CASE STUDY METHODOLOGY: Given the complexity of our operating environment, how are we to proceed in terms of generating and applying ‘practice oriented’ knowledge?  According to Yin (2018), case studies are a powerful tool for proactively handling such complexity.  Yin defines a case study as an empirical method that: i) investigates a contemporary phenomenon (the “case”) in depth and within its real-world context; the method is especially useful when ii) the boundaries between the phenomenon  and context may not be clearly evident.  In our particular field, it is well established that the application of scientific results (i.e. the science policy interface), is a never-ending challenge for environmental disciplines (MacLellan et al 2017; MacLellan 2008b), thus accounting for context in a real-world situation, is a great source of understanding and insight.  Yin (2018) further describes the features of a  case study as: i) coping with a technically distinctive situation in which there are many more variables of interest than data points; ii) which can therefore benefit from the prior development of theoretical propositions to guide design, data collection, and analysis; and from which iii) multiple sources of evidence can be derived and interpreted, in a triangulating fashion.

The implications for our research\application can be summarized as follows: i) given highly complex phenomenon (i.e. the proactive utilisation and application of scientific knowledge to increase the adaptive capacity of communities), in combination with ii) an inability to isolate and study single causal factors, leads us towards the adoption of iii) a case study methodological approach as a means of providing insight into the further development of tools\resources associated with our given purposes. Furthermore, by adding to the pool of practice-oriented case studies that already exist, we (in the broadest sense) can begin to understand the factors that are involved in improving the adaptive capacity of communities.  An apt metaphor is cited by Flyvbjerg (2006) to describe the purpose of case study framework.

In teaching you … I’m like a guide showing you how-to find your way round London. I have to take you through the city from north to south, from east to west, from Euston to the embankment and from Piccadilly to the Marble Arch. After I have taken you many journeys through the city, in all sorts of directions, we shall have passed through any given street a number of times — each time traversing the street as part of a different journey. At the end of this you will know London; you will be able to find your way about like a born Londoner. Of course, a good guide will take you through the more important streets more often than he takes you down side streets; a bad guide will do the opposite … I’m a rather bad guidevi. (Wittgenstein as cited by Flyvbjerg, 2006)

CASE SELECTION: Despite obvious benefits, there are nevertheless, strong criticisms of the case study approachvii  including the inability to produce ‘generalizable’ results, which is the hallmark of the scientific method.  And it is within this context that this assignment is situated.  According to Flyvbjerg (2006) the choice or selection of the case studies themselves, is a strong way to increase the capacity for ‘generalizability.’  In Table 1 we can see that the traditional focus and subject of scientific enquiry (i.e. a representative case or a random sample as seen in Section A) may not always be the most appropriate strategy especially “when the objective is to achieve the greatest possible amount of information on a given problem or phenomenon. This is because the typical or average case is often not the richest in information.”  Extreme, critical and paradigmatic cases on the other hand, may reveal more information because they activate more actors and more basic mechanisms in the situation studied.

STRATEGIES FOR THE SELECTION OF SAMPLES AND CASES

Table 1: Strategies for selection of samples and cases as adapted from Flyvbjerg (2006).

For a full discussion of the suitability of each strategy in Table 1, see Flyvbjerg (2006) page 229 to 233, and\or Yin (2008) page 47-64.  Note that Yin’s treatment of a single case is slightly different, but he is generally consistent in his identification of single-case rationales: critical, unusual, common, revelatory or longitudinal.  In Yin’s typology the critical case is aptly named, because the case is critical to the theory and\or theoretical propositions used to guide the research.  As such, the theory should have a clear set of circumstances within which its propositions are believed to be true; a single case can then be identified to determine if the propositions are supported.  In a like fashion, the extreme or unusual case deviates from theoretical norms, or even everyday occurrences.  Its utilization allows us to comment on normal processes.  Conversely, the common case is used to capture conditions of everyday situations.  A revelatory case provides an opportunity to examine phenomenon that were previously inaccessible to scientific enquiry, and a longitudinal case essential involves studying the same single case at two different points in time.

As one might expect, evidence from multiple cases are generally considered to be more compelling and robust, than from single casesviii.  In this assignment you will be considering both single and multiple case design.  When utilising a multiple case approach, a “replication” design (as opposed to a “sampling” design) is utilized.  Each case is carefully selected so that the individual case studies either: a) predict similar results (a literal replication); or b) predict contrasting results but for anticipatable reasons (a theoretical replication).

“A few case studies (2 or 3) might aim at being literal replications whereas a few other case studies (4 to 6) might be designed to pursue two different patterns of theoretical replications.  If all individual case studies turn out as predicted, these 6 to 10 cases, in

the aggregate, would have provided compelling support for the initial set of

propositions pertaining to the overall multiple-case study.  If the individual case studies are in some way contradictory, the initial propositions must be revised and retested

with another set of case studies. (Yin 2018)

In this exercise you will devise three separate case study designs to progress the research and  application efforts of our partner, Eastern Charlotte Waterways (https://www.ecw.ngo/).  These efforts entail the development and application of various decision and analytical support

systems including the INTERSECT APPix.  This online app is based upon the efforts of Zhou

(2020), Sookhan (2021), MacLellan et al (2015) and Eastern Charlotte Waterways Inc. (Cowie

and Killorn 2020) (https://planetearth.utsc.utoronto.ca/people/dataexplorer/).  INTERSECT is intended to facilitate the exploration of vulnerability to environmental risk, within the

province of New Brunswick, Canada.  In a very real sense, your efforts will be used to inform policy and development decisions within New Brunswick and southern Ontario.

INTERSECT APP: INTERSECT facilitates a coherent exploration of comparative, or relativized, socio-economic and environmental metrics.  It consists of two components: 1) the first represents the minimum foundations for environmental data compatibility, comparability; while 2) the second represents the minimum foundations for socio-economic data compatibility, comparability.  This work was generated subsequent to a workshop held in Nepal (SHRRC: Living with Climate Change, 2014x), in which the task of comparing communities from Low Income Countries (LICsxi) against communities from High Income Countries (HICs) was deemed desirable, yet theoretically insurmountable.  As a result, our colleagues in Eastern Canada began to search for what might be deemed, minimum data compatible metrics between these cases MacLellan et al (2015).

In terms of the environmental data minimum, we found that bioclimatic profiles offered a solid foundation for contextualizing environmental challenges, while allowing us to add the extra dimension  of climate projections (see APPENDIX Figure 1 and 2).  Socio-economic data was much more challenging and resulted in the creation of what we now refer to as M-Chartsxii  (2015)xiii, which are themselves based upon our longitudinal case study in Charlotte Country New Brunswick (2013) and our efforts in Canada,  India, Pakistan and Nepal (i.e. SHRRC Living with Climate Change, 2015).  The challenge was the discordances in the social, economic and cultural conditions that existed between nations; the solution was to accept the differences in data relevance and quality across broader scales, while highlighting the relative nature of data.

Building upon the utility of amoeba (radar) charts (see Figure 3 in APPENDIX) we relativized various geo- political scales so that metrics could be associated in an immediate sense.  Amoeba diagrams abandon the idea of a single index of vulnerability, and instead force users to (relate’ the components of vulnerability for themselves.  To achieve this, Amoeba diagrams (see Figure 3) impose several structures on data which are artificial: 1) relatedness of neighbors — amoeba charts are often used when neighboring variables are unrelated, creating spurious connections; 2) cyclic structure — the first and last variables are placed next to each other; 3) length — variables are often most naturally ordinal: better or worse, though the degree of difference may be artificial; and 4) area — area scales as the square of values, exaggerating the effect of large numbers.  Nevertheless, if we accept the provisional nature of  this form of data representation, the diagrams are highly useful in a comparative, exploratory sense, as is appropriate for a scoping exercise.

The value of Figure 3 lays in its’ representation of multiple amoeba charts, which allow the eye to seek out relative differences between countries.  In other words, they are an effective way to show the relative nature of indicators when used comparatively (i.e. to highlight differences between subjects as seen in Figure 3).  This abiIity to show (reIative differences’ can easiIy be extended.  Imagine three amoeba charts layered over one another (Figure 4).  The different zones of the diagram do not overlap, but rather represent exclusive zones where different informational standards, and\or methodologies can be used to represent different scales of interest.  In this case, the zone at the centre of the Figure 4 is a complete hexagon, the next zone outwards from the centre is more like a donut that surrounds the centre zone but has its own axis, and so on with ever larger donuts.  The centre zone represents the village or community, the outer zone represents the international community.  Data in each zone is unique to the scale of interest and therefore can represent different methodological standards as required.  By presenting data in this way we can (see’ the vuInerabiIity status of a community w.r.t. its local environment, as well as it regional, national and international context.

Vulnerability is clearly not an absolute measure but is tied to local conditions and perceptions.  In our experience with local communities, we found that stakeholders were surprisingly open to using social economic data as long as it was contextualized in a meaningful way.  For instance, in New Brunswick, local community members were little concerned about how they compared to Ontario communities, but they showed a great deal of interest in terms of how their compared to other communities in the AtIantic region for issues such as outmigration.  The insight is that we need to (fIag’ community vulnerability in a comparative sense, over appropriate scales, not only locally but internationally as well. In Figure 5 we can see a very rough exampIe of how a particuIar community couId “see” itseIf in terms of its nearest neighbors (other local communities), across the regional, as well as the national and international landscape.  When augmented with the dynamic functionality of an online platform (see Figure 6), the results are truly immersive, allowing the user to make connections they would normally not have seen, as between environmental factors and socio-economic factors, thereby aiding in (case’ selection.

PROCESS: Generic

In this exercise you will ‘select’ a set of potential cases that will be used to inform researchers and practitioners within New Brunswick.  The ‘selection’ process is itself the object of this assignment.  To cover the range of possible case designs, you will undertake a scoping exercise for ECW in which you consider three possible scenarios: 1) undertaking a single case study; 2) undertaking three cases, in a multiple design framework; and 3) undertaking six cases using a multiple design framework.  The challenge is in justifying your choice of specific ‘cases,’ for each design scenario.

1.    Using the INTERSECT app, you will familiarize yourself with the climatic, and socio-economic context of New Brunswick.  You will note that INTERSECT changes as you move your cursor  across the province, updating both the bioclimatic profiles, as well as the relativized amoeba charts.

2.   You should note that the bioclimatic profiles offer two broad complimentary scales: 1) an

historic profile of climate in New Brunswick; and 2) a future profile of climate in New Brunswick for specific regions across the province.  Although the distinctions can appear slight, they become more pronounced as you move from the Bay of Fundy in the south, to the Quebec boarder in the north, and the St Lawrence coast in the east.  For the purposes of this exercise, we assume that the greater the differential between historic and future climatic profiles, the greater risk to the region.

3.    The socio-economic profiles are derived directly from STATS Canada data.  They are a relative measure and compare individual units within each the three M-CHART zones (i.e. community zone, the county zone, and provincial zone).  The fundamental visual principle at work is, the greater the surface area covered for each zone, the more at risk the community, county, or province is, in a relative sense.  In this respect, communities are compared only to those communities within a specific county; counties are compared to other counties within the province as a whole; and the provinces compared include New Brunswick, PEI and Nova Scotia. As a regulative baseline, a completely filled-in zone would indicate that the subject under consideration (a community) ranks as the worst overall example of vulnerability in terms of all categories.  For a broad and in-depth overview of the use of socio-economic indicators please see Wall and Marzall (2015).

4.    Extending the previous discussion of case study selection, you will identify three case study

design scenarios, and justify your choices in a short report by making specific reference to the steps above: a) for a single case design; b) for a multiple case design with 3 cases; and c) for a multiple case design with 6 cases.

5.    In the subsequent Assignment 2, you will iteratively extend your analysis of the 3-case design to account for more detailed socio-economic and environmental data.

Although this assignment is intended to be self-contained, the main readings associated with the exercise are Flyvbjerg (2006), Yin (2006) and Wall (2006) which are readily available through the library (i.e. Yin (2006) is available online through the followingLINKxiv).  Furthermore, a series of videos will be posted in QUERCUS outlining a step by step overview of the process.  Keep in mind that for the second assignment, you will use the data from the first assignment to undertake a deep dive into the actual circumstances surrounding a subset of your chosen ‘cases.’