Causal effects of affordance change on communication behavior: Empirical evidence from organizational and leadership social media use

Highlights

• The causal effect of an affordance change on digital media users’ communication patterns is examined with an instrumental variable approach.

• On the population level, Twitter users generally complied with the technological change and altered their online behavior.

• On the individual level, Twitter users vastly deviated from each other on the degree of treatment compliance and behavioral change.

• Previous perceptions of communication constraint predict later behavioral changes, which highlights the role of human agencies in ICTs.

Abstract

This article provides empirical evidence for two hypotheses in the affordance literature. First, by leveraging a small affordance change — Twitter increasing its character limit from 140 to 280 on November 7, 2017, employing an instrumental variable approach, and examining 143,771 original tweets published by organizational and leadership accounts half a year before and after the intervention, we showed the direct causal relationship between affordances and communication behavior on digital media platforms. Second, by exploring what factors could explain the heterogeneity of causal effects, we showed that previous endogenous perceptions of communication constraint predicted later behavioral changes, despite the same exogenous intervention. These findings highlight the role of human agencies in the face of technological changes and provide empirical support for the affordance approach to information communication technologies (ICTs) as a reconciliation between technological determinism and social constructivism.

Introduction

Much literature on how digital technology influences human communication has focused on one central concept: affordances (e.g., Taipale, 2014, Halpern and Gibbs, 2013). An affordance approach to digital media argues that interpersonal communication and computer-mediated communication differ in one significant way, that the latter requires inhuman objects to encode and decode messages through technical infrastructure and channels, and the characteristics of these technical infrastructures and channels afford and affect communication outcomes, including but not limited to the creation of media content, the management of organizational knowledge, and the emergence of social interactions (Sun et al., 2020, Park and Yang, 2017, Treem and Leonardi, 2013). This affordance approach provided a theoretical lens for scholars to characterize different action possibilities on mediated channels and understand human-computer interactions in newly developed information communication technologies (ICTs) (e.g., Sundar, 2008, Evans et al., 2017, Shin, 2017, Kaptelinin and Nardi, 2012).

Underlying this stream of research, the assumption lies that communication behavior on digital media platforms, such as social networking sites (SNSs), is closely related to and directly affected by affordances of these platforms. However, if we examine our knowledge critically, it becomes obvious that we do not actually have strong, rigorous, and documented evidence on the direct relationship between digital media affordances and human communication behavior. For example, we still do not know the degree to which a change in affordances can affect communication behavior in a causal manner. Further, in contrast to strong theoretical arguments that human agency affects how information technologies are used for communicative purposes (e.g., Nagy and Neff, 2015), we also lack empirical evidence on whether and how varying human perceptions affect digital media behavior.

This gap in the literature was indeed difficult to fill, because affordances in most cases stay the same across time on digital media platforms. In lab experiments, researchers can artificially change affordances but it threatens external validity. Most existing empirical research used a qualitative and comparative lens to study the effect of affordances on mediated communication (e.g., Fox and Warber, 2015, Halpern and Gibbs, 2013).

The current study tries to fill the knowledge gap by taking advantage of a recent event of affordance change to provide causal evidence on the relationship between affordances and communication behavior. Twitter, notorious for its restriction on text length, suddenly changed its character limit from 140 to 280 on November 7, 2017, which provided a great opportunity for causal identification on the effect of affordance change on communication behavior. Supplementing recent case studies on the event (e.g., Jaidka et al., 2019, Gligorić et al., 2018, Gligorić et al., 2020, Gligorić et al., 2019), we report statistical details on how the switch induces changes in content creation and underline the role of platform affordances in shaping user behavior in the digital world. Analyzing the exogenous intervention with an instrumental variable (IV) approach and using organizational settings (i.e., top organizations and leadership Twitter accounts) as the research context, we show how the change in affordances leads to various degrees of change in social media use, i.e., the use of language and interactive features on Twitter. Estimating user engagement level and perceptions of communication constraint with numeric proxies, we further explore what endogenous factors, if any, can explain the various degrees of behavioral change. The result highlights the role of human agency in the face of technological changes, and provides empirical augmentation to the affordance approach to ICTs as a reconciliation between technological determinism and social constructivism.

The paper is structured as follows. First, we review affordance theory and its application in literature to provide background on how the theoretical framework has developed and evolved in empirical studies. By doing so, we highlight how scholarly knowledge has assumed the direct connection between affordances and digital communication behavior without explicitly evaluating it. In the review, we also touch upon the debate between technological determinism and social constructivism to illustrate the value of rigorously evaluating human agencies during affordance change. Then, we introduce the case of Twitter character limit change, detail the empirical strategy for causal identification, justify the research context of organizational social media use, and present numeric variables and their respective hypotheses. After presenting results on behavioral changes, we discuss theoretical implications of the study as to how the affordance approach — recognizing agencies from both developers and users — helps researchers understand ICTs in society at large. We also discuss new ways forward to quantitatively and computationally analyze the role of digital media affordances in computer-mediated communication.