LongTermCausalInference

Codes for the paper "Long-term causal inference under persistent confounding via data combination"

Description of files

Code for numerical experiment with GAIN dataset

In the ./real_data/ folder, it contains a Readme.txt which describes how to use the codes in that folder to generate Table 1 in the main text as well as Tables 2, 5-9 in the supplement.

Code for singular value decompositions in linear and discrete models

In the ./auxiliary/ folder, it contains a Readme.txt which describes how to generate Tables 1, 4-5 and Figure 3 in the supplement.

Code for simulation study

In the ./minimax/ folder, it contains a Readme.txt which describes how to generate Table 2 in the main text as well as Table 10 in the supplement.

Reference

Imbens, G., Kallus, N., Mao, X., & Wang, Y. (2022). Long-term causal inference under persistent confounding via data combination. arXiv preprint arXiv:2202.07234