• Memo Proposal.This is a short summary of the question you wish to explore. It is sort of draft of yourmemo; use the panel data provided in the Memo Datasets file. Each observation shouldinvolve measurements taken at (more or less) the same time.The proposal should establish one variable that’s your dependent variable and three otherindependent variables that should cause that dependent variable. You must include a logicalexplanation for why each of the three independent variables should cause the dependentvariable. This should be the bulk of the proposal.The independent variables should be very different from one another. Don’t use bothmurder rates and assault rates, for example. Be creative!Avoid making a state’s or country’s population a variable. That variable is there so you canadjust other variables for population. While population can be an appropriate dependent orindependent variable, it’s tricky to do right and best be avoided for the purposes of thiscourse.Avoid obvious connections (e.g. the birth rate causing the death rate).• Tables.Using the different variables you will use for your analysis, provide three tables:The descriptive statistics. Include the following information for each variable:number of observations, arithmetic mean, median, standard deviation, the minimum, andthe maximum. Make sure to put the information into a table (label row 1 “Variable”, row2 “Observations,” etc).Regression analysis. Include the relevant parts of the regression: coefficients, pvalues, R2, adjusted R2, significance F, and the number of observations.Correlation table. Your paper will need a correlation table of your independentvariables to show there’s no multicollinearity. Include a correlation table in this and thefinal assignment.The tables should be easy to read. Variable names should be brief (you’ll explain them insufficient detail in the prose of the memo), commas should be used for large numbers, useno more than three decimal places, the table should be single spaced, etc.All of these tables will be incorporated into your memo; I suggest you make them in Wordand submit the Word file.• Memo.By now, your question should have turned into an argument. Instead of asking “Whatcauses crime?” you should be arguing “X causes crime.” This argument will the basis ofyour memo.Your memo will involve using panel data; it will not be a time series. That means that eachobservation will be a different state or country with measurements taken at (more or less)the same time.Be sure to:o Discuss why your argument is worth makingo Explain why you think this causation existso Include descriptive statistic, as aboveo Support your argument with a regression including implications (“my regressionsuggests that for every additional year of school, criminal activity falls by X”). Alsoinclude a table of the critical information from your regression.▪ Remember to use logical rounding. For example, p values should berounded to four decimal places.o Discuss any weaknesses in your approach, by which I mean to identify aconfounding variable. It’s possible that there is no confounding variable in youranalysis in which case you should identify some other major problem but I’ve foundthe vast majority of analyses could reasonably have a confounding variable.Solving such complications is beyond the scope of this course though it’s importantthat you discuss them.o Proofread!o Remember, you are limited to three (3) pages. This includes any tables you use butdoes not include a works cited page.
