Linear Regression Model

1364 Words3 Pages

There has been increased interest in the deaths of United States citizens while in the custody of officers of the peace. The news seems to purport epidemic levels of arrest related deaths, suggesting a dramatic increase in the number of individuals killed by police officers. Furthermore, the news reports appear to insinuate that minority groups disproportionately bear the brunt of the affects of this dramatic escalation. These veiled assertion are not borne out by the data, however the illumination of this issues does warrant further consideration. The foundational question this article seeks to address is this; is the frequency of arrest related deaths contingent upon some external aggravating factor? Alternatively, is the frequency of arrest-related …show more content…

A linear regression model will be utilized for this analysis. Linear regression models assists researchers in understanding the direction of an assumed linear relationship between variables. Simply, it suggests if the relationship is positive or negative. One hypothesis was tested. This hypothesis states that states with higher frequencies of homicides will be more likely to experience increased arrest related deaths. The dependent variable is arrest related deaths, while the independent variable is frequency of homicides. The data expresses that for each additional one unit increase in homicides arrest related deaths increase by .023. We are assured of this value due to a significance score of .000, well below our threshold of .05. Therefore, this hypothesis is …show more content…

However, there are some analyses that were ran that warrant stating. A linear regression was utilized comparing arrest-related deaths and percentage of African Americans within a state. The results were inconclusive due to a significance score well above the maximum threshold of acceptance. However, this raises a question. If the sample size were the actual population, would that reduce the p-value to an acceptable level? Additionally, in retrospect, higher quality variables could have replaced many in this exercise. If this exercise were recreated, variables such as poverty levels within states and frequency of police encounters would be utilized. Nevertheless, there is solace in knowing the ancillary variables chosen exhibit no interconnection with arrest-related

Open Document