# Software

`dyadclust`

: Estimates cluster-robust standard errors for dyadic data using multiway decomposition as described in Aronow et al. (2015). The package also supports parallel computing using`parallel`

.

To install the package and update it the following command can be used in Stata:

`net install dyadclust, from (https://raw.githubusercontent.com/cfbalcazar/dyadclust/main/dyadclust/) replace force all`

`nopomatch`

: Estimates the decomposition of a binary treatment using non-parametric matching as described in Nopo (2008). I worked with this methodology many years ago and made an extension of the do-file that allows the user decomposing the difference in means for binary outcomes.

To install the package and update it the following command can be used in Stata:

`net install nopomatch, from (https://raw.githubusercontent.com/cfbalcazar/nopomatch/main/nopomatch/) replace force all`

`represent`

: Computes and provides diagnostics over the effective regression weights as defined in Aronow and Samii (2016). Causal effects estimated via multiple regression differentially weight each unit’s contribution. The “effective sample” that regression uses to generate the estimate may bear little resemblance to the population of interest, and the results may be non-representative in a manner similar to what quasi-experimental methods or experiments with convenience samples produce. This command allows the user to explore this problem.

To install the package and update it the following command can be used in Stata:

`net install represent, from (https://raw.githubusercontent.com/cfbalcazar/represent/main/represent/) replace force all`

`fwlgraph`

: Allows one to visualize the relationship between a dependent variable and one independent variable of interest. The approach is motivated by the Frisch-Waugh-Lovell theorem - one of my favorite theorems!. It is quite useful to visualize what the regression with control variables is actually representing, which can help one determine whether a linear or other specification might make sense, as well as visually identifying outliers.

To install the package and update it the following command can be used in Stata:

`net install fwlgraph, from (https://raw.githubusercontent.com/cfbalcazar/fwlgraph/main/fwlgraph/) replace force all`