Because the interpretation of the effects in limited dependent models differs from that in linear models, we followed Wiersema and Bowen’s (2009) procedure to test the true marginal effect over all values of the model’s variables. In the patent applications model, the true marginal effects range from 20.04 to 20.47, with corresponding z-values from 25.06 to 25.77; in the patent infringement model, the true marginal effects range from 0.00 to 0.05, with corresponding z-values from 1.57 to 3.54. Because the true effects do not change signs, we can rely on the true marginal effect at the mean level of all model variables as a summary indicator. The results (patent applications: mean effect 5 20.15, p , .001; patent infringement: mean effect 5 0.01, p , .001) provide strong support for Hypotheses 1a and 1b.

The true interaction effects range from 0.00 to 0.03 (z 5 [2.68, 3.34]; mean effect 5 0.01, p , .01) in the patent applications model and from 20.01 to 20.00 (z 5 [21.12, 23.18]; mean effect 5 20.001, p , .05) in the patent infringement model. These results fully support Hypotheses 2a and 2b.

参考文献：The Impact of Communist Ideology on the Patenting Activity of Chinese Firms. AMJ. -Link-

Catalog

Last update: 2025-7-29

Hi, welcome to my blog!

本人设计了一款商科外文期刊信息查询小软件，可以查询期刊基本信息、英国ABS评级信息、澳洲ABDC评级信息、JCR分区及影响因子信息、中科院分区信息、FMS评级信息等，只需输入期刊名或其ISSN值即可，支持模糊搜索。目前已经更新至2025年中科院分区，可以点击此处下载、体验！

本人构建了一个基于 Python 的事件研究法（Event Study）计算与分析项目，用于计算上市公司在特定事件发生前后的异常收益率（AR）和累积异常收益率（CAR），并进行统计检验。本项目支持多进程并行计算，能够高效处理大规模的事件数据。点击此处可查看详情