Existing statistics are also based on the data that someone else has collected. Unlike secondary data analysis, however, where you get the raw data and compute the statistics you are working with the statistics produced by someone else. Existing statistics are less likely to come from surveys, and more likely to come from the data collection of governmental agencies. The department of education for example does not have to survey people in order to determine how many students are enrolled in California schools these data are part of the department’s bureaucratic record-keeping. Sometimes you might use those statistics to produce new statistics. For instance, you might gather existing statistics on high school exit exam scores and school spending in order to determine how much money a state needs to spend on education in order for most students to reach a particular level of achievement on the exams. Other times, you might conduct a new analysis not by producing new statistics, but by bringing together statistics that have not previously been analyzed in relation to one another. For example, you might take the existing statistics on the number of violent offenses and property crimes, county unemployment rates,median home prices, school enrollment rates, and income statistics to investigate the relationship between socioeconomic conditions in an area an the crime rate. In order for the use of existing statistics to be considered research by our definition, however, you must do more than simply report the already existing statistics; you must produce a new analysis, even if you did not gather the data yourself.