Interpretation of uncertainty expressions: a cross-national study

Description
This study investigates the effect of language-culture and linguistic translation on the interpretation of verbal
uncertainty expressions found in International Accounting Standards. Data are collected from US Certified Public
Accountants and German-speaking Wirtschaftspru¨ fer (chartered or certified accountants) to test three hypotheses. One
group of German speakers evaluated uncertainty terms expressed in German and another group in English. The results
indicate significant differences in interpretation across the three groups. Some differences are attributed to a languageculture
effect and others to a translation effect, with the language-culture effect being more pervasive.

Interpretation of uncertainty expressions:
a cross-national study
Timothy S. Doupnik
a,
*, Martin Richter
b
a
The Darla Moore School of Business, University of South Carolina, Columbia, SC 29208, USA
b
Wirtschafts- und Sozialwissenschaftliche Fakulta¨t, Universita¨t Potsdam, 14439 Potsdam, Germany
Abstract
This study investigates the e?ect of language-culture and linguistic translation on the interpretation of verbal
uncertainty expressions found in International Accounting Standards. Data are collected from US Certi?ed Public
Accountants and German-speaking Wirtschaftspru¨ fer (chartered or certi?ed accountants) to test three hypotheses. One
group of German speakers evaluated uncertainty terms expressed in German and another group in English. The results
indicate signi?cant di?erences in interpretation across the three groups. Some di?erences are attributed to a language-
culture e?ect and others to a translation e?ect, with the language-culture e?ect being more pervasive. These results
raise the question of whether International Accounting Standards can be applied consistently across language-cultures.
#2002 Elsevier Science Ltd. All rights reserved.
Uncertainty and judgment are inherent in both
the ?nancial reporting and auditing domains.
Auditing standards use verbal probability (uncer-
tainty) expressions to establish thresholds for
when su?cient evidence has been gathered, and
auditors frequently use and interpret such expres-
sions (Amer, Hackenbrack, & Nelson, 1994).
Financial reporting standards also use uncertainty
expressions in establishing criteria for the recogni-
tion, measurement, or disclosure of items, and
both accountants and auditors are required to
attach meaning to those expressions. The expres-
sions ‘‘remote,’’ ‘‘reasonably possible,’’ and
‘‘probable’’ in the FASB’s Statement of Financial
Accounting Standards No. 5 are examples.
Davidson and Chrisman (1994) identi?ed some 33
di?erent uncertainty expressions used in Cana-
dian accounting and auditing standards. Even the
basic elements of ?nancial statements may be
de?ned in terms of probability as exempli?ed by
the FASB’s (1985) SFAC 6 de?nition of an asset
as ‘‘probable future economic bene?ts obtained or
controlled by a particular entity. . .’’ (par. 25,
emphasis added).
International Accounting Standards (IAS),
developed for worldwide usage to enhance the
comparability of ?nancial reports across coun-
tries, also include a number of uncertainty expres-
sions.
1
Necessary conditions for cross-national
comparability include the use of a single set of
0361-3682/02/$ - see front matter # 2002 Elsevier Science Ltd. All rights reserved.
PI I : S0361- 3682( 02) 00010- 7
Accounting, Organizations and Society 28 (2003) 15–35
www.elsevier.com/locate/aos
1
International Accounting Standards are developed by the
International Accounting Standards Committee (IASC) based
in London. The IASC was restructured as the International
Accounting Standards Board (IASB) in 2001.
* Corresponding author. Tel.: +1-803-777-7450; fax: +1-
803-777-0712.
E-mail address: [email protected](T.S. Doupnik).
standards and consistent interpretation of those
standards across countries. To the extent that
uncertainty expressions used in IAS are inter-
preted di?erently by accountants and auditors in
di?erent countries, the cross-national compar-
ability of ?nancial reports will su?er. This study
addresses the research question of whether inter-
pretations of IAS uncertainty expressions by pro-
fessional accountants in the United States and in
German-speaking countries have similar meaning.
To our knowledge, this is the ?rst study of uncer-
tainty expressions to use professional accountants
in more than one country.
Hypotheses based on linguistic relativism and
translation e?ects are tested using data gathered
through a mail survey. The answer to the research
question is important to the viability of using a
single set of accounting standards worldwide. The
selection of the United States and Germany as
countries to study is relevant because these two
countries are important for the process of world-
wide harmonization. The study is also timely from
a US perspective as the Securities and Exchange
Commission (SEC) debates whether foreign regis-
trants should be allowed to use IAS in preparing
?nancial statements without reconciliation to US
GAAP.
The remainder of this paper is organized into six
sections. The ?rst section provides a review of
related studies in the psychology and accounting
literatures. This section also includes a discussion
of culture and linguistic relativism from which
preliminary hypotheses are derived. The second
section describes the uncertainty expressions
examined in the study. The third section presents
the speci?c research questions and hypotheses tes-
ted. The fourth section describes the research
methodology, and the ?fth section presents the
analysis and results. The ?nal section provides a
summary and o?ers conclusions.
1. Related literature
1.1. Psychology research
A considerable amount of research has been
conducted and reported in the psychology litera-
ture investigating the quantitative meanings of
verbal probability expressions (e.g. Beyth-Marom,
1982; Budescu & Wallsten, 1985; Lichtenstein &
Newman, 1967; Reagan, Mosteller, & Youtz,
1989; Simpson, 1963). More than 280 di?erent
probability expressions have been examined with
only a small subset examined in any one study.
The predominant method of assessing the mean-
ing of a probability expression has been to ask
subjects to provide a percentage from 0 to 100 that
corresponds to it (Reagan et al., 1989). Results
consistently show a large degree of between-sub-
ject variability (Budescu & Wallsten, 1985). This
variability has been found to be inversely related
to the distance from the center of the scale (Wall-
sten, Budescu, Rapoport, Zwick, & Forsyth,
1986). The variability has been found to be lower
in groups from homogeneous backgrounds (Brun
& Teigen, 1988), and expressions embedded in
context exhibit more variability than those in iso-
lation (Beyth-Marom, 1982).
Results reported in the psychology literature
also show a lack of symmetry (Budescu & Wall-
sten, 1985). That is, probabilities assigned to mirror-
image pairs such as ‘‘probable’’ and ‘‘improbable’’
do not sum to 100% (Lichtenstein & Newman,
1967). Negative probability expressions tend to be
further away from the 50% midpoint than the
related positive expressions (Reagan et al., 1989).
The variability in responses to uncertainty
expressions can be due to (a) the vague meaning of
the expressions or (b) the fact that the meaning of
probability terms is not constant across indivi-
duals. The results of Wallsten et al. (1986) support
the notion that non-numerical probability expres-
sions convey vague uncertainties; the vagueness
can be attributed to the use of a verbal expression
and not to the perceived uncertainty. They suggest
that people not only understand uncertainty
expressions as representing amounts of probability
but also as representing degrees of con?dence in
that probability.
Phillips and Wright (1977) introduced the
notion that culture can in?uence the cognitive
processes involved in probability assessment. They
hypothesized that English people (who have a
‘‘probabilistic’’ world-view) would make ?ner dis-
criminations in degrees of uncertainty than would
16 T.S. Doupnik, M. Richter / Accounting, Organizations and Society 28 (2003) 15–35
Chinese people (who have a ‘‘fatalistic’’ world-
view) and that numerical assessments of prob-
abilities would be more meaningful for the English
than for the Chinese. Their experimental results
generally support their hypotheses.
1.2. Accounting research
Much of the research on uncertainty expressions
in the accounting domain relates to the inter-
pretation of the uncertainty expressions used in
FASB Statement of Financial Accounting Stan-
dards (SFAS) No. 5 for determining when a con-
tingency should be recognized or disclosed. SFAS
5 requires no disclosure when the contingent loss
is ‘‘remote,’’ footnote disclosure when it is ‘‘rea-
sonably possible,’’ and ?nancial statement recog-
nition when the contingent loss is ‘‘probable.’’
Schultz and Reckers (1981) found that auditors’
interpretations of SFAS 5 expressions were a?ec-
ted by the materiality of the potential loss, and
that the variability of responses was reduced after
individuals were involved in group processing of
the disclosure issue. Jiambalvo and Wilner’s
(1985) results show considerable between subject
variability in assignment of probability ranges to
the words ‘‘remote,’’ ‘‘reasonably possible,’’ and
‘‘probable.’’ Subsequent analysis implied that the
variability was due to di?erent interpretations of
the words rather than to an inability to express
feelings in terms of probabilities. In contrast to
Schultz and Reckers, they did not ?nd materiality
of loss to a?ect decisions regarding disclosure.
Harrison and Tomassini (1989) examined audi-
tors’ interpretations of probability thresholds for
‘‘remote,’’ ‘‘reasonably possible,’’ and ‘‘probable’’
across di?erent types of contingencies. They found
little di?erence in thresholds across the various
contingencies although there was less consensus
about the threshold between ‘‘remote’’ and ‘‘rea-
sonably possible.’’
Chesley (1986) conducted two experiments with
accounting students to address several questions
related to the interpretation of uncertainty
expressions. Consistent with results in the psy-
chology literature, he found a low degree of group
consistency for most expressions and a lack of
symmetry in complementary words.
Reimers (1992) conducted an inter-group study
to determine whether samples of auditors, engi-
neering managers, marketing managers, and
graduate students interpret 30 uncertainty expres-
sions, including those of SFAS 5, in the same way.
She found that many of the expressions were
interpreted as synonyms. She also found a large
gap in the 0–100% probability range between the
upper bound on ‘‘remote’’ (15%) and the lower
bound on ‘‘reasonably possible’’ (48%). Using
both practicing accountants and accounting stu-
dents, Davidson (1989) found similar results and
he concluded that the SFAS 5 set of probability
expressions is not optimal. He suggested that
‘‘reasonably possible,’’ which is perceived as quite
similar to ‘‘probable,’’ should be replaced with a
term such as ‘‘sometimes’’ that might better con-
vey a level of probability that is closer to the mid-
point between ‘‘remote’’ and ‘‘probable.’’
Amer et al. (1994) asked auditing managers to
provide numerical interpretations of 23 uncer-
tainty phrases placed in an auditing context. They
also found that multiple phrases have similar
numerical interpretations. Similar mean results
were obtained for six phrases common with
Reimers, even though her study was not in con-
text. Amer et al. (1994) also found that the inter-
subject variability inherent in assigning probability
to uncertainty expressions decreased when moving
from phrases that communicate low probabilities
to those that communicate high probabilities.
To test whether the mental representation of
uncertainty phrases is a?ected by the language in
which they are expressed, Davidson and Chrisman
(1993) examined the interpretation of uncertainty
expressions found in International Accounting
Standards between Anglophone and Francophone
accounting students in Canada. They suggested
that di?erences in mental representations can exist
in two ways. Either the mean probability asso-
ciated with an uncertainty expression is not the
same in each language, indicating a di?erence in
the concept conveyed by the expression, or the
degree of consensus on the probability associated
with an uncertainty expression is not the same in
each language, indicating a di?erence in the preci-
sion of the expression. Davidson and Chrisman
(1993) compared mean probabilities assigned to
T.S. Doupnik, M. Richter / Accounting, Organizations and Society 28 (2003) 15–35 17
the English original and French translation and
found di?erences in 13 of 27 expressions exam-
ined. In addition, they found signi?cant di?er-
ences in the variance of the probabilities for 14 of
the 27 expressions, with the English expressions
generally having lower variance. The authors infer
from this result that the English expressions con-
vey a more precise meaning than the French
equivalent. In a related study, Davidson and
Chrisman (1994) found similar results for uncer-
tainty expressions utilized in Canadian accounting
and auditing standards.
Davidson and Chrisman’s (1993, 1994) results
show that uncertainty phrases expressed in di?er-
ent languages can be interpreted di?erently by two
di?erent linguistic groups located in the same
country. Their research design does not allow
them to determine whether the di?erence in inter-
pretation is a result of the process of translating
English expressions into French or exists because
English-speakers and French-speakers have a fun-
damentally di?erent understanding of the under-
lying concepts expressed through uncertainty
expressions. The current study attempts to disen-
tangle these two factors. The following section
considers literature in the areas of culture and
psycholinguistics to develop arguments as to why
and howlanguage might a?ect these interpretations.
1.3. Culture and linguistic relativism
Culture is thought to be an important environ-
mental factor in?uencing a country’s accounting
system (Gray, 1988; Harrison & McKinnon, 1986;
Mueller, 1967; Violet, 1983) and empirical studies
show this to be generally true (Doupnik & Salter,
1995; Frank, 1979; Salter & Niswander, 1996;
Zarzeski, 1996). Culture has also been shown to
a?ect the design and/or e?ectiveness of manage-
ment control systems (Chow, Kato, & Merchant,
1996; Chow, Shields, & Wu, 1999; Harrison,
1993). (See Harrison & McKinnon, 1999, for a
review of the literature on management control
systems and culture.) The issue at hand, however,
is whether culture a?ects the interpretation by
accountants of accounting standards, in general,
and uncertainty expressions within those stan-
dards, in particular.
Riahi-Belkaoui and Picur (1991) developed a
theoretical justi?cation for the e?ect culture might
have on the perception of accounting concepts.
Using a cognitive orientation to culture, they sug-
gest that national cultures act as networks of sub-
jective meanings or frames of reference shared by
members of the culture. As such, national culture
could in?uence the way members of that culture
perceive basic accounting principles such as
‘‘going concern’’ and ‘‘matching.’’ Perceptions of
accounting principles varied among samples of
US, Canadian, and British auditors in their
empirical tests. Di?erences were found even
though the groups share a similar language.
Bagrano?, Houghton, and Hronsky (1994) sug-
gested that cross-cultural di?erences may a?ect
the meaning associated with, and hence judgment
in applying, accounting standards. They found
di?erences in cognitive structures related to the
concept ‘‘extraordinary items’’ between US and
Australian auditors.
These studies show that national culture can
a?ect perceptions of accounting speci?c principles
and concepts. This is not surprising given that a
country’s authoritative literature can de?ne a spe-
ci?c accounting concept in such a way that it dif-
fers from a similar (but not identical) concept in
another country. Accountants ‘‘learn’’ accounting
concepts within the context of the accounting fra-
mework and traditions speci?c to an individual
country. The results of these studies do not neces-
sarily imply, however, that national culture also
will a?ect interpretations of more general words
and phrases commonly found in uncertainty
expressions such as ‘‘probable’’ and ‘‘remote.’’
The concept of ‘‘probable,’’ for example, is not
learned by an accountant the same way as the
concept of an ‘‘extraordinary item.’’ The link
between national culture and the perception of
general concepts can be made by considering one
very important aspect of culture, namely language.
Language and culture are interrelated. Sapir
(1964) argues that the perfection of language is a
prerequisite for the development of a culture as a
whole, and that language is the verbal expression
of a culture. Belkaoui (1989) suggests that lan-
guage is indicative of the ‘‘metaphysics’’ of a cul-
ture which consist of ‘‘unstated premises which
18 T.S. Doupnik, M. Richter / Accounting, Organizations and Society 28 (2003) 15–35
shape the perception and thought of those who
participate in that culture and predispose them to
a given mode of perception’’ (p. 283). Indeed, in
previous accounting research, language has been
used as a surrogate for culture (Frank, 1979; Nair
& Frank, 1980).
Linguistic relativism relates to the role language
plays in our understanding of the world. The
grammatical forms and categories provided by a
language are thought to a?ect the manner in
which speakers of a given language interpret the
world (Sapir, 1964; Whorf, 1956). In other words,
a given language inclines its users to distinct
beliefs (Belkaoui, 1989).
Based on the concept of linguistic relativism,
Monti-Belkaoui and Belkaoui (1983) tested the
hypothesis that di?erent languages result in dif-
ferent meaning being attached to basic accounting
principles. Using Anglophone and Francophone
students in Canada as subjects, their results
support the notion that speakers of di?erent lan-
guages perceive basic accounting concepts di?er-
ently, even though both groups were members of
the same national culture.
Related to the issue of concern in the current
study, linguistic relativism suggests that a speci?c
language could predispose its speakers to distinct
interpretations of uncertainty expressions and that
di?erent languages could lead to di?erent inter-
pretations of uncertainty expressions. This leads
to the notion that speakers of di?erent languages
could di?er in the meaning attached to uncertainty
expressions.
National culture is a function of the shared
experience of the individuals comprising that
nation. Factors shaping a national culture include
geography, history, climate, religion, and lan-
guage. Although each of the English-speaking
countries has a unique national culture, they share
a similar language-culture. The same is true for
German-speaking countries, Spanish- speaking
countries, and so on. The discussion thus far leads
to the general hypothesis that di?erences in inter-
pretation of uncertainty expressions can exist
across countries that are members of di?erent
language-cultures. The corollary hypothesis is that
interpretations of uncertainty expressions should
not di?er across countries that are members of the
same language-culture. In other words, nationality
alone should not a?ect interpretations. This study
concentrates on the ?rst hypothesis but also makes
an attempt to examine the corollary.
The countries selected to address the ?rst
hypothesis in this study are the United States and
Germany. These two countries have di?erent lan-
guages (English and German) and have been clas-
si?ed as being members of two distinctly di?erent
cultural areas (Anglo and Germanic) (Hofstede,
1980). To address the second issue, two additional
members of the German language-culture, Austria
and German-speaking Switzerland are included in
the study. The primary question addressed is
whether the di?erent language-cultures in which
US Certi?ed Public Accountants (CPAs) and
German-speaking Wirtschaftspru¨ fer (WPs) live
and work a?ects their perception and interpreta-
tion of uncertainty expressions found in IAS. The
IAS uncertainty expressions examined in the cur-
rent study are described in the next section.
2. Identi?cation of uncertainty expressions
The o?cial language of the IASC is English and
IAS are published in that language. In 1997, the
IASC produced an o?cial German translation of
the 33 extant IAS (Scha¨ ?er-Poeschel, 1998). This
was the ?rst and, at the time, only o?cial transla-
tion of IAS into another language. An examina-
tion of these standards resulted in the set of 16
uncertainty expressions used in this study (see
Table 1). Table 1 indicates that there is some dif-
?culty in translating certain English expressions
into German. For example, the single word
‘‘remote’’ is translated into the three-word phrase
‘‘Wahrscheinlichkeit a¨ ußerst gering’’ (literal
translation=‘‘probability extremely small’’). This
apparent lack of direct equivalence of expressions
in the two languages suggests that there may be a
lack of equivalence in the underlying concepts.
There is also some lack of consistency in the
translation to German with multiple expressions
used to translate four of the English expressions.
For example, the expression ‘‘likely’’ is translated
as both ‘‘voraussichtlich’’ and ‘‘wahrscheinlich.’’
This inconsistency could be attributed to two pos-
T.S. Doupnik, M. Richter / Accounting, Organizations and Society 28 (2003) 15–35 19
sible causes. One, the translators felt that these
two German words convey similar meaning and
may be used interchangeably. Two, one German
expression was deemed to be more appropriate
than the other within the context in which it was
being used. It is possible that ‘‘voraussichtlich’’
and ‘‘wahrscheinlich’’ convey subtle di?erences in
probability or degree of consensus that are lost by
using the single term ‘‘likely.’’
Two di?erent English expressions are translated
into the same German word in two situations.
‘‘Likely’’ and ‘‘expected’’ are both translated as
‘‘voraussichtlich,’’ and ‘‘likely’’ and ‘‘probable’’
are both translated as ‘‘wahrscheinlich.’’ The
interchangeability of these expressions suggests
that the German translators believe that the Eng-
lish terms ‘‘likely,’’ ‘‘expected,’’ and ‘‘probable’’
have similar meaning, which indeed may be true.
The translation problems noted above suggest a
possible source other than language-culture that
could lead to di?erences in perceptions of IAS
uncertainty expression between US and German
accountants, namely that translation into another
language distorts the underlying meaning that the
IASC wished to convey in the original English.
For example, German WPs’ perception of
‘‘Wahrscheinlichkeit a¨ ußerst gering’’ could di?er
from US CPAs’ perceptions of ‘‘remote’’ because
the two expressions simply are not equivalent.
Therefore, there are two competing explanations
for why interpretation of uncertainty expressions
might di?er between accountants from di?erent
linguistic groups: language-culture and translation.
3. Research questions and hypotheses
The primary question addressed in this study is
whether di?erences exist between US and German
accountants in the interpretation of uncertainty
expressions found in IAS. If so, this could have
negative consequences for the comparability of
Table 1
Source of uncertainty expressions in International Accounting Standards (IAS)
English German IAS (#–Paragraph)
virtually certain so gut wie sicher 10–17
reasonable assurance angemessene Sicherheit 20–7
assurance Gewibheit 16–10
expected 1. erwartet 9–19; 11–22, 36; 16–7
2. voraussichtlich 4–4, 7
a
su?cient certainty hinreichende Sicherheit 16–10
reasonably likely nach vernu¨ nftigen Annahmen wahrscheinlich 22–48
probable 1. wahrscheinlich 9–17; 10–8, 16;
11–11, 23, 24, 32, 36;
12–24, 34; 16–8,24;
22–27, 52, 55, 58
2. hinreichend wahrscheinlich 18–14, 20, 29, 34
likely 1. voraussichtlich 4–11
a
2. wahrscheinlich 11–34
with the prospect mit der Aussicht 9–6
insu?cient certainty unzureichende Sicherheit 9–18
not probable nicht wahrscheinlich 11–34
no longer probable 1. nicht mehr wahrscheinlich 12–56
2. nicht mehr erwartet 22–47
3. voraussichtlich nicht mehr 9–25
unlikely aller Wahrscheinlichkeit nach nicht 12–36
not expected nicht erwartet 16–61
seriously in question sehr zweifelhaft 11–34
remote Wahrscheinlichkeit a¨ ußerst gering 10–9
a
Note that ‘‘voraussichtlich’’ is used consistently in IAS 4; ‘‘expected’’ and ‘‘likely’’ are used interchangeably in IAS 4.
20 T.S. Doupnik, M. Richter / Accounting, Organizations and Society 28 (2003) 15–35
?nancial statements between these two countries
even though those statements are prepared using a
single set of standards. The ?rst hypothesis exam-
ined is:
H1 Interpretations of uncertainty expressions
di?er between US and German accountants
because of di?erences in language-culture and/
or because of linguistic translation.
If di?erences in the interpretation of uncertainty
expressions exist, the second research question is
what is the cause: Are di?erences due to the dif-
ferent language-cultures of the two groups or are
di?erences due to the e?ects of translation or are
di?erences due to both? Knowing whether lan-
guage-culture or translation is the cause of di?er-
ing interpretations could be important because
one cause might be easier to overcome than the
other in improving the comparability of ?nancial
statements. The second and third hypotheses
examined in this study are:
H2 Interpretations of uncertainty expressions
di?er between US and German accountants
because of di?erences in language-culture.
H3 Interpretations of uncertainty expressions
di?er between US and German accountants
because of linguistic translation.
The corollary question addressed in this study is
whether di?erences exist between German-speak-
ing WPs in di?erent countries (Germany, Austria,
and Switzerland) in the interpretation of uncer-
tainty expressions found in IAS. In other words:
Does nationality alone a?ect interpretation?
Because German-speakers in these three countries
share a common language-culture, we do not
expect to be able to reject the null hypothesis of no
di?erence in interpretations across di?erent
national groups of German-speaking WPs.
4. Methodology
4.1. Research instrument
To examine the research questions, a mailed
survey questionnaire was employed to obtain sub-
jects’ interpretations of the IAS uncertainty
expressions listed in Table 1. The questionnaire
consisted of three parts and four versions of the
questionnaire were developed: an all-English ver-
sion (E), an all-German version (G), and two
mixed-language versions (GE1 and GE2). Exhibit
1 presents a summary of the four di?erent versions
of the questionnaire. The all-English version of the
questionnaire is provided in the Appendix.
Part 1 of the all-English questionnaire (E) con-
tained the 16 expressions listed in Table 1. Sub-
jects were asked to assign point estimate
probabilities to the uncertainty expressions using a
scale of 0–100. A non-accounting example was
provided in the instructions to enhance under-
standing of the task. To mitigate an order e?ect,
two versions of the questionnaire were created in
which the uncertainty expressions were placed in
di?erent random arrangements. Part 2 of the
questionnaire asked respondents to indicate the
range of probabilities associated with 6 of the 16
expressions in Part 1. Pitz (1980) suggests that
uncertainty expressions best describe a range of
numerical probabilities rather than single values.
Demographic information was collected in Part 3
of the questionnaire.
Other than language, the all-German version
(G) of the questionnaire was identical to the Eng-
lish version with the following exceptions. Due to
there being two translations for ‘‘probable’’ and
three for ‘‘no longer probable,’’ Part 1 of ques-
tionnaire G included 19 expressions corresponding
to the 16 expressions in English. Additional
demographic questions in Part 3 asked about
English ?uency and professional experience work-
ing in an English-speaking country. These ques-
tions were asked to examine whether exposure to
the English language and/or the Anglo culture
might in?uence perceptions of uncertainty expres-
sions.
Expressions in Parts 1 and 2 of the all-German
questionnaire were presented in the same random
order as in the English version. To ensure equiva-
lence between the English and German versions of
the questionnaire, the German version was pre-
pared ?rst and then translated into English by the
researcher with English as his ?rst language. Ger-
man colleagues then checked the English version
T.S. Doupnik, M. Richter / Accounting, Organizations and Society 28 (2003) 15–35 21
for consistency with the German version. One of
the advantages of including the German language-
culture in this study is that an o?cial translation
of IAS exists and any subjectivity introduced by a
translation of the uncertainty expressions by the
researchers could be avoided.
The third version of the questionnaire (GE1)
was constructed to elicit responses from German
speakers to uncertainty expressions stated in Eng-
lish. GE1 consisted of a combination of instruc-
tions and demographic questions in German, and
uncertainty expressions in English. Part 1 of GE1
asked respondents to provide point estimate
probabilities for the 16 English-language uncer-
tainty expressions in Part 1 of questionnaire E.
Part 2 of GE1 requested a range of probabilities
for the six English-language expressions in Part 2
of questionnaire E. Part 3 was the same as in
questionnaire G.
In the fourth version of the questionnaire
(GE2), Part 1 consisted of the 19 German-lan-
guage uncertainty expressions included in Part 1
of version G, the all-German questionnaire. Part 2
of GE2 contained the 16 uncertainty expressions
from Part 1 of version E, the all-English ques-
tionnaire. Thus, respondents to GE2 assigned
probabilities to the same uncertainty terms
expressed in English and translated into German.
4.2. Subjects
Questionnaire E was mailed to CPAs in the
United States. The mailing list was obtained from
the American Institute of Certi?ed Public Accoun-
tants (AICPA) and the sample was randomly
drawn from that subset of AICPA members who
indicate that they work in public accounting and
have auditing as their professional interest. Ques-
tionnaires G, GE1, and GE2 were mailed to WPs
throughout Germany randomly drawn from the
WP Directory published by the Institut der Wirt-
schaftspru¨ fer. In addition, to examine whether
di?erences exist between di?erent nationalities
that speak the same language, questionnaire G
was also sent to a sample of WPs in Austria and
Switzerland. A follow-up mailing was conducted
in an attempt to increase response rates.
Sample sizes and response rates are reported in
Table 2. For subsequent analysis, respondents
with internally inconsistent responses were removed
from the data set. Inconsistent responses were
identi?ed by comparing responses to those uncer-
tainty expressions that are direct opposites of each
other, such as, ‘‘likely/unlikely’’ and ‘‘wahrschein-
lich/nicht wahrscheinlich.’’ Those respondents
assigning a probability to the second expression in
the pair greater than the probability assigned to
the ?rst expression were removed from the data
set. These respondents apparently did not cor-
rectly understand the task.
Though not large in an absolute sense, the
usable response rates to questionnaires E and G
are consistent with previous mail surveys in the
United States and Germany and resulted in a
sample large enough to conduct tests of statistical
signi?cance.
2
The German WPs’ response rate to
Table 2
Sample sizes and response rates for six groups
Nationality/questionnaire US German Swiss Austrian German German
E G G G GE1 GE2
Sample size 500 206 150 150 217 211
Number of responses 157 84 41 10 37 60
Response rate 31.4% 40.8% 27.3% 6.7% 17.1% 28.4%
Did not respond to Part 1 0 0 0 1 2 19
a
Inconsistent respondents 29 14 10 1 6 8
Usable responses 128 70 31 8 29 33
Usable response rate 25.6% 34.0% 20.7% 5.3% 14.3% 15.6%
a
Number who did not evaluate the English expressions contained in Part 2 of questionnaire GE2.
2
For example, Agacer and Doupnik (1991) obtained
response rates of 26.5 and 23% from CPAs in the United States
and WPs in Germany, respectively.
22 T.S. Doupnik, M. Richter / Accounting, Organizations and Society 28 (2003) 15–35
questionnaire GE1 and to Part 2 of GE2 were well
below the German WPs’ response rate to ques-
tionnaire G. A probable self-selection bias exists
for these two questionnaires as only those German
WPs con?dent in their English ?uency were likely
to respond. Such a bias is welcome, however, as
responses to English-language expressions pro-
vided by WPs with inadequate English compre-
hension would be meaningless. Because of the
relatively low response rates, non-response bias
cannot be ruled out.
3
Generalizing the results to
the populations of accountants from which the
random samples were drawn should be done with
caution.
Respondent demographics are reported in
Table 3. Each of the six groups of respondents, on
average, has more than 10 years of professional
experience as a CPA (WP) indicating a relatively
high level of experience. The smaller number of
years of experience for the various WP groups as
Table 3
Pro?les of respondents by group
Nationality/questionnaire US German Swiss Austrian German German
E G G G GE1 GE2
Experience in years (mean) 16.8 13.9 14.0 11.3 11.1 10.9
Primary specialty
Audit 54.8% 54.3% 72.4% 25.0% 50.0% 59.4%
Tax 34.9% 35.7% 0% 75.0% 39.3% 34.4%
Consulting 4.8% 8.6% 13.8% 0% 3.6% 6.3%
Other 5.6% 1.4% 13.8% 0% 7.1% 0%
Number of CPAs (WPs) in ?rm
1–5 41.3% 61.4% 31.0% 42.9% 37.0% 48.5%
6–20 24.6% 15.7% 13.8% 28.6% 14.8% 6.1%
21–100 11.1% 1.4% 6.9% 28.6% 3.7% 0%
over 100 23.0% 21.4% 48.3% 0% 44.4% 45.5%
Familiarity with IAS
Very familiar 0.8% 10.0% 13.3% 12.5% 3.6% 0%
Familiar 6.3% 38.6% 36.7% 50.0% 53.6% 46.9%
Somewhat familiar 23.8% 44.3% 46.7% 25.0% 39.3% 50.0%
Not familiar 69.1% 7.1% 3.3% 12.5% 3.6% 3.1%
Refer to IAS in practice
Often 0.8% 22.9% 30.0% 37.5% 21.4% 18.2%
Seldom 31.7% 61.4% 60.0% 50.0% 67.9% 60.6%
Never 67.5% 15.7% 10.0% 12.5% 10.7% 21.2%
English comprehension n/a
Excellent 15.7% 16.7% 25.0% 14.3% 12.1%
Good 31.4% 43.3% 50.0% 46.4% 39.4%
Satisfactory 44.3% 33.3% 25.0% 35.7% 39.4%
Rudimentary 8.6% 6.7% 0% 3.6% 9.1%
Audit experience in English-speaking country n/a
Yes 18.8% 20.0% 0% 21.4% 21.2%
No 81.2% 80.0% 100.0% 78.6% 78.8%
Mean years 5.9 2.0 n/a 8.5 5.8
3
Comparisons of the responses provided by early respon-
dents and late respondents for the three groups (CPA/English,
WP/English, WP/German) that serve as the basis for analysis in
Table 4 found only two signi?cant di?erences in the inter-
pretation of uncertainty expressions, both in the WP/German
group. Although inconclusive, the general lack of di?erence
between early and late respondents suggests that a systematic
non-response bias is not present.
T.S. Doupnik, M. Richter / Accounting, Organizations and Society 28 (2003) 15–35 23
compared to the US CPA group is consistent with
the greater number of years of experience required
to attain the WP designation in German-speaking
countries as compared with the experience require-
ment in the USA. The most interesting result from
the demographic questions is the extent to which
the di?erent groups are familiar with and refer to
IAS in their work. The majority of US CPAs
indicated that they are not familiar and that they
never refer to IAS, whereas only relatively small
percentages of the various German-speaking WP
groups indicated the same. Although this di?er-
ence has no direct bearing on the current research,
it is somewhat surprising to discover that US
CPAs have so little contact with IAS. There is no
systematic di?erence in English comprehension
level between the German WPs who evaluated
uncertainty expressions in German (questionnaire
G) and the German WPs who evaluated uncer-
tainty expressions in English (questionnaires GE1
and GE2).
4.3. Research design
To examine whether nationality alone, control-
ling for language-culture, a?ects interpretation of
uncertainty expressions, the responses to ques-
tionnaire G made by German WPs were compared
with responses to questionnaire G made by Aus-
trian and Swiss WPs. To test whether interpreta-
tions of uncertainty expressions di?er between US
CPAs and German-speaking WPs (Hypothesis 1),
the mean responses to Part 1 of questionnaire E
(CPA/English language) were compared with the
mean responses to Part 1 of questionnaire G (WP/
German language). To test the e?ect of language-
culture (Hypothesis 2), controlling for any trans-
lation e?ect, responses to Part 1 of questionnaire
E (CPA/English language) were compared with
the combined responses to Part 1 of questionnaire
GE1 and Part 2 (the English-language part) of
GE2 (WP/English language).
To test the e?ect of translation (Hypothesis 3),
controlling for language-culture, the responses to
Part 1 of questionnaire G (WP/German language)
were compared to the responses to Part 1 of GE1
(WP/English language)—a between subjects com-
parison. Hypothesis 3 was also tested by compar-
ing the responses to GE2-Part 1 (WP/German)
with the responses to GE2-Part 2 (WP/English)—
a within person comparison. A comparison of
WPs’ perceptions of those expressions where two
di?erent German words have been used for one
English word provides additional insight into
the e?ect translation has on the interpretation of
IAS.
Responses to Part 2 of questionnaires E, G, and
GE1 were also compared across the three groups
(CPA/English, WP/German, and WP/English) to
test for overall, culture, and translation e?ects
with regard to the range of probabilities assigned
to uncertainty expressions.
5. Analysis and results
5.1. Point estimate probabilities
The ?rst step in the analysis was to test for dif-
ferences in the point-estimate interpretations (Part
1 of the questionnaire) of uncertainty expressions
made by the German respondents and those made
by the Swiss and Austrian respondents. One-way
ANOVA found only one expression (‘‘mit der
Aussicht’’) for which a signi?cant di?erence exists.
Post hoc comparisons revealed a signi?cant dif-
ference between the German and Austrian groups
of WPs on this expression. Because of the small
number of Austrian respondents (n=8), treating
the Austrians as a separate group in ANOVA is
tenuous. Therefore, the Austrian and Swiss
responses were combined and compared with
those of the German group. There was only one
signi?cantly di?erent mean response between
these two groups (‘‘erwartet’’). When the Austrian
and German groups were combined and compared
with the Swiss group, no signi?cant di?erences
were found. The general lack of signi?cant di?er-
ences among these three groups of German-
speaking WPs allows us to conclude that nation-
ality alone does not cause di?erences in inter-
pretation of uncertainty expressions. Indeed, there
appears to be a common interpretation among
German-speaking WPs regardless of nationality.
Because of this, the responses of these three
groups are combined for subsequent hypothesis
24 T.S. Doupnik, M. Richter / Accounting, Organizations and Society 28 (2003) 15–35
testing. This combined group is referred to as WP/
German (n=109).
Part 1 of questionnaire GE1 and part 2 of
questionnaire GE2 were the same. In both cases,
German-speaking WPs were asked to assign point
estimate probabilities to English expressions. T-
tests of the means of these two groups found no
signi?cant di?erences across the 16 expressions.
The responses to these two questionnaires are
combined into a group referred to as WP/English
(n=62) for subsequent testing.
Table 4 reports the mean probability assigned to
each English uncertainty expression by the CPA
and WP/English groups (Columns 3 and 4) and to
the German translations by the WP/German
group (Column 5). Several expressions included in
this study have been examined repeatedly in pre-
vious research involving US subjects. The mean
CPA responses to ‘‘probable’’ (71.37%) and
‘‘likely’’ (70.89) are similar to the responses to
these terms in prior studies.
4
In addition, as
appears to be the case in the current study, several
studies have found these two terms to be syno-
nyms. The mean probability associated with
‘‘remote’’ (16.38%) in the current study is higher
than has been found in previous accounting stud-
ies. Reimers (1992) had a mean of 9.4% for her
auditor group and Amer et al. (1994) had a mean
of 12.33% for their in-context study.
The CPA responses also exhibit more symmetry
with regard to mirror-image pairs than has been
found in previous research reported in the psy-
chology literature. The mean probabilities
assigned to ‘‘probable’’ and ‘‘not probable’’ sum
to 104%, and ‘‘likely’’ and ‘‘unlikely’’ sum to
98%, as compared with sums of approximately
85% in other studies. The negative expressions in
the pairs are assigned probabilities much closer to
the 50% midpoint than has been found in pre-
vious studies.
As has been found in prior research, there is
considerable between-subject variability in the
CPA responses. Standard deviations range from
9.70 (‘‘reasonable assurance’’) to 23.54 (‘‘no
longer probable’’). Standard deviations are much
smaller for the positive expressions (those with a
mean probability > 50%; average 13.79) than for
the negative expressions (average 20.91). This is
consistent with results found by Amer et al. (1994).
To determine whether familiarity with IAS
in?uences interpretation of uncertainty expres-
sions used in IAS, respondents in each of the WP/
German and WP/English groups were split into
two groups (high=very familiar or familiar, and
low=somewhat familiar or not familiar).
5
No
signi?cant di?erences were found between the high
and low groups for the WP/German respondents.
However, the high IAS familiarity group for the
WP/English respondents assigned signi?cantly
higher mean probabilities than the low familiarity
group on three uncertainty expressions: ‘‘not
probable,’’ ‘‘no longer probable,’’ and ‘‘remote.’’
The WP/English respondents were also split into
two groups based on their level of English compre-
hension (high=excellent or good, low=satisfactory
or rudimentary). No signi?cant di?erences were
found between these two groups in mean prob-
abilities assigned to the 16 English expressions to
which they responded. Level of English compre-
hension did not a?ect their responses to the uncer-
tainty expressions in English. Similarly, there were
no signi?cant di?erences in mean responses
between those WP/English respondents with audit
experience in English-speaking countries and
those without such experience.
5.2. Tests of hypotheses
The ANOVA results in Table 4 indicate sig-
ni?cant di?erences across the three groups for 14
of the 21 uncertainty expression comparisons
(Column 6). To test Hypothesis 1 (overall e?ect),
the mean point-estimates for the CPA and WP/
German groups were compared. Bonferroni post-
hoc comparison tests
6
indicate signi?cant di?er-
4
Reagan et al. (1989) reported that, across seven prior
studies, the range of mean probabilities assigned to ‘‘probable’’
was 70–77% and to ‘‘likely’’ was 67–75%.
5
There was an insu?cient number of CPAs indicating a
high level of familiarity with IAS to warrant a similar test for
this group.
6
The Bonferroni procedure controls the overall error rate
by setting the error rate for each post hoc comparison test to
the experimentwise error rate (set at 0.05) divided by the total
number of tests.
T.S. Doupnik, M. Richter / Accounting, Organizations and Society 28 (2003) 15–35 25
Table 4
Point estimate probabilities: group means, ANOVA results, and post hoc comparison results
Col. 1 Col. 2 Col. 3 Col. 4 Col. 5 Col. 6 Col. 7 Col. 8 Col. 9
Post hoc comparisons (Sig. 2-tailed)
Group CPA CPA WP/German
CPA WP/English WP/German ANOVA vs. vs. vs.
n=128 n=62 n=109 Sig. WP/German WP/English WP/English
English expression German translation Mean Mean Mean F (2-tailed) (overall) (culture) (translation)
virtually certain so gut wie sicher 91.75 86.24 91.87 5.498 0.005 ** 1.000 0.008 ** 0.009 **
reasonable assurance angemessene Sicherheit 81.38 77.35 81.48 3.497 0.032 * 1.000 0.050 * 0.051
expected erwartet 80.16 72.95 72.88 7.290 0.001 ** 0.002 ** 0.004 ** 1.000
expected voraussichtlich 80.16 72.95 71.99 10.905 0.000 ** 0.000 ** above ** 1.000
assurance Gewibheit 79.46 90.18 96.73 59.514 0.000 ** 0.000 ** 0.000 ** 0.003 **
su?cient certainty hinreichende Sicherheit 78.17 80.86 81.77 2.330 0.099 0.106 0.544 1.000
reasonably likely nach vernu¨ nftigen Annahmen
wahrscheinlich
71.97 67.92 72.30 2.101 0.124 1.000 0.218 0.168
probable hinreichend wahrscheinlich 71.37 67.15 70.49 1.635 0.197 1.000 0.186 0.508
probable wahrscheinlich 71.37 67.15 68.14 2.319 0.100 0.268 above 1.000
likely voraussichtlich 70.89 67.27 71.99 2.213 0.111 1.000 0.331 0.118
likely wahrscheinlich 70.89 67.27 68.14 1.675 0.189 0.452 above 1.000
with the prospect mit der Aussicht 53.28 59.27 58.17 3.202 0.042 * 0.117 0.098 1.000
insu?cient certainty unzureichende Sicherheit 42.60 40.84 44.64 0.560 0.572 1.000 1.000 0.907
not probable nicht wahrscheinlich 32.61 21.60 21.76 13.415 0.000 ** 0.000 ** 0.000 ** 1.000
no longer probable voraussichtlich nicht mehr 29.38 19.66 24.75 4.473 0.012 * 0.300 0.003 ** 0.412
no longer probable nicht mehr wahrscheinlich 29.38 19.66 19.59 9.201 0.000 ** 0.000 ** above ** 1.000
no longer probable nicht mehr erwartet 29.38 19.66 15.51 16.507 0.000 ** 0.000 ** above ** 0.506
unlikely aller Wahrscheinlichkeit nach nicht 27.13 18.34 24.93 4.094 0.018 * 1.000 0.014 ** 0.117
seriously in question sehr zweifelhaft 23.96 19.77 13.05 11.378 0.000 ** 0.000 ** 0.372 0.049 *
not expected nicht erwartet 23.79 16.58 16.83 5.186 0.006 ** 0.015 * 0.039 * 1.000
remote Wahrscheinlichkeit a¨ uberst gering 16.38 27.07 11.46 9.994 0.000 ** 0.089 0.017 ** 0.000 **
**Denotes signi?cance at the 0.01 level. *Denotes signi?cance at the 0.05 level.
2
6
T
.
S
.
D
o
u
p
n
i
k
,
M
.
R
i
c
h
t
e
r
/
A
c
c
o
u
n
t
i
n
g
,
O
r
g
a
n
i
z
a
t
i
o
n
s
a
n
d
S
o
c
i
e
t
y
2
8
(
2
0
0
3
)
1
5
–
3
5
ences for eight of 21 comparisons (shown in Col-
umn 7 of Table 4), providing partial support for
H1. Di?erences are not concentrated on either side
of the probability scale. Three of the expression-
pairs with signi?cant di?erences have mean prob-
abilities greater than 50% and ?ve have means less
than 50%.
To test Hypothesis 2 (language-culture e?ect),
the mean point-estimates for the CPA and WP/
English groups were compared. Post-hoc compar-
ison tests show that means are signi?cantly di?er-
ent for nine of 16 comparisons (Table 4, Column
8), providing partial support for H2.
7
Because
both groups were evaluating uncertainty expres-
sions in English, these di?erences can be attrib-
uted to di?erences in language-culture, and not to
translation.
A de?nite pattern is present in the results re?ec-
ted in Table 4, Column 8. No di?erences exist
between the CPA and WP/German groups for
those expressions for which the mean CPA
response is between 40 and 79%. At the same
time, di?erences exist for all but one of the
expressions for which the mean CPA response
falls outside this range (each of the four highest
probability expressions and all but one of the six
lowest probability expressions). In seven of nine
signi?cant di?erences, the WP’s mean responses
are lower than those of the CPAs. It would be
interesting to investigate whether this pattern
exists in comparisons with other language-cul-
tures, or whether it is speci?c to the English and
German language-cultures.
To test Hypothesis 3 (translation e?ect), the
responses from the WP/English and WP/German
groups were compared. Means are signi?cantly
di?erent for four expression-pairs (see Table 4,
Column 9), providing partial support for H3. The
four German expressions in these pairs were
assigned the two highest probabilities (‘‘Gewis-
sheit’’ and ‘‘so gut wie sicher’’) and the two lowest
probabilities (‘‘Wahrscheinlichkeit a¨ ußerst gering’’
and ‘‘sehr zweifelhaft’’) by the WPs. In all four
cases, the German expression is assigned a prob-
ability that is more extreme than the equivalent
English expression. This result may indicate that
individuals are more able to use the full prob-
ability range and make subtle distinctions in their
own language than they are able to make in
another language. Although the Germans in the
WP/English group may be pro?cient English
speakers, it is unlikely that they have the same
ability to make subtle distinctions in English that
they would have in the German language. This
tendency could well result in a greater ability and
con?dence in assigning extreme probabilities to
verbal expressions in their own language, but less
con?dence to do so in another language.
To further test for a translation e?ect, the
responses to Part 1 (German expressions) and to
Part 2 (English expressions) of questionnaire GE2
were compared for those individuals who respon-
ded to both parts.
8
The results of paired-samples t-
tests reported in Table 5 indicate eight expression-
pairs with signi?cant di?erences, including the
four pairs found signi?cant in Table 4, Column 9.
Again, in each of these cases, the WPs assigned a
more extreme probability to the German expres-
sion than to its English equivalent. These results
reinforce the argument made earlier that indivi-
duals may be more comfortable in assigning
extreme probabilities to terms expressed in their
native language than in a foreign language.
Extending this line of research to other language-
cultures may uncover whether this argument can
be made for groups other than English-speaking
Germans.
To summarize, of the eight signi?cant overall
e?ect di?erences between CPA and WP/German
responses (Table 4, Column 7), six are the result of
a language-culture e?ect alone, but these six dif-
ferences relate to only three di?erent English
expressions (‘‘expected,’’ ‘‘not probable,’’ and ‘‘no
longer probable’’). One di?erence is the result of a
translation e?ect alone (‘‘seriously in question/
sehr zweifelhaft’’), and one is the result of both
language-culture and translation e?ects (‘‘assur-
7
A signi?cant di?erence exists for 12 of 21 items in Column
8. However, three of these signi?cant items are duplicates
(‘‘expected’’ is translated into two German expressions, ‘‘no
longer probable’’ is translated three di?erent ways), thus there
are really only nine signi?cant di?erences in Column 8.
8
A total of 33 responded to both parts, but two were
eliminated because of inconsistent responses in Part 1.
T.S. Doupnik, M. Richter / Accounting, Organizations and Society 28 (2003) 15–35 27
ance/Gewißheit’’). It is interesting to note that for
two expression pairs, the signi?cant language-cul-
ture and translation e?ects cancel each other out
such that no overall e?ect is observed (‘‘virtually
certain/so gut wie sicher’’ and ‘‘remote/Wahr-
scheinlichkeit a¨ ußerst gering’’).
As noted earlier, several English expressions
were translated into German in two or three dif-
ferent ways. To investigate whether the di?erent
translations of a single English expression were
assigned similar probabilities by the German-
speaking WPs, paired-samples t-tests were con-
ducted on responses provided by the WP/German
group. The results in Table 6 indicate that the two
di?erent translations into German for ‘‘expected’’
and ‘‘probable’’ are interpreted as equivalent by
the German WPs. However, the German transla-
tions for ‘‘likely’’ and ‘‘no longer probable’’ are
interpreted di?erently. For these expressions, the
speci?c translation from English to German could
a?ect the manner in which the related standard is
applied.
5.3. Range of probabilities
In Part 2 of questionnaires E, G, and GE1,
respondents indicated the range of probabilities
they associated with six of the uncertainty expres-
sions from Part 1. ANOVA results to test for dif-
ferences in mean probability ranges are reported
in Table 7. Of 329 respondents to questionnaires
E, G, and GE1, 21 were eliminated from this ana-
lysis either because they did not respond to this
part of the questionnaire (n=7), they had one or
more negative ranges (n=3), or they provided
logically inconsistent responses (n=11). Incon-
sistency was de?ned as the lower range value for
‘‘not probable/nicht wahrscheinlich’’ being greater
Table 5
Paired-samples t-tests of di?erences in interpretation of English expression and their German translation for the group of German
WPs evaluating both
Col. 1 Col. 2 Col. 3 Col. 4 Col. 5
English German t-test
n=31 n=31 Sig.
English expression German translation Mean Mean t (2-tailed)
assurance Gewißheit 89.50 95.77 2.731 0.011 *
virtually certain so gut wie sicher 87.00 91.42 5.498 0.029 *
su?cient certainty hinreichende Sicherheit 79.06 80.55 0.802 0.429
reasonable assurance angemessene Sicherheit 77.00 76.32 0.463 0.646
expected erwartet 71.48 70.48 0.403 0.690
expected voraussichtlich 71.48 69.42 0.699 0.490
likely wahrscheinlich 68.10 70.39 1.325 0.195
likely voraussichtlich 68.10 69.42 0.491 0.627
reasonably likely nach vernu¨ nftigen Annahmen wahrscheinlich 66.48 68.90 1.089 0.285
probable wahrscheinlich 65.52 70.39 2.761 0.010 **
probable hinreichend wahrscheinlich 65.52 69.48 2.227 0.034 *
with the prospect mit der Aussicht 60.19 56.84 1.278 0.211
insu?cient certainty unzureichende Sicherheit 38.19 37.68 0.162 0.873
remote Wahrscheinlichkeit a¨ ußerst gering 22.45 9.24 2.986 0.006 **
not probable nicht wahrscheinlich 22.06 13.84 2.606 0.014 *
no longer probable voraussichtlich nicht mehr 18.58 21.00 0.687 0.498
no longer probable nicht mehr erwartet 18.58 15.84 0.703 0.487
no longer probable nicht mehr wahrscheinlich 18.58 11.97 2.622 0.014 *
seriously in question sehr zweifelhaft 18.48 9.32 2.468 0.019 *
unlikely aller Wahrscheinlichkeit nach nicht 15.73 10.63 1.988 0.056
not expected nicht erwartet 14.16 11.61 1.138 0.264
**Denotes signi?cance at the 0.01 level. *Denotes signi?cance at the 0.05 level.
28 T.S. Doupnik, M. Richter / Accounting, Organizations and Society 28 (2003) 15–35
than the upper range value for ‘‘probable/wahr-
scheinlich.’’
Probability ranges di?er among the three
respondent groups for two expression pairs (see
Table 7, Column 6). Bonferroni post-hoc compar-
ison tests show that there is a signi?cant di?erence
in probability range assigned by the WP/English
and WP/German groups on ‘‘virtually certain/so
gut wie sicher’’ (translation e?ect). There is also a
signi?cant di?erence between the CPA and WP/
English groups on ‘‘probable’’ (language-culture
e?ect). But there is no overall e?ect for either
expression pair. The magnitudes of the mean ran-
ges suggests that the expressions ‘‘probable’’ and
‘‘not probable’’ convey less precise concepts of
probability than do expressions such as ‘‘virtually
certain’’ and ‘‘remote.’’ The same is true for the
German translations of these expressions.
Although only one statistically signi?cant dif-
ference exists between the WP/English and WP/
German groups, a de?nite pattern emerges. For
each uncertainty expression, the range of prob-
abilities assigned by the WPs to the English
expression is broader than the range assigned to
Table 6
Paired samples t-tests of di?erences in interpretation of multiple German translations of a single English expression
Col. 1 Col. 2 Col. 3 Col. 4
WP/German t-test
n=109 Sig.
English expression German translation Mean t (2-tailed)
expected erwartet 72.88
voraussichtlich 71.99 0.564 0.574
probable hinreichend wahrscheinlich 70.49
wahrscheinlich 68.14 1.824 0.071
likely voraussichtlich 71.99
wahrscheinlich 68.14 2.920 0.004 **
no longer probable voraussichtlich nicht mehr 24.75
nicht mehr wahrscheinlich 19.59 2.982 0.004 **
no longer probable nicht mehr wahrscheinlich 19.59
nicht mehr erwartet 15.51 3.279 0.001 **
no longer probable voraussichtlich nicht mehr 24.75
nicht mehr erwartet 15.51 4.518 0.000 **
**Denotes signi?cance at the 0.01 level. *Denotes signi?cance at the 0.05 level.
Table 7
Range of probabilities: group means, ANOVA results, and post hoc comparison results
Col. 1 Col. 2 Col. 3 Col. 4 Col. 5 Col. 6 Col. 7 Col. 8 Col. 9
Group ANOVA Post hoc comparisons (Sig. 2-tailed)
English
Expression
German
Translation
CPA
n=148
Mean
WP/English
n=34
Mean
WP/German
n=126
Mean
F Sig.
(2-tailed)
CPA
vs.
WP/German
(overall)
CPA
vs.
WP/English
(culture)
WP/German
vs.
WP/English
(translation)
virtually certain so gut wie sicher 8.59 11.44 7.41 3.737 0.025 * 0.626 0.156 0.021 *
reasonable assurance angemessene Sicherheit 12.18 14.15 11.95 1.402 0.248 1.000 0.406 0.302
su?cient certainty hinreichende Sicherheit 12.72 13.76 10.90 2.838 0.060 0.105 1.000 0.247
probable wahrscheinlich 15.93 22.41 16.84 3.729 0.025 * 1.000 0.021 * 0.066
not probable nicht wahrscheinlich 16.58 19.24 15.67 0.911 0.403 1.000 0.930 0.539
remote Wahrscheinlichkeit
a¨ uberst gering
11.68 13.76 10.90 1.660 0.192 1.000 0.554 0.218
**Denotes signi?cance at the 0.01 level. *Denotes signi?cance at the 0.05 level.
T.S. Doupnik, M. Richter / Accounting, Organizations and Society 28 (2003) 15–35 29
the German expression. (It is also broader than
the range assigned by the CPAs to each English
expression.) This suggests that the WPs may be
less certain of the probability to assign to expres-
sions stated in English than to expressions stated
in their native German.
6. Summary and conclusions
This study has found that nationality alone (at
least among German-speaking countries) does not
result in signi?cant di?erences in probabilities
assigned to uncertainty expressions used in Inter-
national Accounting Standards. However, sig-
ni?cant di?erences exist between English-speaking
US CPAs and German-speaking WPs for a large
number of the uncertainty expressions included in
the study. The results indicate that for some
expressions, the di?erence in mean probability
assignments can be attributed to a di?erence in the
language-culture of the respondent groups. These
two ?ndings (a language-culture e?ect and no
nationality e?ect) together provide support for the
general hypothesis derived from the theory of lin-
guistic relativism that language-culture a?ects the
interpretation of uncertainty expressions. Of
course, these results may be speci?c to the lan-
guage-cultures examined in this study. Future
research should be conducted to determine whe-
ther this e?ect holds for other language-cultures,
initially focusing on economically signi?cant
countries important for international harmoniza-
tion. Extending this line of research to other lan-
guage-cultures also would provide evidence as to
whether the patterns of di?erences found in this
study between the English and German language-
cultures exist between other language-cultures as
well.
Results also indicate that for extreme prob-
ability expressions (highest and lowest), the trans-
lation from English to German results in
signi?cant di?erences in interpretation. This raises
the question whether this e?ect is a result of poor
translation or whether the English expression has
no direct counterpart in German. For example, is
‘‘Gewissheit’’ not the best translation of ‘‘assur-
ance’’ or is there no direct linguistic mapping of
‘‘assurance’’ into German? Poor translation is the
easier problem to solve. Results related to the
various German translations of the phrase ‘‘no
longer probable’’ indicate that, at least for this
uncertainty expression, some translations are bet-
ter than others. Further research could be con-
ducted investigating alternative translations of
other uncertainty expressions for which this study
has found a translation e?ect to exist to determine
whether that e?ect can be avoided.
The translation e?ect found in this study is spe-
ci?c to the German translation of IAS approved
by the IASC in 1997. Since then, IAS have been
translated into a number of languages.
9
Addi-
tional research investigating the e?ect of translat-
ing IAS to other languages might be useful in
assessing the magnitude of a potential translation
problem.
The results suggest that non-native English
speakers may be less con?dent in their interpreta-
tion of uncertainty terms expressed in English
than in their native language. This points to the
importance of translating IAS from English into
other languages. If people can make ?ner and
subtler distinctions and interpretation in their own
language (as the results imply and logic supports),
then application by non-native English speakers of
a good translation should lead to better inter-
pretation of IAS than application of the original
English version. However, the key is a good
translation, which is an important consideration
for the IASC.
On the other hand, while IAS are being trans-
lated into more languages, there will be instances
where the original English versions will continue
to be used by non-native English speakers. In
these instances, this study’s ?ndings are relevant in
that people working in a second, non-native lan-
guage appear to be less comfortable and less con-
?dent in assigning extreme probabilities to
uncertainty expressions in the foreign, English
language than in their native language.
9
The IASC has published translations of International
Accounting Standards in several languages. As of January
2001, approved translations were available in Chinese, Czech,
German, Polish, Romanian, Russian, Slovak, Spanish, and
Ukrainian (Source: www.iasc.org.uk).
30 T.S. Doupnik, M. Richter / Accounting, Organizations and Society 28 (2003) 15–35
This study discovered di?erences between two
language-cultures in the interpretation of uncer-
tainty expressions ‘‘in isolation.’’ A potential out-
come of these interpretation di?erences is a lack of
consistency by members of the two language-cul-
tures in applying IAS. However, this study has not
demonstrated that those di?erences necessarily
would result in di?erent decisions being made by
accountants and auditors in applying IAS. Uncer-
tainty expressions found in IAS must be interpreted
within the context in which they are used. Future
research could examine whether the interpretation
of uncertainty expressions ‘‘in context’’ di?ers
across language-cultures.
10
A judgment task
experimental method could be used in which parti-
cipants from two or more countries are asked to
apply in a hypothetical scenario the provision of an
International Accounting Standard that requires
the interpretation of uncertainty expressions.
11
Laboratory experiments could be used to avoid
the potential non-response bias present in mail
survey approaches. If the results of such future
research con?rm the results of the current study,
then true comparability of ?nancial reporting
worldwide may require the IASC to avoid the use
of vague uncertainty expressions in developing
International Accounting Standards. The potential
problem for cross-national comparability asso-
ciated with the interpretation of uncertainty
expressions is exacerbated by the sheer number of
instances in which verbal probability expressions
are used in IAS.
Acknowledgements
We wish to thank Stefan Schreiber and
Michaela Donle for their expert assistance in
designing and collecting data for this study. We
would also like to thank an anonymous reviewer
for helpful comments and suggestions, and the
Center for International Business Education and
Research at the University of South Carolina for
?nancial support.
10
It would also be interesting to investigate whether uncer-
tainty expressions are interpreted in the same way across con-
texts within a language-culture. For example, is the term
‘‘probable’’ interpreted the same way when evaluating the
potential capitalization of an intangible asset as when evaluat-
ing the possible recognition of a contingent liability. If not,
then there must be factors other than language-culture that
operate in determining how uncertainty expressions are inter-
preted.
11
Because previous research has found greater within group
variability when uncertainty expressions are embedded in con-
text rather than presented in isolation, in context studies may
be less likely to ?nd signi?cant di?erences.
T.S. Doupnik, M. Richter / Accounting, Organizations and Society 28 (2003) 15–35 31
Appendix A
Uncertainty Expressions in International Accounting Standards
PART 1: Listed below are expressions that relate to a level of probability (degree of certainty or uncer-
tainty). All of these expressions are used in International Accounting Standards. Please indicate the level of
probability that best corresponds, in your opinion, to each expression. There are no ‘‘right’’ or ‘‘wrong’’
answers. We are only interested in your perceptions. Please indicate the probability in percentage terms on
a scale of 0% to 100%!
Please indicate the probability in percentage terms that best corresponds, in your opinion, to each of the
following expressions:
32 T.S. Doupnik, M. Richter / Accounting, Organizations and Society 28 (2003) 15–35
PART 2: In this part, please indicate the range of probabilities that best corresponds, in your opinion, to
each of the following expressions. Please indicate in percentage terms both the upper and lower limits to
the probability range.
PART 3: Please respond to the following questions so that we may develop a pro?le of respondents
.
T.S. Doupnik, M. Richter / Accounting, Organizations and Society 28 (2003) 15–35 33
References
Agacer, G. M., & Doupnik, T. S. (1991). Perceptions of auditor
independence: a cross-cultural study. The International Jour-
nal of Accounting, 26, 220–237.
Amer, T., Hackenbrack, K., & Nelson, M. (1994). Between-
auditor di?erences in the interpretation of probability phra-
ses. Auditing: A Journal of Practice & Theory, 13(1), 126–136.
Bagrano?, N. A., Houghton, K. A., & Hronsky, J. (1994). The
structure of meaning in accounting: a cross-cultural experi-
ment. Behavioral Research in Accounting, 6(Supplement), 35–
57.
Belkaoui, A. (1989). Accounting and language. Journal of
Accounting Literature, 8, 281–292.
Beyth-Marom, R. (1982). How probable is probable? Numer-
ical translation of verbal probability expressions. Journal of
Forecasting, 1, 257–269.
Brun, W., & Teigen, K. (1988). Verbal probabilities: ambig-
uous, context-dependent, or both?. Organizational Behavior
and Human Decision Processes, 39, 390–404.
Budescu, D. V., & Wallsten, T. S. (1985). Consistency in inter-
pretation of probabilistic phrases. Organizational Behavior
and Human Decision Processes, 36, 391–405.
Chesley, G. R. (1986). Interpretation of uncertainty expres-
sions. Contemporary Accounting Research, 2(2), 179–199.
Chow, C. W., Kato, Y., & Merchant, K. A. (1996). The use
of organizational controls and their e?ects on data
manipulation and management myopia: a Japan vs. US
comparison. Accounting, Organizations and Society, 21,
175–192.
Chow, C. W., Shields, M. D., & Wu, A. (1999). The importance
of national culture in the design of and preference for man-
agement controls for multi-national operations. Accounting,
Organizations and Society, 24, 441–461.
Davidson, R. A. (1991). Practical di?culties encountered in
selecting uncertainty words: the example of accounting stan-
dards. Applied Psychology: An International Review 40 (4),
353–363.
Davidson, R. A., & Chrisman, H. H. (1993). Interlinguistic
comparison of international accounting standards: the case
of uncertainty expressions. The International Journal of
Accounting, 28(1), 1–16.
Davidson, R. A., & Chrisman, H. H. (1994). Translations of
uncertainty expressions in Canadian accounting and auditing
standards. Journal of International Accounting, Auditing &
Taxation, 3(2), 187–203.
Doupnik, T. S., & Salter, S. B. (1995). External environment,
culture, and accounting practice: preliminary test of a general
model of international accounting development. The Inter-
national Journal of Accounting, 30(3), 189–207.
Exhibit 1. Summary of four di?erent versions of the questionnaire
Questionnaire
Component E G GE1 GE2
Instructions English German German German
language language language language
Part 1 Point estimate Point estimate Point estimate Point estimate
probabilities for probabilities for probabilities for probabilities for
16 English- 19 German- 16 English- 19 German-
language language language language
expressions expressions expressions expressions
Part 2 Range of Range of Range of Point estimate
probabilities for probabilities for probabilities for probabilities for
6 English- 6 German- 6 English- 16 English-
language language language language
expressions expressions expressions expressions
Part 3 English German German German
language language with language with language with
questions on questions on questions on
English ?uency English ?uency English ?uency
and experience and experience and experience
34 T.S. Doupnik, M. Richter / Accounting, Organizations and Society 28 (2003) 15–35
Financial Accounting Standards Board. (1985). Statement of
Financial Accounting Concepts No. 6, Elements of Financial
Statements.
Frank, W. G. (1979). An empirical analysis of international
accounting principles. Journal of Accounting Research, 17(2
Autumn), 593–605.
Gray, S. J. (1988). Towards a theory of cultural in?uence on
the development of accounting systems internationally. Aba-
cus, 24(1), 1 - 15.
Harrison, G. L. (1993). Reliance on accounting performance
measures in superior evaluative style-the in?uence of
national culture and personality. Accounting, Organizations
and Society, 18, 319–339.
Davidson, R. A., & McKinnon, J. L. (1986). Cultural and
accounting change: a new perspective on corporate reporting
regulation and accounting policy formulation. Accounting,
Organizations and Society, 11, 233–252.
Davidson, R. A., & McKinnon, J. L. (1999). Cross-cultural
research in management control systems design: a review of
the current state. Accounting, Organizations and Society, 24,
483–506.
Harrison, K. E., & Tomassini, L. A. (1989). Judging the prob-
ability of a contingent loss: an empirical study. Con-
temporary Accounting Research, 5(2), 642–648.
Hofstede, G. (1980). Culture’s consequences: international di?er-
ences in work-related values. Beverly Hills: Sage Publications.
Jiambalvo, J., & Wilner, N. (1985). Auditor evaluation of con-
tingent claims. Auditing: A Journal of Practice & Theory 5
(1), 1–11.
Lichtenstein, S., & Newman, J. R. (1967). Empirical scaling of
common verbal phrases associated with numerical prob-
abilities. Psychonomic Science, 9(10), 563–564.
Monti-Belkaoui, J., & Belkaoui, A. (1983). Bilingualism and
the perception of professional concepts. Journal of Psycho-
linguistic Research, 12(2), 111–127.
Mueller, G. G. (1967). International accounting. New York:
Macmillan.
Nair, R. D., & Frank, W. G. (1980). The impact of disclosure
and measurement practices on international accounting
classi?cations. The Accounting Review LV(3), 426–450.
Phillips, L. D., & Wright, C. N. (1977). Cultural di?erences in
viewing uncertainty and assessing probabilities. In
H. Jungermann, & G. de Zeeuws (Eds.), Decision making and
change in human a?airs. D. Dordrecht, Holland: Reidel
Publishing Company.
Pitz, G. F. (1980). Critique of ‘procedures for the communica-
tion of uncertainty in auditor’s working papers’. In T. J.
Burns (Ed.), Behavioral experiments in accounting II.
Columbus, OH: The Ohio State University.
Reagan, R. T., Mosteller, F., & Youtz, C. (1989). Quantitative
meanings of verbal probability expressions. Journal of
Applied Psychology, 74(3), 433–442.
Reimers, J. L. (1992). Additional evidence on the need for dis-
closure reform. Accounting Horizons, 6(1), 36–41.
Riahi-Belkaoui, A., & Picur, R. D. (1991). Cultural determin-
ism and the perception of accounting concepts. The Interna-
tional Journal of Accounting, 26(2), 118–130.
Salter, S. B., & Niswander, F. (1995). Cultural in?uence on the
development of accounting systems internationally: a test of
Gray’s [1988] theory. Journal of International Business Stud-
ies, 26(2), 379–397.
Sapir, E. (1964). Culture, language, and pesonality: selected
essays. In D. G. Mandelbaum. Berkeley, CA: University of
California Press.
Scha¨ ?er Poeschel Verlag. (1998). International Accounting
Standards 1997: Deutsche Fassung. Stuttgart: International
Accounting Standards Committee (Hrsg.) Scha¨ ?er Poeschel
Verlag.
Schultz, J. J. Jr., & Reckers, P. M. J. (1981). The impact of
group processing on selected audit disclosure decisions.
Journal of Accounting Research, 19(2), 482–501.
Simpson, R. H. (1963). Stability in meanings for quantitative
terms: a comparison over 20 years. Quarterly Journal of
Speech, 49, 146–151.
Violet, W. J. (1983). The development of international account-
ing standards: an anthropological perspective. International
Journal of Accounting Education and Research, 18(2), 1–12.
Wallsten, T. S., Budescu, D. V., Rapoport, A., Zwick, R., &
Forsyth, B. (1986). Measuring the vague meanings of prob-
ability terms. Journal of Experimental Psychology: General,
115(4), 348–365.
Whorf, B. L. (1956). Language, thought and reality: selected
writings. In J. B. Carroll. Cambridge, MA: MIT Press.
Zarzeski, M. T. (1996). Spontaneous harmonization e?ects on
culture and market forces on accounting disclosure practices.
Accounting Horizons, 10(1), 18–37.
T.S. Doupnik, M. Richter / Accounting, Organizations and Society 28 (2003) 15–35 35

doc_983869466.pdf