Suggested Stata Downloads

Suggested Stata Downloads

Here are some of the programs that I personally like and often use in my classes or research. My handouts often assume/require that these routines be installed. I will no doubt add to this list from time to time. From within Stata, you can use the findit command to locate and install programs, e.g. findit pcorr2.

Program Description

quest7 These are the Stataquest 7 menus and routines. These are no longer as necessary as they were in earlier versions of Stata, but I still find them helpful. Note that, once you find quest7, you need to install quest1, quest2, and quest3. Once you have installed these, you may want to add the command quest on to your profile.do file, so that the Stataquest menus are automatically available to you (they will be under the User menu in Stata 8 & 9). To even find these files, you must first type
net from http://www.stata.com/quest7

fre Excellent program for getting frequency counts on multiple variables, including information on missing values.

zscore and/or center Compute z-score transformations of variables.

tab_chi This is a collection of routines that do various nonparametric tests. chitest and chitesti carry out chi-square tests for univariate frequency distributions. tabchi and tabchii produce two-way tables of frequencies with chi-square and residuals, including observed and expected frequencies and raw, Pearson and adjusted residuals.

ipf Iterative Proportional Fitting. Used for categorical data analysis. This program computes chi square goodness of fit statistics for loglinear models, e.g. the models of independence and conditional independence, etc.

xi3 Powerful enhancement to Stata's xi command. Lets you easily do dummy variable coding, effect coding, and several other types of sometimes-useful contrast coding.

pcorr2 Reports both partial and semipartial correlations. Enhanced version of Stata's pcorr command. Not necessary if you have Stata 11 or higher.

collin Multicollinearity diagnostics. More informative than Stata's vif command and does not require that you run a regression first.

extremes Lists the extremely high and extremely low values of a variable. Can be useful for detecting outliers. Has some of the same functionality as SPSS's EXAMINE command.

spost9 Several commands for the post-estimation interpretation of regression models. Very useful for advanced work. J. Scott Long and Jeremy Freese use these commands in their book, Regression Models for Categorical Dependent Variables Using Stata

lrdrop1 Drops 1 variable at a time from a logistic regression, giving you Likelihood Ratio tests for each coefficient. These can be more accurate than the Wald statistics that are reported by default.

bicdrop1 Similar to lrdrop1, except it gives you BIC tests for each coefficient.

gologit2 & oglm These are programs for the analysis of ordinal data. Since I wrote them, there is a fair chance that they will get used in my advanced courses!

mfx2 & margeff These programs make it easy to compute marginal effects after multiple-outcome models like gologit2 and ologit. They also make it easy to format the results with commands like outreg2, estout and esttab. margeff is faster and more powerful but mfx2 works with more programs. In Stata 11 and higher, the margins command is often but not always superior.

estout/esttab, outreg2 These are very useful routines for formatting the output from estimation commands, e.g. you can use them to show the results from a series of models. You really really really should learn how to use one of these, unless you like recopying thousands of numbers into tables.

Program	Description
quest7	These are the Stataquest 7 menus and routines. These are no longer as necessary as they were in earlier versions of Stata, but I still find them helpful. Note that, once you find quest7, you need to install quest1, quest2, and quest3. Once you have installed these, you may want to add the command quest on to your profile.do file, so that the Stataquest menus are automatically available to you (they will be under the User menu in Stata 8 & 9). To even find these files, you must first type net from http://www.stata.com/quest7
fre	Excellent program for getting frequency counts on multiple variables, including information on missing values.
zscore and/or center	Compute z-score transformations of variables.
tab_chi	This is a collection of routines that do various nonparametric tests. chitest and chitesti carry out chi-square tests for univariate frequency distributions. tabchi and tabchii produce two-way tables of frequencies with chi-square and residuals, including observed and expected frequencies and raw, Pearson and adjusted residuals.
ipf	Iterative Proportional Fitting. Used for categorical data analysis. This program computes chi square goodness of fit statistics for loglinear models, e.g. the models of independence and conditional independence, etc.
xi3	Powerful enhancement to Stata's xi command. Lets you easily do dummy variable coding, effect coding, and several other types of sometimes-useful contrast coding.
pcorr2	Reports both partial and semipartial correlations. Enhanced version of Stata's pcorr command. Not necessary if you have Stata 11 or higher.
collin	Multicollinearity diagnostics. More informative than Stata's vif command and does not require that you run a regression first.
extremes	Lists the extremely high and extremely low values of a variable. Can be useful for detecting outliers. Has some of the same functionality as SPSS's EXAMINE command.
spost9	Several commands for the post-estimation interpretation of regression models. Very useful for advanced work. J. Scott Long and Jeremy Freese use these commands in their book, Regression Models for Categorical Dependent Variables Using Stata
lrdrop1	Drops 1 variable at a time from a logistic regression, giving you Likelihood Ratio tests for each coefficient. These can be more accurate than the Wald statistics that are reported by default.
bicdrop1	Similar to lrdrop1, except it gives you BIC tests for each coefficient.
gologit2 & oglm	These are programs for the analysis of ordinal data. Since I wrote them, there is a fair chance that they will get used in my advanced courses!
mfx2 & margeff	These programs make it easy to compute marginal effects after multiple-outcome models like gologit2 and ologit. They also make it easy to format the results with commands like outreg2, estout and esttab. margeff is faster and more powerful but mfx2 works with more programs. In Stata 11 and higher, the margins command is often but not always superior.
estout/esttab, outreg2	These are very useful routines for formatting the output from estimation commands, e.g. you can use them to show the results from a series of models. You really really really should learn how to use one of these, unless you like recopying thousands of numbers into tables.