1 Quantitative Sustainability Assessment, Department of Management Engineering, Technical University of Denmark2 Department of Management Engineering, Technical University of Denmark3 Innovation and Sustainability, Department of Management Engineering, Technical University of Denmark
- addressing issues of validity and usability
This thesis seeks to add to the development of the Social Life Cycle Assessment (SLCA), which can be defined as an assessment method for assessing the social impacts connected to the life cycle of a product, service or system. In such development it is important to realise that the SLCA is only appealing to the extent that it does what it is supposed to do. In this thesis, this goal of SLCA is defined as to support improvements of the social conditions for the stakeholders throughout the life cycle of the assessed product, system or service. This effect should arise through decision makers following the ‘advise’ of the assessment. In order for a positive effect to arise from following a decision, the preferred alternative has to be associated with more favourable social impacts than the other assessed alternatives, indicating that the assessment has to validly represent the consequences of each alternative. But to create an effect, validity is not enough; the SLCA furthermore has to be usable in a decision making context. It has been the aim of this thesis to identify the issues which may hinder the validity and usability of the SLCA and to propose procedures to incorporate in the SLCA alleviating the problems. With regards to the usability of SLCA, a study was conducted addressing 8 Danish companies’ interest and possibility in using SLCA. Here it was shown that the interest in SLCA was limited to external purposes, most notably comparative assertions for marketing purposes. However, it was also shown that the companies’ ability to obtain data throughout their products’ life cycles was very limited, for example because suppliers were unwilling to hand over this information to the companies or because the goods were bought on open markets furnished by a large number of unidentified suppliers. These issues were found to potentially limit the use of SLCA in companies to applications with very limited life cycle perspective. Mitigation of this data availability problem may show to be very difficult for companies, since the only way seem to be to lower the demand for the validity of the data included in the SLCA. If the SLCA is then used for external purposes, the company would run the risk of taking credit for the results given by a potential untrue assessment, which, if being the case and later discovered, may be highly incriminating for the company. It is furthermore discussed that other user groups, such as governments and intergovernmental organisations, may have other demands for SLCA and therefore also other possibilities. The usability of SLCA is only addressed in this study, whereas the three consecutive studies focus on the validity of SLCA. The reason for this overemphasis on validity is that the usability inevitably will be addressed in the development of SLCA, whereas this in not the case with regards to validity. The first of these studies addresses the validity of impact pathways in SLCA, which denotes the cause effect relationship between indicator and the ‘Area of Protection’ (AoP), representing the underlying issue of importance assessed in the SLCA. The study is based on two examples from the existing work on SLCA: One considers whether the type of indicators included in SLCA approaches can validly assess impacts on the one of the two definitions of AoPs in SLCA, being the well-being of the stakeholder, and the other example addresses whether the ‘incidence of child labour’ is a valid indicator to assess impacts on the AoPs. Both examples show a poor validity of the impact pathways. The first example shows that depending on the definition of ‘well-being’ the assessment of impacts on the well-being of the considered stakeholder can not be performed exclusively with the objective indicators which are presently used in SLCA approaches. Objective indicators are indicators designed to measure impacts which can, at least potentially, be measured without the involvement of the impacted stakeholder. If well-being is understood as something pertaining to the experience of the individual, subjective indicators are needed, which are indicators that focus on the experiences or feelings of the impacted stakeholder. The second example shows that the mere fact that a child is working tells little about how this may damage or benefit the AoPs, implying that the normally used indicator; ‘incidence of child labour’ lacks validity in relation to predicting damage or benefit on the AoPs of SLCA. More valid indicators should rather focus on, among others, the health impacts of child labour and its impacts on schooling outcomes. However, even though the indicators proposed in both examples may improve the validity of the assessment, a problem is that in both cases, the indicators demand more detailed data, which may limit their usability. The second study begins by considering that the SLCA as presented here should assess the consequences of a decision. This can be expressed as the difference between how the world is or will be because of the decision and how the world would look like had it not been for this decision. At this point it is important to realise that social impacts on individuals do not only happen in product life cycles, but in all aspects of their life. Thus, if a decision implies that a worker participates in a product life cycle, the worker will, if the decision is not taken, have to do something else, which will equally impose some impacts on him or her. When assessing the consequence of a decision for the worker it is thus this difference between these two situations, the ‘implemented’ vs. the ‘non-implemented’ decision which should be considered. More or less same argument goes for the product user. The study attempts to model the impacts of the ‘non-implemented’ decision in relation to the worker and the product user and finds that when the non-implemented decision situation means that the product is not produced at all, it is often associated with increased levels of unemployment. Literature on unemployment suggests that unemployment causes decreased health levels, increased poverty, family tension and violence and crime, but that the impacts may vary with context of the unemployed. If the non-implemented decision implies that a product user will no longer use a product the non-implemented decision may lead the user to choose another products associated with another life cycle and thereby other social impacts or choose to spend his or her time on something not related to product life cycles, which will equally impose social impacts on the user. The assessment of the impacts of the non-implemented decision is discussed and found to be difficult due to the complexity of identifying what this non-implemented situation amounts to. However, it is argued that some relatively simple assessments may be performed which may still improve the validity of the assessment in comparison to simply ignoring the impacts of the non-implemented decision, however inaccurate they may be. The third and unfinished study addresses the possible influence of the context on the validity of SLCA. Here two examples are analysed. One relating to the context variability of proposed endpoint categories in SLCA where it is shown on the basis of literature that what influences the well-being of the individual (one of the suggested AoPs in SLCA) differs across respondents and geographical groups, implying that the importance of the various suggested endpoint categories varies with context. The second example addresses the data collection procedures through social audits. Through an interview with a social auditor it is suggested that the auditor varies the procedures for carrying out the audit in order to get the most valid result. For example, the auditor has to take into account the various tricks a company in a given context normally uses to cheat the auditor. However, this conclusion is based on only one interview and must therefore be considered as uncertain. Both cases thus points to that context plays a role for how the methodology in relation to endpoint categories and data collection procedures needs to take account of the context in order to get a valid assessment and it is therefore argued that not only may data be site-specific in SLCA, so may methodology if the context variation should be accounted for. The results of the studies addressing the problems of validity in SLCA all suggest measures of improvement which entail more laborious, and thereby probably also less usable, assessments, whereas the study addressing usability concludes that from a company perspective a less laborious approach is needed. It thus seems that there is a trade-off between validity and usability and it is therefore discussed to what extent compromises can be made. Here it is argued that different users may be imagined who may have different possibilities and demands in terms of requirements to work and validity of the assessment and that several different SLCA approaches should be available fitting these different possibilities and demands to increase the overall use of SLCA. However for all of these different approaches it is argued that the assessment should as a minimum be more accurate than no assessment at all. If this is not the case, SLCA can hardly be regarded as decision support. This minimum requirement is discussed in more detail and it is found that while inclusion of other measures proposed for increasing the validity of SLCA is for the user of the SLCA to decide, the assessment of the impacts of both the implemented and non-implemented life cycle situations, addressed in the second validity related study, must always be included. However, since this validity demand only establishes very few requirements to the user, this methodological ‘openness’ may potentially be used to consciously select indicators or data in favour of one alternative. To mitigate this possibility for manipulation, a more comprehensive demand is considered which is to always include an assessment of the completeness and uncertainties in SLCAs accessible to the public. This, however, requires knowledge about how certainty and completeness is established in SLCA, calling for further studies into the validity of SLCA procedures. Several studies addressing this issue are proposed. A final discussion summarises the findings and concludes that due to raised difficulties in SLCA about data availability and issues like the assessment of the non-implemented decision, SLCA may never gain the same popularity as ELCA.
Main Research Area:
Hauschild, Michael Zwicky, Jørgensen, Michael Søgaard