Logical rules characterization of online consumer trust

• Author: Ana A. Andrade,Vitor V. Lopes,Margarida G. M. S. Cardoso
• DOI: http://doi.org/10.1111/itor.12630
• Published date: 01 May 2021
• Date: 01 May 2021

Intl. Trans. in Op. Res. 28 (2021) 1393–1416

DOI: 10.1111/itor.12630

INTERNATIONAL

TRANSACTIONS

IN OPERATIONAL

RESEARCH

Logical rules characterization of online consumer trust

Ana A. Andradea,∗, Margarida G. M. S. Cardosoband Vitor V. Lopesc

aResearch, Studies and Renewables Division, Directorate-General forEnergy and Geology (DGEG), Lisbon, Portugal

bBusiness Research Unit (BRU-IUL), Instituto Universit´

ario de Lisboa (ISCTE-IUL), Lisbon, Portugal

cCMAF-CIO/Faculdade de Ciˆ

encias, Universidade de Lisboa, Lisbon, Portugal

E-mail: ana.andrade@dgeg.pt[Andrade]; margarida.cardoso@iscte-iul.pt [Cardoso];

vitorvieiralopes@gmail.com [Lopes]

Received 3 April 2017; received in revised form 9 October 2018; accepted 1 January2019

Abstract

In this study, we address the discriminant factors of website trust. We specifically build sets of propositional

rules that can be used to predict the level of trustworthiness of a site. Focusing on initial trust, a survey was

designed to assess site characteristics observed by the respondent and his/her perceptionsaround appearance,

reputation, fulfillment, and security. By exploring data, we look forthe most favorable rules classifiers among

decision trees as well as classical and dominance-based rough sets. A heuristic aiming to derive simpler

classifiers is also proposed. The experimental setup considers diverse groups of attributes (predictors) for the

extraction of rules. Results obtained are compared by taking into account predictiveability and parsimony of

rules’ sets. Finally, the selected sets help bring light on how consumers process site information and suggest

specific recommendations for e-commerce vendors.

Keywords:trust, e-commerce; B2C; information theory; rough sets; tree algorithms; knowledge-based systems

1. Introduction

1.1. Context of the study

In Business to Consumer (B2C) Internet-based commerce (e-commerce), the interface between

vendor—the manufacturer or some intermediary agent—and the potential buyer is an electronic

storefront (site), seen by millions of people worldwide. Through e-commerce, a seller obtains simul-

taneous access to a high number of potential clients and consumers gain access to an increasing

range of products and services, arising from a diversity of geographical locations and suppliers.

However, lack of direct contact that underlies this type of interaction prevents the potential con-

sumer (Riegelsberger et al., 2005) from assessing the usual physical nature cues of trustworthiness,

like vendor body language or store appearance (McKnight et al., 2002) making online trust one

major issue (Huang et al., 2003; Turban et al., 2010).

∗Corresponding author.

C

2019 The Authors.

International Transactionsin Operational Research C

2019 International Federation ofOperational Research Societies

Published by John Wiley & Sons Ltd, 9600 Garsington Road, Oxford OX4 2DQ, UK and 350 Main St, Malden, MA02148,

USA.

1394 A. A. Andrade et al. / Intl. Trans.in Op. Res. 28 (2021) 1393–1416

Reputation mechanisms (McKnight et al., 2002), business rating sites, and trust verification sites

(trust seals) perform site/vendor evaluations based on quality and/or performance (Turban et al.,

2010), helping the potential customer to develop trust even without a previoushistory of interaction.

However, none of these measures takes into account how consumers react to sites, and so they have

limited relevance for e-commerce sellers.

From the beginning of e-commerce, a large amount of literature was produced to study trust

formation and evolution (Shankar et al., 2002; Murphy and Blessinger, 2003; Das and Teng, 2004).

Such studies consist in theoretical model proposals and/or in empirical assessments, both aiming

the identification of factors most contributing to trust.

In the empirical studies, data areobtained mainly through questionnaires concerning perceptions

about the site or the vendor and, generally, data analysis resorts to parametric statistical techniques

(Grabner-Kr¨

auter and Kaluscha, 2003; Sepp¨

anen et al., 2007). Recently, nonparametric techniques

have also been used in this setting further addressing the direct identification of site characteristics

(Andrade et al., 2012; Seckler et al., 2015).

Empirical studies outputs generally producerecommendations for online vendors regarding either

the content and graphical appearance of the electronic interface, its functionality, or the underlying

business model.

Komiak(2004) points out that we need to develop more quantitative modeling of trust attribution.

In this regard, a few studies used the concept of fuzzy logic, acknowledging that the psychological

phenomenon of human trust is complex and unclear enough so that techniques embracing uncer-

tainty are appropriate. Under this trend, Akhter et al. (2005) derived fuzzy logicalrules to quantita-

tively relate trust with security, design and familiarity (Akhter et al., 2005), or with content (Nilashi

et al., 2011; Nilashi et al., 2015). Also based on logical rules, Meziane and Nefti (2007) continue

using fuzzy theory, nevertheless involving a higher number of variables (twelve), grouped according

to existence, affiliation, policy, and fulfillment concepts. These rules, considered as a “stronger and

more objective” tool, were used to predict low, average and high levels of consumer trust.

1.2. Objective

This paper proposes to build sets of logical rules that an e-commerce seller can use to assess the

level of initial trust created by his site on potential clients, as a guidance to improve his site design.

Websitesof unknown sellers come to the knowledge of the potential customer either through search

engines or electronic malls and information portals (Turban et al., 2010). Large companies with

well-established brands have an advantage, since they are somehow already known by a consumer.

The proposed approach is relevant for smaller companies that rely entirely on their sites to initiate

trust, making easier to verify if their sites comply with a trust enhancing design.

From the analysis of these rules, additional knowledge about trust formation mechanisms is

inferred as new input for site development considerations, as also considered in Tanjung and

Dhewanto (2014). This work further aims to provide recommendations for the construction of sites

that promote the level of trust of its users.

Decision models composed of logical rules are transparent and easily understandable (Greco

et al., 2007). The sets of rules proposed are meant to be simple, but induced from an unpreceded

number of variables referring to a detailed site characterization.

C

2019 The Authors.

International Transactionsin Operational Research C

2019 International Federation ofOperational Research Societies