A General Approach for Using Data in the Comparative Analyses of Learning Outcomes

This paper brings together some key concepts and proposes a conceptual framework for understanding and evaluating determinants of learning outcomes. The framework facilitates the process of theorizing and hypothesizing on the relationships and processes concerning lifelong learning as well as their antecedents and consequences. The specific aims are to outline a general approach for evaluating learning outcomes, identify data gaps within this approach, and discuss some of the implications for research and development. Along with more specific frameworks, this general approach can be used to guide policy relevant research in education. In particular, it can be used to inform efforts aimed at establishing a coherent information system on the impact of lifelong learning on both economy and society.


Introduction
It is increasingly acknowledged that people are the most valuable resource of any nation, and that a nation's well-being is dependent on the knowledge, skills, and competencies embodied in its citizens (OECD 2001).A recent publication by Human Resource Development Canada (2002) entitled Knowledge matters: Skills and learning for Canadians highlights that people are the greatest resource of a nation in today's global economy and that their skills, talents, knowledge and creativity are the key to prosperity.Consequently, there is a growing interest in comparatively assessing the knowledge and skills of adult populations.In recent years, much effort has been focused on the development and implementation of large-scale surveys that attempt to directly measure the stock of knowledge and skills in adult populations.These efforts aim to provide reliable information that can be used in the comparative evaluation of human resource development policies, as well as in determining the need for new educational and lifelong learning policies.
Knowledge and skills can be seen as the intermediate outcomes of learning that ultimately lead to increased economic, social, and personal well-being.Adopting this view, researchers and policy-makers are placing an emphasis on the need to identify and prioritize the skills that are critical to well-being.In order to effectively evaluate and reform educational and lifelong learning policies, policy analysts must also identify the determinants of skill development and skill maintenance.Large-scale comparative and representative surveys are important tools for obtaining the data needed to address these broad policy issues.
At least two reasons can be put forth to justify active education and lifelong learning policies.First, the intangible nature of knowledge and skills significantly contributes to informational asymmetries in the marketplace, which implies a host of inefficiencies.For 3 example, informational asymmetries can lead to misallocations of resources, and more particularly to under-investment.Poor matches between employees and employers can arise from a lack of information concerning individuals' knowledge and skills.Moreover, the ones who need to invest the most may not be given the opportunity, or realize that they need to invest, because they lack the necessary information to make those decisions.Interventions can correct these market failures by providing governments, employers, and individuals with the information they need to make good decisions about how to allocate scarce resources.Second, the external benefits or spill-over effects created when individuals acquire knowledge and skills affect collective well-being (Romer 1990;McMahon 1998).What one individual, community, or nation does affects another.Unequal access to opportunities for knowledge and skills acquisition can lead to social exclusion, which has negative implications for social cohesion (Jensen 1998;Ritzen, Easterly, and Woolcock 2000).
Accordingly, this chapter proposes a conceptual framework for understanding and evaluating the determinants of learning outcomes at a national level.More specific frameworks that build on this approach can be used to guide research and development programs aimed at establishing information systems about learning outcomes.The objectives of the chapter are to outline a general approach to evaluating learning outcomes, identify data gaps within this approach, and discuss some of the implications for research and development.

Background
Human capital theory (Schultz 1961) has played a key role in promoting the idea that the knowledge, skills, and other attributes embodied in individuals are crucial factors in economic and social development.This idea coincided with an unprecedented growth of formal education 4 enrolments in the industrialized countries.This was due not only to a general increase in wealth and an appetite for education but to the belief that schooling, by transferring relevant knowledge and skills to individuals and increasing adults' readiness to learn throughout the life span, is a key determinant of economic productivity and quality of life.Learning beyond initial schooling is also widely believed to be an important factor in developing and maintaining people's knowledge and skills (Tuijnman and Van der Kamp 1992).Consequently, an increasing proportion of resources is being devoted in the industrialized nations to encourage learning across the life span.This has created a demand for research and data that can better guide the allocation of these resources.
Supplying reliable and sufficient information about learning is not an easy task, however, for two reasons.It requires a comprehensive model and data that truly measure what is posited.
Too often, studies focus only on initial schooling or job-related training, ignoring learning across the life span and failing to consider the complementarity and substitutability between different forms of learning.There is also too much emphasis in these studies on the market outcomes of learning.Although research on the non-market outcomes is progressing (see Wolfe and Zuvekas, 1997), a more comprehensive measurement and valuation of the learning outcomes are needed (McMahon 1997).
Second, much of the current evidence about the effects of learning on productivity and well-being rests on assumptions that are untestable because the necessary data are lacking (Stern and Tuijnman 1997).Often, the observed measures used as indicators of latent constructs ambiguously represent the true variables in question.As a consequence, erroneous conclusions are drawn.The collection of improved data would provide more precise evidence.This is not to say that research efforts thus far have not been fruitful.Research conducted 5 within the framework of human capital theory (e.g., Mincer 1962Mincer , 1974;;Becker 1962Becker , 1993) has provided very useful information.For example, Psacharopoulos (1973Psacharopoulos ( , 1985) ) provides estimates of private and social returns that not only justify continued high levels of investment in education by both individuals and governments, but also provide guidance to make resource allocation decisions.Using data from the 1960s and 1970s, he finds that there is a declining rate of return to education level, suggesting that investment in primary education is more profitable than investment in secondary or tertiary.Furthermore, studies evaluating the returns to job-related training support the idea that it is advantageous for firms to train their employees (e.g., Cohn and Addison 1998).Even more telling are studies about age-earnings profiles that suggest a continued accumulation of human capital leads to higher economic returns (e.g., Mincer 1974).
Although the information derived from these types of studies meets some of the demand, validity remains an issue because the conclusions are easily contested.For instance, because of a lack of directly observed data, the assumptions guiding research exploring the effects of education on productivity are called into question.In the absence of better data, therefore, it is unlikely that an accurate cost-benefit analysis of education can ever be conducted.
Furthermore, as long as the personal quality of life and social well-being that results from acquiring knowledge and skills are not exchanged or valued in the market place, it will be challenging to achieve a proper cost-benefit analysis.Nevertheless, instruments specifically designed to assess various types of well-being can be developed.Such information would allow for a more comprehensive investigation of the relationships between various learning and quality of life indicators.These types of investigations, however, undoubtedly require an interdisciplinary approach.
To evaluate initiatives that aim to develop and sustain knowledge and skills, it is This is the accepted version of the article: Desjardins, R., & Tuijnman, A.C. (2005).A general approach for using data in the comparative analyses of learning outcomes, Interchange, Volume 36, Number 4, pp.349-370.https://doi.org/10.1007/s10780-005-8164-46 necessary to have access to direct measures of relevant knowledge and skills, as well as a full appreciation of all possible outcomes of learning, and not merely the economic ones.There are at least two reasons for this.First, economic, social, and personal well-being may be independently valued; and second, economic well-being may be a function of social or personal well-being.The former suggests that there may be a trade-off between different types of well-being, and the latter suggests that there may be complementarity.Without this broader appreciation, it is impossible to assess the relationships that may exist between different skills and different forms of wellbeing.Moreover, it is difficult to identify the necessary skills, or the criteria to judge the initiative's success or failure.
For these reasons, approaches to the definition of human capital are changing.It is now recognized that human capital refers not only to individuals' knowledge and skills that lead to increased economic well-being, but also the skills that lead to personal and social well-being (OECD 2001, 18).Individuals invest in themselves to achieve more than future economic wellbeing.The time and resources devoted to acquiring knowledge and skills that lead to other benefits should also be treated as investments.It is challenging to deal with investment concepts beyond those that provide tangible monetary benefits but the analysis of such concepts remain, in the spirit of the study of human behaviour and the allocation of scarce resources (Becker 1965).
While economists have understood that personal and social well-being influence economic wellbeing, both directly and indirectly, it seems that they have been reluctant to explicitly model this relationship for fear that it cannot be cogently assessed.But sociologists and psychologists are measuring some of these concepts using latent trait modelling methodologies (Rost and Langeheine 1997).In recent years, economists have begun to draw on psychometric assessments of cognitive skills in order to estimate labour market outcomes (e.g., Murnane, Willet, and Levy This is the accepted version of the article: Desjardins, R., & Tuijnman, A.C. (2005).A general approach for using data in the comparative analyses of learning outcomes, Interchange, Volume 36, Number 4, pp.349-370.https://doi.org/10.1007/s10780-005-8164-47 1995).These studies, however, tend to adhere to the mainstay of conventional economics by focusing only on those skills that are valued in the marketplace.

The general approach
A general approach to analyzing learning outcomes across the life span is presented in this section.This approach extends beyond conventional human capital theory by explicitly taking into account learning outcomes that are not only economic but also social, psychological, and physical.As such, learning outcomes are treated as multi-dimensional, which allows for an explicit investigation of the direct and indirect effects of learning on various outcomes.
Moreover, learning itself is treated as multi-dimensional.Different forms of learning that span the life-wide spectrum of lifelong learning (Colletta 1996;Longworth and Davies 1996;OECD 1996) are separated to depict how they can potentially relate to each other, and in turn what their potential contributions are to different learning outcomes.
The general approach draws an analogy to the resource conversion process developed by Coleman (1971).Coleman based his approach on a convention commonly used by economists, namely an input-output analysis or activity analysis.Figure 1 presents a model of the resource conversion process.Financial, social, and human resources available at time t0 are converted into other financial, social, and human resources at time t1.Depending on how the conversion process plays out with time, the weight of each resource component can remain the same, grow, or be reduced.According to Coleman, the education system is at the heart of the resource conversion process.Education serves as a means for reproducing wealth, social status, and knowledge and skills.

INSERT FIGURE 2. DETERMINANTS AND OUTCOMES OF LEARNING
Processes of learning consist, at least in part, of converting sensory information into knowledge (Mayer 1996).Reflecting and reasoning, as well as an array of other functions, are also believed to play an important role in learning processes.The focus of this entry, however, is to account for the role of contexts in constructing knowledge (Collins, Greeno, and Resnick 1996), since the nature and quality of learning is profoundly influenced by the context in which it occurs (Brown, Collins, and Duguid 1989).Here, the term context refers broadly to the setting in which individuals experience sensory information ranging from the characteristics of the physical environment to the way information is taught or presented.Revealing the linkages between different learning contexts, providers, and sectors is key.Learning occurs in a variety of contexts and as such it is important to investigate how different contexts interact with each other, and the extent to which they contribute to the acquisition of knowledge.This includes investigations on transitions and pathways between initial schooling, work, and further learning, as well as the complementarity or substitutability among them.
When one attempts to study the effect of learning contexts on learning processes and their subsequent outcomes, however, one is faced with the limitation that the mental actions responsible for constructing knowledge and skills are not directly observable.Nevertheless, psychologists and other researchers have been able to make inferences about these unobservable Educational research often focuses on intermediate outcomes such as performance on a standardized mathematics test (e.g., Mullis et al. 1997).These tests seek to assess an individual's knowledge and skills, and are then used as tools to evaluate the effectiveness of educational contexts.It is generally assumed that the assessed knowledge and skills were acquired as a result of experiencing a specific educational context (i.e., the classroom).By drawing inferences about the effectiveness of learning in a single context, however, the impact of other learning contexts that may have played an instrumental role in the tested outcomes, are ignored.
Learning is the result of complex processes that are unlikely to be captured through an assessment of a single context, such as the classroom.For example, informal contexts such as work, home, and the community at large can play an instrumental role in learning processes by enabling individuals to conceive new bits of knowledge, as well as by prompting them to reflect on previously acquired knowledge-therefore leading the individual to form larger and more complex pieces of knowledge.
By considering the complex relationship between learning processes and learning contexts, the relevance of assessing learning outcomes vis-à-vis multiple learning contexts becomes evident.Research designs should take this into account since ignoring the effects of 11 other contexts may lead to erroneous inferences.Which learning contexts lead to an optimal combination of knowledge and skills?The linkages among different learning contexts have a host of implications for the design of learning systems and are an important area of research.

Lifelong and life-wide learning contexts
Lifelong learning is best understood as a process of individual learning and development across the life span, from learning in early childhood to learning in retirement.It is an inclusive concept that refers not only to education in formal settings, such as schools, universities, and adult education institutions, but also to life-wide learning in informal settings, at home, at work, and in the wider community (OECD 1996).
Some believe that lifelong learning only concerns adult education, defined as organized educational processes whereby people engage in systematic and sustained educational activities.
Others also consider non-formal, self-directed, and experiential forms of adult learning as expressions of lifelong learning.For yet another group, lifelong learning is not confined to adults, but includes the full range of learning extending over the individual's entire life course.
This chapter endorses the latter view.It supports the notion that the settings for learning are both "life long" (cradle to grave) and "life wide" (formal, non-formal, and informal learning), and embrace social and individual development of all kinds.
Lifelong learning implies that learning takes place throughout life, that it is not confined to any specific age group or educational institution.The concept thus refers to all systematically organized learning activities associated with formal education, as well as to learning that takes place in informal or non-formal settings (Coombs and Ahmed 1974;Colletta 1996).Whereas formal education refers to any organized and systematic education provided by schools and other educational institutions, non-formal education may be defined as any organized and systematic The information that one is exposed to in a formal situation and learned as a piece of knowledge may be used to learn or form a relation with another piece of knowledge that may be prompted in another context such as an informal one.Hence there is substitution and complementarity in life-wide learning.But many studies only focus on education as defined by initial schooling.An advantage of examining learning activities that span all three forms of learning-formal, non-formal, and informal (Colleta 1996)-is to explicitly consider the interrelationship between the multiple dimensions of learning as well as their impact on outcomes.Explicitly modelling the multiple dimensions of learning will allow for their 14 substitutability and complementarity to be assessed.
The variety of contexts in which people can learn is virtually infinite.What research questions should be exploring is to identify the ones that lead to an optimal combination of final outcomes in a viable manner.The overarching research question is to ask how different learning contexts affect different learning outcomes.To study this question properly, a longitudinal data set is ideally needed in order to measure all three life-wide learning dimensions over time (Tuijnman 1989).Figure 4 illustrates how such a design could look.

Multiple benefits of learning
Final learning outcomes refer to the effect that intermediate learning outcomes such as knowledge and skills have on individual well-being, as well as their broader impact on social well-being.Assuming the outcomes are positive, the numerous individual and collective benefits that potentially arise can be classified into monetary or non-monetary benefits, and can accrue to individuals, society, or both.McMahon (1998) classifies them into monetary private benefits, monetary social benefits, non-monetary private benefits, and non-monetary social benefits.
Figure 5 depicts the various benefits that result from the acquisition of knowledge and skills from various learning contexts.

INSERT FIGURE 5. BENEFITS OF LEARNING
Various social science theories capture the potentially positive effects of learning and, more particularly, the effects of acquiring knowledge and skills.As noted earlier, human capital theory is the most dominant and well-established theory (Schultz 1961;Becker 1962) and thousands of studies have attempted to assess the impact of learning on economic outcomes, such as productivity, earnings, and economic growth.Most studies focus on the impact of 15 schooling (Sweetland 1996, 341) or job-related training on earnings.But the imperfect correlation between education and skills (OECD and HRDC 1997) is evidence of the gap between the knowledge and skills acquired from formal schooling, and the knowledge and skills embodied in individuals.For this reason, further research is needed to understand the nature and extent of the substitutes and complements to the more traditional schooling context.
Furthermore, most previous studies have only attempted to indirectly assess the impact of knowledge and skills on various benefits by using the learning context, such as years of schooling or educational attainment, as a proxy for acquired knowledge and skills.With regard to labour market outcomes, these studies are perhaps more pertinent to the theory of screening (Arrow, 1973;Spence, 1973), whereby schooling credentials act as validations of knowledge and skills, and supply critical information to labour markets.Furthermore, this approach not only ignores learning processes, but it also ignores the multiple dimensions of knowledge and skills.It treats knowledge and skills as one set, and so does not explicitly take into account its different types.Presumably this is because insufficient resources are allocated to the collection of better data and to the development and application of more complex research designs.Consequently, most studies are conducted at a high level of aggregation and offer little insight for ways to improve learning systems.
Recently, however, psychometric methods used for assessing knowledge and skills have been combined with household survey methods to collect and compile improved data.Examples of such efforts are the Literacy Skills Used in Daily Activities Survey (1989), the National Adult Literacy Survey (1992), the International Adult Literacy Survey (1994)(1995)(1996)(1997)(1998), and the forthcoming Adult Literacy and Life-skills (ALL) Survey ( 2004).These surveys provide direct measures of knowledge and skills that can be used to improve the analysis of learning outcomes.This is the accepted version of the article: Desjardins, R., & Tuijnman, A.C. (2005).A general approach for using data in the comparative analyses of learning outcomes, Interchange, Volume 36, Number 4, pp.349-370.https://doi.org/10.1007/s10780-005-8164-4 16 The direct measures of skill, however, are not comprehensive.Thus far, assessments of adult skills have focused on literacy skills, including prose, document, and quantitative literacy.But in the ALL survey, additional direct measures such as analytical reasoning will be made available.
In the same survey, there will also be modules assessing teamwork skills and information and communication technology (ICT) skills.
These surveys also collect an array of background variables that can be used to describe different learning contexts as well as different final learning outcomes, such as economic and social outcomes.In particular, the ALL survey includes a module measuring aspects of social capital (Coleman 1988(Coleman , 1990) that will allow for an investigation of the relation between social capital and learning on the one hand, and social capital and economic outcomes on the other (Schuller and Field 1998;Schuller 2000).Such improved data allows for richer analyses of both intermediate and final learning outcomes.
Despite a disproportionately high number of studies dealing with the economic benefits of learning, few have investigated how these benefits can arise as a result of other learning benefits.One notable exception is McMahon (1999) who isolates the indirect effects of education on economic outcomes via other outcomes such as the non-monetary social and private benefits of education.This approach follows from the groundbreaking work of Lucas (1988) and Romer (1990), which began to explicitly model the externalities associated with the creation of knowledge and skills.The mechanisms responsible for generating the economic benefits of learning can be better understood by explicitly modelling these indirect effects.
McMahon's (1999) findings suggest that overall, the indirect effects of education on growth are about 40 percent of the total effect.Such findings indicate that the mechanisms behind these indirect effects need further investigation.17 Beyond the economic benefits of learning, there are other benefits of learning that may be independently valued, such as social, personal, and psychological well-being.These also include inequality, political stability and democratization, health and population growth, the environment, and crime, just to name a few.See McMahon (1997) for a comprehensive list of the wider benefits of learning.Because learning may have indirect benefits on economic outcomes via other outcomes, and other outcomes may be independently valued, this area of research has important implications for social welfare functions.
As the wider benefits of learning are becoming better understood and more appreciated, there is growing interest in assessing these types of learning outcomes.The following key research questions are inherent in these assessments: What knowledge and skill combination leads to an optimal combination of benefits?Which learning contexts are forming the knowledge and skills needed to achieve certain benefits?
Studies that consider multiple dimensions of learning, multiple dimensions of knowledge and skills, and multiple dimensions of learning outcomes, can be illuminating in many ways.By using desirable outcomes as a reference point, different knowledge and skills can be identified and prioritized according to their effectiveness in achieving these outcomes.In turn, using prioritized knowledge and skills as a reference point, the effectiveness of different learning contexts in arriving at those intermediate learning outcomes can be assessed.

Dynamic effects of learning: Converting outcomes into resources
A general approach to analyzing learning outcomes is summarized in 18 presented to acknowledge that other factors may interact with and potentially affect learning processes.The focus, however, is on immediate learning environments, such as the classroom or any other situation in which learning takes place, and this is referred to as learning contexts.
These include the location, method, and other relevant descriptors of the particular environment that plays a role in initiating learning processes, either by exposing new information or by prompting reflection.
Learning contexts then interact with an individual's initial state, which consists of one's prior knowledge and skill structure, ability (working memory and processing speed), affective state (attitudes, readiness to learn, motivation) (Shute 1994), as well as one's learning strategies and styles (Pask 1976).Both the learning context and the initial state combine to affect various learning processes that lead to certain intermediate and final outcomes (Shute 1994).
To add to the implicit flow of the table, the vertical columns entitled Resources, Information and Decision Making/Choices are included.By keeping with the resource conversion theme, individuals use all the available Resources at their disposal to generate resources in the future, and as such they are primary inputs into the conversion process.
Furthermore, assuming that learning processes are at the centre of the conversion process, these resources are used to define learning contexts, which ultimately define the nature of the  1992).Moreover, the motivation and readiness to learn, which are potential intermediate learning outcomes, are also important determinants of subsequent learning.This is also implicit in Table 1, where all the listed outcomes can potentially be converted into resources that can create and define future learning contexts and initial learning states, and that can lead to the creation of further resources via learning processes.

Conclusion: Implications for the research agenda
A number of implications arise for a strategic research agenda on knowledge and skills.
First, it is clear that an interdisciplinary approach will be needed, one that not only draws upon conceptual and theoretical frameworks from a range of social behavioural sciences but also relies These desirable skills are possible criteria to evaluate the effectiveness of educational systems and other lifelong learning programs.Periodically taking stock of these skills and the wide range of possible determinants enables rich analyses that can help to identify and prioritize the determinants that are policy relevant.Depending on policy objectives, these determinants can improve well-being through the implementation of focused and well-informed policies.
Moreover, these analyses can be used to improve the design of future assessments, and to continually improve the quality and relevance of the information sought.It is likely, however, that to properly assess the determinants, a long-term research and development agenda needs to be introduced.

Biography
This is the accepted version of the article: Desjardins, R., & Tuijnman, A.C. (2005).A general approach for using data in the comparative analyses of learning outcomes, This is the accepted version of the article:Desjardins, R.,& Tuijnman, A.C. (2005).A general approach for using data in the comparative analyses of learning outcomes, Interchange, Volume 36, Number 4, pp.349-370.https://doi.org/10.1007/s10780-005-8164-48 All too often, however, educational processes are conceived as "black box" interactions without considering their implications for learning contexts or learning outcomes.This may be for good reason since the issue of how people learn has been debated for centuries.But an extensive body of cognitive science research seldom referred to in economic and sociological analyses of outcomes offers insights into learning processes, as well as into how they relate to learning contexts and learning outcomes (Berliner and Calfee 1996; De Corte and Weinert 1996; Figure 2 depicts a simplified causal sequence of the terms introduced.An array of factors is hypothesized to have an influence on various learning processes, which in turn bring about changes in an individual.To reiterate, these changes first arise by affecting an individual's state (intermediate learning outcomes); and secondly, through how these changes will affect individual well-being as well as social well-being (final learning outcomes).
Learning outcomes-particularly knowledge and skills-develop as a result of learners encountering multiple contexts, and not just a single context such as schooling or job-related training.From a constructivist point of view, knowledge is created in a fluid manner where one bit of knowledge builds on another to form larger and more complex relations.As such, learning outcomes, learning processes, and learning contexts are inextricably linked.By formally defining a process to be a series of actions or functions that bring about changes or results, it follows that learning processes are a series of mental actions responsible for bringing about changes or results to an individual.These changes are often referred to as learning outcomes.For this entry, it is convenient to draw a distinction between two types of learning outcomes, intermediate and final outcomes.The former refers to the changes that occur in an individual, such as knowledge and skills or motivational behaviour.Final outcomes extend beyond this and refer to how the changes affect an individual's well-being as well as the broader social well-being.Researchers (such as cognitive psychologists) who investigate learning processes are generally preoccupied with intermediate learning outcomes.More recently, economists and sociologists have expressed an increasing interest in how intermediate outcomes relate to final outcomes.This is the accepted version of the article: Desjardins, R., & Tuijnman, A.C. (2005).A general approach for using data in the comparative analyses of learning outcomes, Interchange, Volume 36, Number 4, pp.349-370.https://doi.org/10.1007/s10780-005-8164-4 9 Bruner 1983;Gardner 1987)ion of the article:Desjardins, R.,& Tuijnman, A.C. (2005).A general approach for using data in the comparative analyses of learning outcomes, Interchange, Volume 36, Number 4, pp.349-370.https://doi.org/10.1007/s10780-005-8164-4 10 mental actions (e.g.,Bruner 1983;Gardner 1987).In general, attempting to account for learning outcomes as the result of an individual having been prompted or stimulated by a controlled context can provide useful insights.But in educational assessments, learning outcomes are often measured in order to draw inferences about the effectiveness of learning contexts, rather than about learning processes.In the latter type of research, it is more difficult to control the context in which individuals learn (De Corte 1999).For this reason, it is important to consider a broader range of learning activities when accounting for learning outcomes.
This is the accepted version of the article: Desjardins, R., & Tuijnman, A.C. (2005).A general approach for using data in the comparative analyses of learning outcomes, Interchange, Volume 36, Number 4, pp.349-370.https://doi.org/10.1007/s10780-005-8164-4 This is the accepted version of the article:Desjardins, R.,& Tuijnman, A.C. (2005).A general approach for using data in the comparative analyses of learning outcomes, Interchange, Volume 36, Number 4, pp.349-370.https://doi.org/10.1007/s10780-005-8164-4 12 educational activity that is carried on outside the formal system.Informal learning refers to a lifelong process of experiential and open learning-a process of informally acquiring certain values, attitudes, skills, and knowledge from experience, from learning resources available in the environment, and through independent, self-directed learning (Livingstone 2000).Lifelong learning is an overarching concept with many supporters.Popular education, formal schooling, adult education, self-directed learning, continuing vocational training, on-thejob training, informal learning in the work place, and social education for senior citizens, are examples of more specific elements (Sutton 1996).Lifelong learning thus embraces all learning that takes place from infancy throughout adult life, in families, schools, vocational training This is the accepted version of the article: Desjardins, R., & Tuijnman, A.C.(2005).A general approach for using data in the comparative analyses of learning outcomes, Interchange, institutions, universities, the work place, and at large in the community.Its merit lies in the challenge it brings to using institutional and age criteria as delimiting factors in educational policy.But in so doing it creates another problem.Because lifelong learning denotes a philosophy and an ideal based on humanistic principles, it has so far evaded precise definition.Conceptually, lifelong learning activities can be classified in various ways.A useful way of thinking about it is in terms of two dimensions.The first is the lifelong dimension.Knowledge depreciates with the passing of time and so it is necessary for individuals to update their skills and competencies.The lifelong perspective invites us to consider the whole range of human chronological age.In contrast, the life-wide dimension presents more difficulties because it sectors are to be monitored over time, or if the attainment of specific policy goals is to be assessed.For various reasons, an undertaking such as this lies outside the scope of what may currently be achieved.It is therefore inevitable that the picture is simplified, and that choices are made about which features of lifelong learning to highlight.By necessity, these choices are based on political grounds-given that decisions on which indicators to use, how to provide data for them, and how to interpret them are never neutral-as well as based on knowledge about what can be measured validly and reliably given the current state of measurement technology in the social sciences and the availability of funding.

Table 1 .
There are three major elements presented in the table-learning contexts, learning processes, and learning outcomes.Implicit in Table 1 is a causal sequence that can be read from left to right.Individual contexts such as background, culture, and other descriptors of one's general situation are This is the accepted version of the article: Desjardins, R., & Tuijnman, A.C. (2005).A general approach for using data in the comparative analyses of learning outcomes, Interchange, Volume 36, Number 4, pp.349-370.https://doi.org/10.1007/s10780-005-8164-4

Table 1 . Learning contexts, processes, and outcomes
Information fed into learning processes.Initially, the learning outcomes are intermediate, whereby individuals have to then Make Decisions/Choices on how to use these intermediate resources to generate final outcomes.This is the accepted version of the article: Desjardins, R., & Tuijnman, A.C. (2005).A general approach for using data in the comparative analyses of learning outcomes, Interchange, Volume 36, Number 4, pp.349-370.https://doi.org/10.1007/s10780-005-8164-419 This is the accepted version of the article: Desjardins, R., & Tuijnman, A.C. (2005).A general approach for using data in the comparative analyses of learning outcomes, Interchange, Volume 36, Number 4, pp.349-370.https://doi.org/10.1007/s10780-005-8164-421 Despite the implicit causal sequence, Table 1 should not be interpreted as suggesting that relationships among variables are static.Over time, variables are reciprocally determined, and this introduces a dynamic aspect into the scheme.
on the use of various statistical tools and methodologies, including psychometrics and econometrics.Second, a longitudinal research design is required, one that allows for the modelling of accumulated learning.Third, because all three life-wide dimensions are relevant, future studies should include measures of formal schooling, non-formal education and training This is the accepted version of the article: Desjardins, R., & Tuijnman, A.C. (2005).A general approach for using data in the comparative analyses of learning outcomes, Interchange, Volume 36, Number 4, pp.349-370.https://doi.org/10.1007/s10780-005-8164-422 and informal learning.Finally, direct measures of multiple dimensions of knowledge and skills are needed.Research in life-span developmental and cognitive psychology identifies life skills that are associated with the ability to solve everyday problems (e.g., Reynolds and Bezruczko 1989; Willis and Schaie 1986).Examples of these skills are literacy and numeracy skills, problem solving skills, analytical reasoning skills, communication skills, interpersonal (or team work) skills, intra-personal skills (or "emotional intelligence"), and information and communication technology skills.These skills are believed to have a positive and critical role in the creation of well-being in advanced economies.Their relative importance in predicting well-being and quality of life, however, is likely to vary across people and, more importantly, across different environments.