The dataset "Econ172_Spring 2021_ProblemSet3_data.csv" is a partial extract of data from the actual project, where each observation (row) in the dataset represents one village [denoted by "vid") in Meatu district, Tanzania observed during one year ("year"), with years ranging from 1992 to 2002. In other words, this is panel data. The data have also been modied somewhat from the actual data. Variables in the dataset include: "witch_murders" (the number of witch murders that occurred in a particular village in a particular year), 1'oth_murders" [the number of non- witch murders), "any_rain" [an indicator variable for whether either a drought or ood occurred in a particular village in a particular year), "any_disease" [an indicator for whether a disease outbreak occurred, e.g., measles, cholera, etc.), "famine" [an indicator for whether there was an extreme shortage of food in a particular village in a particular year), "educat" [average years of schooling in the village, as measured in 2002), and "trad_relig" [the proportion of households in the village practicing traditional religions, also in 2002). a) Start by constructing a variable for the total number of murders in a village in a given year, namely, witch murders plus non-witch murders. Then, create a table of summary statistics, including the mean, standard deviation, minimum value, maximum value, and number of observations, using the descriptive commands in R you have learned in section and used in previous problem sets. For example, \"mean[violence$any_rain) " where "violence" is what we called our data set and we care about computing the mean of the "any_rain" variable. [You can also use Stargazer to generate summary statistics table; see ??stargazer for details.) Do this for all variables in the dataset and discuss any noteworthy patterns, paying particular attention to the murder and rainfall variables. [1 point] b) Now consider the causes of the non-witch murders in the village [\"oth_murders")., Regress the number of non-witch murders [in a village in a particular year) on the indicator for whether a drought or ood occurred in that year. Interpret the relationship. Make sure that error terms are allowed to be correlated (\"clustered") across years for the same village using the lfe package and the felm command, as discussed in section. Remember that to cluster standard errors, you run "felm[y~x1+x2 |0|O| j, data=violence)", where j is the variable by which you want to cluster and "violence" is the name of your dataset [see ??felm for details). In a second regression, also include average years of schooling and the proportion of households practicing traditional religions as additional explanatory variables in the regression [while still clustering the error terms at the village level), and again present and interpret the results. [1 point] c) Next consider the effect of disease epidemics on the number of non-witch number of murders in the village. Run the same two regressions as in part "b" but replace the indicator for whether a drought or ood occurred with the indicator for whether a disease outbreak occurred. Present the results, and contrast them to the results for rainfall shocks