Prediction and Explanation: Towards a Communication Study of Causal Representation
LI Xuelian LIU Dehuan
Author information+
Li Xuelian is a lecturer of School of Journalism and Information Communication, Huazhong University of Science and Technology. Email: lixuelian@hust.edu.cn.
Liu Dehuan is a professor of School of New Media, Peking University. Email: liudehuan@ vip.sina.com.
{{custom_zuoZheDiZhi}}
{{custom_authorNodes}}
{{custom_bio.content}}
{{custom_bio.content}}
Collapse
History+
Published
2023-07-23
Issue Date
2023-12-26
Abstract
Positive communication studies have been focusing on the explanation of media exposure and its effects, while less on the prediction of information dissemination or individual behaviors. Various modelling methodologies under deductive logic confound explanatory modelling with predictive modelling, affecting the accuracy, validity and reliability of causal inferences. The development of computational social science enacts changes in data analysis methods, and it also prompts researchers to focus on the epistemological distinction between explanation and prediction. We believe that a clear distinction between and effective integration of explanation and prediction could assist communication studies to discover better causal knowledge and yield scientific prediction. This paper firstly reviews the empirical research strategies in the field of computational social science to clarify the current status and the differences between the two methods employed in our discipline. Drawing on Pearl’s causal ladder, communication studies based on online data are further extracted into four layers: (1) correlational analysis; (2) intervention studies; (3) explanatory studies; and (4) counterfactual causal reasoning. This paper outlines the analytical obstacles researchers face in causal explanation and prediction based on online data, and makes specific suggestions for achieving effective integration of explanatory and predictive models in causal modeling. Through circular validation on four levels of analysis, we offer a better way to understand the laws of communication and human behaviors.
LI Xuelian LIU Dehuan.
Prediction and Explanation: Towards a Communication Study of Causal Representation. Chinese Journal of Journalism & Communication. 2023, 45(7): 157-176
{{custom_sec.title}}
{{custom_sec.title}}
{{custom_sec.content}}
References
{{custom_fnGroup.title_en}}
Footnotes
{{custom_fn.content}}
Funding
This paper is the result of the National Social Science Foundation of China entitled “Research on Establishing Omnimedia Communication System” (No. 20ZDA057).