The first stage is called pre-sampling. What pre-sampling involves is the partitioning of records and the calculation of the right sample size. A study was conducted where 12 of the largest municipal police departments in the state overtook the population of this study. The population of all these cities combined came out to 285,898 and the number of incidents that were reported and recorded by police was 31,003. The total number excludes incidents that are considered “unfounded”. When an “unfounded” report is taken it’s usually after an initial investigation reveals that the crime in fact did not occur. The total number of incidents also does not include “general incident” reports that are written when police take action and chose to record it, but isn’t related to any certain crime. Each police department were put into categories and selected at random and 3 were picked to participate. A table describes the specific three police departments that were chosen to participate and the 15 categories, including unfounded and general incident reports and the estimated sample size for each category and participating police agency. In order to establish point estimates of statistical accuracy, a sample of records from each of the 15 reporting categories was drawn. Prior to drawing the sample, however, it was necessary to decide on a desired level of …show more content…
With the help of police agency representatives, hard copies of all reports in the sample were manually pulled and read for accuracy. In reviewing each record, definitions provided by the UCR program officials at the FBI were applied.6 A systematic procedure for the assessment of each record was established to ensure a high level of reliability between reviewers.7 For each record, a system of verifying the inter-rater agreement among reviewers was constructed by having multiple reviewers examine those records with a suspected classification error. When a classification error was suspected in a given record, at least three members of the research team (including the first author) reviewed the report to either confirm the classification error or reaffirm the accuracy of the original classification. If it was difficult to determine whether the report was classified correctly, the record was judged not to contain a classification error. Essentially, this procedure gave the reporting officer the benefit of any doubt in the absence of clearer information; thus, the research team relied on the judgment of the officer on the scene when crime classification was difficult to determine because the narrative was vague. This rarely happened in the current study.” Last but not least calculating statistical accuracy, this whole study comes down to this, These results that were conducted are followed by an