|
AIDS in Indiana County Project Two: Correlations
|
||
|
|
Significant Correlations |
|
|
Unmarried-Partner Household Correlations
Sexually Transmitted Disease Rate Correlations |
After researching and gathering U.S. Census data, I found correlation coefficients for each data set in relation to the AIDS rate. I found six significant areas of correlation, to a 95-99% degree of certainty (p≥ ±0.602): unmarried-partner households, race, sexually transmitted disease rate, highest level of educational attainment, place of work, and population density. The certainty and significance of all these correlations greatly decreased, however, when any outlying data points were removed. Additionally, there were many other data sets that I considered that did not have a significant correlation to the AIDS rate (see Insignificant Correlations). In many cases, the strength, significance, and even sign (positive or negative) of the correlation surprised me, contradicting what I had previously assumed and guessed. In all but one of the significant correlations I examined, I observed a cluster of data points with an outlier that through off results. This outlier was regularly Allegheny County, and when I removed it from consideration, the correlation coefficient became exceedingly insignificant. After examining my data and graphs closely, I have come to the conclusion that there are no significant correlations for the other ten Southwestern Pennsylvanian counties because the demographics of each county are so similar that the slightest deviation from linear tendencies renders the correlation insignificant. In other words, the narrow range of data limits confidence in correlations. For example the range of the percentages of population who are Hispanic is less than one percent. The data then, is extremely similar, but since the range is so small, any discontinuities or deviations are greatly magnified. And since Allegheny County is so dissimilar to the homogenous profile of the other ten counties, that is, it is a metropolitan rather than rural area, has a greater minority population, a larger population density, and a much larger AIDS incidence rate, it will of course be an outlier, since its demographics differ from the rest of Southwestern Pennsylvania so greatly.
|
|