Interdisciplinary collaboration improves the accuracy of social science prediction
August 02, 2023 09:53 Source: "China Social Sciences" August 2, 2023 Total Issue 2704 Author: This reporter Liu Yuwei

  

  ■ Igor Grossmann (Igor Grossmann) respondents/confession

For a long time,Social science researchers to collect and analyze data to reveal the patterns and rules behind social phenomena,and propose solutions for social problems。Also because of this,Cause and effect inference,Find the interrelationship between the input conditions (reasons) of the model (reasons) and the output action (result),Occupy the mainstream in social science research。Since the new crown epidemic,Scholars from various countries have published guessing the "new normal" of the world after the epidemic,Experts studying human behavior and human society began to express their opinions on epidemic and social impacts,and propose to govern countermeasures。Soon,These prospects have triggered the concept of the academic circles for "prediction" in the field of social sciences、Discussion of Application and Value。Can social science accurately predict social changes? What are the problems in the prediction results that exist in social science experiments?,A reporter interviewed Igor Grossmann, an associate professor at the Department of Psychology at the University of Waterloo, Canada.。

  bet365 best casino games

Current,The world is in a big change in a hundred years,Global uncertainty is increasing。Social science's precise prediction of future changes,It will improve people's quality of life,Promote the development of the country。In the interview,Grossman lists the precise predicted macro utility from the following aspects。At the social level,Accurate prediction given by social sciences can promote policy formulation、Prevent public health risks and respond to disasters,Enhanced the ability of social disasters,Accelerate recovery speed。Prediction with high accuracy can also cultivate the public’s trust in social sciences,Help them make a wise decision on personal behavior or choose,to improve the Bet365 lotto review well -being of individuals and communities。From a fiscal perspective,Accurate prediction can guide resource allocation,minimize waste expenditure,Stimulate economic growth。For example,Accurate prediction of changes in the employment market can achieve targeted education investment,Provide necessary skills training for future labor,Realize economic prosperity。other,Prediction accuracy of improving public health and other fields can also help improve prevention、Monitor and control the effect of the outbreak of the disease,Save huge funds for the country。

In order to test the ability and accuracy of social science researchers to predict social change,Groasman and its team have formed the "Forecast Collaborative Organization",The largest forecasting activity in the history has been carried out。From the line for science、Researchers in the field of social sciences and data science participated in two "prediction competitions"。They are satisfied with life、Social media emotion、Gender -Occupation and Race prejudice、Ideological preference、Top 12 common social science research fields, including subjective happiness, predict the future development trend of American society。In the first game,86 teams predict the historical data of 39 months (January to March 2020), the value of the next 12 months (May 2020 -April 2021)。After 6 months,120 teams conducted a second round of predictions based on historical data of 45 months after the update (January 2017 -September 2020)。

Groasman and others discovered,Social science researchers perform poorly in the real world during predicting the epidemic period。The results of the competition show,In most fields,Social science researchers predict the accuracy of social changes or are similar to the prediction results of non -expert groups,Either worse。At the same time,Social science researcher's prediction is often worse than predictions produced by simple statistical models。Groasman told reporters,Social science researchers have poor performance in predictions. There are many reasons,For example, lack of training for calculating models,Ignoring bet365 best casino games time sequence data,Confusion analysis level,, etc.。other,Some researchers are habitually adopted by personal preferences、"Pet theory" with only one or two key factors,Ignoring the real world is determined by the common role of multiple factors。Other people only put forward specific judgments and instructions on the specific change among,Lack of "risk" prediction ability that exceeds narrative explanations。again,Expert Excessive obsession Laboratory Experimental Law,Create experimental conditions in the control environment,Method of separating specific variables,Accurate measurement and observation of the effects of specific factors on the results。Considering the control environment cannot accurately reflect the complexity and dynamicity of the world,Grossman suggested that researchers attach great importance to experimental methods with higher ecological validity。In the fields of psychology and social sciences,Ecological validity refers to the extent that the research results can be promoted and applied to real life situation。

Grossman pointed out,Social science researchers are currently predicted,Compared with their authoritative interpretation of different social phenomena, there is a big gap。but,It should not be determined that it is meaningless to predict in the field of social sciences。At the same time,Scientific groups should be stricter、More hard to measure and understand the ability and limitations of social science prediction in the future。Groasman proposes three suggestions for improving the accuracy of social science researchers to predict the accuracy of social science researchers based on the results of the "Forecast Collaborative Organization" competition.。First,Let experts specializing in a certain field predict the development trend of this field,to get more accurate predictions。For example,If someone is an expert in depression,He/she should be better at predicting the social trend of depression。​​Second,The prediction effect of the research team of multi -disciplinary cross -cooperation is better than a single discipline group。In terms of experimental operation,Those groups that use simple models bet365 best casino games and historical prediction data,Its prediction effect is often better than a group that does not adopt such models or data。third,Social science researchers can also from the aspects of developing reliable data sources and extending forecast time spans,Further improving the accuracy of social predictions。

  Organic combination of explanation and prediction

  

  ■ Internationally need a large -scale scientific team,and gather from different fields、Experts and scholars in different branches,Cross -study research in the field of cross -disciplinary fields from a broad perspective,Solve the problem of accuracy of social science prediction。Picture Source: CFP

Groasman talks,Social science predicts the upcoming trend or incident in the real world.。part of the reason is,The emergence of social phenomena is not because of certain properties that play a decisive role。On the contrary,What they follow is a random process,Its complexity exceeds the number of parameters that usually consider social science researchers and policy makers。So far,Researchers who agree with this view are just a minority。Some social science researchers、Policy makers and the public usually believe that,Just find the correct theoretical or predictive factors,All social problems are theoretically predictable。From a technical point of view,A series of cause and effect models used by social science researchers may not be accurate enough,It is impossible to provide prediction results with actual use。For example,The model can only predict the direction of the trend,Can't clearly point out its amplitude。Only recognize these restricted factors,Social science researchers can better plan solutions when facing the evolving social problems,and take corresponding preventive measures。

One aspect,Social Science still has a lot of room for improvement in predicting social changes。On the other hand,Some scholars start from the definition of prediction,Discuss the role played by predictions in improving the experiment of social science theory。Mark Verhagen, a postdoctoral researcher at the School of Population Science, School of Population Science, University of Oxford, pointed out in the article "Guide to Bet365 lotto review Practice in Social Sciences",The academic community seems to believe that the interpretability and predictability of social sciences are not compatible。This view has caused researchers to make a choice in two ways,and formulate different data processing according to the selection、Model construction and post -prediction diagnosis process。This means,In specific experiments,Social science researchers often bid to use social science's interpretation ability。Filhaheng discovered,From 2010 to 2021,In all published in the six famous economies、Paper in Political and Social Sciences Pharmaceuticals,Only 5%of the summary mention the verbs and noun forms of "predictive"。In the article mentioned in these two words,Only 13%of articles have predicted the result variables。

Ferhaheng advocates social science researchers to organically combine the two analysis perspectives of "explanation" and "prediction"。There are three major advantages to integrate predictions into the empirical work process of social science researchers。First,As a benchmark tool,Forecast can regulate more different types of models。By comparing the prediction ability of different models,Researchers can evaluate the applicability and complexity of different styles of models,Find the optimal model complex level,Make the model accurately fit the data。Next,As a descriptive tool,Forecast will help researchers understand the fitness of the model。Felhaheng said,It is very important to establish a set of conceptual system around predictive accuracy。If the prediction ability of the model is significantly lower than expected,Researchers have to reflect,Did they ignore the important decisive factor in the model? Does the functional form that researchers preferences capture its mechanism?,Promote the progress of the project while improving the prediction level。Last,Forecast helps researchers understand the behavior of complex models。Intervention through the variable,Observe its impact on the prediction results,Researchers can more accurately grasp the coefficient estimation in the model、The role and influence of each variable in the model,Bet365 app download to enhance the interpretability of the model。

  Multiple ways to improve prediction accuracy

In addition to exploring the essence and accuracy improvement method of social science prediction,Some scholars are already in action to strengthen the emphasis on prediction and participation in the academic community。Professor Stefano Dellavigna, a professor at the University of California Berkeley Economics, led the team to develop the "Social Science Forecast Platform"。As a free online tool,The platform aims to systematically collect and evaluate a series of、Forecast of the results of unliked social projects。Dravinia Represents,Prediction data of colleagues collected a large -scale collection of experiments is conducive to researchers formulating policies and projects that are more in line with social status。For example,A social science research team is ready to implement a series of intervention measures,To improve the quality of educational quality in some rural areas in Kenya。Researchers have planned a total of 5 implementation plans,But there are only funds that can be operated by 3 solutions。If there is no reliable historical data,Researchers will take great risks to choose。With a public platform such as the "Social Science Forecast Platform",Researchers can predict their predictions on the implementation of different schemes with the prediction of peers,to improve the accuracy of the prediction。Dravinia Represents,Mastering the prediction of the peers on the experimental solution is no less than the feedback information that obtains preliminary experiments。This is greatly helpful for researchers' perfect experiments。Based on this,The experimental scheme error that the researcher finally chose is also smaller,The decision -making effect made better。

Groasman told reporters,Many scholars have recognized,To better grasp the complex social phenomenon,Everyone in the team needs to abandon a single "pet theory",Challenge the limitations of a single social and cultural environment。International internationally need a large -scale scientific team,and gather from different fields、Experts and scholars in different branches,To improve the cognition of major scientific issues。From the perspective of bet365 best casino games a broad sense,There are many benefits to the cross -disciplinary field of cross -disciplinary fields,For example, the development of comprehensive research methods,Provide more comprehensive、Multi -angle real problem solving method,Reduce repeated experiments and waste of resources,Improve the efficiency and quality of scientific research。In this background,Universities and scientific research institutions should be in a multi -pronged way,Cultivate cross -disciplinary research culture,To increase the influence of the subject,Promoting academic progress。

Editor in charge: Chen Jing
QR code icons 2.jpg
Key recommendation
The latest article
Graphics
bet365 live casino games
Video

Friendship link:

Website filing number: Jinggong.com Anxian 11010502030146 Ministry of Industry and Information Technology:

All rights reserved by China Social Sciences Magazine shall not be reprinted and used without permission

General Editor Email: zzszbj@126.com This website Contact information: 010-85886809 Address: 11-12, Building 1, Building 1, No. 15, Guanghua Road, Chaoyang District, Beijing: 100026