We demonstrate our framework through a proof-of-concept empirical design drawing on administrative elites in North China since the formation of local Communist Party organizations. Our core personnel data was collected systematically from the Materials on the organization of Party and government institutions from 1921 through 1987. Such complete coverage allows us to gather detailed career paths of all officials appearing in the Materials and model their mobility and network structure.
We reconstruct the exhaustive network of administrative and political elites in China. We draw on the rich historical materials to trace mobility patterns among political elites and identify the structure and the extent of their networks. In this context, Artificial Intelligence (AI) tools have been widely touted as a revolutionary breakthrough for research in the social sciences, particularly since the introduction of large language models (LLMs) that allow researchers to query and analyze large corpora of textual data.
To analyze such data efficiently, we developed an LLM workflow that streamlines data processing from printed materials to structured datasets suitable for statistical modeling and/or further LLM queries. Our engagement with LLMs is centered on the Retrieval-Augmented Generation (RAG) framework in order to eschew ‘hallucinations’ and carry out queries based on prescribed, reliable data sources. Our framework also advances the principles of open, replicable, and transparent research. To do so, we develop mechanisms to validate the quality of LLM outputs.
We present preliminary findings for the pre-cultural revolution era regarding the pace of recruitment, the ethnic and gender composition of the fledgling elite, as well as an assessment of elite-mobility patterns driven by opportunities that arise and close due to changes in the ratio of new entrants to departures as revealed by our extensive dataset of institutions and their leaders.
