Variation Research Paper - Research Paper Examples

This sample Variation Research Paper is published for educational and informational purposes only. If you need help writing your assignment, please use our research paper writing service and buy a paper on any topic at affordable price. Also check our tips on how to write a research paper, see the lists of research paper topics, and browse research paper examples.

Variation refers to the degree of dispersion, diversity, or inequality in a distribution: the extent to which observations of an attribute are similar to or different from one another. Applications of variation in the social sciences include income inequality, the degree to which incomes are concentrated among a few households or spread broadly across a population, and religious diversity, which describes whether a single religion dominates the cultural landscape or multiple religions exist side by side.

Components And Measures Of Variation

Popular quantitative measures of variation include the following:

Range : The difference between extreme values.

Variance : The sum of the squared distances from each observation to the mean divided by the number of observations (for a population) or divided by one less than the number of observations (for a sample).

Standard deviation : The square root of the variance. The standard deviation is the average distance between a set of observations and the mean.

Coefficient of variation : The standard deviation divided by the mean. Graphically the coefficient of variation describes the peakedness of a unimodal frequency distribution. Because the coefficient of variation scales to the mean, this measure often is used in comparative analysis.

Gini coefficient: Bounded by zero and one, where lower values correspond to greater equality, the Gini coefficient is defined as twice the area between the Lorenz curve and an equality diagonal. (To construct a Lorenz curve, one ranks the observations from lowest to highest on the variable of interest and then plots the cumulative proportion of the population against the cumulative proportion of the variable of interest.)

Theil’s T statistic: Particularly appropriate for hierarchical, nested, or aggregate data, Theil’s T statistic is the product of the population share of each observation, the quotient of the observation and the population average, and the natural logarithm of the quotient of the observation and the population average, summed over all observations. Unlike the Gini coefficient and the coefficient of variation, Theil’s T statistic is sensitive to the number of underlying observations.

In addition to these widely used metrics, researchers sometimes create original measures to isolate a particular type of variation. For example, in American Apartheid (1993), Douglas Massey and Nancy Denton develop five dimensions of variation to identify racial “hypersegregation” in U.S. metropolitan areas.

Whereas measures of central tendency, such as the mean, median, and mode, describe the center of gravity of a distribution, measures of variation express how concentrated the observations are around the average. Both types of measures are necessary to describe most social science data sets. For instance, with regard to the test scores in a school, the median expresses the score of a typical student. Schools desire high median scores. However, a high level of variation in test scores indicates that some students excel, while others lag behind their peers. This could suggest differences among students in family resources, native intelligence, or other characteristics over which a school has little control, but high variation also can result from inconsistent teacher quality or administrative choices to overallocate resources to high-achieving students and/or underallocate resources to low-achieving students.

The ease of computing variation can hide the difficulties in its interpretation. Although variation reflects differences among individuals, it is a property of a group, not an individual. Correlation between variation and another group-level variable does not imply a corresponding relationship across individuals; improperly asserting that it does is an instance of the ecological fallacy. Similarly variation is relative. A measure of variation can only be described as high or low in comparison to the variation of another group, the same group in a different time period, or a predefined standard value.

Use Of Variation In Social Sciences

Whether variation in a particular attribute is preferred is an issue of context and perspective. A high degree of variation may be preferable, as in the case of racial diversity within communities, in which variation may imply integration rather than segregation. A low degree of variation may be preferable; for instance, a good medical intervention consistently will improve the health of those who receive treatment and have few side effects. Alternatively, preference for high or low variation may depend on one’s theoretical or ethical perspective. For instance, one view of income inequality is that higher levels of variation reflect a more efficient economic system that rewards hard work (or talent) and punishes laziness (or ineptitude). An alternative view is that income inequality suggests nondemocratic use of monopoly power or outright oppression of the poor.

One way to reduce subjective judgment is to view variation as a predictor variable rather than an outcome variable. An example of such an application is research that relates economic inequality to population health. In a 1975 article Samuel Preston popularized this issue by pointing out an inherent nonlinearity in the relationship of health to income. For individuals with low incomes, an increase in economic resources will be accompanied by significant health gains, but as an individual’s income rises to the top of the distribution, subsequent increases in income will yield diminishing returns in terms of health. Thus at least among nations with high average incomes, lower variation of incomes should be associated with better health. Although subsequent scholarship seemed to confirm this correlation, there is debate over causation. Some scholars assert that inequality leads to stress and stress leads to poor health. Other researchers point to mediating factors, such as race and access to care, to explain the correlation between inequality and health.

Individuals with low incomes are more likely to have poor health, and individuals with high incomes are more likely to have good health, on average. Thus the suggestion that an increase in inequality that results from the poor getting poorer would decrease population health seems reasonable. However, advocates of the inequality hypothesis also must explain why an inequality increase that occurs when the rich get richer also would reduce health. Until they address this seeming paradox, the burden of proof lies with those who claim that income variation is an operative factor for health. This exemplifies the central challenge in using variation in social science analysis; to be a useful tool, variation must be both measurable and meaningful.

Bibliography:

Kachigan, Sam Kash. 1991. Multivariate Statistical Analysis: A Conceptual Introduction. 2nd ed. New York: Radius.
Massey, Douglas S., and Nancy A. Denton. 1993. American Apartheid: Segregation and the Making of the Underclass. Cambridge, MA: Harvard University Press.
Mullahy, John, Stephanie Robert, and Barbara Wolfe. 2004. Health, Income, and Inequality. In Social Inequality, ed.Kathryn Neckerman, 523–544. New York: Russell Sage Foundation.
Preston, Samuel H. 1975. The Changing Relation between Mortality and Level of Economic Development. Population Studies 29: 231–248.