Research of any kind can be successful only if it is structured and carried out properly. The proper structure of the research includes the focused methodology and the outlined practical tools to use the theoretical framework of the research. In the methodology, sampling strategies, sample size, and statistical tests prove to be the crucial parts as they define who will be sampled, how they will be samples, and for what. Therefore, either in research aimed at studying the political views of the populations of certain geographic areas or in the research of natural conditions like slope failures, sampling and statistics play a significant role in the research basis.
Thus, if the researcher is interested in political views of populations between 18 and 25 years living in two different areas of the city, he or she should pay attention to the peculiarities of the population and should also define the exact number of people surveyed, develop the special questionnaire for the survey, and create the system of the analysis of the results. As well, the relation between the geographical area analyzed and the responses obtained should be established to achieve the very goal of the research, i. e. to see the similarities and differences in political views of people of 18 – 25 years living in the two different areas in one city.
One of the most important parts of every research is the sampling strategy, i. e. “the plan you set forth to be sure that the sample you use in your research study represents the population from which you drew your sample” (Landreneau, 2006). Thus, the sampling strategy should select the people to objectively represent the whole population one tends to study. Selecting from random, systematic, and stratified sampling strategies, this research needs to use the stratified one to include all the layers of the social groups analyzed (Motulsky, 2009). Needless to say, it is physically difficult and almost impossible to survey the whole 18 – 25-year-old population even in a part of the city as this might involve a questioning of thousands of people. This is what the sampling strategy is used for: it helps the researcher to choose the most fitting respondents and include all possible varieties of the population observed in the area to the sample (Wilmot, 2002).
The size of the sample is also vital for geographical research as, being proper, this criterion facilitates the objective results of the research: “The sample size is very simply the size of the sample” (HyperStat, 2009), i. e. the number of people the researcher selected as reflecting the comprehensive picture of the studied population. The sample size should be not excessive but not too small either. If the sample size is large it is difficult to carefully examine the data of all respondents and provide objective results. If the sample is too small, it might fail to represent the studied population to the full extent, and thus also deprive the research of objectivity (Landreneau, 2006).
Finally, the statistical tests are also important to the geographic research as they provide variety to the methods of questioning and ensure the objectivity of results if the latter can be biased under a uniform statistical test implemented. In selecting the test, Motulsky (2009) advises defining the character of the data collected and the goals of the research. Based on this, tests can be parametric, i. e. assuming certain relations in the population distribution, and the non-parametric tests, i. e. the ones that do not assume anything about population distribution (Motulsky, 2009). Tests can also be paired and unpaired, i. e. the ones used to compare two populations or more respectively. In the case of analyzing 18 – 25-year-old populations from two areas in the city, the only paired test is applicable. Thus, the Wilcoxon test and Chi-square test are the most applicable ones for this research (Motulsky, 2009).
The natural geographic research aimed at finding out whether the frequency of slope failures is determined by the slope angles and vegetation density or not should also be structured according to the sampling and statistical norms to contain the sampling strategy with the defined sample size and to operate with the proper fitting number of statistical tests that allow checking the objectivity of research results. The different focus of the research, and namely not comparative but inquiring research, changes the sampling strategy and the statistical test to be used.
The sampling strategy used for this kind of research will dramatically differ from the sampling strategy of people-related geographical research (Wilmot, 2002). First of all, the sampling will be carried out not in a comprehensive manner but in the manner of including the steep and densely vegetated slope types observed in the area under analysis. In other words, the researcher will need to study several slopes with the largest angles and several most densely vegetated slopes to see any tendencies in the slope failures of the studied samples (Wilmot, 2002). The sampling strategy will be based mostly on the assumption that the specific number of slopes will be enough to generalize the findings but not to represent all possible slopes found in the area (Landreneau, 2006). The combination of systematic and random sampling strategies should be used to achieve both objectivities of the research and its comprehensive character (Motulsky, 2009).
Drawing from this, the sample size will also be considerably different. Again, based on the researcher’s assumption, the sample size can include a dozen of slopes of both types discussed to see if there are any trends as for slope failures in steep and densely vegetated slopes (Wilmot, 2002). The sample size can be changed, either to the larger or smaller numbers of studied samples, in case if the study of the initially selected sample does not allow making any conclusions or even assumptions (HyperStat, 2009).
The choice of the statistical tests for the research of slope failure occurrences will also differ from the number of tests possible for application in the research of people’s political views (Motulsky, 2009). Thus, only non-parametrical tests can be used for slope failure causes study as far as no population distribution data or assumptions are presented for consideration in such research (Niles, 2009). The idea is to study the steep slopes and find out what is the frequency of slope failures in them. The second point is to consider the densely populated slopes to see the frequency of their failures. The final step should be to consider the statistics of failure of steep and densely populated slopes to compare it to the research findings and relate it to the question of the research (Niles, 2009). The most fitting statistical tests for this research will be the Mann-Whitney test that allows comparing two unpaired groups of samples and the Spearman correlation test allowing the quantification of the relations between two samples considered (Motulsky, 2009).
To conclude, geographical research demands a clear sampling strategy defining its sample size and the outlined set of statistical tests to succeed in achieving its goals in studying a certain geographical phenomenon related to a certain natural phenomenon or the behaviors of people in specific geographical areas.
HyperStat 2009, Sample Size, Online Contents.
Landreneau, KJ 2006, Sampling Strategies, University of California-San Francisco. Web.
Motulsky, H 2009, Intuitive Biostatistics: Choosing a Statistical Test, Oxford University Press.
Niles, R 2009, Picking the Right Statistical Test, Robert Niles. Web.
Wilmot, A 2002, Designing sampling strategies for qualitative social research: with particular reference to the Office for National Statistics’ Qualitative Respondent Register, ONS.