Through A/B Experiment and data analysis, an approach to validating the effectiveness of state level firearm provisions with respect to murder and population data pulled from the U.S. Census organization.
Firearms continue to be a leading method of murder across the United States. Data shows us the over 73% of murders commited in the year 2016 were linked to firearms.
Hypothesis: A 'Mental Health' provision reduces the amount of firearms murders Mental Health Provision: Background checks for private sales are required for at least some weapons and there is an explicit requirement for search of mental health records.
A decrease in firearm murder count is expected as the states apply the 'Mental Health' provision.
Compare U.S. statistics involving firearm murders, population data, and firearm provisions enacted on the state level. Since firearms provisions are nonstandarized bewtween all states, select two of the most similar states based on number and type of provivsions applied respectively, such that neither state applies the Mental Health provision. Introduce the Mental Health provision to one of state's legislature to be in affect beginning the next year. Observe and record the amount of murders linked to firearms in the following year. Adjust each state's annual the murders quantities based on population growth. Compute and test each state's ratio of annual firearm murder to total murder count with a two proportion Z-test in order to discern in the case of a difference of annual firearm murder, if the difference is statistically significant. Select a confidence level of alpha = .05 to be the indicator for successful reduction in murder total.
Does legislation help decrease the number of firearms fatalities?
Is the Mental Health Provision effective in reducing firearms fatalities?
Arizona 2016 Firearm Provision and Murder Data
Virginia 2016 Firearm Provision and Murder Data
Number of state provisions applied Annual state firearms murders Annual state total murders Annual state population change Two proportion Z-test p-value