Prints an aggTrees.inference object.
# S3 method for aggTrees.inference
print(x, table = "avg_char", ...)aggTrees.inference object.
Either "avg_char" or "diff", controls which table must be produced.
Further arguments passed to or from other methods.
Prints LATEX code.
A description of each table is provided in its caption.
Some covariates may feature zero variation in some leaf. This generally happens to dummy variables used to split some
nodes. In this case, when table == "avg_char" a warning message is produced displaying the names of the covariates
with zero variation in one or more leaves. The user should correct the table by removing the associated standard errors.
Compilation of the LATEX code requires the following packages: booktabs, float, adjustbox,
multirow.
Di Francesco, R. (2022). Aggregation Trees. CEIS Research Paper, 546. doi:10.2139/ssrn.4304256 .
# \donttest{
## Generate data.
set.seed(1986)
n <- 1000
k <- 3
X <- matrix(rnorm(n * k), ncol = k)
colnames(X) <- paste0("x", seq_len(k))
D <- rbinom(n, size = 1, prob = 0.5)
mu0 <- 0.5 * X[, 1]
mu1 <- 0.5 * X[, 1] + X[, 2]
Y <- mu0 + D * (mu1 - mu0) + rnorm(n)
## Construct sequence of groupings. CATEs estimated internally,
groupings <- build_aggtree(Y, D, X, method = "aipw")
## Analyze results with 4 groups.
results <- inference_aggtree(groupings, n_groups = 4)
## Print results.
print(results, table = "diff")
#> \begingroup
#> \setlength{\tabcolsep}{8pt}
#> \renewcommand{\arraystretch}{1.2}
#> \begin{table}[b!]
#> \centering
#> \begin{adjustbox}{width = 1\textwidth}
#> \begin{tabular}{@{\extracolsep{5pt}}l c c c c}
#> \\[-1.8ex]\hline
#> \hline \\[-1.8ex]
#>
#> & \textit{Leaf 1} & \textit{Leaf 2} & \textit{Leaf 3} & \textit{Leaf 4} \\
#> \addlinespace[2pt]
#> \hline \\[-1.8ex]
#>
#> \multirow{3}{*}{GATEs} & -1.456 & -0.251 & 0.053 & 1.452 \\
#> & [-1.783, -1.129] & [-1.013, 0.511] & [-0.249, 0.355] & [ 1.087, 1.817] \\
#> & \{NA, NA\} & \{NA, NA\} & \{NA, NA\} & \{NA, NA\} \\
#>
#> \addlinespace[2pt]
#> \hline \\[-1.8ex]
#>
#> \textit{Leaf 1} & NA & NA & NA & NA \\
#> & (NA) & (NA) & (NA) & (NA) \\
#> \textit{Leaf 2} & 1.20 & NA & NA & NA \\
#> & (0.009) & ( NA) & ( NA) & (NA) \\
#> \textit{Leaf 3} & 1.51 & 0.30 & NA & NA \\
#> & (0.000) & (0.469) & ( NA) & (NA) \\
#> \textit{Leaf 4} & 2.91 & 1.70 & 1.40 & NA \\
#> & (0.000) & (0.000) & (0.000) & (NA) \\
#>
#> \addlinespace[3pt]
#> \\[-1.8ex]\hline
#> \hline \\[-1.8ex]
#> \end{tabular}
#> \end{adjustbox}
#> \caption{Point estimates and $95\%$ confidence intervals for the GATEs based on asymptotic normality (in square brackets) and on the percentiles of the bootstrap distribution (in curly braces). Leaves are sorted in increasing order of the GATEs. Additionally, the GATE differences across all pairs of leaves are displayed. $p$-values testing the null hypothesis that a single difference is zero are adjusted using Holm's procedure and reported in parenthesis under each point estimate.}
#> \label{table:differences.gates}
#> \end{table}
#> \endgroup
#>
print(results, table = "avg_char")# }
#> \begingroup
#> \setlength{\tabcolsep}{8pt}
#> \renewcommand{\arraystretch}{1.1}
#> \begin{table}[b!]
#> \centering
#> \begin{adjustbox}{width = 1\textwidth}
#> \begin{tabular}{@{\extracolsep{5pt}}l c c c c c c c c }
#> \\[-1.8ex]\hline
#> \hline \\[-1.8ex]
#> & \multicolumn{2}{c}{\textit{Leaf 1}} & \multicolumn{2}{c}{\textit{Leaf 2}} & \multicolumn{2}{c}{\textit{Leaf 3}} & \multicolumn{2}{c}{\textit{Leaf 4}} \\\cmidrule{2-3} \cmidrule{4-5} \cmidrule{6-7} \cmidrule{8-9}
#> & Mean & (S.E.) & Mean & (S.E.) & Mean & (S.E.) & Mean & (S.E.) \\
#> \addlinespace[2pt]
#> \hline \\[-1.8ex]
#>
#> \texttt{x1} & -0.018 & (0.073) & -0.041 & (0.188) & 0.001 & (0.078) & -0.029 & (0.082) \\
#> \texttt{x2} & -1.056 & (0.044) & -0.299 & (0.012) & 0.207 & (0.018) & 1.286 & (0.046) \\
#> \texttt{x3} & 0.043 & (0.075) & -0.136 & (0.198) & -0.156 & (0.079) & 0.050 & (0.085) \\
#>
#> \addlinespace[3pt]
#> \\[-1.8ex]\hline
#> \hline \\[-1.8ex]
#> \end{tabular}
#> \end{adjustbox}
#> \caption{Average characteristics of units in each leaf, obtained by regressing each covariate on a set of dummies denoting leaf membership using only the honest sample. Standard errors are estimated via the Eicker-Huber-White estimator. Leaves are sorted in increasing order of the GATEs.}
#> \label{table:average.characteristics.leaves}
#> \end{table}
#> \endgroup
#>