The Effect of the Unemployment Rate on Rape - Report Example

a The Effect of the Unemployment Rate on Rape Research Results For this study, rape is defined as the estimated level of forcible rape per 100,000 people. Unemployment is the unemployment rate calculated from the number of people unemployed compared to the labor force. Raw data for each of these variables was obtained for the 2010 year, for each of the 50 states (Table 1). The average unemployment rate across the states from this information was 8.8% (ranging from 3.9% to 14.9%) with a standard deviation of 2.1. The average number of rapes per 100,000 people was 31.2 (ranging from 11.2 to 75) with a standard deviation of 10.2. Table 1: Raw employment and rape data from each of the 50 states for the year of 2010. Unemployment Rape 9.5 28.2 8 75 10 33.9 7.9 45 12.4 22.4 8.9 43.7 9.1 16.3 8.5 34.7 9.6 31.1 11.5 28.6 10.2 21.6 6.6 26.8 9.3 33.5 10.3 23.6 10.2 27.2 6.1 27.4 7 38.8 7.5 31.8 7.9 27.2 7.5 29.3 8.5 21.3 12.5 26.7 7.3 47.3 10.4 33.9 9.6 31.2 7.2 23.9 4.7 32.4 14.9 36.8 6.1 35.7 9.5 31.3 8.4 11.2 8.6 46.5 10.6 14.3 3.9 21.1 7 35.2 10.1 32.1 7.1 38.7 10.8 31.7 8.7 26.9 11.6 28.1 11.2 31.7 4.8 47.9 9.7 33.7 8.2 33.3 7.7 34.3 6.2 21.1 9.6 31.1 9.1 38.1 9.1 19.1 8.3 20.9 7 29.1 The Pearson Product-Moment Correlation was used to test for a significant correlation between the two variables (Table 2). A two tailed test was used. The Pearson correlation coefficient (r) was -0.137 which indicates that there is a weakly negative relationship. For this study, an α value of less than 0.05 was taken to mean the result was significant (at α=0.05, the result has a 5% likelihood of being incorrect) (Lehman, 2005). The p value for the correlation was 0.336, indicating that there was no statistically significant relationship between the two variables. Table 2: Output of Pearson Product-Moment Correlation test for the raw data given in Table 1. Unemployment Rape Unemployment Pearson Correlation 1 -0.137 Sig. (2-tailed) 0.336 N 51 51 Rape Pearson Correlation -0.137 1 Sig. (2-tailed) 0.336 N 51 51 This can be shown graphically by examining the line of best fit through the data points (Figure 1). As can be seen, the relationship between the two variables is weak. Figure 1: Data from Table 1 plotted on a scatter plot, with a line of best fit added. As a consequence of these results, the null hypothesis, that there is no statistically significant relationship between the rates of rape and unemployment, is not rejected. These results mean that as the unemployment rate increased, the number of rapes per 100,000 people decreased. However, this trend was slight, and not significantly significant. Consequently, the most reliable and accurate conclusion from the data would be that the unemployment rate had no significant effect on the number of rapes. Policy and Research Effects The results of this study, as well as its policy implications, indicate little to no effect of unemployment rate on rape. Furthermore, the small effect that was observed was in the opposite effect than was predicted. It suggests that rape and unemployment are distinct factors that do not influence one another. As a consequence, and if the results of more rigorous studies support this finding, policy should not look at trying to decrease the amount of rape by decreasing unemployment, but rather work at developing solutions to both problems independently. The effect of unemployment on rape have been the focus of many studies, however the results from the current paper indicate that there may not be an effect. I would recommend that studies examine this trend over a longer period of time, as only one year was considered for this paper. In addition, it would be worth widening the focus to include other factors that may correlate with levels of rape, such as the prevalence of internet pornography, which has been suggested to decrease rape due to providing release for potential offenders. Other factors that have also been suggested to affect the amount of rape occurring include: the level of crime overall, availability of alcohol and drugs, poverty and potentially factors such as violence and sex on television and mass media. Limitations of the Analysis For this study, information was collected from databases with one measure for each state for the year 2010. One limitation of the analysis is that it considers only data from a single year, and as a consequence it does not take into account potentially confounding factors. For example, the economy was going through hard times in 2010, and as a consequence even those employed did not have much money. This could reduce the effect that unemployment had on rape, and result in no observed trend. A Pearson Product-Moment Correlation test looks for whether two variables affect one another, but it does not test the direction of this effect. While it is relatively intuitive that rape does not cause changes in unemployment for the most part, one factor that is not considered by this test is whether the two factors are affected by an external variable. For example, as considered above, the economic climate may have an effect on both variables. If the economic environment is harsh, then businesses often need to cut back on their expenditure, and as a consequence there are fewer jobs. Likewise, a difficult economic environment can mean that the level of expendable income for most people decreases, regardless of whether they are employed or not. Consequently, people may become stressed and have an increased tendency to release this stress in various manners, which can include rape. Thus the economic environment could potentially be responsible for both increased levels of rape and decreased availability of jobs. An additional problem with this test is the data that are used to measure rape. The FBI crime statistics considered only forcible rape, which included sexual assault that did not involve rape. In addition, statutory rape was not counted in this definition. Consequently, this is not a true test of the effect of unemployment on rape. Also, the data are an estimate of the level of forcible rape which is occurring based on the level of reported rape. Thus, it is difficult to know whether this is a true indication of the amount of rape that is occurring, as rape is the most underreported crime (Alvarez & Bachman, 2007). There are many different factors that can influence whether a female reports rape, such as how close they are to the perpetrator, their economic status and where the rape occurred. Other factors include embarrassment, shame or the fear of the perpetrator taking revenge if the crime is reported. Because of this the amount of rape that is occurring can only ever be an estimate, and there is no way of being certain how accurate the estimate is. Because of these limitations, it is difficult to generalize from the results of this study and provide estimations of the general effect that unemployment has on rape. The fact that there was no replication (one value for the year for each state) and that only a single year was considered in the analysis means that the results are not reliable. They could easily be the effect of the year that the data were analyzed for rather than rape and unemployment. Overcoming limitations A more effective study would be to look at the same information over a period of at least five years. This would allow for trends to be analyzed over a longer period of time. It would be most effective to determine the Pearson coefficient for each of the years, and then examine whether the correlation changes over time. In addition a Pearson coefficient test over the whole period would determine whether there was any general correlation between rape and unemployment. Because a longer time period is examined, this would reduce the chance of confounding factors affecting the results. Another way of making the study statistically sound would be to use more data. The information used for this paper was gathered from two databases, both of which gave information as an average for the year. However, original source of the data in both cases contained values for each of the twelve months. Using data from the months individually instead of an average, and over multiple years would give the statistical test more power (Ellis, 2010). Consequently, it would more effectively be able to test for differences between the two factors. In addition, an ANOVA test could be used, taking the years as a control variable (Anderson, 2001).This type of test would be able to analyze the direction of the interaction, as well as removing the potentially confounding effect of year. Using the estimated level of rape rather than reported level helps to eliminate some of the confounding factors connected to when people report rape and when they do not. For the results from this study to be adequately able to be compared to results from other studies, a comparative measure of rape needs to be used. It is possible that no correlation was seen between rape and level of unemployment because of the definition of rape that is used in this paper. For example, one particular type of rape, such as statutory, may be affected more by unemployment than other factors. Because of the definition of rape used in this paper, this would not be apparent. To reduce this limitation one of two steps could be taken. The first is to use an alternative set of data that has a definition of rape that is similar to that of other studies. The second is to redefine what is being tested, rather than testing the effect of unemployment on rape, the study could instead test the effect of unemployment on violent sexual assault. This involves no actual change to the study, instead simply a change to the terms being used, as rape is not strictly accurate. There are many other factors may potentially influence the level of rape than unemployment. Correlation tests could be used to test these variables, but this would give little indication of how these factors interact with each other. A more effective analysis would be to examine information for multiple factors at the same time, testing which have a significant effect and which do not. Data could be gathered from similar databases that were used for this analysis. A multifactor ANOVA could be use to determine which factors have a significant effect on the level of rape and which do not, as well as the importance of each factor (Anderson, 2001). This would allow for much clearer elucidation of cause and effect. References Alvarez, A., & Bachman, R. (2007). Violence: The enduring problem. Thousand Oaks: Sage Publications, Inc, Anderson, N. H. (2001). Empirical direction in design and analysis. Mahwah: Lawrence Erlbaum, 240-260. Ellis, P. D. (2010). The Essential Guide to Effect Sizes. New York: Cambridge University Press, 58. Lehman, A. (2005). JMP for basic univariate and multivariate statistics: a step-by-step guide: SAS Publishing, 105-120. Read More