Highly effective systems have a high congruence with the tasks that they are supposed to complete, but they also need to be capable of a certain degree of self-management. A resilient system is one that can survive in multiple stable states. Such systems are either comprised of elements that are highly adaptable or highly survivable. Most flood regions comprise a resilient system, where there are two states of flooded and dry. These systems are able to exist in both states of equilibrium with equal health and balance. Some resiliency comes from organisms that survive in both, but resiliency in other instances means that some elements might go dormant, or move to a different system (or in from a different system).
A self-organized system is one that finds balance itself without any intervention. Most natural systems are self-organized by their nature, but this is differentiated from systems that already need intervention in order to maintain stability. Systems that are self-organized are resilient because they exist in a form that has developed over the centuries. Systems that require intervention from man have typically been damaged by man, or are under threat. These systems are no longer self-organizing and therefore are less resilient. Hierarchy is systems reflects the way that the system is organized, in particular the power structure of the system.
I think that I had little understanding of these concepts so that this conceptualization aligns with what I understand these ideas to mean, but that is because I take my cues from this course and what I am learning here. If we look at floodplains, say regular ones like the Okovango in southern Africa. This is a system that exists in two separate equilibrium states, the wet and dry. It is a resilient system because of this, and because it remains self-organizing. This contrasts to ecosystems where human interaction has occurred, and in many cases human intervention is required just to restore natural water flows, for example.
Part III. One of the articles I've looked at in the course of my research is O'Farrell (1993). In this case, the research was testing a solution to disease based on changing behaviors among the population. The idea with this solution is simple -- they needed to test it in the real world. The study was conducted across countries, because there are differnces in the ability of the subjects to acquire the soap and clean water needed. The subjects were chosen at random to be part of he survey, but they were instructed about what they were supposed to do. Results were self-reported. The project involved a number of medical NGOs working in Africa, those who are experienced in health and who would have access to appropriate subjects for the study.
In a developed country, this subject of male sexual health is not addressed by NGOs who experiment with different solutions to find one that works locally. Instead, it is addressed in part in the health system and in part in the education system. There is probably less need for specialized education and training. Moreover, techniques are more sophisticated as well -- the techniques tested here are geared towards an audience with a lower budget for their health. So while there are definitely similarities in terms of the outcomes, the approach is a little bit different. The starting points in terms of education are different, and there are socioeconomic issues as well that are quite different between Africa and the United States. Thus, it is necessary to test different approaches to see what works best in the African context rather than trying to apply what works in American over there.
Part IV. In looking at water in sub-Saharan Africa, the water system can be visualized in a couple of different ways. The first is just to understand the basics of where the natural water is on the continent:
This is actually quite important because you have to compare this with the international boundaries -- some countries have plentiful fresh water and others that are lacking in this resource -- the politics makes a difference in terms of water policy. The other way to visualize water in sub-Saharan Africa is to understand how the water systems work. I'm not much of a graphic artist, but here's a rough sketch:
What this illustrates, crudely, is where water comes from, and in what form it collects. All three major forms (flowing, standing and underground) are available to humans for use, and it is these three critical resources that need to be measured and conserved. There is natural water loss that should be evaluated as well, but even more important is the wild card element -- natural and anthropogenic pollution. Natural pollution could be organisms in the water that render it unfit for human consumption. This is most likely to be found in standing water. Anthropogenic pollution -- introduced by man -- can be found in any three but in surface water most especially. Underground supplies can be affected, for example by mining or oil extraction activities. Ultimately, understanding the different critical points for influence is key to understanding how best to develop and protect Africa's fresh water supply.
Part V. The first problem that I have with this procedure is that the subjects are self-selecting. That right away adds a certain amount of bias and makes it impossible to extrapolate the results to the population as a whole. There are other problems, too. For example, the principle of informed consent holds that the research subjects should be made aware of all pertinent information about the research. In this case, they might not be at the stage where such disclosure needs to be made, but they are signing up for something they know nothing about. In addition to raising concerns about the researchers' commitment to informed consent, I would also argue that the research might end up with participants who are useless, or maybe even far more participants than they need. Before anybody participates in the study, they need to be made informed about the nature of the research and the form it will take, so that when the subjects consent they fully understand what they are consenting to.
It looks like there is basically no information given. What needs to be given are key details, for example time commitments, some background about the research and more. The subject needs to know what their role in the research is. Further, the subject needs to be made aware of what risks they might face as the result of their participation in the research, so there are a whole host of informed consent issues at play with this ad.
Ideally, the study should be random. If the subjects are "leaders" or "followers," then that needs to be identified, and because those are big groups a target group needs to be specified -- for example employees of fast food chains, so that the study is about leadership in fast food chains. I would gather as many candidates as possible, by getting a few different companies to participate. Only some will be selected, randomly, and they would not be allowed to self-select. You want to get a large enough group for statistical significance, but anything more than that just adds to the costs. The group has to be randomized, however, to ensure that there is no bias in the research.
A Chi-squared test is one that would be used in situations "to compare observed data with data we would expect to obtain according to a specific hypothesis" (Fisher et al., n.d.). Whereas with a normal distribution you are looking at observed data, you are not comparing it with a specific point that a model has identified as a likely outcome. That is what the Chi-squared test is for.
The Mann-Whitney U Test is a test that is used to compare two groups of populations to an alternative hypothesis. This test is more effective than a t-test for this scenario, as it is a non-parametric test.
The Wilcoxon rank-sum test is another non-parametric test. This is used when there are two samples, to see if the means differ between the two populations. It is used specifically in non-parametric situations.
The Kruskal-Wallis one-way analysis of variance is the non-parametric equivalent of ANOVA. This test is used when there are two independent samples, to find out if the samples originate from the same distribution. The null hypothesis would be that they do, the alternative hypothesis being that they do not.
Part VII. Constructing a valid, reliable assessment instrument is fairly easy if you undertake the right steps. You have to be dogmatic, and cannot take shortcuts. For example, if you need a random sample is has to be truly random -- you cannot select from a narrow population and then extrapolate it to the greater population. So the population is very important. It needs to represent what you want to study, and it must be random, not hand-picked and not self-selected.