Electronic copy available at: http://ssrn.com/abstract=1869013 i XBRL-Enabled, Excel or PDF? The Effects of Exclusive Technology Choice on the Analysis of Financial Information Diane Janvrin Associate Professor of Accounting Iowa State University 3365 Gerdin Business Building Ames, IA 50011 Phone: (515) 294-9450 Fax: 515 294 3525 Email: [email protected]Robert Pinsker Assistant Professor of Accounting Florida Atlantic University Boca Raton FL 33431-0991 Phone: (561) 297-3422 Fax: (561) 297-7023 Email: [email protected]Maureen Francis Mascha Assistant Professor of Accounting University of Wisconsin-Oshkosh Email: [email protected]Please do not quote without permission. June 2011 We thank EDGAR-Online for use of their software. We thank Michael Alles, Roger Debreceny, Stephanie Farewell, Carsten Felden, James Hunton, Jon Perkins, Saeed Roohani, Nathan Stuart, and participants at Bryant University, University of Wisconsin – Oshkosh, University of Kansas 2011 XBRL Symposium, 22 nd XBRL International Conference, 2011 South Florida Research Symposium, and Iowa State University Accounting/Finance Research Forum for helpful comments, Sue Childs and Liv Watson for their software assistance and Kevin Den Adel, Jim Kurtenbach, Patrick Wheeler, and Winston Chappell for assisting with the experimental instrument development and data collection.
Transcript
Electronic copy available at: http://ssrn.com/abstract=1869013
i
XBRL-Enabled, Excel or PDF? The Effects of Exclusive Technology Choice on the Analysis of Financial Information
Diane Janvrin Associate Professor of Accounting
Iowa State University 3365 Gerdin Business Building
Maureen Francis Mascha Assistant Professor of Accounting University of Wisconsin-Oshkosh Email: [email protected]
Please do not quote without permission. June 2011
We thank EDGAR-Online for use of their software. We thank Michael Alles, Roger Debreceny, Stephanie Farewell, Carsten Felden, James Hunton, Jon Perkins, Saeed Roohani, Nathan Stuart, and participants at Bryant University, University of Wisconsin – Oshkosh, University of Kansas 2011 XBRL Symposium, 22nd XBRL International Conference, 2011 South Florida Research Symposium, and Iowa State University Accounting/Finance Research Forum for helpful comments, Sue Childs and Liv Watson for their software assistance and Kevin Den Adel, Jim Kurtenbach, Patrick Wheeler, and Winston Chappell for assisting with the experimental instrument development and data collection.
Electronic copy available at: http://ssrn.com/abstract=1869013
ii
XBRL-Enabled, Excel or PDF? The Effects of Exclusive Technology Choice on the Analysis of Financial Information
Abstract
U.S. adoption of search-facilitating technology has been slow and several constituencies question whether investors will choose to use the XBRL-formatted information the Securities and Exchange Commission (SEC) is now requiring companies to provide. Unlike prior research, we use an exclusive choice experimental design to examine (1) which reporting technology nonprofessional investors will choose to complete a financial analysis task (XBRL-enabled, portable document file, or spreadsheet) and (2) why they choose the specific technology. We found 66 percent of nonprofessional investor proxies chose to use XBRL-enabled technology, while 34 percent chose spreadsheets. Participants who chose XBRL-enabled technology perceived it reduces the time required to complete the task (i.e., increases task efficiency) while participants who chose spreadsheets indicated their technology choice was driven by amount of prior experience with that technology. Our findings have implications for the technology choice literature, regulators mandating or considering mandating XBRL-based reporting, and XBRL-enabled technology adoption. Keywords: exclusive technology choice, user choice, XBRL-enabled technology adoption Data Availability: Contact the authors.
1
XBRL-Enabled, Excel or PDF? The Effects of Exclusive Technology Choice on the Analysis of Financial Information
I. INTRODUCTION
Technology is a pervasive and growing component of accounting tasks and has been
shown to change work processes (Mauldin and Ruchala 1999, 328). With regard to financial
analysis tasks, recent technology advancements create the opportunity to analyze financial
information more efficiently. For example, eXtensible Business Reporting Language (XBRL)
proponents argue that data reported in this format will be more transparent to nonprofessional
investors (Baldwin et al. 2006; Gunn 2007; XBRL International Standards Board [XSB] 2010).
With individual financial statement items tagged, investors are able to quickly search for and
analyze the information they perceive to be important during the financial analysis process using
search-facilitating (also known as XBRL-enabled) reporting technology; that is parser software
designed to manipulate financial information tagged in XBRL (Hodge et al. 2004; Locke et al.
2009). Thus, XBRL proponents contend that investors who formerly relied on either
spreadsheets (e.g., Excel) or document exchange software (e.g., PDF) to perform financial
analysis tasks may be able to perform these tasks more efficiently with XBRL-enabled reporting
technology.1
Motivated by the goal of assisting nonprofessional investors in understanding financial
information (Cox 2006), the Securities and Exchange Commission (SEC) recently mandated that
all publicly traded firms furnish financial information tagged in XBRL as exhibits to their
quarterly and annual filings (SEC 2009) and place the tagged financial information on their
1 Although prior research (i.e., Hodge et al. 2004; Blankespoor et al. 2011) use the term ‘search-facilitating’ to describe parser software that manipulates financial information tagged in XBRL, we elect to describe this technology as ‘XBRL-enabled’ since investors can also conduct searches using spreadsheets or PDFs. As noted in Section II, searches using spreadsheets or PDFs are often more labor intensive.
2
Investor Relations web sites. However, despite the significant investment by firms, regulatory
agencies, and academics throughout the world,2 relevant research and anecdotal evidence suggest
that investors may elect to use other reporting technologies rather than XBRL-enabled
technology when performing financial analysis tasks (Hodge et al. 2004). For example, Hodge et
al. (2004) find that despite the purported benefits of XBRL, almost 50 percent of nonprofessional
investor participants did not use XBRL-enabled technology in their information acquisition task.
The extant user choice literature finds that users do not always choose to use the best decision
aid feature for the task (Jones and Schkade 1995; Whitecotton and Butler 1998; Wheeler and
Jones 2003). Further, prior research (e.g., Ghani et al. 2009; Pinsker and Wheeler 2009),
assigned participants one reporting technology rather than directly address the issue of our study:
whether investors will choose an XBRL-enabled or another reporting technology (e.g., Excel or
PDF) to download and analyze financial information. If firms provide their financial information
in XBRL format, but investors choose to analyze the information using other reporting
technologies (e.g., Excel or PDF), then the SEC’s mandate may produce XBRL-formatted
information that investors will not use. Thus, mandatory XBRL adoption may not provide the
expected benefits to investors (see Pinsker and Li 2008 for more details).
We conjecture that nonprofessional investors are likely to choose to use only one
reporting technology when obtaining and analyzing financial information due to time constraints
and the desire to minimize cognitive effort. While the extant user choice literature (Jones and
Schkade 1995; Whitecotton and Butler 1998; Rose 2002; Wheeler and Jones 2003) finds that
2 XBRL supporters include a consortium of over 400 CPA firms, companies (e.g., Microsoft), regulators (e.g., Securities and Exchange Commission), standard setters and accounting bodies (e.g., the Financial Accounting Standards Board, the International Accounting Standards Committee and the Canadian Institute of Chartered Accountants [Trites 1999; Debreceny and Gray 2001; Debreceny et al. 2005]).
3
users may choose a decision aid feature not created for the specific purpose, little is known
regarding why users may make a choice which appears to be ‘irrational.’ We expand the user
choice literature that examines exclusive choice between alternative decision aid features to
investigate exclusive choice between reporting technologies. Accordingly, we define technology
choice as the decision between alternative technologies.
Further, we develop theoretically-driven explanations as to why users may make these
‘irrational’ choices. Based on existing user choice literature, we hypothesize that nonprofessional
investors will choose a reporting technology due to efficiency issues (i.e., users perceive the task
will take less time to complete).3 Further, we employ cognitive fit theory and the Technology
Acceptance Model (TAM; Davis 1989) to develop two alternative explanation hypotheses. The
first, based on cognitive fit theory, suggests that nonprofessional investors may base their
technology choice on prior experience with that technology. The second, based on TAM,
postulates that nonprofessional investors will choose a reporting technology due to its perceived
usefulness and perceived ease of use.
We examine nonprofessional investors’ choice of three reporting technologies (PDF,
Excel, and XBRL-related) since most U.S. companies’ Investor Relations web sites and the
SEC’s web site present data in formats which are usable by these three reporting technologies.4
We trained participants to use each reporting technology to perform a financial analysis task.
Following training, participants were required to choose one of the three technologies to
complete the financial analysis task. The task consisted of selecting specific items from the
3 As explained in greater detail later in the paper, we examine the impact of technology choice for nonprofessional investors since XBRL proponents argue that nonprofessional investors, rather than professional financial analysts, are most likely to benefit from the technology (Hodge et al. 2004; Baldwin et al. 2006). 4 We acknowledge that these technologies can be used to perform non-financial reporting tasks.
4
financial statements of two existing companies, calculating three financial ratios to create
investment decision model estimates for each company, and making an investment choice based
on these estimates. Modifying the experimental design used by Hodge et al. (2004), we selected
experimental companies where one company’s net income was significantly higher than the
second company. However, if the participants searched for the required financial statement items
and created the investment decision model estimates correctly, the second company had a better
investment decision model estimate than the first.
Results indicate that more participants (66 percent) chose to use the XBRL-enabled
technology proxy (i.e., EDGAR-Online’s I-Metrix) to complete their financial analysis task
relative to completing the task using a spreadsheet (34 percent). When examining why users
chose a specific reporting technology, we find the participants who chose the XBRL-enabled
reporting technology indicated that the task was more efficient (i.e., took less time) using XBRL-
enabled reporting technology compared to Excel or PDF. Further, participants who chose Excel
indicated that their technology choice was driven by the amount of prior experience with that
technology. Finally, participants did not (1) choose PDF or (2) base their technology choice on
perceived usefulness or perceived ease of use.
Our study makes both academic and practical contributions to the extant research. While
Hodge et al. (2004) found evidence that XBRL-enabled technology provides better data search
functionality, they noted that almost 50 percent of participants did not choose to use XBRL-
enabled technology. Further, Pinsker and Wheeler (2009) report that investors perceive that
XBRL-enabled technology is beneficial; however, they did not directly address the issue of
which reporting technology investors would prefer to choose to complete a financial analysis
task. We extend Hodge et al. (2004) and Pinsker and Wheeler (2009) by (1) designing a study
5
that directly requires nonprofessional investors to choose between alternative reporting
technologies, and (2) examining why nonprofessional investors choose a particular reporting
technology to complete their financial analysis task. We also extend the extant user choice
literature describing choices between alternative decision aid features (Wheeler and Jones 2003)
to investigate technology choice, that is users’ exclusive choice between alternative technologies.
Our results may be of interest to academics examining technology choice, investor groups and
the SEC who are interested in nonprofessional investor behavior, regulators considering
mandating XBRL reporting requirements, and software designers developing XBRL-enabled
reporting technology for investor use.
II. BACKGROUND AND HYPOTHESIS DEVELOPMENT
XBRL represents a unique financial reporting format relative to previous reporting
formats, offering investors a choice that facilitates searching activities (Hodge et al. 2004).
Clements and Wolfe (2000, 79) observe that “media choice in financial reporting is a new
phenomenon brought on by the widespread use of multimedia-capable computers and financial
reporting on the Internet.” We examine financial report users’ choice of reporting technology for
analyzing financial information. Proponents of XBRL argue that XBRL-tagged financial
statements increase information transparency by allowing users to search for and directly select
the individual financial items they deem to be most important and use these items in financial
analysis tasks without little or no data reentry required (Pinsker and Wheeler 2009; Janvrin and
Mascha 2010). In contrast, investors must sequentially search for specific financial items
reported in Excel and PDF formats and then manually enter this information into their financial
analysis model (Bartley et al. 2011).
6
Research to date has focused either on the practical benefits of XBRL to users (Baldwin
et al. 2006) or technical aspects of XBRL implementation (Plumlee and Plumlee 2008; Janvrin
and Mascha 2010; Bartley et al. 2011). In addition, prior research and debate throughout the SEC
XBRL mandate implementation process has lamented the slow rate of XBRL-enabled
technology adoption by investors (Cox 2006; Gunn 2007; Troshani and Rao 2007; XSB 2010;
Bartley et al. 2011; Blankespoor et al. 2011). Debreceny et al. (2005) suggest that developing
investor applications for XBRL-formatted documents is important to XBRL’s success. Pinsker
and Wheeler (2009) assigned participants one reporting technology to use (either XBRL-enabled
or paper-based) and found that users perceive that XBRL-enabled reporting technology is
beneficial for financial analysis. Further, Ghani et al. (2009) assigned participants one of three
reporting technologies to use: PDF, HTML, or XBRL-enabled. They find that participants using
XBRL-enabled technology provided higher perceived usefulness ratings than did PDF or HTML
participants; however, perceptions of ease of use were similar across the three reporting
technologies. Finally, the XBRL International Standards Board (2010) recently issued a strategic
initiative to catalyze development of more XBRL-enabled technology.
Unfortunately, prior research (e.g., Ghani et al. 2009; Pinsker and Wheeler 2009) does
not require participants to choose between using XBRL-enabled or other reporting technologies
to perform financial analysis even though many proponents of XBRL argue that nonprofessional
investors will rely exclusively on XBRL-enabled technology instead of spreadsheet or portable
document file technology (Cox 2006; Locke et al. 2009). Thus, practitioners and standard setters
assume that nonprofessional investors will choose XBRL-enabled technology for financial
analysis (Pinsker and Wheeler 2009). Examining technology choice is important since prior
research suggests that user choice may be driven by heuristics that lead to suboptimal
7
performance (Wheeler and Jones 2003). Further, Hodge et al. (2004) found that nearly 50
percent of their participants did not choose to use XBRL-enabled technology. Therefore, we first
examine:
RQ1: Which reporting technology (PDF, Excel, or XBRL-enabled) will nonprofessional investors choose to complete a complex, financial analysis task?
Our research assumes that investors must make an exclusive choice of reporting
technology due to time constraints and the desire to minimize cognitive effort. That is, the
investors must choose between using XBRL-enabled or another reporting technology typically
available on firms’ Investor Relations web sites (specifically, Excel and PDF) to perform
financial statement analysis. We expand user choice literature used in prior accounting
information systems’ research that examined exclusive choice between alternative decision aid
features (Wheeler and Jones 2003) to investigate technology choice. We define technology
choice as the decision to use one technology to the exclusion of all other alternative reporting
technologies. Further, we develop theoretically-driven explanations to examine why users’ chose
a specific reporting technology for a financial analysis task.
TechnologyChoice
Nonprofessional investors often encounter uncertainty when choosing among reporting
technologies. Early psychological and social science research commonly relied on expected
utility theory to examine decision-making under uncertainty. However, expected utility theory
often failed to predict choice behavior (see Einhorn and Hogarth 1986) since findings from
examining simple gambles in highly-controlled laboratory settings may not generalize to
complicated uncertainties individuals face in real-world settings. In response, Hogarth and Reder
8
(1987) developed rational choice theory. Rational choice theory relies on rationality to explain
choices and assumes that decision-makers pursue behaviors at their lowest possible costs, given
their beliefs (Pincione and Teson 2006). Yet, multiple studies in varying contexts indicate that
decision-makers did not make rational choices (operationalized as “optimal” choices) due to
ambiguity avoidance (Ellsberg 1961), perceived high domain expertise (Arkes et al. 1986), or
high self-determination (i.e., self-attribution; Becker 1997). These reasons suggest that
individuals make decisions consistent with their own mental models. Thus, individuals may have
difficulty incorporating information (e.g., a new technology) outside their existing mental models
(Lewis et al. 1988).
The user choice literature examining decisions between alternative decision aid features
evolved from rational choice theory. Based on user choice literature, Wheeler and Jones (2003)
find that users provided with alternative decision aid features must engage in some form of
choice behavior. Similarly, to analyze financial information, nonprofessional investors may
choose to rely on (1) one reporting technology to the exclusion of the others; (2) one reporting
technology for one instance of a task, but use other technologies subsequently; or (3) two or
more reporting technologies to varying degrees for the same task (e.g., analyze the same task
separately with each technology). Case studies and discussions with practitioners suggest that
option (2) is unlikely since once most participants adopt a reporting technology, they use it when
completing related instances of the same or similar tasks. Further, prior research indicates that
option (3) is unlikely because of time constraints, the desire to minimize cognitive effort, the
opportunity costs incurred by using additional technologies, and the belief that the use of more
than one reporting technology is unnecessary due to personal competence (Wheeler and Jones
2003). Consequently, we examine option (1).
9
Dos Santos and Bariff (1988, 461) state that there exists “the common assumption that
greater flexibility and choice in software aids will promote improved user performance.”
However, research does not fully support this prediction. Wheeler and Jones (2003) found that
user choice systematically resulted in sub-optimal decisions, because many users’ chose not to
use the decision aid feature. Rather, users evoke heuristics and over-attribute their own abilities
as a means to circumvent the aid. Whitecotton and Butler (1998) find similar results.5
Heuristics also play a role in the relationship among problem representation, cognitive
effort, and perceived task efficiency. For example, Hoch and Schkade (1996) find evidence that
users choose the technology that most efficiently utilizes their cognitive ability in a demanding
environment. More specifically, Jones and Schkade (1995) state that decision makers will apply
the representativeness heuristic when they encounter an unknown or unfamiliar problem
format/representation if they perceive that doing so outweighs the cost of increased cognitive
effort. The representativeness heuristic helps the decision makers reformat the problem into a
more familiar or representative mental model, which increases task efficiency.
Although the extant user choice literature finds uncertainty with regard to the relationship
between choice and performance (Wheeler and Jones 2003), the representative heuristic noted in
Jones and Schkade (1995) suggests that perceived user efficiency gains play a role in technology
choice. Thus, decision makers are more likely to invoke the representativeness heuristic and
choose the technology which they perceive can help them complete the task more efficiently by
reducing time needed to perform the task. Accordingly, we propose the following hypothesis.
5 A robust finding in decision aid research is that users rarely rely on the decision aid in an optimal manner, systematically either under-relying or over-relying (cf. Rose 2002 for review).
10
H1: The reporting technology nonprofessional investors choose to complete the financial analysis task will be directly related to perceived task efficiency (i.e., perceived less time to complete the task).
The user choice literature cited to support hypothesis one represents one potential
explanation for technology choice. Yet, this literature is still being developed and is fairly narrow
in scope. Specifically, the prior user choice literature is restricted to investigating (1) whether or
not technology, in the form of features within a researcher-created decision aid, would be
chosen, and (2) if users would rely on the decision aid feature to form a decision. Consequently,
we look to other potentially-related information systems theories, specifically cognitive fit and
TAM, to develop additional explanations regarding why users choose a particular reporting
technology.
Cognitive Fit Theory
Cognitive fit theory suggests that task effectiveness increases as the three-way match
among (1) the problem representation, (2) the problem-solving task, and (3) the users’ problem-
solving skill set increases (Vessey 1991; Vessey and Galletta 1991). The theory suggests that
decision makers correctly perceive improvements in cognitive fit (Vessey and Galletta 1991;
Vessey 2006). Prior psychology literature finds that users may not select decision rules that
would allow them to choose correctly on 70 percent of judgment tasks (Arkes et al. 1986).
Specifically related to technology usage, Hayes (2004) notes that users typically resist giving up
software they are experienced with to learn new software, even though the new software is
perceived to be better. Thus, we argue that prior experience with a technology represents a
potential explanation for technology choice. Our argument is supported by Jones and Schkade’s
(1995) results indicating that decision makers tend to choose the problem representation with
11
which they have the most experience and to a lesser extent by Arkes et al.(1986)’s finding that
those with perceived higher levels of domain expertise tend not to use a given decision rule.
Further, the cognitive effort literature (e.g., Jones and Schkade 1995) suggests that users
choose the problem representation with which they are most experienced. Prior experience is a
heuristic in its own right, because decision makers know they can rely on it to make decisions
and save cognitive effort. Decision makers have a choice to either expend cognitive effort to
reformat a new problem representation into their mental models or not. Consistent with research
examining cognitive fit and cognitive effort, we expect that some decision makers will choose to
not exert the necessary cognitive effort to reformat the problem representation into their mental
models, and instead base their technology choice on prior experience.6
H2: The reporting technology nonprofessional investors choose to complete the financial analysis task will be directly related to their prior experience with that technology.
The Technology Acceptance Model
TAM postulates that when users are provided new technology (e.g., XBRL-enabled
technology) to assist them in performing assigned tasks, their perceptions of the technology’s
usefulness and ease of use significantly influence their acceptance and assumed usage of the
technology (Davis et al. 1989; Bagozzi et al. 1992). Perceived usefulness is the degree to which a
person believes that using a technology could enhance his or her job performance, while
perceived ease of use is the degree to which a person believes that using a technology could be
6 While we acknowledge that some prior studies include experience as a covariate (e.g., Pinsker and Wheeler 2009), cognitive fit theory and our unique research design identify experience as a variable pertinent to our research question.
12
free from effort (Davis et al. 1989). However, TAM does not postulate that a technology’s
functional superiority to other technologies will necessarily result in its perception by users as
superior. In fact, the literature indicates that under-utilization of technology due to users’
negative misperceptions is a major problem in business (Mun and Hwang 2003).
TAM research and related findings have important implications for technology choice.
Pinsker and Wheeler (2009) find a difference in favorability perceptions (relating both to their
own analysis and to the firm providing XBRL-enabled financials) between XBRL-enabled users
and PDF users. Investigating the perceptions of information technology users is important when
examining expectations about the diffusion of technology into the business community, since
TAM research attempts to link user perceptions of technology to technology use (Pinsker and
Wheeler 2009).
Our reporting technology choices allow users to search and retrieve financial information,
but at different thresholds (i.e., XBRL’s drill-down possibilities are at a more detailed level than
Excel or PDF). Therefore, we expect that technology choice should be related ot the level of
perceived usefulness. Further, our reporting technology choices are likely to result in higher
perceived ease of use, because all three choices (PDF, Excel, and XBRL-enabled) allow
nonprofessional investors the ability to gather, integrate, and compare firm data more rapidly,
and therefore, at a lower cost, compared to using paper-based data. We propose the following
hypotheses.
H3a: The reporting technology nonprofessional investors choose to complete the financial analysis task will be directly related to perceived usefulness. H3b: The reporting technology nonprofessional investors choose to complete the financial analysis task will be directly related to perceived ease of use.
13
III. METHOD
Participants
We focus on technology choices made by nonprofessional rather than professional
investors since prior research suggests that search-facilitating technology such as XBRL will
most likely benefit nonprofessional investors. Specifically, prior research documents that
experienced financial analysts generally possess well-defined valuation models and search for
specific financial statements items of interest (Frederickson and Miller 2004; Hunton and
McEwen 1997). In contrast, nonprofessional investors are more likely to possess less developed
valuation models and sequentially search for financial statement information (Hunton and
McEwen 1997; Frederickson and Miller 2004). Thus, nonprofessional investors are more likely
to benefit from using reporting technologies that may facilitate searching activities.
Our nonprofessional investor participants were 53 graduate business students enrolled in
a financial statement analysis course at two medium-sized state universities. Each participant
received course credit for completing the experiment. Table 1 provides descriptive statistics for
our sample. Libby et al. (2002) contend that experiments that focus on the judgments of
nonprofessional investors only require participants who possess basic accounting and investing
knowledge. Our student participants had completed at least two undergraduate financial
accounting courses (not displayed in the table) and over 30 percent had previously bought or sold
common stock or mutual funds. Further, two-thirds of the participants intend on buying stock in
the next five years. On average, participants had analyzed financial statements over five times.
Thus, these demographics suggest that the participants had the requisite course and “hands-on”
experience to perform the financial analysis task.
[Insert Table 1 about here]
14
Procedures and Financial Analysis Task
We used software to conduct the experiment online. As illustrated in Figure 1, the first
screen in phase I provided a brief introduction to the experimental task. Participants then
navigated to an overview of the task and three reporting technologies: PDF, Excel, and XBRL-
enabled.7 Next, the participants were trained to use each reporting technology to complete the
financial analysis task. The financial analysis task involved creating investment decision model
estimates for two companies.8 Similar to Wheeler and Arunachalam (2008), the investment
decision model included three common financial ratios. To create each investment decision
model, participants were required to search for selected items from each company’s financial
statements and use these items to calculate price-to-book, revenue to assets, and operating
income margin ratios. Participants were given weights to apply to each ratio to form their
investment decision models and asked to determine which of the two companies they would
invest in.
[Insert Figure 1 – Experimental Procedure – about here]
The training guided participants through step-by-step instructions of how to use each
reporting technology to calculate the investment decision model for two sample firms, Abaxis
and Blackbaud. All participants viewed the same training tasks although the order of presentation
7 We used EDGAR-Online’s I-Metrix tool to proxy for XBRL-enabled technology. To ensure that our task consisted of a test of the XBRL-enabled technology in general, rather than a test of the specific tool we chose, our task was designed to be completed using many of the XBRL-enabled tools currently available in the marketplace. Further, we pilot tested the experiment using both I-Metrix and Savanet tools to proxy for XBRL-enabled technology and found no significant differences in our results. 8 Consistent with Hodge et al. (2004), we define our task as “complex,” because participants were required to acquire information from financial statements, integrate this information into investment decision models and use the results to form investment judgments.
15
(PDF, Excel, and XBRL-enabled) was randomized. Participants could go back and retrain if they
felt uncomfortable with any reporting technology.
In phase II, participants navigated to an overview screen, which presented summary
financial statement information for two new companies, Concur Technologies and CommVault,
in a neutral Word format.9 Participants were asked to review the summary financial information
and make judgments regarding which of these companies to invest in. Next, participants were
asked to calculate the investment decision models for each company using only one of the three
reporting technologies. This technology choice decision represents the dependent variable
TECHNOLOGY CHOICE. Once participants made their technology choice, the software
prevented them from backtracking and changing their choice. After calculating and recording the
investment decision model estimates for both companies, participants indicated which company
they would prefer to invest in. Modifying the experimental design used by Hodge et al. (2004),
we designed the experiment so that one company’s reported net income was higher while the
other company had a higher (i.e., better) investment decision model estimate.
In phase III, participants completed a post-test questionnaire containing demographic
questions (including prior experience with PDF, Excel, and XBRL-enabled technologies; TAM
questions on perceived usefulness and perceived ease of use, and both open- and close-ended
questions on why the participants chose the particular reporting technology).
Some participants completed the task in the classroom as part of a three-hour, ‘in-lab’
class, while others completed the task outside the class to simulate a typical day-trader. Chi-
9 We chose two unfamiliar companies for the training and two other unfamiliar companies for the actual test in order to avoid familiarity demand effects. Results from demographic questions suggest that participants had no familiarity with either Concur or CommVault.
16
square tests did not show any differences between TECHNOLOGY CHOICE responses between
in-lab and out-of-lab participants (χ2 = 0.03; p-value = 0.87). The in-lab participants took an
average of approximately 45 minutes to complete the entire experiment with a range of 35
minutes to approximately 90 minutes. We were unable to collect this information for the out-of-
lab participants, but do not believe it is a threat to our research design given the insignificant
difference between groups reported above.
Pilot Studies
Using both graduate and undergraduate students, we conducted two years of pilot testing
in order to examine many factors: clarity of the research instrument; understandability and
appropriateness of single-item constructs; time of task; task complexity; generalizability of task
to multiple XBRL-enabled tools; and familiarity with sample companies and technologies
employed. Based on feedback obtained, we attempted to equalize the amount of time necessary
to train on each technology. All three reporting technologies took approximately the same
amount of time for the participants to complete the training, but the XBRL-enabled technology
took up more screen space than the others due to including the requisite number of screen
shots.10 We believe this slight difference is realistic given the relative newness of XBRL
(compared to the other two reporting technologies) and its software tools.
Additional pilot testing included the administration of surveys in order to collect more
process-oriented data. Using a series of t-tests and separating participants based on which
reporting technology they chose, the survey data did not find significant differences (using an
alpha of five percent) between XBRL-enabled technology and Excel choosers with regard to
10 Initially, our design included both a simple task (calculating the current ratio) and a complex task (the current study’s task). Pretests showed an average completion time of approximately 90 minutes; thus, due to time constraints, we decided to include only the more realistic financial analysis task for this study.
17
perceived complexity, length, or confusion of the experimental directions for any of the reporting
technologies. Pilot study survey data also allowed us to be more precise in operationalizing our
explanatory variables.
Dependent Variable
Similar to related user choice research (Wheeler and Jones 2003), we use
TECHNOLOGY CHOICE as our dependent variable. To operationalize TECHNOLOGY
CHOICE, participants were required to choose one reporting technology (PDF, Excel, or XBRL-
enabled) to use to complete the financial analysis task following training for all three reporting
technologies. We chose these three reporting technologies since several U.S. firms’ Investor
Relations web sites (e.g., Microsoft [2011]) and the SEC’s web site allow users to download
financial information in applicable formats. We coded TECHNOLOGY CHOICE as follows: 0 =
As noted above, our design consisted of training in three reporting technologies, followed
by the experimental task of choosing one technology to perform a financial analysis task. All
participants viewed the same training tasks. Based on theory and our pilot study results, our
explanatory variables consisted of the potential explanations why participants would choose a
particular technology for their financial analysis task: efficiency (i.e., takes less time to complete
task), prior experience, perceived usefulness and perceived ease of use.11 Each variable was
measured on an 11-point scale (0-10) with higher scores indicating higher perceived importance.
11 We also included two other potential explanations for choosing a reporting technology for the task: (1) “because the researchers wanted me to choose it” and (2) “because I thought it was the latest state-of-the-art technology.” Both relate to potential demand effects in the experiment, but neither was statistically significant in our upcoming analyses. Thus, we do not discuss them further.
18
Given the uniqueness of a choice setting relative to experimenter-assigned conditions, we
were careful to tie our operationalizations of explanatory variables closely to the most applicable
prior research studies. For instance, we operationalized efficiency as taking less time to complete
the task in order to be consistent with related financial analysis (Hodge et al. 2004) and user
choice (Wheeler and Jones 2003) research. The participants’ prior experience with each
reporting technology was measured consistent with Hoch and Schkade (1996) incorporating 11-
point scales ranging from “novice” to “expert.” For the TAM constructs of perceived usefuleness
and perceived ease of use, we adopt single-item measures consistent with Pinsker and Wheeler
(2009), rather than using the scales developed in Davis’ (1989) seminal work. Single-item (rather
than multiple-item) measures are appropriate when the constructs measured are understandable
per se by participants (Ilgen et al. 1981; Wanous et al. 1997).12 Given the previously-cited pilot
tests, open-ended responses as to why a particular reporting technology was chosen, and the
participants’ self-reported experience with examining firms’ financials (presented in Table 1), we
believe that our single-item TAM measures are appropriate and sufficiently unambiguous to
participants.
IV. RESULTS
Manipulation Checks
Two manipulation check questions were used. The first question asked whether the
participants understood that they had a choice of reporting technology to use to complete the
12 According to Pinsker and Wheeler (2009, 261), prior research notes multi-dimensional constructs (e.g., “personality factors”) as those being too complex for single-item measures. These constructs are typically misunderstood by participants and involve processes of which the participants are unaware. Constructs for which prior research has empirically verified to be appropriate include “job satisfaction” and “probability that an effort affects performance of a task.” These latter constructs have unambiguous meanings to participants and participants are fully aware of their beliefs and perceptions regarding them.
19
investment task (after the training phase). Using a binary “True/False” response, participants
responded to the following statement in the post-test questionnaire, “In the second part of the
study, I was allowed to choose whether I wanted to perform the task using Excel-only, PDF, or I-
Metrix (please respond by choosing True or False).”13 Forty-nine out of 53 participants (92.5
percent) correctly responded “True.” We conclude, therefore, that participants properly
understood they could choose one of three technologies to complete the task.14
The second item was a statement related to understanding the type of task performed.
Specifically, the statement read, “I was asked to make investment judgments.” All 53
participants correctly answered “True.” Thus, the participants appeared to understand the task at
hand. In sum, both manipulation check questions were successful.15
Research Question
We begin our tests by investigating our research question: which reporting technology do
nonprofessional investors’ choose to complete the complex, financial analysis task? Using our
dependent variable, TECHNOLOGY CHOICE, we explored whether one reporting technology
technology (I-Metrix) to perform their task; 18 chose Excel (34 percent), and none chose PDF. A
13 We note that I-Metrix uses Excel sparingly as part of its interface. Thus, we clearly specified Excel on its own versus XBRL-enabled (I-Metrix and its occasional Excel interface) in all experimental materials. 14 Removing the four participants who did not answer this question correctly did not materially affect our results. Thus, we include these participants in our analyses. 15 We also inquired as to the salience of the course credit and if the participants could identify that I-Metrix was an XBRL-enabled reporting technology. Only one participant (1.9 percent) did not respond correctly about receiving course credit. Forty-six participants (86.8 percent) correctly identified I-Metrix as an XBRL-enabled decision-making technology, rather than a database, spreadsheet, or word processor. Removing the one incorrect respondent from the course credit questionnaire item or the seven incorrect respondents on the I-Metrix item did not materially change our results. Thus, we include all 53 participants in our analyses.
20
Chi-Square test shows a critical majority of participants favored the XBRL-enabled technology
relative to Excel (χ2 = 5.45, p-value = 0.02).
Next, we examine why nonprofessional investors’ chose a certain reporting technology
(Excel or XBRL-enabled). We examine the correlations between our four explanatory variables
and TECHNOLOGY CHOICE (see Table 2) to gain an initial understanding of the data.16
Examining Table 2, and using one-tailed tests we find that PRIOR EXPERIENCE is
significantly correlated with TECHNOLOGY CHOICE (corr. = -0.48, p-value < 0.01). Given
our dependent variable coding described earlier, we interpret the negative correlation to suggest
that participants who chose Excel (coded as zero) are more likely to base their decision on prior
experience with the reporting technology than are participants who chose XBRL-enabled
technology. The only other significant explanatory variable was TAKES LESS TIME (corr. =
0.26, p-value < 0.05). Given our coding, we interpret this correlation to suggest that participants
who chose XBRL-enabled technology (coded as one) perceive that they could perform the task
in less time using XBRL-enabled technology than using either PDF or Excel. Neither TAM
construct was correlated with TECHNOLOGY CHOICE.
[Insert Table 2 about here]
Hypothesis Testing
We employ Kruskal-Wallis (KW) to test our hypotheses. The KW test rank-orders the
technology choice explanatory variables to provide insights regarding why nonprofessional
16 Unless otherwise specified, we use two-tailed tests. We also investigated the effect of demographic differences identified in Table 1, with the exception of the last two items related to task realism and task difficulty, as possible covariates in our analyses. Excluding the PRIOR EXPERIENCE variables, no demographic characteristics were significant at a p-value < 0.10. Additionally, we considered risk preference as a factor in technology choice. We used three questions adapted from Kahneman and Tversky (1979), but they did not correlate well together (Cronbach’s Alpha = 0.44), and did not clearly affect TECHNOLOGY CHOICE.
21
investors chose their particular reporting technology. H1 uses extant user choice literature to
predict that participants will choose a reporting technology based on efficiency (time) issues.
Examining Table 3, TAKES LESS TIME has significantly different mean ranks (mean rank of
the Excel choice = 21.33; mean rank of the XBRL-enabled technology choice = 29.91; χ2 = 3.80,
p-value < 0.05); thus supporting H1. Thus, we find that the mean ranks are higher for the XBRL-
enabled choosers (the majority reporting technology chosen). Therefore, we provide evidence
that TAKES LESS TIME is the main factor determining TECHNOLOGY CHOICE for XBRL-
enabled choosers.
H2 incorporates cognitive fit theory to predict that TECHNOLOGY CHOICE will be
directly related to PRIOR EXPERIENCE. As shown in Table 3, the mean ranks for PRIOR
EXPERIENCE are significantly different (mean rank of the Excel choice = 37.81; mean rank of
Thus, our results suggest that for the Excel choosers, PRIOR EXPERIENCE is the main
explanation for their reporting technology choice.
Lastly, H3a and H3b predicted that TECHNOLOGY CHOICE will be directly related to
a reporting technology’s perceived usefulness and perceived ease of use, respectively. Results
shown in Table 3 indicate that neither of the mean rank differences for these TAM variables are
significantly different (for PERCEIVED USEFULNESS [mean rank of the Excel choice = 22.83;
mean rank of the XBRL-enabled technology choice = 29.14; χ2 = 2.05, p-value = 0.15] and
PERCEIVED EASE OF USE [mean rank of the Excel choice = 23.92; mean rank of the XBRL-
enabled technology choice = 28.59; χ2 = 1.11, p-value = 0.29]). This evidence rejects both H3a
22
and H3b and suggests that the TAM measures are not appropriate explanatory variables of
TECHNOLOGY CHOICE.17
[Insert Table 3 about here]
Further Analysis
Single Explanation for Technology Choice
We asked our participants to identify which explanatory variable (TAKES LESS TIME,
PRIOR EXPERIENCE, PERCEIVED USEFULNESS, or PERCEIVED EASE OF USE) was
their most important explanation for their reporting technology choice. There were no significant
differences among the top two explanations chosen (TAKES LESS TIME and PERCEIVED
USEFULNESS; χ2 = 0.01, p-value = 0.94). We then separated the responses by technology
choice (i.e., Excel vs. XBRL-enabled technology) and found mixed results. Specifically, even
though the Excel choosers selected PRIOR EXPERIENCE as their most important explanation
most often (eight times relative to only three times for the next highest selection choice), it was
not statistically significantly higher than the other choices (χ2 = 7.56, p-value = 0.11). Despite
not being able to distinguish the XBRL-enabled technology choosers most common explanation,
we find that PRIOR EXPERIENCE represented the least important explanation (only two
17 As a supplemental analysis, we examined the technology choice explanations using principal component analysis. Our initial analysis revealed six factors with eigenvalues greater than one, but no clear pattern regarding factor loadings. Therefore, we employed Varimax and Promax rotation method. Both rotations resulted in three factors, but the Promax rotation, being oblique, provided greater differences between variable loadings (SAS 2005). PERCEIVED USEFULENESS, PERCEIVED EASE OF USE, and TAKES LESS TIME each load on one factor which we refer to as EFFICIENCY (0.46, 1.00, and 0.99, respectively). PRIOR EXPERIENCE loads almost exclusively (1.00) on a second factor. Finally, confirming the earlier non-significant results noted in footnote #11 and our decision to exclude from subsequent analysis, the explanations ‘researcher wanted me to choose it’ and ‘latest state-of-the-art technology’ load on third factor (0.71 and 1.00 respectively). Thus, principal component analysis provides results consistent with our main analyses.
23
participants chose this explanation; χ2 = 11.29, p-value = 0.01). Thus, we provide weak
incremental evidence to supplement the hypothesis testing results.
Relating the Technology Choice to Investment Choice
Modifying the experimental design used by Hodge et al. (2004), we chose our sample
companies such that one company seemed to be the “better” investment choice when considering
only the face financials, but the other company would be chosen if the participants correctly
calculated and relied upon the investment decision model (a proxy for accuracy). The experiment
was run over two years with different graduate business students and we used actual firm data
reported using the actual reporting technologies employed by the companies. Interestingly, the
companies reversed as to the “correct” company to invest in (using the investment decision
model estimates). In the first year, CommVault had the “better” financials, but Concur
Technologies had the higher investment decision model estimate. The opposite was true in the
second year. Thus, this reversal of firm performance biases us against finding an effect.
Rather than only coding the individual companies, we used a binary coding of 0 = the
“incorrect company” and 1 = the “correct” company to invest in given the investor decision
model estimates. Twenty-four participants chose the “correct” company to invest in; 24 chose the
“incorrect” company; five participants did not provide a choice. Chi-Square testing (χ2 = 0.87; p-
value = 0.35) shows no relationship between TECHNOLOGY CHOICE and choosing the
“correct” company (i.e., no accuracy differences in favor of a given technology).18
18 Since we initially asked participants to choose a company solely given their financial data, we attempted to link this initial decision to TECHNOLOGY CHOICE. A Chi-Square test indicated no statistical difference in initial accuracy choice between those choosing XBRL-enabled and Excel (χ2 = 1.22, p-value = 0.27). Further testing did not find any accuracy differences for those who changed their mind relative to participants who did not change their mind from their initial company chosen (χ2 = 0.43, p-value = 0.51) or when comparing participants who changed from an incorrect (correct) initial company choice to a correct (incorrect) final investment company choice (χ2 = 0.38, p-value = 0.83).
24
Investment Decision Model and Acquisition Technology
We explored the perceived importance (or lack thereof) of TECHNOLOGY CHOICE on
the investment decision model. We used an 11-point scale (0-10) with the higher scores meaning
higher perceived importance. Using a t-test, we found that the mean of the Excel choosers (7.17,
S.D. = 1.25) was marginally higher than the XBRL-enabled technology choosers (6.23, S.D. =
1.86; t = 1.92, p-value = 0.06). Therefore, we find only weak evidence linking the importance of
reporting technology choice to the investor decision model.19
Next, motivated by Hodge et al. (2004)’s argument that XBRL-formatted data will
facilitate information acquisition, we asked participants which reporting technology (PDF, Excel,
or XBRL-enabled) they would prefer to use to acquire a company’s financial information. Thirty
participants chose XBRL-enabled (56.6 percent), 21 chose Excel (39.6 percent), and two chose
PDF (3.8 percent). A Chi-Square test did not show any significant differences between XBRL-
enabled and Excel choices for acquiring financial information (χ2 = 1.59, p-value = 0.21).
Prior Technology Experience
Our hypothesis testing indicated that PRIOR EXPERIENCE was the key factor as to why
a minority of participants chose Excel over XBRL-enabled technology in their financial analysis
task. We performed an additional paired-samples t-test to provide empirical support for this
finding. We examined the participants’ reported prior experiences with both Excel and XBRL-
enabled reporting technologies as presented in Table 1. Both of these measurements were on 11- 19 Further, we asked participants to confirm which reporting technology chosen when calculating the investment decision model, The participants’ choices included paper-and-pencil, calculator, spreadsheet (Excel), mental methods, I-Metrix (XBRL), PDF, and other. Only 22 participants responded to this question. Of the 22 respondents, 11 chose spreadsheet and nine chose I-Metrix, with one answer each for calculator and PDF. These descriptive results add credence to the dominant choices of spreadsheet and XBRL-enabled technology found in the hypothesis testing.
25
point scales (0-10) with higher scores indicating higher levels of experience. Results indicated
that the participants’ mean PRIOR EXPERIENCE with Excel (7.21, S.D. = 1.92) was
significantly higher than their mean PRIOR EXPERIENCE with XBRL-enabled technology
Slow acceptance of XBRL-enabled technology has prompted recent regulatory mandates
(e.g., the SEC) designed to assist both the regulators themselves and relatively less sophisticated
market participants. Accordingly, we use an exclusive choice experimental design to examine
whether nonprofessional investors will choose XBRL-enabled or another reporting technology to
complete a financial statement analysis task. We also extend the extant user choice literature to
introduce the concept of technology choice and provide specific explanations why
nonprofessional investors choose to use XBRL-related or an alternative reporting technology
when performing financial analyses. Our work is important since recent evidence suggests
investor adoption of XBRL-enabled technology is slow despite the regulation requirements (SEC
2009) and emphasis on XBRL-formatted information (Hodge et al. 2004; Locke et al. 2009;
Bartley et al. 2011). For example, almost 50 percent of participants in Hodge et al. (2004)
research did not use XBRL-enabled reporting technology. Furthermore, some U.S. firms view
XBRL implementation as a compliance exercise and question the benefits of providing XBRL-
formatted statements (Janvrin and No 2011). In other words, XBRL and its underlying
technology may be under-utilized in terms of its main intended purposes: increased efficiency,
increased reusability, and increased transparency of data (SEC 2009; Bartley et al. 2011;
Vasarhelyi et al. 2011).
26
We trained nonprofessional investors to use three reporting technologies available to
analyze data formats commonly found on both U.S. public companies’ Investor Relations’
websites and the SEC’s own site (PDF, Excel, and XBRL-enabled) to perform a financial
analysis task. Using an exclusive choice experimental design, we then asked participants to
choose which reporting technology they would use to complete the same financial analysis task
for two different companies. Results indicate 66 percent of participants chose to use XBRL-
enabled technology while 34 percent chose spreadsheets (i.e., Excel). Further, no participants
chose document exchange software (i.e., PDF).
Results from hypothesis testing provide evidence that: (1) efficiency in performing the
task (i.e., takes less time) was the major factor for participants who chose the XBRL-enabled
technology and (2) prior experience with the reporting technology was the major factor for
participants who chose Excel. Our findings suggest that investors are likely to choose XBRL-
enabled technology due to efficiency issues; a positive sign for XBRL advocates.
We do not find empirical support linking the TAM measures of perceived usefulness and
perceived ease of use to technology choice. Prior TAM literature (e.g., Davis 1989, etc.) does not
consider choice in the decision making process. Our results suggest that acceptance (the typical
TAM dependent variable) may not imply usage even though prior research has assumed this to
be the case (i.e., Pinsker and Wheeler 2009). Thus, acceptance may be a requisite, but not
sufficient factor of choice.
Further, we do not find a single explanation for reporting technology choice. Specifically,
spreadsheet choosers’ documented prior experience with the technology as the dominant factor,
but not sufficiently enough to be statistically more relevant than the other explanations; while the
only conclusion from the XBRL-enabled choosers was that prior experience was not a factor.
27
Our lack of a single explanation finding provides future researchers an opportunity to re-examine
relevant theory and identify other variables that may impact technology choice. We also do not
find any accuracy differences between the reporting technologies chosen. Future research is
needed to validate potential accuracy measures.
Our findings build upon the existing user choice literature that investigates choice
between alternative decision aid features to examine technology choice. We define technology
choice as choice between alternative (reporting) technologies. Further, we develop theoretically-
based explanations as to why individuals chose the specific reporting technology. We argue that
examining technology choice in the context of nonprofessional investors’ choice between using
XBRL-enabled technology relative to other reporting technologies is appropriate since
proponents of XBRL-enabled technology suggest that investors will use it exclusively instead of
other reporting technologies.
As with all experimental work, we acknowledge research limitations. Several limitations
may suggest future research opportunities. First, our participants were required to choose one of
three reporting technologies. While we believe we have chosen the three most common
technologies to display/report financial information, we may have missed a technology. Second,
while we pilot tested our experiment using both Savanet and I-Metrix as our proxies for XBRL-
enabled technology, we used only I-Metrix which provides users with additional financial
statement items not available via XBRL for the actual experiment. Third, our experiment
assumes a cost-free investing environment. Whereas nonprofessional investors may already have
access to PDF and Excel technology, they may need to purchase XBRL-enabled tools. This
consideration was not part of our research design, but represents an interesting avenue for future
research. Fourth, due to time constraints, we use a single task to measure nonprofessional
28
investor technology choice. Future research should consider multiple tasks of varying complexity
to determine if our results generalize or if reporting technology choice rationale varies by task
complexity. Finally, Lowe and Locke (2011) argue that tagging individual financial statement
items with XBRL tags will allow investors to view the financial items outside the underlying
financial statement format and its related linkages.20 Our financial analysis task assumed
participants had a working knowledge of the underlying financial statement format. Future
research could examine exclusive technology choice for tasks where financial items are
presented in ways that do not require knowledge of the financial statement format.
20 For example, when viewing long term debt securities separately in an XBRL-enabled technology, investors do not automatically view the reference point indicating the total amount of long term liabilities that would be shown if the long-term debt securities amount was presented in a financial statement format.
29
REFERENCES
Arkes. H. R., R. M. Dawes, and C. Christensen. 1986. Factors influencing the use of a decision rule in a probabilistic task. Organizational Behavior and Human Decision Processes 37: 93-110.
Bagozzi, R. P., F. D. Davis, and P. R. Warshaw. 1992. Development and test of a theory of technological learning and usage. Human Relations 45 (7): 660-686.
Baldwin, A. A., C. E. Brown, and B. S. Trinkle. 2006. XBRL: An impacts framework and
research challenge. Journal of Emerging Technologies in Accounting 3 (1): 97-116. Bartley, J. W., A. Y. S. Chen, and E. Z. Taylor. 2011. A comparison of XBRL filings to
corporate 10-Ks – Evidence from the voluntary filing program. Accounting Horizons 25 (2): forthcoming.
Becker, D. A. 1997. The effects of choice on auditors’ intrinsic motivation and performance. Behavioral Research in Accounting 9: 1-19. Blankespoor, E., B. P. Miller, H. D. White. 2011. The initial impact of mandatory search-
facilitating technology: Evidence from the XBRL mandate. Working paper, University of Michigan and Indiana University.
Clements, C. E., and C. J. Wolfe. 2000. Reporting financial results with the video medium: An
experimental analysis. Journal of Information Systems 14 (2): 79-94. Cox, C. 2006. The interactive data revolution: Improved disclosure for investors, less expensive
reporting for companies. Available at: http://www.sec.gov/news/speech/2006/spch053006cc.htm.
Davis, F. D. 1989. Perceived usefulness, perceived ease of use, and user acceptance of
information technology. MIS Quarterly 13 (3): 319-340. _______, R. P. Bagozzi, and P. R. Warshaw. 1989. User acceptance of computer technology:
A comparison of two theoretical models. Management Science 35: 982-1003. Debreceny, R., and G. L. Gray. 2001. The production and use of semantically rich accounting
reports on the Internet XML and XBRL. International Journal of Accounting Information Systems 2 (1): 47-74.
_______, A. Chandra, J. Cheh, D. Guithues-Amrhein, N. Hannon, P. Hutchison, D. Janvrin, R. Jones, B. Lamberton, A. Lymer, M. Mascha, R. Nehmer, S. Roohani, R. Srivastava, S. Trabelsi, T. Tribunella, G. Trites, and M. Vasarhelyi. 2005. Financial reporting in XBRL on the SEC’s EDGAR system: A critique and evaluation. Journal of Information Systems 19 (2): 191-210.
30
DeLong, J., A. Shleifer, L. Summers, and R. Waldman. 1989. The size and incidence of the
losses from noise trading. Journal of Finance 44: 681-696. Dos Santos, B. L., and M. L. Bariff. 1988. A study of user interface aids for model-oriented
decision support systems. Management Science 34 (4): 461-468. Einhorn, H. J., and R. M. Hogarth. 1986. Decision making under ambiguity. Journal of Business 59 (4): S225-S250. Ellsberg, D. 1961. Risk, ambiguity, and the Savage axioms. Quarterly Journal of Economics 75: 643-669. Frederickson, J., and J. Miller. 2004. The effects of pro forma earnings disclosures on analysts’
and nonprofessional investors’ equity valuation judgments. The Accounting Review 79 (3): 667-686.
Ghani, E.K., F. Laswad, and S. Tooley. 2009. Digital reporting formats: Users’ perceptions,
preferences and performances. International Journal of Digital Accounting Research 9: 45-98.
Gunn, J. 2007. XBRL: Opportunities and challenges in enhancing financial reporting and
assurance processes. Current Issues in Auditing 1 (1): A36-A43. Hayes, F. 2004. Chaos is back. ComputerWorld November 8. Hoch, S. J., and D. A. Schkade. 1996. A psychological approach to decision support systems. Management Science 42 (1): 51-64. Hodge, F. D., J. J. Kennedy, and L. A. Maines. 2004. Does search-facilitating technology
improve the transparency of financial reporting? The Accounting Review 79 (3): 687-703.
Hogarth, R. M., and M. Reder. 1987. Rational Choice. University of Chicago press.
Hunton, J., and R. McEwen. 1997. An assessment of the relation between analysts’ earnings forecast accuracy, motivational incentives and cognitive information search strategy. The Accounting Review 72 (4): 497-0515.
Ilgen, D., D. Nebeker, and R. Pritchard. 1981. Expectancy theory measures: An empirical comparison in an experimental simulation. Organizational Behavior and Human Performance 28: 189-223. Janvrin, D., and M. F. Mascha. 2010. The process of creating XBRL instance documents: A
research framework. Review of Business Information Systems 14 (2): 11-34.
31
_______, and W. G. No. 2011. XBRL implementation: A field investigation. Working paper,
Iowa State University. Jones, D. R., and D. A. Schkade. 1995. Choosing and translating between problem
representations. Organizational Behavior and Human Decision Processes 61 (2): 214-223.
Kahneman, D., and A. Tversky. 1979. Prospect theory: An analysis of decision under risk. Econometrica 47 (2): 263-291. Lewis, B. L., J. M. Patton, and S. L. Green. 1988. The effects of information choice and
information use on analysts’ predictions of municipal bond rating changes. The Accounting Review 63 (2): 270-282.
Libby, R., R. Bloomfield, and M. W. Nelson. 2002. Experimental research in financial accounting. Accounting, Organizations & Society 27 (8): 775-810.
Locke, J., A. Lowe, and A. Lymer. 2009. Sociotechnical alignment within the interactive data
constituency. Working paper, University of Birmingham, U.K. Lowe, A, and J. Locke. 2011. Digital business reporting: Epistemic culture clash for
accountants. Working paper, Aston University and University of Birmingham. Maudlin, E. G., and L. V. Ruchala. 1999. Towards a meta-theory of accounting information systems. Accounting, Organizations & Society, 24 (4): 317-331. Mircosoft. 2011. Investor Relations. Available at:
http://www.microsoft.com/investor/SEC/default.aspx. Mun, Y., and Y. Hwang. 2003. Predicting the use of web-based information systems: Self-
efficacy, enjoyment, learning goal orientation, and the technology acceptance model. International Journal of Human-Computer Studies 59 (4): 431–449.
Pincione, G., and F. R. Teson. 2006. Rational Choice and Democratic Deliberation: A Theory of Discourse Failure. Cambridge University Press. Pinsker, R., and S. Li. 2008. Costs and benefits of XBRL adoption: Early evidence. Communications of the ACM, 51 (3): 161-166. _______, and P. R. Wheeler. 2009. Nonprofessional investors’ perceptions of the efficiency and
effectiveness of XBRL-enabled financial statement analysis and of firms providing XBRL-formatted information. International Journal of Disclosure and Governance 6 (3): 241-262.
32
_______, T. Pitre, and R. Daigle. 2009. An investigation of nonprofessional investors’ qualitative materiality judgments incorporating SEC listed vs. non-listed events. Journal of Accounting and Public Policy 28 (5): 446-465.
Plumlee, R. D., and M. A. Plumlee. 2008. Assurance on XBRL for financial reporting.
Accounting Horizons 22 (3): 353-368. Rose, J. 2002. Behavioral decision aid research: Decision aid use and effects. In Research
Accounting as an Information Systems Discipline, edited by V. Arnold and S. Sutton, 111-133. Sarasota FL: American Accounting Association.
SAS. 2005. Handbook for SAS users. Securities and Exchange Commission (SEC). 2009. Interactive data to improve financial
reporting. Available at: http://www.sec.gov/rules/final/2009/33-9002.pdf. Trites, G. 1999. The Impact of Technology on Financial and Business Reporting. Canadian Institute of Chartered Accountants, Toronto, Canada. Troshani, I., and S. Rao. 2007. Drivers and inhibitors to XBRL adoption: A qualitative approach
to build a theory in under-researched areas. International Journal of E-Business Research 3 (4): 98-111.
Tuttle, B., M. Coller, and F. G. Burton. 1997. An examination of market
efficiency: Information order effects in a laboratory market. Accounting, Organizations, and Society 22 (1): 89-103.
Vasarhelyi, M., D. Chan, and J. P. Krahel. 2011. XBRL: Consequences to financial reporting,
data analysis, decision support, and others. Working paper, Rutgers University. Vessey, I. 1991. Cognitive fit: a theory-based analysis of the graphs versus tables literature.
Decision Sciences 22 (2): 219-241. _______, and D. Galletta. 1991. Cognitive fit: an empirical study of information acquisition.
Information Systems Research 2 (1): 63-84. _______. 2006. The theory of cognitive fit: One aspect of a general theory of problem solving?
In Human-computer Interaction and Management Information Systems: Foundations, Advances in Management Information Systems Series, edited by P. Zhang and D. Galletta. Armonk, NY: M.E. Sharpe.
Wanous, J., A. Reichers, and M. Hudy. 1997. Overall job satisfaction: How good are single-item measures? Journal of Applied Psychology 82 (2): 247-252.
33
Wheeler, P. R., and D. R. Jones. 2003. The effects of exclusive user choice of decision aid features on decision making. Journal of Information Systems 17 (1): 63-83.
_______, and V. Arunachalam. 2008. The effects of decision aid design on the information
search strategies and confirmation bias of tax professionals. Behavioral Research in Accounting 20 (1): 131-145.
Whitecotton, S., and S. Butler. 1998. Influencing decision aid reliance through involvement in information choice. Behavioral Research in Accounting 10 (Supplement): 182-200. XBRL International Standards Board (XSB). 2010. Preserve. Promote. Particpate. Moving
XBRL forward. Strategic initiative issued October. Available at: http://www.xbrl.org/P3%20docs/XBRL2010Initiatives.pdf
34
TABLE 1 Participant Demographics
Demographics Frequencies Mean or % (Std. Dev.)
Age 30.06 (8.13) Years of business work experience 4.68 (7.61) Gender Male = 22 41.5% Female = 31 58.5% Bought or sold individual company’s No = 37 69.8% common stock or mutual fund Yes = 16 30.2% Bought or sold individual company’s common stock No = 46 86.8% or mutual fund as part of class exercise Yes = 7 13.2% Plan to invest in individual company’s common stock No = 18 34.0% in the next five years Yes = 35 66.0% Number of times evaluating a company’s performance 5.53 by analyzing its financial statements (16.94) Used an investment model to buy a stock/mutual fund No = 48 90.6% Yes = 5 9.4% Experience level:a Working with spreadsheets
7.21
(1.92) Working with XBRL-enabled technology
3.38
(2.90) Working with portable document files
6.26
(2.46) Difficulty of financial analysis taskb
5.08
(1.88) Rate the realism of this taskc
6.81
(1.68) a Participants were asked to rate their experience on an 11 point scale where 0 = novice and 10 = expert. b Participants were asked to rate the difficulty on an 11 point scale where 0 = extremely simple and 10 = extremely complex c Participants were asked to rate the realism of the task on an 11 point scale where 0 = unrealistic and 10 = very realistic
35
TABLE 2 Pearson Correlations for All Hypothesized Variables
