This article has been accepted for inclusion in a future issue of this journal.

Content is final as presented, with the exception of pagination.

IEEE TRANSACTIONS ON PROFESSIONAL COMMUNICATION, VOL. 00, NO. 00, 2017 1

Research Article

Diagrams in Contracts: Fostering Understanding in Global

Business Communication
STEFANIA PASSERA, ANNE KANKAANRANTA, AND LEENA LOUHIALA-SALMINEN

Abstract—Research problem: Business-to-business contracts are complex communication artifacts, often considered
“legal stuff” and the exclusive domain of lawyers. However, many other stakeholders without a legal background are
involved in the negotiation, drafting, approval, and implementation of contracts, and their contributions are essential for
successful business relationships. How can we ensure that all stakeholders in the global business context—whatever
their native language or professional background—easily and accurately understand contract documents? This study
suggests that integrating diagrams in contracts can result in faster and more accurate comprehension, for both native and
non-native speakers of English. Literature review: We focused on the following research topics: (1) ways to integrate
text and visuals to create more effective instructions, since we conceptualize contracts as a type of business instructions;
(2) cognitive load theory, as it may help explain why contracts are so hard to understand and why text-visuals
integration may ameliorate their understandability; (3) cognitive styles, as individual differences may affect how
individuals process verbal and visual information, thus allowing us to explore the limitations of our suggested approach;
(4) the English lingua franca spoken by business professionals in international settings, their needs and challenges,
and the fact that pragmatic approaches are needed to ensure successful communication. Methodology: We conducted
an experiment with 122 contract experts from 24 countries. The research participants were asked to complete a series
of comprehension tasks regarding a contract, which was provided in either a traditional, text-only version or in a version
that included diagrams as complements to the text. In addition to measuring answering speed and accuracy, we asked
the participants to provide information about their educational background, mother tongue, and perceived mental effort
in task completion, and to complete an object–spatial imagery and verbal questionnaire to assess their cognitive style.
Conclusions: We found that integrating diagrams into contracts supports faster and more accurate comprehension;
unexpectedly, legal background and different cognitive styles do not interact with this main effect. We also discovered
that both native and non-native speakers of English benefit from the presence of diagrams in terms of accuracy,
but that this effect is particularly strong for non-native speakers. The implication of this study is that adding diagrams
to contracts can help global communicators to understand such documents more quickly and accurately. The need for
well-designed contracts may open new opportunities for professional writers and information designers. Future research
may also go beyond experimental evaluations: by observing this new genre of contracts in vivo, it would be possible
to shed light on how contract visualizations would be perceived and interpreted in a global communication environment.
Index Terms—Contract visualization, contract design, cognitive load, diagrams, global communicative competence
(GCC), international business communication (IBC).

INTRODUCTION BellAliant and Rogers Telecom, two Canadian firms,

went through a bitter 18-month dispute, worth
T hink about contracts: long, complex, demanding $1 million CDN, on whether BellAliant was allowed
documents written in “legalese,” not really an to terminate a contract by simply giving notice
archetype of effective, clear, and easy professional or whether the agreement was iron-clad until the
communication. Misunderstandings may lead end of the contract period. The source of confusion
to litigation, which damages businesses in terms boiled down to a single comma in the English version
of wasted resources and reputation. For instance, of the agreement, which suggested that the first
interpretation was the correct one, while the French
Manuscript received May 15, 2015; revised September 8, 2016; version of the agreement (Canada is a bilingual
accepted September 10, 2016. Date of publication; date of
current version. country) supported the latter interpretation [1].
S. Passera is with the Department of Industrial Management Plain language scholars have been addressing the
and Engineering, Aalto University School of Science, Espoo shortcomings of traditional legal writing for years.
17700, Finland (email: [email protected]).
A. Kankaanranta and L. Louhiala-Salminen are with the Even attorneys and legal scholars have repeatedly
Department of Management Studies, Aalto University School of warned that legalese is a waste of money for
Business, Helsinki 21210, Finland (email: anne.kankaanranta@ businesses and governments in the form of claims,
aalto.fi; [email protected]).
Color versions of one or more of the figures in this paper are administrative and compliance costs, and time
available online at http://ieeexplore.ieee.org. lost in writing, reading, and managing confusing
documents [2]–[4].
IEEE 10.1109/TPC.2017.2656678

0361-1434 C 2017 IEEE. Personal use is permitted, but republication/redistribution requires IEEE permission.

See http://www.ieee.org/publications standards/publications/rights/index.html for more information.

This article has been accepted for inclusion in a future issue of this journal. Content is final as presented, with the exception of pagination.

2 IEEE TRANSACTIONS ON PROFESSIONAL COMMUNICATION, VOL. 00, NO. 00, 2017

It is also erroneous to think of contracts as merely (1) The benefits of integrating texts and visuals in
“legal stuff,” the exclusive, specialized domain of instructional texts (and how contracts can be
lawyers [5]. In today’s global, networked economy, intended as a particular type of instructions)
contracts represent key business instruments for (2) Cognitive load theory (CLT) and the effects of
companies to synchronize, collaborate and individual cognitive styles on learning
undertake new ventures. They are used in all (3) The specific needs of professional NNSs of
spheres of business, from narrow aspects such as English for pragmatic solutions and practices
the working conditions of a certain employee, to that can improve their overall communicative
large corporate-wide issues such as mergers and competence
acquisitions. Moreover, after signature, contracts
The overall problem of this study can be formulated
need to be implemented by responsible teams, who
as follows: can a visually enhanced style of
need to carefully follow contracts as instructions
communication help professionals to understand
and translate them into concrete actions. Thus, we
complex contractual information? This leads to our
see contracts as a business communication genre.
first two research questions:
Managers and executives, in addition to lawyers,
are the key stakeholders of contracts, and should RQ1. To what extent can professionals’
be able to understand what contracts say and how understanding of contracts be enhanced by
they impact day-to-day business. employing diagrams in these documents?

The challenge is even greater in international RQ2. To what extent can diagrams in contracts
business, as parties do not share the same reduce the time taken to understand these
language and may lack common interpretive documents?
frames. The highly formal register of English If we find that diagrams indeed support contract
“legalese,” which is modeled on the English used by comprehension, we want to investigate whether
its legally trained native speakers (NS), is the their presence equally benefits individuals with
standard language of international contracts. different cognitive styles, language abilities, and
However, this is not the real-life English used in knowledge of specific jargon. For instance, could a
international business, which is better visual approach increase the communicative
characterized as a lingua franca, typically used by competence of NNSs dealing with English legalese
non-native speakers (NNS) of English, to get work and reduce the gap with NSs? This leads to two
done. As shown in the BellAliant versus Rogers further research questions:
case, the long and syntactically complex sentences
typical of legalese may be ambiguous even for NS RQ3. Which individual characteristics (e.g.,
managers. In light of the concept of global cognitive style, educational background, age,
communicative competence (GCC) [6], [7]—which, language . . . ) may affect the extent to which a
drawing on the lingua franca aspect of English, professional benefits from diagrams in contract
stresses pragmatic communication skills rather documents?
than flawless native proficiency—we propose that
RQ4. In light of the possible effects of individual
the threat of ambiguity to international contracts
characteristics, what are the implications of
must be addressed by going beyond the use of
using diagrams in contracts in an international
textual language. After all, the legal principle of
business setting?
freedom of contract does not limit in any way
contract formats, style, and aspect (even a verbal The remainder of the paper is structured as
contract is valid, although it may be hard to follows. In the literature review, we describe our
enforce). theoretical orientation and the selection process for
the literature, and then present the identified key
In this study, we show how the comprehension of topics. At the end of that section, we suggest eight
agreements among international contract hypotheses. Then, we introduce the methodology of
professionals can be enhanced by presenting this study, after which the results are presented
information in a more visual fashion: contracts and discussed. Finally, we address limitations and
complemented with explanatory diagrams and suggestions for future research.
presented in clearly structured layouts. The
argument in favor of visualizations as a way to
clarify the meaning of complex textual information
LITERATURE REVIEW
is based on these areas of earlier research: Insights from past research motivate our interest
in complementing international contracts with
This article has been accepted for inclusion in a future issue of this journal. Content is final as presented, with the exception of pagination.

PASSERA ET AL: DIAGRAMS IN CONTRACTS: FOSTERING UNDERSTANDING IN GLOBAL BUSINESS COMMUNICATION 3

explanatory diagrams as a way to make them more Our concern with user-centeredness and effective
understandable. The reviewed studies help us to communication led us to deepen our
formulate the hypotheses of our study and identify understanding of individuals and the context in
the relevant variables to investigate. This section which they communicate. Although contracts
begins with an introduction of our theoretical encode complex information, understanding is not
orientation, followed by sections focusing on visual dependent only on the intrinsic ease or difficulty of
representations in instructions, CLT and instruc- the document; the cognitive resources available to
tional design, and cognitive styles and the chal- the readers should also be accounted for. For this
lenges of English legalese within the frame of GCC. reason, we looked into cognitive theories about
learning and comprehension, as well as individual
Theoretical Orientation This study is cognitive variables affecting the processing of verbal
multidisciplinary in nature, combining theories of and visual information. We focus in particular on
visual communication and of cognition and CLT, which acknowledges the importance of
language. Although the study focuses on business information presentation in promoting or hindering
contracts, it looks at them from the perspective of understanding, and on individual cognitive styles,
design and communication rather than from the that is, how each individual prefers to process,
tradition of legal studies. Communication encode, and express information. Additionally, we
effectiveness in a global business setting is the reviewed literature on communicative competence
main concern, and we zoom in on how and why in global settings, and since our study focused on
complementing texts with diagrams enhance the comprehension needs of contract experts
information search and comprehension. operating in an international setting, we paid
particular attention to the research streams of
Traditionally, the efforts of clarifying and English as (business) lingua franca and English for
simplifying contracts have aimed at “lifting the fog specific purposes (ESP).
of legalese” [8]. Rethinking legal and bureaucratic
language has been (and still is) the main battle of The reviewed literature also affected our
the proponents of Plain Language— methodological choices: we designed the study as
an experiment, but paired it with a self-reported
Language that reflects the interests and needs of questionnaire. The experimental setup follows the
the reader and consumer rather than the legal, tradition of evaluating and comparing design
bureaucratic, or technological interests of the solutions in their ability to satisfy users’ needs,
writer or of the organization that the writer and improve comprehension performance; the
represents. [9, p. 7] questionnaire allows us to introduce individual
cognitive and language abilities as variables in the
The focus on visual communication is thus a study, and understand how they interact with the
novelty in contracts, even though a number of presentation format of the contract, and how they
plain language authors have acknowledged the key affect comprehension.
role of information design in communicating
content clearly [10], [11]. A small but growing body
Selection of Literature for the Review To
of legal scholarship has been proposing the use of
address the questions and concerns of this paper,
visual representations—for instance, diagrams [12],
we opted for a convenience review of past studies
[13], photographs [13], comics [14], and icons and
focusing thematically on the following:
document design [9], [15]—as means of increasing
the understandability of contracts [5], [9], [12]–[14], (1) How to integrate text and visuals to create more
[16]–[19] and other legal documents [20]. effective instructions (we conceptualize
contracts as a type of business instructions)
One issue in the literature, however, is that several (2) CLT, as it may help explain why contracts are
studies do not rely on actual user testing [5], [14], so hard to understand and why text-visuals
[16]. The studies that provide some quantification integration may ameliorate their
of the comprehension benefits of visual understandability
communication do so in an insufficiently rigorous (3) Cognitive styles, as individual differences may
manner—for example, the samples are too small affect how individuals process verbal and visual
[17], [12], [19], and statistical analysis is information, thus allowing us to explore the
insufficient [17], [19]. A theoretical explanation of limitations of our suggested approach
why and to what extent visual communication (4) The English lingua franca spoken by business
supports contract comprehension is also missing. professionals in international settings, their
This article has been accepted for inclusion in a future issue of this journal. Content is final as presented, with the exception of pagination.

4 IEEE TRANSACTIONS ON PROFESSIONAL COMMUNICATION, VOL. 00, NO. 00, 2017

needs and challenges, and the reasons why from purely textual information, doing so requires
pragmatic approaches are needed to ensure more effort and induces a heavier cognitive load
successful communication compared to a multimodal approach [36]: visual
representations allows for a more direct
Visual Representations in Instructions A construction of mental models, since external and
pragmatic approach to effective communication in internal representations share a quasi-spatial
contracts leads us to conceptualize contracts as nature (e.g., [22], [37])
instructions for correctly delivering a given promise
to the other party. This metaphor is not completely However, these observations do not mean that
new, as in fact legal scholars have previously visual representations alone are sufficient to
proposed contracts as “blueprints for [business] ensure comprehension. Cumulative evidence
performance” [21, p. 92], “user’s guides and shows that the most effective instructions combine
instruction manuals” [20, p. 21], and “visible text and visuals [36], either with explanatory
scripts for the parties to follow” [16, p. 57]. Taking visuals accompanying text [38], [39], or as visually
inspiration from this metaphor, we explore structured diagrams that include text, such as
research on multimodal instructions, which have flowcharts, decision trees, and tables [30], [40],
previously tackled the problem of supporting user [41]. Alexander [42] summarized the benefits of
comprehension by integrating texts and visuals. integrating text and visuals as faster task
Most research on instructions focuses on how to completion, fewer comprehension mistakes, faster
effectively provide guidance to use appliances [22], learning to use a device, and more positive user
use software [23], [24], perform procedures [25], attitudes toward the instructions. Several scholars
[26], and solve scientific/logical problems [27], [28]. have also pointed out how visual elements play an
Even though these uses of instructions are quite indispensable rhetorical role in relation to text:
different from contracts—where the to-do actions they better organize and structure the content [43],
are more abstract (e.g., when and how to pay the they guide readers’ attention to precise topics and
supplier; how to calculate liquidated damages for a arguments [44], and they affect interpretation of
delay in product delivery)—we believe that the key the overall document [45] by providing hints about
results of this research area could inform and be its tone, credibility, function, and relevance. [46].
applied to our area of inquiry.
Finally, several studies indicate that understanding
First, visuals in instructions—such as screenshots, and applying instructions is a cognitively intense
technical illustrations, photos, and diagrams—offer activity, which imposes high demands on working
visual relief from dense pages of text [24], which memory (WM) [36], as readers need to explore the
users are not enticed to read [29]. Moreover, document and find relevant information,
visualizations reduce complexity, for instance by understand it, integrate it with previous and
preselecting and giving salience to the most situational knowledge, and finally apply the
important details needed to accomplish the goal instruction. Spatially integrating texts and visuals
[30]; making more explicit relationships that need allows reduction of the cognitive load by creating a
to be inferred from the text [22]; presenting degree of redundancy which is beneficial to
procedures in an easy-to-follow step-by-step comprehension [35] and by reducing the need to
approach [30]; illustrating more saliently split attention between sources of information
alternative routes and conditions, as well as located in different places [47].
conditional information (if X then Y) [30], [31]; and
offering a visual reference that helps users to verify In the next section we take a closer look at CLT,
that they are proceeding correctly and to recover which is concerned with the learning of complex
from errors [32]. information and the ways that learners often
struggle to process the many elements needed for
Second, visual representations facilitate readers in understanding because of the capacity limits of
creating a mental model of themselves in action, WM.
executing the procedure [33]. In respect of
successful mental model creation, research has CLT and Instructional Design CLT [48], [49]
shown the importance of presentation order in postulates that information that is too complex or
which instructions should be carried out [34] and abundant overloads our cognitive systems, as
the type of perspective—for example, user-centric capacity limitations prevent adequate processing of
or system-centric—to be adopted in visualizations all inputs and ultimately learning. CLT is based on
[25], [35]. While mental models can also be created three assumptions:
This article has been accepted for inclusion in a future issue of this journal. Content is final as presented, with the exception of pagination.

PASSERA ET AL: DIAGRAMS IN CONTRACTS: FOSTERING UNDERSTANDING IN GLOBAL BUSINESS COMMUNICATION 5

(1) Human MW possesses different processing

systems for visual and verbal information (dual
channel assumption; e.g., [50], [51]).
(2) Each channel has a limited processing capacity
(limited capacity assumption; e.g., [48], [50]).
(3) Meaningful learning requires substantial
cognitive activities in scanning, organizing and
integrating information from different channels
(active processing assumption; e.g., [52]).
Additionally, Schnotz and Bannert [28] point out,
comprehension is an active, goal-oriented process:
humans construct multiple mental representations
with the goal of accomplishing a specific task (e.g.,
being able to perform a calculation, solving a Fig. 1. Three components of cognitive load, according to
problem, behaving accordingly to a rule . . . ). Sweller [48].
Different ways to present the same content affect
how we construct it mentally, and these mental
representations may facilitate or hinder problem
resolution. usually cannot be affected by instructional
means.
The superior performance associated with the (2) Extraneous cognitive load is caused by the
integration of visuals and texts in instructions, manner in which information is presented, and
compared to texts or visuals alone, can be thus its reduction through different designs has
explained in light of CLT. For example, a long traditionally been the focus of instructional
running text may feel difficult to understand designers. Intrinsic and extraneous cognitive
because it is processed through only one channel loads are additive, and both impose demands to
with limited capacity; if the same information were WM. Extraneous load interferes with learning,
presented as a diagram, information processing especially when information is intrinsically
would be routed to different channels, as text complex [48], [54].
conveys separate bits of meaning, while the visual (3) Germane cognitive load is the processing
communicates the relationship that exists between devoted to schema formation and
those bits. The cost of interpretation of such automation—that is, the cognitive processing
relationships is offloaded to the external that positively contributes to understanding
representation, as they simply need to be “read off” and learning. Usually there is an opposite
instead of requiring effort to be inferred [53]. relationship between extraneous and germane
cognitive load [55]: for example, devoting much
In the case of contracts, one might argue that they mental effort to understand something
are difficult because they carry intrinsically unclearly written (extraneous load) will not
complex messages; however, information result in increased learning, but thoroughly
presentation matters. CLT conceptualizes cognitive processing and actively working through an
load as being composed of three separate elements, example (germane load) will [56], [57].
intrinsic, extraneous, and germane load, which are
Comprehension errors and slow reading/task
differently affected by changes in how the
completion are typical symptoms of cognitive
information is designed [48] (summarized in Fig. 1).
overload [58]. However, to exactly assess and
(1) Intrinsic cognitive load derives from the compare the efficiency of alternative instructional
inherent complexity of the information, and it designs in supporting learning, Paas and
mainly depends on whether the different Merriënboer [59] and Paas et al. [60] suggest the
elements to be processed can be attended to need to consider both performance and mental
individually or need to be attended to all effort because learners may maintain a good
together [54]. For example, performing a series performance in cognitively demanding tasks by
of simple, unrelated numerical additions does simply making a greater mental effort. Considering
not cause overload, while solving a literal both performance and mental effort allows us to
equation does, as it requires knowing where distinguish between desirable and undesirable
and how to apply several algebraic rules, which instructional designs: low-efficiency instructions
need to be used in concert. Intrinsic load will require high mental effort and still yield poor
This article has been accepted for inclusion in a future issue of this journal. Content is final as presented, with the exception of pagination.

6 IEEE TRANSACTIONS ON PROFESSIONAL COMMUNICATION, VOL. 00, NO. 00, 2017

performance, while high-efficiency instructions will analytic style dimension (whether an individual
yield good performance with little mental effort [60]. processes information as wholes or in parts) and
the verbal-imagery style dimension (whether an
These considerations open the possibility to individual better processes and represents
improve the efficiency of contracts—in terms of information as words or images) [71]. Given the
learning and comprehension—by changing how focus of this paper, we will concentrate on the
they are designed. In the previous section, we saw verbal-imagery style dimension.
how effective instructions integrate texts and
visuals. There is also considerable evidence for Recent developments on the study of the verbal-
these results in the educational psychology imagery cognitive style have shown how this is not,
literature, which shows how diagrams—which as previously believed, a bipolar continuous
integrate visual representations with construct, according to which individuals are
texts—support learning (e.g., [27], [61], [62]). categorized as either verbalizers or visualizers [73],
However, other studies suggest that the effects of [74]. In fact, three relatively independent
visual presentation modes in enhancing learning dimensions exist [68], [73], [75]:
are strong for novices but tend to disappear and
(1) Object imagery: ability in building and
even reverse as expertise grows (e.g., [62],
processing vivid, colorful, detailed images of
[63]–[66]). Furthermore, some studies conclude
individual objects
that only high visual–spatial learners can fully
(2) Spatial imagery: ability in representing and
benefit from integrating visual and verbal materials
processing schematic images, spatial
[27], [67]. Our study thus needs to assess whether
relationship between objects, movement, and
adding diagrams to contracts will result in higher
spatial transformations
efficiency instructions and whether the
(3) Verbal: ability in verbal, logical, sequential, and
hypothesized benefits are limited to only some
analytical reasoning
groups of readers.
With this understanding, we can see that an
Cognitive Styles In addition to CLT, individual individual with a distinctively favorite cognitive
cognitive styles should be taken into consideration style may or may not struggle with other types of
when optimizing information for learning and information representations, as a high score on one
understanding. “Cognitive style” is a term used to dimension does not imply low scores on the other
describe the consistent way in which an individual two.
acquires and processes information (e.g., [68], [69]).
According to Ausburn and Ausburn [69], cognitive In the same way in which the savvy global
styles are stable characteristics, resistant to communicator has to be sensitive to different
change by training and in no way correlated with values and cultures, a higher appreciation of
intellectual ability: if a learner struggles with different cognitive styles may lead to more effective
learning, the problem might be in how the communication, given that a contract may already
information is presented, and a different approach be intrinsically complex, and even more so for a
to conceptualizing, encoding, and displaying NNS of legalese or English. In the next section, we
information can result in successful learning. will consider the lingua franca spoken by global
Thus, for training to be more inclusive, they business professionals, their practical
recommended providing alternative complementary communication needs, and the ways such needs
representations. For instance, an individual can be supported through a pragmatic approach.
scoring low on verbal ability might struggle to
interpret a complex text and integrate its parts; if The Challenges of English Legalese within the
we were to complement this text with a diagram, Frame of GCC In the past 20 years, English has
and were the learner to possess high scores on become the shared language of most international
visual cognitive styles, the problematic information organizations, both at the macrolevel of the
processing would be supplanted by an appropriate organization as “the corporate language,” and at
instructional design, thus facilitating the microlevel of the individual professionals
understanding [69]. working within the organization (e.g., [76]–[79]). For
this reason, we must examine what this extremely
While several labels and tools to describe and fast spread of English to international
assess cognitive styles exist, it has been suggested companies—and, hence, to contracting—might
[70]–[72] that they can be all grouped under two mean from the point of view of communicative
main cognitive style dimensions: the wholist- competence.
This article has been accepted for inclusion in a future issue of this journal. Content is final as presented, with the exception of pagination.

PASSERA ET AL: DIAGRAMS IN CONTRACTS: FOSTERING UNDERSTANDING IN GLOBAL BUSINESS COMMUNICATION 7

The majority of the international professionals who As was pointed out earlier, although legal English
use English as a shared language are NNSs, causes challenges for both NS and NNS of English
making use of a resource often called English as a because of its highly specialized, formal register, it
lingua franca (ELF) or more specifically, English is particularly demanding for NNSs (see, e.g., [87]).
as a business lingua franca (BELF). This The needs of NNSs are addressed in the paradigm
resource-based view of language differs from of ESP, focusing on the pedagogy of teaching
English as a native language (ENL), English as a specialized language, in this case legal English.
second language (ESL, e.g., Singaporean English) Although the English needed is defined as “legal,” it
or English as a foreign language (see, e.g., [80]), is still conceptualized as the English of the NS,
which all reflect the conceptualization of language whose language use the learner is supposed to
as an independent entity and a pre-existing imitate. The difference from the ELF/BELF
system. paradigm is distinct: in ELF/BELF, the language
user is a professional using the language at work,
ELF research looks at language differently: ELF is whereas in the ESP paradigm, the focus is on
not “owned” by anyone, nor it is directly linked to a learning a language.
particular cultural context (see, e.g., [81]). Instead
the ELF approach assumes a constructionist view Lawyers can be seen as “NSs of legalese,” but they
that emphasizes the particular ELF situation and are far from being the only ones involved with
its participants. ELF communicators tend to avoid contracts. As suggested by Haapio [5, pp. 54–55],
complex lexis or structures, and they incorporate several contract users exist at the same time, in
their own native communicative practices and different contexts of use, with different goals and
situation-specific requirements into the situation at needs. On the one hand, traditional legal
hand [82]. scholarship focuses on judges, arbitrators,
litigating lawyers, and authorities, and the
The conceptualization of ELF as an independent prototypical scenario of use is the courtroom: the
linguistic resource also entails that ELF speakers main interest is to predict how courts might
are not assumed to imitate ENL as closely as interpret contracts, so contract drafting is seen as a
possible, but rather identify themselves as speakers communication and persuasion activity happening
in their own right (see, e.g., [83], [84]). Neither is between legal professionals. This is the paradigm
ELF considered deficient in comparison to ENL, where contracts written in “native legalese” are the
albeit rather different, nor is resorting to other unchallenged norm and where communication
languages during interaction regarded as evidence might break down if the correct grammar, registers,
of a gap in the knowledge of English but rather as a rhetorical devices, and norms are not followed.
bilingual pragmatic resource (see, e.g., [85]).
On the other hand, Haapio identifies a
Drawing from ELF research, but contextualizing managerial-legal paradigm, focusing on the needs
their study in international business, of in-house counsels, managers (sales,
Louhiala-Salminen et al. [86] identified BELF to procurement, risk management), financial officers,
emphasize the significance of the domain—the and engineers, where the main driver is to reach
“B”—with its goal-oriented nature and shared business objectives through commercial
business fundamentals. When the employees of relationships, proactively minimizing legal and
globally operating companies attend to their daily business risks, and always balancing them with
practices—including contracting—they thus draw rewards. These contract users, who may well be
from their BELF resource, but also need business English NSs, are the “NNSs of legalese,” for whom
know-how and multicultural competence to do communicative success can be assessed according
their jobs. Although the “English” in BELF may not to criteria similar to those relevant to GCC [6]:
be grammatically perfect, BELF competence multicultural (in this case, also multidisciplinary)
involves the use of highly specialized, shared competence or the awareness and sensitivity of the
terms, concepts, and genres so that the speaker communicator toward how different cultures/
can adapt to the business situation at hand. It also disciplines get things done, specific business
includes an element of strategic skills; as shared know-how, and BELF competence.
understanding cannot be taken for granted,
successful BELF communicators ask questions, In conclusion, findings on BELF and GCC can be
repeat utterances, and use more than one channel used as a lens to look at any context where
or mode to achieve shared understanding. Finally, imposing a native English language model (e.g.,
clarity, directness, and politeness are essential. legalese) on NNSs (e.g., internationally operating
This article has been accepted for inclusion in a future issue of this journal. Content is final as presented, with the exception of pagination.

8 IEEE TRANSACTIONS ON PROFESSIONAL COMMUNICATION, VOL. 00, NO. 00, 2017

managers) not only causes difficulties in H2. Contracts including diagrams allow for more
comprehension, but does not serve the practical accurate comprehension than traditional, text-only
goals of the professionals, that is, getting their contracts.
work done.
Although improved and rapid comprehension is the
ultimate practical goal for contract readers, this
Conclusions from Literature Review and alone cannot tell us whether different designs
Hypotheses The earlier studies discussed above actually reduce cognitive load. When performing a
suggest a way forward to answer our research difficult task, people may still achieve good
questions. Diagrams and other strategies to performance by investing even more mental effort
integrate texts and visuals offer a way to provide as a reaction to an increase in cognitive load [59].
effective instruction to individuals tasked with For a more meaningful comparison of different
accurately following a procedure—a scenario akin communication solutions, we need to assess how
to using a contract in day-to-day business. The mentally efficient they are—how much mental
difficulty in understanding contracts accurately effort they require to achieve a certain level of
and quickly, moreover, hints at a problem of performance. Estimating efficiency allows us to
cognitive overload, in particular the presence of determine whether one document requires more
extraneous load, as typically contracts do not mental effort than the other (meaning that it
follow any best practice in document design and imposes a higher cognitive load) and what the
have been criticized by plain language supporters performance outcomes of such mental effort
for their use of legalese. Changing the design of the investment are. As a result, we propose this
information is a way to reduce cognitive load, and hypothesis:
support better understanding.
H3. The mental efficiency of the version of the
We also established that the risks associated with contract including diagrams is higher than that of
ineffective contract communication are particularly the text-only version, meaning that participants
high in international settings, as the English lingua require less mental effort to achieve an equal level
franca used by NNSs differs from the convoluted of performance.
register of native-language-informed English
An appreciation for individual and different ways to
legalese. Reframing and redesigning contracts as
process, learn, and represent information—as
instructions, in our view, is a solution well in line
shown in studies on cognitive styles—might
with the pragmatic approach to communication
warrant a more finely tuned assessment of the
needed by global communicators. However, given
benefits that visualizations might bring to
the novelty of this approach in contracts, we need
comprehension. As illustrated in Appendix I, the
to discover the extent of benefits that may be
“visual contract” used in our study includes
brought forward by integrating text and visuals,
explanatory diagrams, which complement rather
and define how this solution may also
than substitute for the text. We know that
accommodate individual characteristics, such as
spatial-visualizer learners benefit more from
cognitive styles, language, and professional
multimodal instructions [67], and we assume that
proficiency of native and non-native contract
verbalizers may easily process either version, given
professionals alike. Below, we have formulated a
the high content of text. Predominantly object
series of testable hypotheses that will help us find
visualizers may somewhat improve their
answers.
performances with a more visual contract, but
since contract understanding requires analytic
Based on the consistent evidence indicating that
skills, they may still have a comparatively weaker
instructions integrating visuals and texts support
performance. However, since verbal and visual
better comprehension, and on the predictions of
skills are not mutually exclusive, we need to study
CLT (that errors and slow reading/answering are
all possible interactions of cognitive styles and
symptoms of cognitive load, and that the
treatment in an exploratory manner, as it is
extraneous cognitive load can be reduced through
difficult to predict what their final effects on
design solutions, arguably reducing errors and
performance will be and whether a specific class of
time taken), we put forth two initial hypotheses:
learners is the fastest and most accurate. Thus, we
have developed these hypotheses:
H1. Contracts including diagrams allow for faster
information finding than traditional, text-only H4a. Cognitive styles interact with experimental
contracts. treatment. Verbalizers are equally fast with either
This article has been accepted for inclusion in a future issue of this journal. Content is final as presented, with the exception of pagination.

PASSERA ET AL: DIAGRAMS IN CONTRACTS: FOSTERING UNDERSTANDING IN GLOBAL BUSINESS COMMUNICATION 9

contract and are the fastest group with the textual METHODOLOGY
contract; spatial visualizers are the fastest with
contracts including diagrams; object visualizers In this section, we explain how we designed our
become faster with contracts including diagrams study—an experiment combined with a
than textual ones, but remain the slowest questionnaire aimed at collecting control
performing group. variables—and analyzed the results. First, we
introduce our choice of a research methodology
H4b. Cognitive styles interact with experimental and the way that we sought out participants. Then,
treatment. Verbalizers are equally accurate with we review how we collected data, administered the
either contract, and are the most accurate group study, and analyzed the data. Finally, we present
with the textual contract; spatial visualizers are how we ensured the reliability and validity of the
the most accurate with contracts including study.
diagrams; object visualizers become more
accurate with contracts including diagrams than Choice of a Research Methodology In line with
textual ones, but remain the least accurate group. the research traditions analyzed in our literature
Finally, we need to discriminate the effects of review, we opted for an experimental design,
visualization for native and NNSs of English, the integrating it with a self-reported questionnaire.
globally shared language used in international This combination of methodologies allowed us to
contracting. Arguably, NSs (especially if they have a collect objective comprehension performance data
legal background) should be more proficient in with the use of two different versions of a contract
understanding contracts than NNSs. Studies on (experiment), and to integrate those results with
CLT have consistently demonstrated an “expertise individual psychometric and self-reported data. The
reversal effect,” in which more knowledgeable experiment and the questionnaire were previously
readers do not benefit from additional instructional piloted during two smaller scale studies [17], [19],
aids, and actually experience higher cognitive load and their design and measures were discussed
because of the presence of redundant, more explicit with several researchers not involved in the study.
information (see, e.g., [62]–[66]). Although NNSs
using English—or rather BELF—in their everyday Participants We sought to recruit participants
work may be highly competent global who
communicators, English legalese is not within that (1) Currently held a position in which they
type of pragmatic competence. As a result, as routinely wrote, negotiated, or implemented
nonexperts they are more likely to benefit from international contracts
more explicitly presented information, and so we (2) Were currently involved in international deals
propose these hypotheses: (3) Represented both sell-side or buy-side of
H5a. Language interacts with experimental business transactions
treatment: NSs of English are faster and more (4) Represented both native and non-native
accurate than NNSs with the traditional version of English speakers
the contract (text-only), while NNSs perform faster (5) Represented different organizational functions
and more accurately with the version of the involved with contracts (commercial and
contract including diagrams. contract management, sales, procurement,
legal)
H5b. Profession-specific knowledge interacts with (6) Had a university degree, so as to avoid low
experimental treatment: subjects with a legal literacy as a confounding factor
background are faster and more accurate with the
traditional version of the contract (text-only), while The participants were thus recruited from the
subjects without legal background perform faster members of the International Association for
and more accurately with the version of the Contract and Commercial Management (IACCM), a
contract including diagrams. global, industry-independent professional
association promoting best practices and standards
H5c. Language interacts with profession-specific in contract and commercial management. At the
knowledge and experimental treatment: NSs of time of the study, IACCM had 34,000 members
English with a legal background are the fastest from 158 countries, working in private and public
and most accurate with the traditional version of organizations alike. Collaborating with IACCM
the contract (text-only), while NNSs without a legal allowed us to access a highly diverse community,
background are fastest and most accurate with without the need to individually contact a variety of
the version of the contract including diagrams. organizations worldwide.
This article has been accepted for inclusion in a future issue of this journal. Content is final as presented, with the exception of pagination.

10 IEEE TRANSACTIONS ON PROFESSIONAL COMMUNICATION, VOL. 00, NO. 00, 2017

IACCM sent out our request for participants as to the text-only version, there are wider margins
part of their newsletter, and we recruited 122 and more white space on the page. In addition
participants. Even though the newsletter was sent to the diagrams, key information (e.g., deadlines,
to all IACCM members, their webinars (which last prices) is displayed in boldface in the text of the
approximately 1 h, just like our study) are most agreement to give it more salience. Excerpts from
typically attended by 200 people. Since 200 people the visualized contract are shown in Appendix I.
was, most likely, the best participation rate we
could aspire too, we calculate response rate on its Comprehension performance on the experimental
basis (62% response rate). As an incentive, the task was measured as follows:
volunteers were offered the possibility to receive
(1) Answering speed—The sum of the time taken to
their detailed, individual results of the experiment,
answer the six comprehension questions. The
information about how they compared with the
time taken for each answer was measured
overall average, and an assessment of their
individually, and then summed up.
cognitive style.
(2) Answering accuracy—The sum of correct
answers. Each fully correct answer was
How Data Were Collected Here we describe the
assigned one point. Partially correct answers
methods that we employed to collect quantitative
were graded with 0.25, 0.5 or 0.75 points,
data for our study. We begin with a description of
depending on the magnitude of the mistakes
the experimental task, followed by an explanation
and imprecisions (see grading rationale in
of the questionnaire that we used to collect data
Appendix II), and thus the scores vary between
about the control variables.
0 (all completely incorrect) and 6 (all completely
correct). To avoid bias, we asked two colleagues
Experimental Task: We asked participants to
not involved in this study and blind to the
answer six comprehension questions about a
experimental condition in which the answers
contract, randomly assigned to the participants in
were given to grade the accuracy of the
its traditional, text-only version or in a visually
answers. The graders agreed on 98% of the
enriched version which featured explanatory
answers, and were asked to agree and reassess
diagrams, and clearer layout and typography.
together the answers that they had graded
differently.
The contract used in the experiment was a
business-to-business (B2B) agreement for the
purchase of machinery and equipment. We Instrument: Before the experiment, we asked the
obtained the permission from Ruukki (now SSAB), participants to complete a questionnaire, thus
a European metal and engineering company providing data on the following variables (see
operating globally, to use an anonymized version of details in Table I):
their B2B machinery and equipment purchase (1) Age
agreement, as we were aware that they had (2) Gender
collaborated with a graphic designer to develop a (3) Educational background
new, visualized version of it. The visual version and (4) Legal background
the older, text-only version of the agreement had (5) Nationality
exactly the same text, and differed only in the way (6) Level of English language proficiency
in which it was presented. The order and (7) Native language
numbering of the clauses was also the same. All (8) Current professional role
sensitive details such as the names of the (9) Years of experience in working with contracts
contracting parties, prices, places, and scope of (10) Industry
delivery were fictionalized. (11) Object–spatial–verbal cognitive style
(12) Mental effort
In this agreement, diagrams are presented below
the relevant clause that they seek to clarify. The While most of these were control variables, we will
diagrams can be categorized mostly as timelines or explain “native language,” “legal background,”
charts representing a process. All diagrams contain “object–spatial–verbal cognitive style,” and “mental
textual labels, repeating terms from the clause effort” in more detail, as they were instrumental in
nearby, so as to connect texts and diagrams through our analysis and hypothesis testing.
a strategy of repetition [22]. The layout of the agree-
ment is a single column, with clause headings on English Language Proficiency/Native Language: As
the left-hand side to facilitate searching. Compared it would have been impossible, because of time
This article has been accepted for inclusion in a future issue of this journal. Content is final as presented, with the exception of pagination.

PASSERA ET AL: DIAGRAMS IN CONTRACTS: FOSTERING UNDERSTANDING IN GLOBAL BUSINESS COMMUNICATION 11

TABLE I
VARIABLES COLLECTED THROUGH THE QUESTIONNAIRE

Variable Description Role

1. Age Continuous variable Control

2. Gender Categorical variable: female (1); male (0) Control
3. Educational background Categorical variable: Engineering, manufacturing, construction; Law; Background information
Business and administration; Life sciences, physics, math, computer
science; Education; Arts; Humanities; Social sciences; Other
4. Legal background Categorical variable: yes (1); no (0) (variable obtained from “Educational Analysis of main and
background”) interaction effects
5. Nationality Categorical variable: US; UK; Canada; Australia; Germany; Italy; Spain; Background information
Finland; Norway; Ireland; Czech Republic; Denmark; Poland; France;
Hungary; Netherlands; Ukraine; Croatia; China; India; Pakistan; Senegal;
South Africa; Nigeria
6. Level of English language Categorical variable: No proficiency (0); Elementary proficiency - (1); Background information
proficiency Elementary proficiency (2); Elementary proficiency + (3); Limited working
proficiency (4); Limited working proficiency + (5); Professional working
proficiency (6); Professional working proficiency + (7); Full professional
proficiency (8); (9); Full professional proficiency + (10); Native or bilingual
proficiency (11)
7. Native language Categorical variable: English (1); Other (0) Analysis of main and
interaction effects
8. Current professional role Categorical variable: Contract management; Legal; Sales/Commercial; Background information
Procurement/Strategic sourcing; Combination sales/sourcing; IT;
HR/Recruitment/Placement; Finance; Operations; Academic; Other
9. Years of experience in working Continuous variable Control
with contracts
10. Industry Categorical variable: IT/Telecoms; Oil/Gas/Minerals/Utilities; Background information
Banking/Insurance/Finance; Aerospace/Defense; Technology;
Services/Outsourcing/Consulting; Transportation/Logistics; Automotive;
Engineering/Constructions; Electronics; Healthcare/Pharma/Chemicals;
Public sector; Manufacturing/Processing; Retail; Other
11a. Object-visualizer cognitive Continuous variable: value ranging from 0 to 5; obtained as average of the Analysis of main and
style 15 “O” items in the OSIVQ questionnaire interaction effects
11b. Spatial-visualizer cognitive Continuous variable: value ranging from 0 to 5; obtained as average of the Analysis of main and
style 15 “S” items in the OSIVQ questionnaire interaction effects
11c. Verbalizer cognitive style Continuous variable: value ranging from 0 to 5; obtained as average of the Analysis of main and
15 “V” items in the OSIVQ questionnaire interaction effects
12. Mental effort Categorical ordinal variable ranging between: very, very low mental effort Analysis of main and
(1) and very, very high mental effort (9) interaction effects

constraints, to objectively test the participants for participants with legal knowledge should be able to
English language proficiency, we asked skim through and interpret legal texts with more
participants for a self-evaluation of their language ease as “legalese NSs,” we needed to analyze
skills on a rating scale from 0 (no proficiency) to 11 whether this variable had a simple or interaction
(native or bilingual proficiency), using the number effect on either answering speed or accuracy.
and type of proficiency levels as measured by the
Interagency Language Roundtable’s scale [88]. We
Object–Spatial–Verbal Cognitive Style: The
clearly instructed the participants to mark 11 only
verbal–visual cognitive style of the participants (the
if English was actually their mother tongue. In this
individual preference in relying on visual imagery
way we were able to discriminate between NSs and
versus verbal strategies in thought processes) was
NNSs in our analysis. Information on the English
measured according to the Object–Spatial Imagery
proficiency of NNSs was kept as background
and Verbal Questionnaire (OSIVQ) developed by
information.
Blazhenkova and Kozhevnikov [68]. The OSIVQ is
an efficient self-report instrument composed of 45
Legal Background: Based on the educational items that uses a 5-point Likert scale, and provides
backgrounds of the participants, we created a three separate scores: object-imagery skills (ability
dummy variable to single out everyone with a legal to build and process vivid, colorful, detailed images
background. Because we assumed that of individual objects); spatial-imagery skills (ability
This article has been accepted for inclusion in a future issue of this journal. Content is final as presented, with the exception of pagination.

12 IEEE TRANSACTIONS ON PROFESSIONAL COMMUNICATION, VOL. 00, NO. 00, 2017

to represent and process schematic images, spatial to the survey was shared with participants during
relationships between objects, movement, and the webinar.
spatial transformations); and verbal skills (ability
in verbal, logical, sequential, and analytical Upon accessing the webpage for the experiment
reasoning). The questionnaire is easy to administer and the survey, a script randomly assigned the
and, as shown by Blazhenkova and Kozhevnikov participants into control and treatment groups.
[68], it displays a relatively good internal reliability First, the participants completed the questionnaire
(Cronbach’s alpha of 0.74 for the verbal scale, 0.83 collecting data for the control variables. Then, they
for the object scale, and 0.79 for the spatial scale). proceeded to the experimental setup, and had up
to 4 min to get acquainted with the assigned
The OSIVQ is superior to previous instruments version of the document. Next, each of the six
assessing individual differences in the ability to comprehension questions were presented one at a
process visual and verbal information (e.g., time. After having carefully read the question, the
Individual Differences Questionnaire [51] and participants had to click a button to start the time
Verbalizer–Visualizer Questionnaire [74]) because it countdown and gain access to the document to
corrects some misconceptions. First, the search for the information needed to provide the
visual–verbal dimension is not a bipolar construct, correct answer. Once they had completed their
and different processing abilities need to be answers, the participants stopped the timer, and
evaluated individually and separately [89]. Second, could proceed to the next comprehension question.
neuroscientific evidence shows that visual After stopping the timer, participants could not
information is processed in two different imagery modify their answers any further.
subsystems, the object imagery system and the
spatial imagery system [90], resulting in two The participants were given a maximum of 7 min to
distinct visual cognitive styles rather than one [68]. answer each comprehension question. This
restriction was imposed for two reasons: first, to
limit the length of the research session (experiment
Mental Effort: Although a self-assessed measure of
plus questionnaire) to about 1 h; second, because
perceived mental effort may appear questionable,
our questions were quite practical and expressed
especially on a unidimensional scale, several
operational doubts, and did not require legal
studies on CLT had employed this simple and quick
interpretation (Appendix III): if contracts are to be
method (for a review, see [60]), as these scales have
seen as instructions or “blueprint[s] for
been demonstrated to be as sensitive and reliable
performance” [21], it must be possible for readers
as more intrusive and complex methods [91].
to find information and understand in a relatively
Moreover, given the virtual setup of our experiment,
short time what needs to be done in business-
it would have been impossible to use physiological
as-usual situations. If doing so is not possible, then
measures (e.g., heart rate and pupillary responses).
the contract is not an effective communication tool.
Following Paas and Merriënboer [59], we chose a
9-point scale, ranging from “very, very low mental
After the experiment, the participants were asked
effort” (1) to “very, very high mental effort” (9).
to self-assess the mental effort they experienced in
Mental effort was used as a control variable in our
completing the tasks.
regression models, and as a means to calculate
mental efficiency (see “How Data Were Analyzed”).
How Data Were Analyzed To test H1 and H2,
relating to speed and accuracy, we analyzed the
Process for Administering the Study: The comprehension performance scores obtained in the
participants had the chance to get instructions on experiment. The difference in mean answering
the study and on its procedure by attending one of speed in the two experimental groups was analyzed
three webinars, organized at different times to with a Welch t-test, an adaptation of the classical
accommodate the time zones of our international t-test to be alternatively used when two samples
participants and to ensure wide participation. One have unequal variances, as in this case. The
of the authors was available online for the duration difference in answering accuracy, however, had to
of the experiment to answer any further questions be analyzed through the nonparametric
from the participants in regards to the procedure. Mann–Whitney U-test, because the scores of the
The webinars were also recorded for participants experimental group were non-normally distributed,
who could not attend live (2% of the participants), thus violating one of the assumptions required for
and we gave them the possibility to ask further a t-test. The Mann–Whitney U-test could be
questions of one of the authors via email. The link interpreted as a test of medians because its two
This article has been accepted for inclusion in a future issue of this journal. Content is final as presented, with the exception of pagination.

PASSERA ET AL: DIAGRAMS IN CONTRACTS: FOSTERING UNDERSTANDING IN GLOBAL BUSINESS COMMUNICATION 13

main assumptions were respected: the two samples Regression analysis also helped us to prove that
had equal variance, as assessed through a Levene’s the experimental manipulation was the main cause
test, and their distributions were similar in shape for changes in comprehension performance, and
(as assessed through visual inspection of the thus made it possible to discard other variables as
plotted data-points). alternative explanations for differences in
performance between groups, ensuring internal
In order to test H3, we followed the methodology validity. Linear regression was appropriate as it
described in [59] on how to calculate the mental allowed us to include all control variables—
efficiency of alternative instructional conditions. categorical, continuous, and interaction terms—in
The method puts individual performance in the the same analysis. Since our study sample was too
context of related subjectively experienced cognitive small to generate sufficient predictive power if all
load, and vice versa [59]. This procedure was variables and interaction terms had been included
deemed appropriate as it provides an overall score at once, we chose a stepwise method based on
for each document alternative, helping to probability of the F value. We included the
determine which one requires less mental effort variables in the regression model one at a time,
and yields better performance. starting with those with lowest p-values. Variables
already in the equation would be eliminated if their
The first step of the procedure is to standardize p-value became not significant (higher than 0.05)
answering speed, answering accuracy, and mental because of the inclusion of a new variable.
effort scores into z-scores:
(individual score − M) We used a two-way ANOVA to investigate further
z= . the only interaction term, native language ×
SD
experimental treatment, that showed statistical
significance in the regressions. ANOVA was a
Performance is then obtained as the mean of suitable follow-up method because the interaction
zAnswering Speed and zAnswering Accuracy. Mean z-scores of term that we discovered was between categorical
performance and mental effort constitute the independent variables on one continuous
Cartesian coordinates to plot a result point for each dependent variable, and we could more precisely
condition (x-axis = performance; y-axis = mental assess the main effect of each independent variable
effort). Efficiency (E ) of each condition is finally and the way that they interacted.
calculated as the distance of the plotted points
from the bisecting diagonal: To conclude, we also analyzed the simple main
Z Performance − Z Mental effects through univariate post hoc tests and
Effort
E= √ . Bonferroni-adjusted pairwise comparisons. The
2 post hoc tests reveal whether there are simple main
effects at each combination of variable levels (e.g.,
The statistical difference between efficiency scores whether “treatment” has a significant effect on NSs
is then assessed through a t-test, as the data were and NNs), and provides an effect size through the
normally distributed and had homogeneous F-value. The pairwise comparisons establish
variance, and samples were independent. whether the mean difference between two variable
levels within a group is statistically significant (e.g.,
To test H4a, H4b, H5a, H5b, and H5c, related to whether the mean score difference of NSs and NNSs
the interaction of the cognitive style and language within the “visual” treatment group is significant).
with experimental treatment, we employed stepwise
linear regression and two-way analysis of variance All statistical tests and analyses were performed
(ANOVAs). with the statistical software SPSS.

First, we examined the interactions of objective

Assuring Reliability and Validity In this section,
performance measures (speed and accuracy) by
we examine the efforts we took to assure reliability
different verbal skills (OSIVQ verbal skill score,
and validity.
native language, and legal background) and by
different visual skills (OSIVQ object-imagery skill
To ensure ecological validity,
and spatial-imagery skill scores). All variables
involved in the interaction terms were centered to (1) We used an authentic agreement, for which
avoid multicollinearity between these terms and both text-only and visually-enhanced versions
their individual components [92, pp. 37–38]. existed.
This article has been accepted for inclusion in a future issue of this journal. Content is final as presented, with the exception of pagination.

14 IEEE TRANSACTIONS ON PROFESSIONAL COMMUNICATION, VOL. 00, NO. 00, 2017

(2) We asked experts from IACCM to check whether Most participants were from North America
the type of contract we chose, its language and (n = 56) and Europe (n = 45). The most represented
structure, and our comprehension tasks would professions were contract manager (n = 71) and
be plausible and “business as usual” for the lawyer/in-house counsel (n = 16). The most
average contract professional. represented industries were information technology
(3) We ran the experiment with actual contract (n = 36), services/outsourcing/consulting (n = 16),
professionals, rather than using a convenience oil/gas/utilities (n = 15), and technology (n = 11).
sample of master’s students. In terms of cognitive styles, we had 53 predominant
object visualizers, 44 predominant verbalizers, 28
To ensure population validity,
predominant spatial visualizers, and 3 with mixed
1) We sought to increase the cultural and cognitive styles (their highest score was the same
professional diversity of the participants as for 2 or 3 cognitive styles). The average scores of
much as possible by recruiting participants the whole sample were 3.370 for object imagery,
through IACCM, an international association, 2.950 for spatial imagery ability, and 3.230 for
rather than through individual national verbal ability. Sixty-seven people were randomly
associations (some countries do not even have allocated to the control group (answering the
national associations of contract professionals). comprehension questions with the text-only,
We also organized explanatory sessions at original contract), and 55 to the treatment group
different times to eliminate hurdles to (answering the comprehension questions with the
participation due to unfavorable time zones. redesigned visually enhanced contract).
Finally, we avoided preselecting participants
based on specific criteria. However, we Hypothesis Testing The results of our
acknowledge that our participants are hypothesis testing follow.
volunteers, and may also be considered a
convenience sample because to facilitate H1. Contracts including diagrams allow for faster
recruiting, we sought to recruit the members of information finding than traditional, text-only
an international association, rather than trying contracts.
to recruit participants directly from individual
firms randomly chosen from all around the Experimental data on the answering speed of the
world. participants confirmed H1. The treatment group,
To ensure reliability of the instruments, using the visually enhanced version of the contract,
was significantly faster than the control group in
(1) We sought to use scales and procedures finding the relevant information and solving the
documented and validated in previous studies. comprehension tasks. Comparing the sums of
(2) In grading answering accuracy, we followed an answering times of the two groups, in average the
established rationale (Appendix II) and checked treatment group was 182.760 s faster (treatment
interrater reliability. group mean = 897.070 ± 305.750; control group
mean = 1079.830 ± 396.440). This difference is
statistically significant, as assessed through a
RESULTS two-tailed Welch’s t-test (t = −2.940, p = 0.004).
Fig. 2 shows the mean of the time taken to answer
In this section, we report detailed information each question and the mean of the sum of all
about the participant sample, as well as the individual scores.
analyses performed to test our hypotheses.
H2. Contracts including diagrams allow for more
accurate comprehension than traditional, text-only
Participants The sample was composed of 122 contracts.
participants, of whom 59 were women, 31 had a
legal background, and 71 had English as their Experimental data on the answering accuracy
mother tongue and were thus NSs. Among the confirmed H2: the treatment group was
NNSs, the average self-assessed proficiency in significantly more accurate than the control group,
English was 8.570 on an 11-point scale with respectively a mean score of 4.350 correct
(SD = 1.375). The average age of the participants answers versus only 3.010 correct answers. This
was 44.46 years old (SD = 10.12), and on average difference is statistically significant, as assessed
they had been working with contracts for 12.611 through a two-tailed Mann–Whitney U-test
years (SD = 9.032). (U = 2835, p < 0.001). Fig. 3 shows the mean
This article has been accepted for inclusion in a future issue of this journal. Content is final as presented, with the exception of pagination.

PASSERA ET AL: DIAGRAMS IN CONTRACTS: FOSTERING UNDERSTANDING IN GLOBAL BUSINESS COMMUNICATION 15

Fig. 2. Mean answering speed for each comprehension question (left), and total mean answering speed (right).
Results of Welch’s t-test (table at the bottom).

Fig. 3. Mean answering accuracy for each comprehension question, where 0 indicates a wrong answer and 1
indicates a correct answer (left). Total mean answering accuracy, where 0 indicates that participants’ answers were
all wrong, and 6 indicates that they were all right (right). Results of Mann–Whitney U-test (table at the bottom).
This article has been accepted for inclusion in a future issue of this journal. Content is final as presented, with the exception of pagination.

16 IEEE TRANSACTIONS ON PROFESSIONAL COMMUNICATION, VOL. 00, NO. 00, 2017

H4a. Cognitive styles interact with experimental

treatment. Verbalizers are equally fast with either
contract and are fastest with the textual contract;
spatial visualizers are the fastest with contracts
including diagrams; object visualizers become
faster with contracts including diagrams than
textual ones, but remain the slowest performing
group.

H4b. Cognitive styles interact with experimental

treatment. Verbalizers are equally accurate with
either contract, and are most accurate with the
textual contract; spatial visualizers are the most
accurate with contracts including diagrams; object
visualizers become more accurate with contracts
including diagrams than textual ones, but remain
the least accurate group.
We then proceeded to run regression analyses
(descriptive statistics and correlations are provided
in Table II). The results show that cognitive styles
do not interact with experimental treatment, alone
Fig. 4. Graph of efficiency (E) for visual and textual or in concert, and do not have a statistically
conditions. significant effect in predicting answering accuracy
or speed

accuracy scores for each question and as a sum of The regression model explaining answering speed
all individual scores. (see Table III) shows that the only statistically
significant coefficient is experimental treatment
(β = −0.271, p = 0.03), and the model predicts
H3. The mental efficiency of the version of the
answering speed with statistical significance
contract including diagrams is higher than that of
(R 2 = 0.073, F (1, 119) = 9.404, p = 0.03).
the text-only version, meaning that participants
require less mental effort to achieve an equal level
If we look at the regression model explaining
of performance.
answering accuracy (see Table IV), treatment is
also statistically very significant (β = 0.468,
H3 is strongly confirmed as the participants in the p < 0.001). The model is statistically significant in
treatment group not only exerted less mental effort predicting answering accuracy (R 2 = 0.27,
but also achieved higher performance. Given the F (3, 117) = 14.438, p < 0.001), and features two
formula presented earlier, the visual contract’s more significant coefficients—mother tongue, alone
efficiency (0.282) is higher than the textual and in interaction with experimental
contract’s (–0.232), and the difference is treatment—which we will discuss in the next
statistically significant (t = 3.399, p = 0.001). section.
Moreover, the visual contract’s score indicates high
efficiency, while the textual contract’s indicates low To conclude, we can confidently associate the
efficiency (see Fig. 4). We can also exclude that cause of lower response times and increased
performance in the visual group was caused by the comprehension accuracy to the design of the
motivation of investing extra effort to succeed. visualized contract, for visualizers and verbalizers
alike. The results provide further support to accept
Since the intrinsic cognitive load was arguably the H1 and H2, while H4a and H4b are rejected.
same in each group (because all participants had
the same contract content and the same tasks), we H5a. Language interacts with experimental
can infer that the overall decrease in mental effort treatment: NSs of English are faster and more
was caused by a reduction in extraneous cognitive accurate than NNSs with the traditional version of
load, which is the load imposed by inefficient ways the contract (text-only), while NNSs perform faster
of communicating the content and representing a and more accurately with the version of the
problem. contract including diagrams.
This article has been accepted for inclusion in a future issue of this journal. Content is final as presented, with the exception of pagination.

PASSERA ET AL: DIAGRAMS IN CONTRACTS: FOSTERING UNDERSTANDING IN GLOBAL BUSINESS COMMUNICATION 17

TABLE II
DESCRIPTIVE STATISTICS AND CORRELATIONS

Variable Mean SD 1 2 3 4 5 6 7 8 9 10 11

1. Answering 997.438 360.193

speed
2. Answering 3.606 1.431 0.059
accuracy
3. Experimental 0.450 0.500 –0.254∗∗ 0.464∗∗
treatment
4. Age 44.460 10.118 0.021 –0.041 –0.001
5. Gender 0.480 0.502 –0.015 –0.156 –0.086 –0.162
6. Legal 0.270 0.464 –0.028 –0.038 0.183∗ –0.128 0.072
background
7. Years of 12.611 9.03 0.011 0.017 –0.048 0.674∗∗ –0.299∗∗ 0.047
experience in
working with
contracts
8. Mother tongue 0.580 0.495 0.161 0.151 –0.034 0.230∗∗ –0.211∗ –0.079 0.293∗∗
9. Object-imagery 3.375 0.595 –0.039 −0.148 –0.057 –0.093 0.273∗∗ 0.049 –0.145 –0.226∗∗
skill score
10. Spatial- 2.951 0.646 –0.001 0.069 –0.080 –0.065 –0.281∗∗ –0.052 0.025 –0.178∗ –0.001
imagery skill score
11. Verbal skill 3.230 0.499 –0.013 –0.022 –0.028 0.175 0.051 .251∗∗ 0.151 0.161 –0.183 -0.190∗
score
12. Mental effort 4.668 1.556 0.237∗∗ –0.208∗ –0.258∗ –0.048 0.128 –0.046 –0.098 0.065 –0.153 0.076 –0.132

n = 121 ∗ p < 0.05 ∗∗ p <0.01.

As already illustrated above, the experimental this superiority disappears in the treatment group:
treatment is the only factor explaining the native and NNSs are virtually equally accurate, and
answering speed scores, so H5a is partially rejected there is no statistically significant difference in
in regards to this dependent variable: experimental their accuracy (F (1, 118) = 0.67, p = 0.796).
subjects do not differ in their answering speed
because of their native language, but because of In fact, the average of correct answers by NNSs
the contract version they use. (4.385) is even slightly higher than the average for
NSs (4.289) (see Fig. 5). However, this difference is
However, there is a difference in terms of answering statistically insignificant (see Table IX): NNSs
accuracy: the results of the regression analysis “catch up” on NS, performing equally accurately,
show not only a statistically significant coefficient but a full expertise reversal effect is not observed
for experimental treatment, but also for mother because both groups significantly improve their
tongue (β = 0.167, p = 0.037) and for the accuracy with visual contracts, even though the
interaction effect of mother tongue × treatment effect is stronger for NNSs (mean difference = 1.95,
(β = −0.179, p = 0.026). Thus, answering accuracy 95% CI [1.265, 2.635], F (1, 118) = 31.771,
is affected both by the version of the contract used p < 0.001) than for NSs (mean difference = 0.905,
in the comprehension task and by whether the 95% CI [0.320, 1.489], F (1, 118) = 9.397,
respondent is a native English speaker. The p = 0.003).
statistically significant interaction term suggests
that the performance of both native and NNSs is We can conclude that H5a is partially confirmed
affected by the type of contract, but to a different when considering answering accuracy. As
degree. We further investigated this difference and postulated by the expertise reversal effect
assessed whether an expertise reversal effect is prediction, instructional visual aids are more
present using a two-way ANOVA (see Table V), with effective for those with lower expertise (in this case,
post hoc tests (see Tables VI and VII) and pairwise native language expertise). However, a full expertise
comparisons (see Tables VIII and IX). reversal effect is not observed, because even NSs
benefit from visualization—just to a lesser extent
In the control group, NSs are on average than NNSs.
significantly more accurate than NNSs (mean
difference = 0.959, 95% CI [0.350, 1.567], H5b. Profession-specific knowledge interacts with
F (1, 118) = 9.737, p = 0.002), but surprisingly, experimental treatment: subjects with a legal
This article has been accepted for inclusion in a future issue of this journal. Content is final as presented, with the exception of pagination.

18 IEEE TRANSACTIONS ON PROFESSIONAL COMMUNICATION, VOL. 00, NO. 00, 2017

TABLE III
RESULTS OF THE REGRESSION ANALYSIS (ANSWERING SPEED AS THE DEPENDENT VARIABLE)

Variables Coefficients

Variables included in the model

Experimental treatment –0.271∗
Variables excluded from the model, p > 0.05
Age 0.021
Gender –0.021
Legal background 0.012
Years of experience in working with contracts –0.008
Mother tongue 0.136
Spatial-imagery skill score –0.025
Object-imagery skill score –0.089
Verbal skill score –0.011
Mental effort 0.175
Spatial-imagery skill score × Experimental treatment 0.004
Object-imagery skill score × Experimental treatment –0.002
Verbal skill score × Experimental treatment 0.114
Spatial-imagery skill score × Verbal skill score 0.018
Object-imagery skill score × Verbal skill score –0.087
Object-imagery skill score × Spatial-imagery skill score –0.065
Legal background × Mother tongue 0.086
Legal background × Experimental treatment 0.086
Mother tongue × Experimental treatment 0.084
Object-imagery skill score × Spatial-imagery skill score × Experimental treatment –0.129
Object-imagery skill score × Verbal skill score × Experimental treatment –0.066
Spatial-imagery skill score × Verbal skill score × Experimental treatment –0.089
Object-imagery skill score × Spatial-imagery skill score × Verbal skill score –0.137
Legal background × Mother tongue × Experimental treatment –0.029
Object-imagery skill score × Spatial-imagery skill score × Verbal skill score × Experimental treatment 0.051
Adjusted R2 0.065
R2 0.073
F (1, 119) 9.404
Significance of the model, p 0.003

n = 121 ∗ p = 0.003.

background are faster and more accurate with the CONCLUSIONS, LIMITATIONS, AND SUGGESTIONS
traditional version of the contract (text-only), while FORFUTURE RESEARCH
subjects without legal background perform faster
and more accurately with the version of the In this section, we elaborate on our results, and
contract including diagrams. their implications for practice and theory. We close
by considering the limitations of the study and by
suggesting areas for future research.
H5c. Language interacts with profession-specific
knowledge and experimental treatment: NSs of Conclusions In this study, we argued for
English with a legal background are the fastest integrating diagrams in B2B contracts as a way to
and most accurate with the traditional version of enhance comprehension speed, accuracy, and
the contract (text-only), while NNSs without a legal ultimately communication in an international
background are fastest and most accurate with business setting. Given the novelty of the
the version of the contract including approach, we proposed four research questions to
diagrams. explore the merit of this suggestion: to what extent
this approach may be beneficial, whether
As illustrated in Tables III and IV, all regression individual characteristics may alter its
coefficients involving the variable “Legal effectiveness, and what the implications for
background”—alone or in interaction with other international businesses might be.
terms—are not significant in accounting for
answering speed and accuracy results. Thus, H5b We discovered that adding diagrams to contracts
and H5c are rejected. indeed supports more accurate and faster
This article has been accepted for inclusion in a future issue of this journal. Content is final as presented, with the exception of pagination.

PASSERA ET AL: DIAGRAMS IN CONTRACTS: FOSTERING UNDERSTANDING IN GLOBAL BUSINESS COMMUNICATION 19

TABLE IV
RESULTS OF THE REGRESSION ANALYSIS (ANSWERING ACCURACY AS THE DEPENDENT VARIABLE)

Variables Coefficients

Variables included in the model

Experimental treatment 0.468∗∗
Mother tongue 0.167∗∗∗
Mother tongue × Experimental treatment –0.179∗∗∗
Variables excluded from the model, p > 0.05
Age –0.083
Gender –0.055
Legal background –0.095
Years of experience in working with contracts –0.012
Spatial-imagery skill score 0.139
Object-imagery skill score –0.085
Verbal skill score –0.023
Mental effort –0.075
Spatial-imagery skill score × Experimental treatment 0.029
Object-imagery skill score × Experimental treatment 0.070
Verbal skill score × Experimental treatment 0.025
Spatial-imagery skill score × Verbal skill score –0.005
Object-imagery skill score × Verbal skill score –0.046
Object-imagery skill score × Spatial-imagery skill score –0.015
Legal background × Mother tongue 0.016
Legal background× Experimental treatment –0.092
Object-imagery skill score × Spatial-imagery skill score × Experimental treatment –0.076
Object-imagery skill score × Verbal skill score × Experimental treatment 0.130
Spatial-imagery skill score × Verbal skill score × Experimental treatment –0.155
Object-imagery skill score × Spatial-imagery skill score × Verbal skill score 0.074
Legal background × Mother tongue × Experimental treatment 0.052
Object-imagery skill score × Spatial-imagery skill score × Verbal skill score × Experimental treatment –0.038
Adjusted R2 0.251
R2 0.270
F (3, 117) 14.438
Significance of the model, p < 0.001
∗∗ ∗∗∗
n = 121 p < 0.01 p < 0.05.

TABLE V
TWO-WAY ANOVA: TEST OF BETWEEN-GROUPS EFFECTS ON ANSWERING ACCURACY

Source of Variance Sum of Squares df Mean Squares F p-value Partial Eta Squared

Treatment 59.945 1 59.945 39.410 <0.001 0.250

Mother Tongue 5.587 1 5.587 3.673 0.058 0.030
Treatment × Mother Tongue 8.041 1 8.041 5.286 0.023 0.043
Error 179.485 118 1.521
Total 247.740 121

TABLE VI
POSTHOC UNIVARIATE TESTS ON SIMPLE MAIN EFFECT OF TREATMENT WITHIN
“MOTHER TONGUE” GROUPS

Source of Variance Sum of Squares df Mean Squares F p-value Partial Eta Squared

NSs 14.293 1 14.293 9.397 0.003 0.074

Error 179.485 118 1.521
NNSs 48.326 1 48.326 31.771 <0.001 0.212

Error 179.485 118 1.521

This article has been accepted for inclusion in a future issue of this journal. Content is final as presented, with the exception of pagination.

20 IEEE TRANSACTIONS ON PROFESSIONAL COMMUNICATION, VOL. 00, NO. 00, 2017

TABLE VII
POSTHOC UNIVARIATE TESTS ON SIMPLE MAIN EFFECT OF “MOTHER TONGUE”
WITHIN TREATMENT GROUPS

Source of Variance Sum of Squares df Mean Squares F p-value Partial Eta Squared

Treatment group (visual) 0.102 1 0.102 0.067 0.796 0.001

Error 179.485 118 1.521
Control group (textual) 14.811 1 14.811 9.737 0.002 0.076
Error 179.485 118 1.521

TABLE VIII
PAIRWISE COMPARISONS: MEAN DIFFERENCE OF “TREATMENT” WITHIN MOTHER TONGUE GROUPS

Mother-tongue Mean Score Treatment Mean Score Control Mean Difference p-value 95% Confidence Interval
Groups Group (visual) Group (textual) for Difference

Lower Bound Upper Bound

NSs 4.298 3.394 0.905 0.003 0.320 1.489

NNSs 4.385 2.435 1.950 <0.001 1.265 2.635

TABLE IX
PAIRWISE COMPARISONS: MEAN DIFFERENCE OF “MOTHER TONGUE” WITHIN TREATMENT GROUPS

Treatment Groups Mean Score NSs Mean Score NNSs Mean Difference p-value 95% Confidence Interval for Difference

Lower Bound Upper Bound

Treatment group (visual) 4.298 4.385 –0.087 0.769 –0.757 0.577

Control group (textual) 3.394 2.435 0.959 0.002 0.350 1.567

comprehension of their content, and the increase

in accuracy is especially significant for NNSs of
English, and positive also for NSs, thus reinforcing
our belief that this approach may be highly
beneficial in international business, where English
is the de facto lingua franca. Moreover, we have not
observed negative reversal effects caused by any
individual characteristic of respondents (e.g.,
cognitive styles, expertise, and professional
background), leading us to conclude that this
approach will not potentially backfire with specific
user groups.
Implications for Practice: The implications for
global business are important, because through a
multimodal approach to contract design, firms
could ensure that professionals from different
backgrounds and cultures understand each other
and collaborate more easily. This mutual
understanding and collaboration, in return, could
lead to more effective negotiations, prevent legal
troubles resulting from contract misinterpretation,
Fig. 5. Mean answering accuracy for NS and NNS in the
control group and in the treatment group.
and improve business relationships. Technical and
professional communicators, as well as information
designers, could apply their skills in the field of
This article has been accepted for inclusion in a future issue of this journal. Content is final as presented, with the exception of pagination.

PASSERA ET AL: DIAGRAMS IN CONTRACTS: FOSTERING UNDERSTANDING IN GLOBAL BUSINESS COMMUNICATION 21

contracts and help global firms conduct representations that are more explicit and readily
transactions more effectively and successfully. applicable in relation to a given task (we will return
to this point in the “Limitations” below). There is no
Visualization skills can be seen as part of the such thing as the correct diagram, only diagrams
strategic competence of the global communicator, well-suited to answer specific questions and
whose goal is to communicate effectively achieve specific goals.
to “get the job done” [6]. In Louhiala-Salminen
and Kankaanranta’s Global Communication These results also suggest avoiding
Competence framework [7], in fact, communication predeterministic assumptions about what modes of
proficiency goes beyond grammatical and discourse presentation may work well for particular cognitive
proficiency. One of the key abilities of competent styles. Counterintuitively, individual cognitive
global communicators is the ability to continuously styles did not have an effect on how the
check for understanding (their own and that of participants processed either version of the
their communication partners) through questions, contract. Object and spatial visualizers were not at
repeated utterances, and the use of more than one a particular disadvantage in understanding the
channel or modality. We suggest that a competent traditional contract (it was equally difficult for
global communicator should craft messages verbalizers), and verbalizers did not benefit less
in whichever way offers the best possibilities from diagrams in the experimental version of the
of constructing shared understanding: proficiently contract. The “fitness-for-purpose” of a design
using different media and modalities is a critical skill apparently trumps cognitive style in predicting the
for professionals involved in international business. comprehension performance of an individual
processing particularly complex information: dense
Implications for Theory: The results from the prose will remain dense even for the best
experiment show how a supplantation strategy [69] verbalizer, and an irrelevant diagram will not be
based on diagrams helps international managers helpful even for the most skilled spatial visualizer.
understand contracts more accurately and more
quickly. In particular, we discovered that diagrams Limitations This study is not without limitations.
can support NNSs of English in developing First, while experiments provide a way to control
a more accurate understanding of a complex for variables, they are carried out in an artificial
document written in English, the shared business setting. In real life, managers might overcome their
language of today’s professional communicators. difficulties in understanding a document by
This effect will ultimately close the accuracy discussing and sharing knowledge with colleagues,
gap in regards to NSs. These results are partially so in addition to individual cognitive mechanisms,
explained in terms of expertise reversal effect there are social mechanisms at play in
(see, e.g., [62]–[66]); while NNSs benefit more than sensemaking that our study does not consider.
NSs from visual aids in contracts, the latter also
see an improvement in their answering accuracy. Second, the participants in our experiments
constituted a convenience sample: at best, the
These results can be further explained in light of results can be generalized to the population of the
CLT [48], [49]: when information is already members of IACCM Association. The members of
intrinsically complex and requires substantial this association are very active in learning
processing, inefficient or confusing presentation contractual and commercial best practices, and all
formats need to be minimized as much as possible; participants were volunteers. Arguably, the
better instructional design prevents extraneous participants were interested from the start in trying
cognitive load by offering a meaningful problem out their contract comprehension skills or having a
representation. When this problem representation chance to try out a visual contract. This interest
offers a ready solution to the task at hand, the may affect the results and participants’
design is more mentally efficient: fewer cognitive characteristics in terms of nonresponse bias, as we
resources need to be invested, and understanding do not know how the respondents may compare
will be better and faster. with nonrespondents.

However, we believe that the presence of There is good evidence that volunteers in
visualizations is not enough to ensure psychological research are, for instance, more
comprehension [28], [62], as for some tasks we educated and have higher IQs than nonvolunteers
found that there were no performance [93]. If this factor was at play, it may have affected
improvements. Diagrams must offer problem the representativeness of our sample and the mean
This article has been accepted for inclusion in a future issue of this journal. Content is final as presented, with the exception of pagination.

22 IEEE TRANSACTIONS ON PROFESSIONAL COMMUNICATION, VOL. 00, NO. 00, 2017

overall performance results. Moreover, the fact that

they were invited to participate in the study
through the IACCM may have biased them in
unconsciously considering the visual contract a
superior solution from the very start. On the other
hand, running the experiment without the help of
the Association would have made it almost
impossible to recruit such a varied and
international sample, which was necessary to
assess the possible benefits of our proposal for
global communicators.
Fig. 6. Suggestion for improving the “Warranty Period”
diagram (see Appendix I).
Third, to further increase validity, the sample
might have been bigger, or the experiment might
have been replicated within different communities. the diagrams did not provide a problem
Within the scope of this study, however, we felt representation that was sufficiently helpful for the
confident to present only one experiment, because respondents and readily applicable.
converging results have been obtained in other
studies where visual communication has been used Question 3 (Appendix III) asks the participant to
to improve speed and accuracy in contract calculate the correct duration of the warranty
interpretation (see, e.g., [9], [12], [94]). period. The diagram (Appendix I) does suggest how
to perform the calculation correctly but does not
Fourth, some variables may be more rigorously provide a ready-made answer and does not reframe
measured. We acknowledge that actual English the problem. The readers still needed to apply a
language proficiency could have been more “tricky” rule. A better approach would have been,
thoroughly assessed. Since a self-assessment perhaps, to create a simple table proposing an
without an objective test may not be reliable, we easier-to-apply rule, as we suggest in Fig. 6.
did not include self-assessed English proficiency as
a control variable in our analysis. However, Question 5 (Appendix III) had very high correct
differences in English proficiency among response rates in both groups because, we
NNSs—and in literacy among NSs—may yield more discovered, it is customary that the Supplier
nuanced results. Moreover, we considered only the carries the risk for the equipment during its
verbal-imagery dimension of cognitive style and testing. Most managers were probably familiar with
found no evidence for their effect on contract this convention and already possessed the relevant
comprehension. Future studies might need to knowledge to answer the question very quickly. The
additionally consider different dimensions of diagram (Appendix I) as well as the text thus was
cognitive style, for instance the wholist-analytical rather irrelevant, as it was not the main driver
dimension [70]–[72]. Better controls for the culture behind people’s good performance.
of the participants could also be included, as visual
representations are interpreted differently in Suggestions for Future Research This study
high-context and low-context cultures [95]. establishes the usefulness of diagrams in
simplifying contract interpretation and supporting
Finally, it is to be noted that the results of our effective global communication. We suggest that
study apply only to diagrammatic visualizations. further research should go beyond experimental
The contracts used in the experiment did not evaluation, by possibly observing in vivo how visual
include other types of visuals such as photographs, representations support comprehension and
illustrations, or cartoons that possess a more knowledge sharing. Such a study could shed light
pictorial or metaphorical nature. Moreover, more on how global communicators might interpret the
attention should be paid to the type of diagram (or very presence of visualizations in a contract: would
visual representation used) in relation to the task the benefits in comprehension be accompanied by
that participants need to complete. More research an increase in trust and collaboration? Or would
is needed to discover the actual doubts and people feel patronized? Would they fear that images
practical goals of contract professionals, as could be used insincerely, thus leading to lower
identifying those tasks is the first step to envision mutual trust between the parties?
truly helpful diagrams. For example, in tasks 3 and
5 there was no significant difference in accuracy or Other studies looking at both experimental and
speed across treatments because, we hypothesize, observational data could help us shed light on
This article has been accepted for inclusion in a future issue of this journal. Content is final as presented, with the exception of pagination.

PASSERA ET AL: DIAGRAMS IN CONTRACTS: FOSTERING UNDERSTANDING IN GLOBAL BUSINESS COMMUNICATION 23

what diagrams more effectively foster competent Finally, one possible avenue to take would be to
contractual communication, and how and when investigate the relationship between cognitive load,
they should be used in an international business information design, and cognitive styles more
setting. Is there such a thing as an archetypical, thoroughly, as the current study counterintuitively
best way to communicate certain clauses, or does showed a weak link between the verbal-imagery
communication effectiveness depend on the goal dimension of cognitive style and comprehension of
readers want to achieve with that information? Are differently presented materials. Such an
certain types of representations consistently clearer investigation might reinforce the argument that
and easier to process than others, or does their individual cognitive styles ultimately do not
effectiveness depend on individual and cultural significantly impact learning [96]; on the other
variables? Future research could take into hand, it may suggest a mechanism between
consideration other types of visuals in addition to cognitive load and cognitive styles that is
diagrams, for instance comics and photographs, undiscovered in this study and that would explain
which have been reportedly used in contracts [13], these results.
[14], but have not been systematically evaluated.

complement rather than substitute for the

APPENDIX I
text.
EXCERPTS FROM THE VISUALIZED CONTRACT
Figs. 7–11 illustrate the “visualized contract” used
in our study, including explanatory diagrams that

APPENDIX II

RATIONALE FOR GRADING THE ACCURACY OF PARTICIPANTS RESPONSES

The following rationale is copyrighted (
C 2013) by
0.75 points = The participant provided a correct,
Stefania Passera. precise, and clear response in his/her own words,
eventually referencing the clause or rule s/he
0.25 points = The participant mentions only the applied. However, there is a minor imprecision
clause number where the correct answer can be (e.g., if answering correctly would require to
found, or copy-pastes the relevant excerpt. provide a precise date such as 4th March 2016, the
However, s/he does not provide a correct, precise, participant wrote 5th March 2016).
and clear response in his/her own words (e.g., if
answering correctly would require to provide a 1 point = The answer is fully correct. The
precise date, the date is not given). participant provided a correct, precise, and clear
response in his/her own words, eventually
0.50 points = The participant wrote down the referencing the clause or rule s/he applied (e.g., if
relevant clause or rule in his/her own words. answering correctly would require to provide a
However, s/he does not provide a correct, precise, precise date, the date is given).
and clear response in his/her own words (e.g., if
answering correctly would require to provide a
precise date, the date is not given).

APPENDIX III Question 1 – What are the consequences if the

Test Run of the Equipment fails for the second
COMPREHENSION QUESTIONS time?
The following questions are copyrighted (
C 2013)
Question 2 – Given that the date of provisional
by Stefania Passera. acceptance is 6th March 2012, which date marks
the end of the availability measurement period?
This article has been accepted for inclusion in a future issue of this journal. Content is final as presented, with the exception of pagination.

24 IEEE TRANSACTIONS ON PROFESSIONAL COMMUNICATION, VOL. 00, NO. 00, 2017

Fig. 7. Sample spread of the visualized agreement.

Fig. 8. This excerpt from “Test Runs” (clause + diagram) provides an answer to Question 1.
This article has been accepted for inclusion in a future issue of this journal. Content is final as presented, with the exception of pagination.

PASSERA ET AL: DIAGRAMS IN CONTRACTS: FOSTERING UNDERSTANDING IN GLOBAL BUSINESS COMMUNICATION 25

Fig. 9. This excerpt from “Warranty” (clause + diagram) provides an answer to Question 3.

Fig. 10. This excerpt from “Liquidated Damages” (clause + diagram) provides an answer to Question 4.

Question 3 – Given that the date of the delivery Question 4 – The first batch of the Equipment
of the Equipment is 15th July 2011, and the date delivery was delayed 6 weeks. The second batch
of provisional acceptance is 6th March 2012, on of the Equipment delivery was delayed 3 weeks.
what date does the Warranty Period The third batch of the Equipment delivery was
expires? again delayed 3 weeks. What is the total amount
This article has been accepted for inclusion in a future issue of this journal. Content is final as presented, with the exception of pagination.

26 IEEE TRANSACTIONS ON PROFESSIONAL COMMUNICATION, VOL. 00, NO. 00, 2017

Fig. 11. This excerpt from “Transfer of risk and ownership” (clause + diagram) provides an answer to Question 5.

of liquidated damages that the Supplier is liable brought back to the Purchaser on 20th May
to pay to the Purchaser? 2015. Given that the date of provisional
acceptance is 6th March 2012, on what date does
Question 5 – Which of the Parties shall bear the the Warranty Period for such reconditioned part
risk for deterioration and damage to the
expire?
Equipment during the Test Runs?
Question 6 – A spare part, included in the
original scope of supply, is taken into service and

REFERENCES
[1] I. Austen. (2006 Oct. 25). The comma that costs 1 million dollars (Canadian). New York Times. [Online].
Available: http://www.nytimes.com/2006/10/25/business/worldbusiness/25comma.html
[2] M. Salzwedel, “Face it – Bad legal writing wastes money,” Mich. Bar J., vol. 92, no. 3, pp. 52–53, 2013.
[3] J. Kimble, “Writing for dollars, writing to please,” Scribes J. Legal Writing, vol. 6, pp. 1–38, 1996–1997.
[4] J. Kimble, “Answering the critics of plain language,” Scribes J. Legal Writing, vol. 5, pp. 51–85, 1994–1995.
[5] H. Haapio, “Next generation contracts: A paradigm shift,” Ph.D. dissertation, Univ. Vaasa, Vaasa, 2013.
[6] A. Kankaanranta and L. Louhiala-Salminen, “‘English? – Oh, it’s just work!’: A study of BELF users’
perceptions,” Engl. Specific Purposes, vol. 29, no. 3, pp. 204–209, 2010.
[7] L. Louhiala-Salminen and A. Kankaanranta, “Professional communication in a global business context: The
notion of global communicative competence,” IEEE Trans. Prof. Commun., vol. 54, no. 3, pp. 244–262, Sep. 2011.
[8] J. Kimble, Lifting the Fog of Legalese. Essays on Plain Language. Durham, NC, USA: Carolina Academic Press,
2005.
[9] S. Passera, “Beyond the wall of text: How information design can make contracts user-friendly,” in Design, User
Experience, and Usability: Users and Interactions. A. Marcus, Ed. Berlin, Germany: Springer International
Publishing, 2015, pp. 341–352.
This article has been accepted for inclusion in a future issue of this journal. Content is final as presented, with the exception of pagination.

PASSERA ET AL: DIAGRAMS IN CONTRACTS: FOSTERING UNDERSTANDING IN GLOBAL BUSINESS COMMUNICATION 27

[10] C. Stephens. (2016 Dec. 6). “Building plain language from the ground up, part IV – Layout and design principles
applied.” Plain Language Association International. Sep. 19, 2011. [Online]. Available:
http://shelternetbc.com/wp-content/
uploads/2012/08/Layout-and-Design-Cheryl-Stevens.pdf.
[11] C. M. Baldwin, Plain Language and the Document Revolution. Washington, DC, USA: Lamplighter Press, 1999.
[12] T. Mamula and U. Hagel. (2015). The design of commercial conditions. Jusletter IT. [Online]. Available:
http://jusletter-it.weblaw.ch/issues/2015/IRIS/the-design-of-commer_ca59099de3.html__ONCE
[13] K. Rekola and K. Boucht, “Visualizing service contracts – The case of an equipment manufacturer,” in Proc. 15th
Int. Conf. Inf. Visual., 2011, pp. 359–364.
[14] H. Haapio, D. A. Plewe and R. de Rooy. (2016). “Next generation deal design—Comics and visual platforms for
contracting,” Jusletter IT. [Online]. Available: http://jusletter-it.weblaw.ch/issues/2016/IRIS/next-generation-
deal_560222777a.html__ONCE
[15] R. Waller, J. Waller, H. Haapio, G. Crag, and S. Morrisseau, Cooperation through clarity: Designing simplified
contracts, J. Strat. Contract. Negot., vol. 2, no. 1–2, pp. 48–68, 2016.
[16] G. Berger-Walliser, R. C. Bird, and H. Haapio, “Promoting business success through contract visualization,” J.
Law Bus. Ethics, vol. 17, pp. 55–75, 2011.
[17] S. Passera, S. Pohjonen, K. Koskelainen, and S. Anttila, “User-friendly contracting tools—A visual guide to
facilitate public procurement contracting,” in Proc. IACCM Acad. Forum Integrating Law Contract Manage.:
Proactive, Preventive, Strat. Approaches, 2013, pp. 74–94.
[18] M. Kay and M. Terry, “Textured agreements: Re-envisioning electronic consent,” in Proc. 6th Symp. Usable
Privacy Security, 2010, pp. 13:1–13:13.
[19] S. Passera, “Enhancing contract usability and user experience through visualization—An experimental
evaluation,” in Proc. 16th Int. Conf. Inf. Visual., 2012, pp. 376–382.
[20] J. A. Mitchell. (2013). Putting some product into work-product: Corporate lawyers learning from designers.
Social Sci. Res. Netw. Working Paper 2325683. [Online]. Available: http://dx.doi.org/10.2139/
ssrn.2325683
[21] L. A. DiMatteo, G. J Siedel, and H. Haapio, “Strategic contracting: Examining the business-legal interface,” in
Proactive Law in a Business Environment. G. Berger-Walliser and K. Østergaard, Eds. Copenhagen, Denmark:
DJOF Publishing, 2012, pp. 59–106.
[22] T. Tenbrink and A. Maas, “Efficiently connecting textual and visual information in operating instructions,” IEEE
Trans. Prof. Commun., vol. 58, no. 4, pp. 346–366, Dec. 2015.
[23] W. Horton, “Dump the dumb screen dumps,” Tech. Commun., vol. 40, pp.146–148, 1993.
[24] M. Gellevij, H. van der Meij, T. de Jong, and J. Pieters, “The effects of screen captures in manuals: A textual and
two visual manuals compared,” IEEE Trans. Prof. Commun., vol. 42, no. 2, pp. 77–91, Jun. 1999.
[25] D. Roy, “Designing procedural graphics for surgical patient-education modules: An experimental study,” Tech.
Commun. Quart., vol. 17, no. 2, pp. 173–201, 2008.
[26] D. Roy and R. Grice, “Helping readers connect text and visuals in sequential procedural instruction: developing
reader comprehension,” Tech. Commun., vol. 51, no. 4, pp. 517–525, 2004.
[27] R. E. Mayer, W. Bove, A. Bryman, R. Mars, and L. Tapangco, “When less is more: Meaningful learning from
visual and verbal summaries of science textbook lessons,” J. Educ. Psychol., vol. 88, no. 1, pp. 64–73, 1996.
[28] W. Schnotz and M. Bannert, “Construction and interference in learning from multiple representation,” Learn.
Instruct., vol. 13, no. 2, pp. 141–156, 2003.
[29] M. J. Albers, Human-Information Interaction and Technical Communication: Concepts and Frameworks. Hershey,
PA, USA: IGI Global, 2012.
[30] A. Boekelder and M. Steehouder, “Selecting and switching: some advantages of diagrams over tables and lists for
presenting instructions,” IEEE Trans. Prof. Commun., vol. 41, no. 4, pp. 229–241, Dec. 1998.
[31] V. M. Holland and A. Rose, “A comparison of prose and algorithms for presenting complex instructions,” TR 17,
American Institutes for Research, Washington, DC, USA, 1981.
[32] M. Gellevij, “Visuals in instruction,” Ph.D. dissertation, Univ. Twente, Twente, the Netherlands, 2002.
[33] R. Krull, S. J. D’Souza, D. Roy, and D. M. Sharp, “Designing procedural illustrations,” IEEE Trans. Prof.
Commun., vol. 47, no. 1, pp. 27–33, Mar. 2004.
[34] C. J. M. Jansen and M. F. Steehouder, “The sequential order of procedural instructions: Some formal methods
for designers of flow charts,” J. Tech. Writing Commun., vol. 26, no. 4, pp. 453–473, 1996.
[35] A. A. Maes and H. Lenting, “How to put the instructive space into words,” IEEE Trans. Prof. Commun., vol. 42,
no. 2, pp. 100–113, Jun. 1999.
[36] F. Ganier, “Factors affecting the processing of procedural instructions: Implications for document design,” IEEE
Trans. Prof. Commun., vol. 47, no. 1, pp. 15–26, Mar. 2004.
[37] V. Khatri, I. Vessey, S. Ram, and V. Ramesh, “Cognitive fit between conceptual schemas and internal problem
representations: The case of geospatio-temporal conceptual schema comprehension,” IEEE Trans. Prof.
Commun., vol. 49, no. 2, pp. 109–127, Jun. 2006.
[38] W. Fukuoka, Y. Kojima, and J. H. Spyridakis, “Illustrations in user manuals: Preference and effectiveness with
Japanese and American readers,” Tech. Commun., vol. 46, no. 2, pp. 167–176, 1999.
[39] D. E. Stone and M. D. Glock, “How do young adults read directions with and without pictures?” J. Educ.
Pyschol., vol. 73, no. 3, pp. 419–426, 1981.
[40] P. Wright and F. Reid, “Written information: Some alternatives to prose for expressing the outcomes of complex
contingencies,” J. Appl. Psychol., vol. 57, no. 2, pp. 160–166, 1973.
This article has been accepted for inclusion in a future issue of this journal. Content is final as presented, with the exception of pagination.

28 IEEE TRANSACTIONS ON PROFESSIONAL COMMUNICATION, VOL. 00, NO. 00, 2017

[41] A. M. Glenberg and W. E. Langston, “Comprehension of illustrated text: Pictures help to build mental models,” J.
Mem. Lang., vol. 31, no. 2, pp. 129–151, 1992.
[42] P. Alexander, “The usability of print and online video instructions,” Tech. Commun. Quart., vol. 22, no. 3,
pp. 237–259, 2013.
[43] R. H. W. Waller, “Graphic aspects of complex texts: typography as macropunctuation,” in Processing of Visible
Language. P. A. Kolers, M. E. Wrolstad, and H. Bouma, Eds. New York, USA: Plenum, 1980, pp. 247–253.
[44] S. A. Bernhardt, “Seeing the text,” Coll. Comp. Commun., vol. 37, no. 1, pp. 66–78, 1986.
[45] C. Kostelnick, “Typographical design, modernist aesthetics, and professional communication,” J. Bus. Tech.
Commun., vol. 4, no. 1, pp. 5–24, 1990.
[46] C. Kostelnick, “Supra-textual design: The visual rhetoric of whole documents,” Tech. Commun. Quart., vol. 5,
no. 1, pp. 9–33, 1996.
[47] M. Betrancourt and A. Bisseret, “Integrating textual and pictorial information via pop-up windows: An
experimental study,” Behav. Inf. Technol., vol. 17, no. 5, pp. 263–273, 1998.
[48] J. Sweller, “Cognitive load theory, learning difficulty, and instructional design,” Learn. Instruct., vol. 4, no. 4,
pp. 295–312, 1994.
[49] J. Sweller, J. J. G. van Merrienboer, and F. G. W. C. Paas, “Cognitive architecture and instructional design,”
Educ. Psychol. Rev., vol. 10, no. 3, pp. 251–296, 1998.
[50] A. Baddeley, “Working memory: Looking back and looking forward,” Nature Rev. Neurosci., vol. 4, no. 10,
pp. 829–839, 2003.
[51] A. Paivio, Imagery and Verbal Processes. Oxford, UK: Holt, Rinehart & Winston, 1971.
[52] M. C. Wittrock, “Generative learning processes of the brain,” Educ. Psychol., vol. 27, no. 4, pp. 531–541, 1992.
[53] M. Scaife and Y. Rogers, “External cognition: How do graphical representations work?” Int. J. Comput. Studies,
vol. 45, no. 2, pp. 185–213, 1996.
[54] F. Paas, A. Renkl, and J. Sweller, “Cognitive load theory and instructional design: recent developments,” Educ.
Psychol., vol. 38, no. 1, pp. 1–4, 2003.
[55] N. Debue and C. van de Leemput, “What does germane load mean? An empirical contribution to the cognitive
load theory,” Frontiers Pyschol., vol. 5, pp. 1–12, 2014.
[56] R. Moreno, “When worked examples don’t work: Is cognitive load theory at an impasse?” Learn. Instruct., vol. 16,
no. 2, pp. 170–181, 2006.
[57] F. Paas and T. van Gog, “Optimising worked example instruction: Different ways to increase germane cognitive
load,” Learn. Instruct., vol. 16, no. 2, pp. 87–91, 2006.
[58] C. Jarrett, J. C. Redish, K. Summers, and K. Straub. (2010). “Design to read: Guidelines for people who do not
read easily user experience mag.,” User Experience Mag., vol. 9, no. 2. [Online]. Available:
http://uxpamagazine.org/people_who_do_not_read_easily/
[59] F. G. W. C. Paas and J. J. G. V. Merriënboer, “The efficiency of instructional conditions: an approach to combine
mental effort and performance measures,” Human Factors: J. Human Factors Ergon. Soc., vol. 35, no. 4,
pp. 737–743, 1993.
[60] F. Paas, J. E. Tuovinen, H. Tabbers, and P. W. M. Van Gerven, “Cognitive load measurement as a means to
advance cognitive load theory,” Educ. Psychol., vol. 38, no. 1, pp. 63–71, 2003.
[61] N. Marcus, M. Cooper, and J. Sweller, “Understanding instructions,” J. Educ. Pyschol., vol. 88, no. 1, pp. 49–63,
1996.
[62] R. E. Mayer and J. K. Gallini, “When is an illustration worth ten thousand words?” J. Educ. Psychol., vol. 82,
no. 4, pp. 715–726, 1990.
[63] J. G. Lambiotte and D. F. Dansereau, “Effects of knowledge maps and prior knowledge on recall of science
lecture content,” J. Exper. Educ., vol. 60, no. 3, pp. 189–201, 1992.
[64] R. E. Mayer, K. Steinhoff, G. Bower, and R. Mars, “A generative theory of textbook design: Using annotated
illustrations to foster meaningful learning of science text,” ETR&D, vol. 43, no. 1, pp. 31–41, 1995.
[65] R. E. Mayer, Multimedia Learning. Cambridge, MA, USA: Cambridge University Press, 2001.
[66] H. Lee, J. L. Plass, and B. D. Homer, “Optimizing cognitive load for learning from computer-based science
simulations,” J. Educ. Psychol., vol. 98, no. 4, pp. 902–913, 2006.
[67] R. E. Mayer and R. Moreno, “Nine ways to reduce cognitive load in multimedia learning,” Educ. Psychol., vol. 38,
no. 1, pp. 43–52, 2003.
[68] O. Blazhenkova and M. Kozhevnikov, “The new object-spatial-verbal cognitive style model: Theory and
measurement,” Appl. Cog. Pyschol., vol. 23, no. 5, pp. 638–663, 2009.
[69] L. J. Ausburn and F. B. Ausburn, “Cognitive styles: Some information and implications for instructional design,”
ECTJ-Educ. Commun. Technol., vol. 26, no. 4, pp. 337–354, 1978.
[70] R. Riding and I. Cheema, “Cognitive styles—An overview and integration,” Educ. Psychol., vol. 11, no. 3–4,
pp. 193–215, 1991.
[71] R. J. Riding, “On the nature of cognitive style,” Educ. Psychol., vol. 17, no. 1–2, pp. 29–49, 1997.
[72] S. Rayner and R. Riding, “Towards a categorisation of cognitive styles and learning styles,” Educ. Psychol.,
vol. 17, no. 1–2, pp. 5–27, 1997.
[73] M. Kozhevnikov, S. Kosslyn, and J. Shephard, “Spatial versus object visualizers: A new characterization of visual
cognitive style,” Memory Cog., vol. 33, no. 4, pp. 710–726, 2005.
[74] A. Richardson, “Verbalizer-visualizer: A cognitive style dimension,” J. Mental Imagery, vol. 1, no. 1, pp. 109–125,
1977.
[75] O. Blazhenkova and M. Kozhevnikov, “Visual-object ability: A new dimension of non-verbal intelligence,”
Cognition, vol. 117, no. 3, pp. 276–301, 2010.
This article has been accepted for inclusion in a future issue of this journal. Content is final as presented, with the exception of pagination.

PASSERA ET AL: DIAGRAMS IN CONTRACTS: FOSTERING UNDERSTANDING IN GLOBAL BUSINESS COMMUNICATION 29

[76] R. Marschan, D. Welch, and L. Welch, “Language: The forgotten factor in multinational management,” Eur.
Manage. J., vol. 15, no. 5, pp. 591–598, 1997.
[77] R. Fredriksson, W. Barner-Rasmussen, and R. Piekkari, “The multinational corporation as a multilingual
organization: The notion of a common corporate language,” Corporate Commun. Int. J., vol. 11, no. 4,
pp. 406–423, 2006.
[78] D. Maclean, “Beyond English: Transnational corporations and the strategic management of language in a
complex multilingual business environment,” Manage. Dec., vol. 44, no. 10, pp.1377–1390, 2006.
[79] L. Louhiala-Salminen and A. Kankaanranta, “Language issues in international internal communication: English
or local language? If English, what English?” Public Rel. Rev., vol. 38, no. 2, pp. 262–269, 2012.
[80] B.B. Kachru, “Englishization and contact linguistics,” World Engl., vol. 13, no. 2, pp. 135–154, 1994.
[81] W. Baker, “Culture and complexity through English as a lingua franca: Rethinking competences and pedagogy
in ELF,” J. Engl. Lingua Franca, vol. 4, no. 1, pp. 9–30, 2015.
[82] C. Meierkord, Interactions Across Englishes. Linguistic Choices in Local and International Contact Situations.
Cambridge, UK: Cambridge University Press, 2012.
[83] J. Jenkins, The Phonology of English as an International Language. Oxford, UK: Oxford University Press, 2000.
[84] J. Jenkins, A. Cogo, and M. Dewey, “Review of developments in research into English as a lingua franca,” Lang.
Teaching, vol. 44, no. 10, pp. 281–315, 2011.
[85] T. Virkkula-Räisänen, “Linguistic repertoires and semiotic resources in interaction: A Finnish manager as a
mediator in a multilingual meeting,” Int. J. Bus. Commun., vol. 47, no. 4, pp. 380–407, 2010.
[86] L. Louhiala-Salminen, M. Charles, and A. Kankaanranta, “English as a lingua franca in Nordic corporate
mergers: Two case companies,” Engl. Specific Purposes, vol. 24, no. 4, pp. 401–421, 2005.
[87] C. Candlin, V. Bhatia, and C. Jensen, “Developing legal writing materials for English second language learners:
Problems and perspectives,” Engl. Specific Purposes, vol. 21, no. 4, pp. 299–320, 2002.
[88] Interagency Language Roundtable. (2016 Dec. 6). Descriptions of proficiency levels. [Online] Available:
http://www.govtilr.org/Skills/ILRscale1.htm
[89] G. Thierry and C. J. Price, “Dissociating verbal and nonverbal conceptual processing in the human brain,” J.
Cog., Neurosci., vol. 18, no. 6, pp. 1018–1028, 2006.
[90] R. Cabeza and L. Nyberg, “Imaging cognition II: An empirical review of 275 PET and fMRI studies,” J. Cog,
Neurosci., vol. 12, no. 1, pp. 1–47, 2000.
[91] F. G. W. C. Paas, J. J. G. V. Merriënboer and J. J. Adam, “Measurement of cognitive load in instructional
research,” Perceptual Motor Skills, vol. 79, no. 1, pp. 419–430, 1994.
[92] L.S. Aiken and S.G. West, Multiple Regression: Testing and Interpreting Interactions. Newbury Park, CA, USA:
Sage, 1991.
[93] R. Rosenthal and R. L. Rosnow, The Volunteer Subject. New York, USA: Wiley, 1976.
[94] S. Passera, “Flowcharts, swimlanes, and timelines– Alternatives to prose in communicating legal-bureaucratic
instructions to civil servants,” J. Bus, Tech. Commun., to be published.
[95] S. Bresciani, “Visualizing knowledge for organizational communication within and across cultures,” Ph.D.
dissertation, Università della Svizzera Italiana, Lugano, Switzerland, 2011.
[96] M. H. Dembo and K. Howard, “Advice about the use of learning styles: A major myth in education,” J. Coll.
Reading Learn., vol. 37, no. 2, pp. 101–109, 2007.

Stefania Passera received the M.A. degree in Graphic Design Leena Louhiala-Salminen received the M.Sc. degree in
from the Aalto School of Art, Design, and Architecture, Aalto Economics from the Helsinki School of Economics, Helsinki,
University, Helsinki, Finland, in 2011. She is currently working Finland, in 1983, and the Lic.Phil. degree in English and Ph.D.
toward the Ph.D. degree in Organizational Behavior at the Aalto degree in Applied Linguistics from the University of Jyväskylä,
University School of Science, Aalto University. She also works as Jyväskylä, Finland, in 1995 and 1999, respectively. She is
a Freelance Information Designer and is the Initiator of Legal currently a Senior Fellow in the Department of Management
Design Jam. Her main research interest includes information Studies, Aalto University School of Business, Helsinki. She is
design applied to legal information, and her doctoral research also a Program Director of the Master’s Program in Corporate
focuses on how visual communication can be used to enhance Communication, Aalto University School of Business. Her main
the usability and user experience of contracts. research interests include the various genres of business
communication, the use of English as business lingua franca,
Anne Kankaanranta received the M.Sc. degree in Economics and corporate communication in international contexts.
from the University of Vaasa, Vaasa, Finland, in 1984, the
EMBA degree from the Helsinki School of Economics, Helsinki,
Finland, in 1997, and the Ph.D. degree in Applied Linguistics
from the University of Jyväskylä, Jyväskylä, Finland, in 2005.
She is a Senior University Lecturer in Organizational
Communication in the Department of Management Studies,
Aalto University School of Business, Helsinki. She has been a
Visiting Scholar at the University of Michigan, Ann Arbor, MI,
USA; Southampton, U.K.; and Aarhus, Denmark. Her research
interests include language policies in MNCs, including the use
of English as business lingua franca, intercultural business
communication, and business communication genres.