Bo Strangert (RD7)

On the choice of "perspectives"

when investigating complex phenomena 

Complex phenomena cannot be completely or sufficiently described from a single point of view or perspective. This depends on the intrinsic meaning of complexity, which involves a diverse set of interactive components in dynamic and uncertain (probabilistic) states. An additional characteristic is that complex phenomena often are only partially transparent, i.e., less accessible to direct observation and measurement.

However, the need to describe and understand complex phenomena has generated numerous approaches for partial and simplified solutions. One is to restrict the description or analysis to a few different points of view or perspectives and combine them to get an overall description of a phenomenon. For example, the distinction between 'objective' and 'subjective' aspects is often used to combine two mutually exclusive perspectives into an alleged exhausted description of a phenomenon. But in most cases, such a classification practice will not be enough to capture the complexity issue.

Another, more sophisticated approach, is to treat a phenomenon as a system that can be partitioned into subsystems that are possible to investigate separately. The assumption that the subsystem results can be integrated into a complete picture of the system may be correct for advanced physical and technical systems. But these types of complicated systems normally do not correspond to the definition of complex systems (see A2, Section 7 for comments about the difference between complex and complicated systems). There is no guarantee that advanced technical systems will keep their validity within larger interactive systems and practical settings.

When a phenomenon also includes social aspects, the conditions for adequate conceptual modeling are still worse. A broad bottom-up approach of selecting perspectives or preliminary frameworks is to apply correlative methods or multidimensional scaling (e.g., Keren & Lewis, 1993). The resulting models commonly show a structure of a few essential attributes, sometimes with added parameters of time and probability of occurrence. Unfortunately, such conceptualizations do not even satisfy moderate needs to understand complex processes. The conclusion is that there are no shortcuts but to carefully examine the unique options for each project and try to foresee the consequences for concept development and experimentation, CD&E.

A professional consultant's dilemma

The commissioner of a project commonly has a certain perspective on the task in mind and expects the provider to share this view. In this context "perspective" is an expression of the client's view or selected cognition of a task or phenomenon. If the project concerns a really complex problem with many dimensions and uncertainties that the client probably is not aware of, then the consultant has a dilemma. Two connected challenges have to be confronted: What is the most appropriate professional approach to the task? And how should one persuade the client about the necessity of providing an extended approach, possibly deviating from the client's expectation?

Let us suppose that the commissioner will not give a carte blanche and that the consultant is inspired by a belief in professional quality. Of course, this quality requires proficiency to inform and persuade the client about difficult and perhaps uncomfortable matters. Nevertheless, the most crucial matter of expertise is the capability to work out inventive approaches to complex problems. Let us take this question first.

What kind of expertise?

One pertinent question of expertise is how you can use science as a standard for the R&D of complex tasks in practical contexts. What about conceptual modeling and experimentation? In principle, there are an infinite number of approaches to model and observe real complex phenomena. The diversity of preconditions is enormous, though the logic of sound scientific conceptual and empirical procedures is limited, transparent, and the obvious rational alternative. Which special arguments are there for a strictly scientific approach to complex tasks as possible? Here are a few brief  suggestions:

A crucial early step is to try to "translate" fuzzy perspective(s) to conceptual models. Why not instead use a bottom-up approach of data collection to gradually identify and refine concepts? The arguments for CD revolve around the utility of constructing logically testable models as a basis for development.

A concomitant action is to figure out hypothetical consequences by constructing corresponding concrete empirical representations. That can be accomplished by using methods from a broad repertory of e.g. observation, subjective reporting, experiments, scenario construction, simulation, and gaming.

A systematic strategy for short and long-term follow-up and evaluation is necessary because investigation and action design of complex tasks are tantamount to control a process of trial and error.

In science, progress in knowledge is possible by its culture of critical reexamination of facts and principles as well as methods over time. In practical R&D, there is no corresponding institutionalization. It is up to the individual professional practitioner to critically assess the value of different approaches. Such an undertaking is what most professional consultants try to do when confronted with a new R&D mission.

However, I suppose that many consultants’ perspectives on a project task or goal often is established too quickly and conventionally. This readiness also recurs at the end of a project; the post-analysis tends to be superficial and does not examine alternative possibilities of planning and design. There are some understandable reasons for that, like shortage of time and resources, besides the frenetic demands on continual renewal and quick fixes in a diverse practical context. As a consequence, some necessary consequences for further professional methodological development would be missing.

The challenge of promoting a scientific approach

In practical contexts, science is surely respected but thought to be unrealistic to apply to social systems. A bottom-up approach based on data collection may be accepted by the commissioner, especially if it includes subjective reports from clients or personnel. Therefore, a straight theoretical basis for mutual project planning may be possible only if the consultant is extraordinarily charismatic or the client has got serious training in science. My experience is that a practician’s ability and willingness to participate actively in a complex endeavor is strongly associated with a scientific background.

The challenge is to participate in discourse, starting with a fuzzy perspective, which finally leads to testable results through a process of CD&E. This is not something that can be achieved easily for complex tasks.

Methodological consequences of choosing project perspectives:

Two short examples

The previous arguments for CD&E in practical work are too general to include any hints about what inventive or non-conventional approaches could mean. Two examples are briefly commented below; they may illuminate my reasoning somewhat. The first case is about planning for a given task. The second one is about how to post-evaluate a previous project.

1. Perspectives on military effectiveness

and risk of collateral damage.

This example is taken from a case study, including exercises on initial project planning (RD1, A2). The project task concerned means to reduce risks of collateral damage without sacrificing crucial military aims. Collateral damage was defined as unintentional damage to both own forces and non-combatants and their properties.

The case context was similar to the situation the American General McChrystal encountered when he 2009 personally surveyed the state of war in Afghanistan. He found that the soldiers’ risky business provoked misconduct and low trust towards the Afghan people. The stress on effectiveness and security for own forces led incidentally to collateral damage, which indicated the difficulties of command and control of the mission. McChrystal stated that actions had to be taken on all levels of command and operation to counter the negative humanitarian and political effects.

A conventional bottom-up approach to such a project is to systematically collect data from a broad array of incidents — their types, frequency, specific circumstances, etc — and then try to calculate correlations between conditions and effects as a basis for causal explanations. The interpretations of the results are then used to develop various actions that should possibly fulfill both humanitarian and military aims. The diverse applications are evaluated as new results and experiences aggregate over time.

A contrasting, theoretically motivated approach is exemplified by the case study referred to in RD1 and A2. It begins with a perspective that tries to capture the complex characteristics of the task. Instead of collecting data about various incidents of collateral damage and its possible causes as a basis for examination, this perspective proceeds from the conceptual relation between collateral damage and operational effectiveness. It starts with the observation that `Intelligence capability’ is a common precondition for both concepts; it can reduce the risk of collateral damage and improve the effects of operations. This capability is a joint product of technical, tactical, and strategic qualities and should be developed on all levels.

A final component in the structure of core concepts is ’interoperability’. This concept captures the meaning of how components interact in a context of dynamics and uncertainty on the battlefield. In this particular case, the components included technical, tactical, and strategic levels on a vertical dimension, and the operations of different military branches, units, and other components on a horizontal one.

This case of a quite general CD perspective focused in-depth on a restricted empirical application to incidents in urban operations with a coordinated army and air operations (see A2). However, it is easy to generalize the approach to different applications of other types of incidents and areas of collateral damage.

_______________

What can we learn about the consequences of choosing certain perspectives? This is the theme in some forthcoming papers (see T17). Looking retrospectively on earlier projects may hint at why some types of initial perspectives will lead to stable but conventional solutions, while others will either contribute to change or produce dead ends. Of course, it is generally not possible to casually explain why a ”project perspective" leads to a certain complex event sequence. That would require truly contrasting cases for comparison. But it seems possible to speculate about motives and reasons for choosing one or another perspective.

2. Perspectives on supervising work environment risks

at workplaces

The complex task of supervising. This next example deals with early projects on developing inspection methods for the National Board of Occupational Safety and Health during the Nineties (from 2001 for the Swedish Work Environment Agency). The Authority issues Provisions (AFS) and General Recommendations specifying the requirements to be met by the work environment.

Our focus is now on the development of methods to conduct inspections regarding the Provisions on Internal Control of the Working environment (IK 1996), later substituted by the very similar Provisions on Systematic Work Environment Management (SAM 2001). These provisions pertain  ”to systematically planning, carrying out and follow-up activities in such a way that requirements concerning the working environment are satisfied”(1996), and ”to the work done by the employer to investigate, carry out and follow up activities in such a way that ill-health and accidents at work are prevented and a satisfactory working environment achieved”(2001).

Both versions include provisions on IK/SAM requirements concerning the natural part of work, participation, policy, and routines; allocation of tasks, knowledge; risk assessment, measures, and follow-up; occupational health services. The inspection of workplaces is carried out by inspectors from the Authority, who ”check that the employer has an effective organization”. They also ”check the work environment in the holistic perspective of the risks)(physical, mental, and social) which the operation entails” (see SWEA, Report 2008:01).

The inspection districts started projects in the early nineties to meet the need for new inspection methods for the provisions on internal control of the working environment, IK 1996. Our focus is on the project activities during that early period.

Common or diverging perspectives on the inspection task?

The provisions 1996 and 2001 aimed at advancing work environment management as a means to improve the quality of work environments. This necessitates acceptance, implementation, compliance, and compliance control of the provisions. Such a huge undertaking presupposes a ”common perspective” among legislators, target groups, and inspectors. It should be established by creating an efficient communicative chain between all parties involved, based on precise and understandable texts, talks, and acting.

The early projects for developing new inspection methods can roughly be separated into three categories according to their main perspective or orientation.

– Most inspection districts started their projects from an ”administrative perspective” and derived requirement specifications directly from the text of the provisions, by formulating specific questions and observation tasks about e.g. routines, allocation of tasks, competence and resources, types of measurement, and follow-up, etc. The main emphasis was on the preconditions and functionality of work environment management. The inspection data had a distinct objective stance, and overall judgments could be constructed as an index based on aggregated data.

– Another approach used an ”effect-based perspective” for the inspection. The core component was the inspector’s investigative dialogues with employers and employees focussing on organizational and social problems in the working environment. Essential issues were identified through preliminary examinations. A major emphasis was on psychosocial conditions and their relations to the management of work — as a potential cause of the increasing rates of sickness absences. The data from the inspection had a substantial subjective stance due to its perspective and content.

– A third approach used a ”process perspective”. The process involved was the transaction between the inspector and the employer to achieve the two related goals of control and improvement of work environment management. This required CD&E, and therefore a control model was used to represent the structure of the IK provisions as well as of the inspector’s procedure and the employer’s work environment management. The complex model structure was the basis for the design of procedures, tools, and training sessions for inspectors.

About theoretical diversity among inspection methods

I would characterize the first two types of perspectives as bottom-up approaches. They use low-level concepts which lend themselves to be easily operationalized into data during the first round of observation and description. An ”administrative perspective” allows you to recognize a string of related concepts from the provision text to a corresponding observable statement or action about the work management environment. An ”effect-based perspective” allows you to collect immediate data of colloquial speech in an investigative dialogue about a focal work environment problem.

However, the concept-data relations are deceptively simple and they rely strongly on a consensus about interpretation. Also, the direction of interpretation possibly differs between the two approaches. The administratively defined preconditions would dominate the interpretations of data in the first case, while the data from the effect-based dialogue may be a result of an interaction between parties that is only vaguely anchored in the provisions’ definitions.

From a top-down ”process perspective” the inspection is a complex interaction that needs to be translated into a strict conceptual model to avoid messy interpretations of observations. The structure of the model must correspond to a complex data structure that includes all important aspects of inspection events. There are many steps from model to inspection procedure and even more from action to data interpretation and feedback. The question is whether a prospect of advanced CD&E can supersede the prevailing praxis of promoting administrative simplicity among authorities?

Is the path from a ’perspective’ to advanced conceptual modeling and operationalization worth the effort?

Now it’s time to try to generalize from these examples. Think about the contrasting scenarios of development. An ”administrative approach” proceeds smoothly from the text of provisions to a broad survey of preconditions and functions. To investigate the state of management is straightforward regarding data collection. For an ”effect-based approach," there is a corresponding clear line between ubiquitous organizational and social problems and the dialogue method as a natural means of reporting them. Thus, it is easy to reach a consensus about low-level concepts and simple data reports, objective as well as subjective, particularly when there exist common stereotypes about cause and effect.

This reasoning sounds like a reasonable thesis about the advantage of simplification: Those perspectives will be preferred that present a new, sufficiently simple but coherent view of old, conventional, and well-known facts.

What about a contrasting advanced CD&E alternative? Most authorities should probably prefer to make a cautious and conservative choice. Pragmatic arguments about e.g. simplicity, consensus-based philosophy, need for transparency and a realistic view of competence levels point towards using the viable thesis of simplicity in development approaches.

However, reality sometimes discloses the need to confront complex phenomena with complex scientific endeavors. The first example of military effectiveness and collateral damage illustrates some arguments for an advanced approach to combined physical and social challenges. Furthermore, knowledge growth without compromises is the strong drive in the natural and technical sciences that advances applications. (See T1 on this website for comments on physical sciences compared to social and behavioral sciences.)

The professional consultant’s dilemma revisited

What could we as consultants learn from an examination of ongoing and previous projects about complex tasks and the promises of R&D? Presumably a lot, though the present comments were limited to the start of projects: How should we choose a productive perspective of a task and adjust us to an alternative perspective?

The conclusion that most organizational and social practices rest on superficial theoretical rationales should not prevent us from exploring more adequate ways when confronting complex matters. Hence, a strategy congruent with preliminary CD&E would be desirable. Does this need some kind of serendipity to succeed? I think not. A conventional definitional procedure may assist in sharpening the meaning of different ”perspectives” — and hopefully their deficiencies as well.

Just define what constitutes a given perspective. One straightforward and obvious procedure is to try to define which aspects should be considered in a project, and what kind of structuring of aspects could be made. However, to define a ’relational structure’ among aspects may indeed involve a challenge!

One simple example of describing a ’perspective’ is to define a system as a structured set of relevant elements concerning their context or environment. A corresponding description of ’productive perspective’ could be to formulate how the system operates effectively to transform input into output according to rules and goals. A comparison between your own preferred perspective and other old or competing ones may be made as a matching of alternative systems in operation or through data review, scenario construction, or simulation – either physically or by imagination. In which specific ways (concerning elements and relations) are the perspectives congruent and incongruent?

Further exploration into types of project strategies is planned on this website.

References

Keren, G. & Lewis, C. (Eds.)(1993). A Handbook for Data Analysis in the Behavioral Sciences. Methodological Issues. Hillsdale, N.J: Lawrence Erlbaum.

Swedish National Board of Occupational Safety and Health (1996). Ordinance containing Provisions on Internal Control of the Working Environment, together with General Recommendations on the implementation of the Provisions.

Swedish Work Environment Authority (2001). Provisions on Systematic Work Environment Management together with General Recommendations on the implementation of the Provisions.

Swedish Work Environment Authority (2008). Rules of Inspection. SWEA Supervision Series. Report 1/2008.

Papers  on this website  arborg.se:

A2 – Applied research methods for development projects

RD1 – A case approach to study applied research methods