final two types of fit lead to a dropout without the individual themselves being directly involved. The type intrainstitutional non-fit shows that the dropout process takes place on a meso-level and the individual is affected by it on a micro-level, without participating themselves in the process. This type of dropout can occur for instance when an institution registers non-fit within itself, e.g., between the institution and a lecturer tasked with running a course who in the course of the program quits this teaching assignment. In that case, a course can perhaps not be continued which would entail the participants' dropping out. 7. The type inter-institutional non-fit shows that non-fit may also occur between institutional bodies, thus, for instance, between an employer funding the training and the institution or between the employment agency as funding agency and the institution of adult education. In those instances, too, a decision leading to a dropout may be taken without involving the individual, but by which he or she is affected. This clearly manifests itself when, e.g., an employer terminates a current contract with the provider without considering the adult education interests of the person participating.
As already indicated in the description of the types, these seven types can be assigned to two different dimensions of comparison. By interlinking these two comparison dimensions in a crosstab, it becomes clear how the seven types of nonfit are related to one another in their systematic reference system (sc. Figure 1).

Figure 1: Typology of non-fit in cases of dropout from adult education
On the one hand, the fit-related dropout is evoked on different levels: On the institutional just as on the individual level. This results in the first comparative dimension, which we have recorded in the table (Figure 1) in the row category and provided with the following key question: 'In which contexts is dropout evoked?'. On the other hand, in these fit-related cases, different actors are involved in the adult education process in different ways, directly or indirectly. We have mapped these in the second comparative dimension. As can be seen in the illustration in the columns, the main question in the second comparative dimension is 'Which actors are involved?'. Here, we differentiate between involved actors who can be allocated to the lifeworld of the individuals and who are not involved in the adult education process and actors who, due to their institutional affiliation, are directly (e.g., as staff members of the institution of adult education) or indirectly (e.g., third bodies, such as employers) involved in the training. With regard to the individual's lifeworld, categorized phenomena of dropout are also differentiated by looking at whether the overall context of the interview allows to conclude that the reasons for dropping out can be traced back to lasting causes (e.g., the individual's ways of dealing with the world, such as subordination or structure aversion) or whether they were caused by an ever-changing context in which the individual find themselves respectively (e.g., through loss of a funding source). This typology constitutes a useful supplement to the models by Tinto and by Boshier, also because the types inter-institutional non-fit, intra-institutional non-fit, and intraindividual as well as life context-related non-fit cannot be explained by Tinto's model, but are brought to the fore here. The type inter-institutional non-fit provides a new perspective with regard to Boshier's model because here it becomes evident that a dropout may occur although there is a pronounced 'growth motivation' among the participants, which does not prevent a dropout grounded on reasons outside of the individual. The typology presented here differs from Boshier's and Tinto's considerations in that it provides a meaningful broadening of the perspective on dropout, especially in the field of adult education.

A proposal for an integrative model for the genesis of dropout in the field of adult education
Instead of the strongly psycho-therapeutically determined concept of congruence, as proposed by Boshier, and instead of Tinto's concept of integration, which emphasizes the assimilation of the individual in an institutional context, we would like to propose the concept of fit. This concept implies a status description which detaches itself from the normative demand for (unilateral) assimilation and which focuses on the processuality of dropout. Furthermore, the concept of fit as a purely descriptive-analytical category is initially free from attributions of responsibility and open to the dynamics of change inherent in the prerequisites on the part of the learners and their living conditions as well as on the part of the education offers. Fit is not to be understood as congruence, rather, fit may be perceived even under conditions of diversity. We do not assume a phenomenon which can be observed objectively, but rather, we conceive of fit as a relational construct (Emirbayer, 1997) which may be experienced as such subjectively by the individuals or institutions involved. If this fit, which will be further elaborated as a construct in the next sentences based on the empirics of the research project, is deemed insufficient by at least one side (individual or institution) and if it is felt to be impossible to (re-)create a sufficient fit, then it comes to tolerance of this or dropout ensues. Aspects that may not be a fit between these two contexts (individual and institution) and their inherent groups of actors include mutual expectations, motives, desires, needs, contents, interests, requirements and prerequisites for participation and completion of a continuing education program, teaching/learning cultures, imparted contents, and other framework conditions of the offering (such as times, costs, teaching/learning setting, etc.). Based on all these aspects and the evaluation of these by the various groups of actors, fit assessments are ultimately formed on the part of the individuals as well as the institutions. These can have an objective character (e.g. formulated criteria such as access requirements) as well as a subjective character (e.g. non-formulated criteria such as benefit expectations). They are by no means rigid, but are subject to dynamic, situational negotiation processes against the backdrop of the individual's own socialization, prior experiences, evaluation schemes, etc., and are accompanied by cognitive/rational as well as emotional components. Our typology thus also implies an interlacing of perspectives of different actors. The following graphic representation (see Figure 2) illustrates these aspects within a proposed integrative model for the genesis of dropout in the field of adult education. The proposed model for the genesis of dropout in adult education (see Figure 2) expands the models by Tinto and Boshier by including other institutions involved directly or indirectly in the training and by also emphasizing the temporal dimension of the dropout case. The complexity of the model thus generated is also due to its development and its empirical grounding in qualitative data. In contrast to Tinto's rather heuristic model, which, although it includes numerous empirically grounded studies, is based on theoretical-systematic considerations, and to Boshier's concept, which is based on psychological theory models and large quantitative surveys, our model builds on these two approaches and expands them by using a number of empirical data. Nonetheless, the data base of this model is subject to limitations. Thus, we lack longitudinal studies, which would be much better suited to depict the progressions and dynamics of dropout cases. The longitudinal data available, e.g., through the National Educational Panels Study (NEPS) in Germany (Blossfeld, Rossbach & Maurice, 2011), record only a section of the relevant variables and, above all, by no means all cases of dropout from adult education (Hoffmann et al., 2020).
The model presented here allows to map, in the form of a graphic depiction, the different types and their contextual and processual embedding. Causal relations and mechanisms of cause and effect cannot be detected with this model. The proposed typology as well as the accompanying model are located between empiricism and theory; they could represent an intermediate step towards a more elaborate theoretical model. They might stimulate adult empirical studies and, at the same time, they could be of relevance to the practical work of institutions of adult education and of adult education counselling, for the reflection on and both preventive and situational handling of dropout.

Conclusion
The paper was devoted to the question, whether and in how far the models of Tinto und Boshier can be contributed to the explanation of dropout in adult education. In conclusion, the following can be stated: Because Boshier's congruence model is based on empirical data that relate exclusively to liberal adult education without crediting, it cannot be applied to all areas of continuing education. It is particularly suitable for explaining reasons for dropping out, which refer to the individual dispositions of the learner. Its explanatory power is exhausted when third actors are involved in decisions about the training and / or when participation is linked to certain obligations (e.g. examinations). In contrast, the importance of expectations is in Tinto´s integration model very central and without them the model is only partially meaningful. In adult education, performance expectations are also relevant in some cases, but for participants they have a different meaning in life than a degree; in-process tests or certification-related tests are often completely absent. The model is particularly applicable when exam formats are integrated into the training, as these can then be relevant for dropping out. Even if Tinto´s integration model offers potential for the transferability of its explanatory power from the university sector to the field of adult education, it is not entirely suitable for explaining drop-out in adult education.
The empiricism presented here, the typology of dropout in the field of adult education deduced from it and the model for the genesis of dropout in adult education, follow the models by Tinto and Boshier and moreover they expand these models by including constellations of dropout which had not been integrated before. Our typology shares with these models' -although comprehended differently -the focus on processuality, on aspects of fit and on the interaction of individuals and institutions. Both Tinto's integration model and Boshier's congruence model clearly reveal the high significance of other actors involved in the interplay between those participating in adult education. Our typology furthermore emphasizes that dropout may occur without the individual being involved (thus, without the individual experiencing incongruence and without integration issues). Neither Boshier's psychological approach nor Tinto's sociological model can sufficiently explain this constellation of dropouts (of the type inter-institutional non-fit or intra-institutional fit). The same goes for the changes in the relation with the learning topics. While, in Tinto's model, individual interest in and attitudes towards learning topics are at best indirectly conceivable as a facet of the goal commitments (or as a facet of academic integration), Boshier completely suppresses the level of learning topics in favor of the intra-and inter-personal level of relations, thus reducing the triad of I, we, and content, described by Ruth Cohn (Gary, 2010), to its two primary components. The integrative model proposed her for the genesis of dropout in adult education extends the models of Tinto and Boshier, in particular by emphasizing the temporal dimension of dropout.
Although the dropout phenomenon can never be empirically illuminated in all its complexity for a selected sector of education, it is still possible to reveal clear research desiderata. For one thing, we lack longitudinal studies which would capture dropout in the field of adult education systematically and which would allow to test models, like the one designed here, empirically, also with regard to their temporal dynamics. Ultimately, it is only thus that we will be able to examine causal relations, as they are suggested by the qualitative data, within a larger sample. On the other hand, we lack further studies offering a more theory-based analysis of dropout in diverse contexts. The model proposed here provides, at least for the field of adult education, such a theory-based approach. A comparison of the empirical-theoretical models, too, might also be worthwhile for the different educational sectors, in order to thus reveal similarities or peculiarities of dropout cases in the respective areas. A challenge for further studies would be to link, with regard to content, the typology and the model presented here to empirical findings from research on participation in adult education, e.g., to the clusters revealed on the basis of the NEPS data. Furthermore, it would be interesting to carry out more in-depth detailed analyses of the processual nature of dropout, also from the perspective of cumulative non-fit.

Notes
1 We use the term adult education as in the Adult Education Survey (AES) for the participation of adults especially in non-formal education, be it work related or not (Kuwan & Larsson, 2008). 2 Likewise, we have intensive scientific analyses of the dropout phenomenon in school research (cf. Stamm 2012). 3 The project "Analysis of drop-out in further education (distribution, influencing factors, effects): Development of an object-based theory perspective on drop-out" was funded by the German Research Foundation from 10/2017 to 10/2019 under the following funding number: SCHM 2391/6 -1 and HI 1599 / 4-1. 4 A comparison with the state of research made it clear that the generation of types by Stamm (2012) with regard to dropout from school had developed similar comparative categories (actors, contexts), also in an inductive-deductive manner. The difference, however, resides in the types themselves, which, in our case, focus non-fit, whereas Stamm focusses on dropouts from school as a typology of persons.