# Rape Statistics

Working at a Historic Black College/University (HBCU) for the past three (3) years, I just recently came in contact with a rape victim. I’ve seen the shows but never thought, I would be a person to come in contact with one. Talking to the different individuals involved with the student case, I decided to do my research paper on rap victim. Specifically, I am looking at how the likelihood of being the victim of a violent crime in the United States is related to gender and race of the victim. The raw data is readily available from the US department of Justice min cooperation with the US Department of Health and Human Services. The Bureau of Justice Statistics and the Centers for Disease Control and Prevention issued their preliminary study in June of 2001. Thomas Simon, Ph.D. and James Mercy, PhD preformed the research, both are scientists at the CDC in association with Craig Perkins, a BJS statistician.

In order to make such a study possible, the researchers have to look at a large collection of data, sometimes incomplete that will vary from doctor to doctor. One of the first problems that arise is the underreporting of rape in the United States. It is widely believed in the United States that only one (1) out of three (3) rapes are reported to authorities. The second problem is that different doctors will characterize the injuries differently among no injury, severe injury, and minor injury. For example, a bruise for one doctor may be a minor injury, but for another doctor, no injury. The measurement scale is an interval, severe, some (minor) injury, or no injury. Attached is a copy of the data from the Bureau of Justice statistics.

The best statistical tests best suited for this type of data is correlation. Correlation is a bivariate measure of association (strength) of the relationship between two variables. It varies from 0, which indicates not relationship or a random relationship to one, which is a perfect linear relationship or to –1, which is a perfect inverse relationship. Correlation is usually reported as the r squared, which can be interpreted as a percent. For example, if r squared is .25, then the independent variable explains 25% of the variance in the dependent variable. In this case, we have data for a multiple correlation, that is, a correlation of multiple independent variables 9race, gender, and age are not dependent on each other, but the likelihood of being the victim of a violent crime could rest on all three.

Using multiple correlation, the statistician ends up with a percentage of exactly how closely related the cause as to the effect. For example, a look at table seven will show that males are more likely to be the victim of a violent crime than a female is. By using a multiple correlation, the statistician can look at percentages. The question how much of a role does gender play in whether or not a person will be a victim of a violent crime in the United States is answerable. The idea behind the thesis of correlation is one of the most elementary in statistics. Finding a percent correlation is a simple as finding how close the scattered data match up with a straight line. In this case, Spearman’s rho is the best formula to determine R. In this case

rho = 1 - [(6*SUM(d2)/n(n2 - 1)]

This equation holds true for use with two ordinal variables or one ordinal variable and an interval variable. Rho for ranked data is equaled to Pearson’s r for ranked data. D is equal to the difference in rank. This works well if the statistician wants to compare just one independent variable with the dependent variable. For example, if the researcher just wants to compare gender and the likelihood of being the victim of a violent crime.

However, the idea that more than one factor plays a role must be taken into account. For example, the likelihood of being the victim of a violent crime in the United States with respect to that person’s gender, age, and...

Please join StudyMode to read the full document