Screenreader Navigation - [ Skip to Content | Skip to Main Navigation ]



This page contains open access information and links to files that aid in political science research and data collection/processing.




Automating Clarify & Predicted
Values for
Chart Generation
The CLEAR-PLOT package includes a modified version of the CLARIFY program created by Tomz, Wittenberg and King (see CLARIFY: Software for Interpreting and Presenting Statistical Results, 2003), which was combined with the automated .do file featured in "Understanding Interaction Models" by Brambor, Clark and Golder.

The .ado and .do files allow users to create marginal effect/predicted probability plots for all models covered by Tomz et al.'s CLARIFY software. Users must replace the existing "simqi.ado" file with the provided file of the same name before running the .do file. See the instructions file (right) for more.

Included files are meant to be utilized on the Stata platform.
Instructions: Using 

Automated Automated CLARIFY,
Ex. basic
Automated CLARIFY,
Ex. Logit
Automated CLARIFY,
Ex. with
Modified simqi.ado
CLEAR-PLOT with Histogram The accompanying example .do file allows users to underlay a histogram plot behind the marginal effects/predicted probability plots detailed above. The example is for a OLS regression, but the code can be changed to cover all of the models featured in Tomz et al.'s CLARIFY software.

Note: the distribution of the variable of interest may lead to compression effects (see the .do file). Included files are meant to be utilized on the Stata platform.
CLEAR-PLOT with Rare Events Logit CLARIFY (Tomz, Wittenberg and King, 2003), was expanded to include rare events models (see King and Zeng 2001) with the creation of the ReLogit and ReLogitq packages.

The following files allow users to create marginal effect/predicted probability plots for rare events models by adding a new program (relogit_clear).

Included files are meant to be utilized on the Stata platform.
CLEAR-PLOT with Multiple Outcomes For models with multiple outcomes (mlogit, oprobit, ologit, and sureg) the current CLEAR-PLOT program only reports the last predicted outcome (Ex. In an mlogit model where y = 1, 2, or 3, CLEAR-PLOT would only report Pr(y=3)). Until the program is updated to record all outcomes, users can use this example to generate multiple predicted outcomes.
Automated CLARIFY, 
Creating Predicted Probability Graphs with Multiple Predicted
This Stata .do file allows users to generate marginal effect/predicted probability plots with multiple marginal effect/predicted probability lines in a single graph (and accompanying confidence intervals). Note to users: if the generated predicted values are similar, the graph will appear cluttered; in other words, the package is best utilized for highly contrasted scenarios.

The .do file utilizes the CLEAR-PLOT package (above), and requires the replacement of CLARIFY's simqi.ado with the one featured here.
Modified simqi.ado
Calculating Log-Likelihood, AIC, and BIC using RELogit RELogit is a Rare Events Logit model used when a binary dependent variable is inflated with an extremely high number of 1's or 0's (the original program for Stata and Gauss can be found here). This modified .ado file adds another command to Stata: relogitll. This modified program will both run a relogit model, and report the log-likelihood, AIC, and BIC when the user types "scalar list" after model estimation.
Instructions &

Generating Predicted Probability Graphs for Tobit Models
in Stata
These Stata .do files generate predicted outcomes for E[y|y>c], where c is the censor point, Pr(y>c) [rarely used], and E[y] in tobit models. There currently exists no automated means of generating these values, however these .do file allows users to adjust the censorship and specification values to manually generate the corresponding plots.
Further details are given in the .do files.
State Abbrivation Transformations (to FIPS) in Stata These Stata .do files transform various state abbrevation codes into FIPS (Federal Information Processing Standard) codes.


State Two-Letter Abbr.
(string) to
Full State Name (string)