Total Survey Error in Practice

2Total Twitter Error: Decomposing Public Opinion Measurement on Twitter from a Total Survey Error Perspective

Yuli Patrick Hsieh and Joe Murphy

Survey Research Division, RTI International, Chicago, IL, USA

2.1 Introduction

2.1.1 Social Media: A Potential Alternative to Surveys?

Social scientists investigating public opinion trends typically begin their research by seeking national estimates from representative surveys, such as the General Social Survey (GSS), or those conducted in the U.S.A. by the Pew Research Center for the People and the Press and Gallup. These reputable sources are useful if a national estimate and broad trends are sufficient for their analyses. However, surveys are limited in their ability to produce very timely and rapid estimates in response to current events, since large‐scale survey data collection is very time‐ and resource‐intensive. Additional challenges have arisen in recent years that have made it more costly and difficult to obtain accurate survey estimates (i.e., the erosion of landline telephone coverage and declining response rates). Such limitations may also indicate an emerging need to look for alternative methods to study public opinion.

As new information and communication technologies (ICTs) like mobile phones and social media become widely adopted and deeply integrated into contemporary daily routines, they are changing the nature of the public sphere—many users share thoughts and information to express their attitudes and opinions ...