The value-free ideal of science: a useful fiction?
A review of non-epistemic reasons for the research integrity community
Jacopo Ambrosj (KU Leuven)
Kris Dierickx (KU Leuven)
Hugh Desmond (CNRS/Paris I-Sorbonne)
Even if the “value-free ideal of science” (VFI) were an unattainable goal,
one could ask: can it be a useful fiction, one that is beneficial for the
research community and society? This question is particularly crucial for
scholars and institutions concerned with research integrity (RI), as one
cannot offer normative guidance to researchers without making some
assumptions about what ideal scientific research looks like. Despite the
insofar little interaction between scholars studying RI and those working
on values in science, the overlap of topics and interests make
collaboration between the two fields promising for understanding
research and its ethics.
Here, we identify—for the use of RI scholars—the non-epistemic reasons
(societal, political, professional) for and against the VFI considered in the
literature. All of these are concerned with the beneficial or detrimental
consequences that endorsing the VFI would have on society, policy-
making, or the scientific community, with some authors appealing to the
same principles to argue for opposite positions. Though most of the
reviewed articles do not endorse the VFI, it is generally agreed that some
constraints have to be put on the use of non-epistemic values.
Disagreement on the utility of the VFI lies both on the different
epistemic-descriptive positions taken by different authors, and on the
scarcity of relevant empirical studies. Engaging critically with the reasons
here identified and more in general with the values in science debate will
help the RI community decide whether the VFI should be included in
future codes of conduct.
In just a few decades, the scholarly community has produced a voluminous and highly
fragmented academic literature on research integrity (RI). The empirical data and
reflections so produced have contributed to inform the many codes of conduct that
have been published. These codes spell out the values that should guide research, and
give guidance to researchers on how to stay true to those values, especially in cases
where other interests may be at stake. In addition to the RI scholarly community, also
philosophers of science have been increasingly debating which values should guide
research, which ones should be avoided, and, more in general, what should be our ideal
of research and scientific community. Particular attention has been paid to the so-
called “value-free ideal” of science (VFI), according to which during the epistemic
phases (data analysis, inference, theory assessment, etc.) of research, scientists should
not be guided by “non-epistemic values”, i.e. moral, cultural, commercial, or political
. This means that scientific decisions, whether about how to collect and analyze
data or what conclusions to draw from them, should only be influenced by “epistemic
values” such as simplicity, broadness of scope, accuracy, consistency, or fruitfulness
Should the VFI be endorsed? Even a cursory inspection reveals this question to be
complex and many-faceted. On the one hand, an oft-cited reason for endorsing the VFI
is the distorting influence of commercial interests. Such distortion does not necessarily
involve overt falsification or fabrication, but can lead to a corporation’s scientists to
prioritize certain research questions over others, or avoiding publication of
unfavorable results (Smith, 2005). On the other hand, it has been argued that scientific
research, in practice, cannot be entirely independent of non-epistemic values. The
most well-known of these arguments is the one from inductive risk: when drawing a
conclusion from data, the scientists must evaluate the costs of being wrong, and this
There are many different interpretations of the VFI. The one we employ here comes very close to
Heather Douglas’ (Douglas, 2009).
The very distinction underpinning the VFI, that between epistemic and non-epistemic values, has been
criticized by authors like Phyllis Rooney and Helen Longino (Longino, 1996; Rooney, 1992, 2017).
However, scholars involved in the debate have been keeping using this distinction, or similar ones. As
we aim at offering an overview of the debate, we keep using it in our study, with no pretentions to
evaluation cannot but be based on non-epistemic values (Douglas, 2000; Rudner,
Despite the overlap of themes, concerns, and aims, only little collaboration between
the scholars studying RI and those studying values in science has taken place to the
present day. We believe—together with Kevin Elliott and David Resnik (2019)—that
our general understanding of research, its values, and ethics, will benefit from the
collaboration of these two academic communities.
Rationale and objectives
In this article, we take a first step towards this collaboration by offering to the RI
community an overview of the reasons discussed by philosophers of science for and
against the VFI. Given the aims of RI, instead of focusing on whether the VFI describes
a real state of affairs or not, we focus on its normative dimension. Even if the actual
practice of scientific research were unavoidably influenced by non-epistemic values, it
may be still desirable to endorse the VFI for the beneficial consequences it may have
for the research community and society in general. Affirming the VFI as a community-
wide norm may be beneficial for public trust in science, or it may encourage scientists
to refrain from conflicts of interest. In other words, even if the VFI were a fiction, not
accurately describing the reality of scientific research, to what extent could it still be
considered a “useful fiction”?
To shed light on this, we investigate some of the main non-epistemic reasons both for
and against endorsing the VFI as a community-wide norm. By non-epistemic reasons
we mean all those reasons that are not primarily concerned with epistemic desiderata
(objectivity, accuracy, consistency, etc.), but are grounded on political, social, moral,
and other non-epistemic considerations. Does endorsing the VFI enhance public trust
or the integrous behavior of scientists—or conversely, does it discourage scientists from
being engaged with their communities? We aim to map the most important benefits
and downsides that can be expected from an endorsement of or a distancing from the
Part of the relevance of our research question resides in the collaboration between the
two fields we aim to foster. In particular, it is important to consider the implications
that an endorsement of the VFI both at the individual and institutional level has for the
values and norms that should govern research.
We follow the PRISMA guidelines (Page et al., 2021) to report our review of reasons
(Sofaer & Strech, 2012; Strech & Sofaer, 2012).
We include publications satisfying the following criteria:
1. Publication type: peer-reviewed; journal article; written in English; published after
the year 2000 (included).
This excludes books, contributions to books, non-research articles (e.g. book reviews),
articles published before 2000, and articles not written in English. Though books are
still very influential in the philosophical literature, a great deal of the most influential
contributions to the debate are mainly focused on epistemic and descriptive arguments
against the VFI. For instance, Heather Douglas’ work is based on the descriptive
observation that, because of inductive risk, scientific inferences requires non-epistemic
value judgments (Douglas, 2009), and Helen Longino’s work is largely based on the
descriptive observation that scientific theories are underdetermined by the evidence
(Longino, 1990). Further discussion on this is offered in the limitations section.
2. Publication content: one or more non-epistemic reasons for or against the VFI;
concerns science in general, not just a specific scientific domain.
This excludes articles containing solely descriptive claims about the VFI (e.g. science
is value-free), or epistemic reasons (e.g. if not value-free, science is not objective).
Moreover, it excludes articles that highlight the role of values in a specific science by
contrasting it to other sciences (e.g. taxonomy—unlike molecular biology—is not value-
Information sources and search strategy
First, we searched Web of Science Core Collection and Philosopher’s Index using
combinations of keywords related to four core concepts: non-epistemic values,
epistemic values, values in science, and value-free science (the full search strings are
available online). Then, we used the snowball method, and ran a related search on
Google Scholar. The databases were last searched on April 20, 2021.
After we had agreed on the eligibility criteria, the records (title, abstracts, and other
publication data) were screened. We discarded articles dealing with values in non-
scientific contexts (e.g. articles on how different kinds of values influence customer
choice, as Assarut & Eiamkanchanalai, 2015). We also discarded any articles not
matching the desired publication type that were returned notwithstanding our search-
filters. Remaining articles were fully read. In this phase, the selection was made on the
basis of content-related considerations. We included articles offering at least one non-
epistemic reason for or against the VFI. For instance: we should endorse the VFI
because it increases public trust; or: we should reject it because science should pursue
social values. We excluded articles that, in evaluating the VFI, considered epistemic
reasons only, such as objectivity or accuracy. For instance, though it deals with non-
epistemic values and the VFI, Ludwig, 2016 was eventually excluded because we were
not able to identify a single non-epistemic reason in it.
Data collection and extraction process
We considered all the reasons in the reviewed articles, regardless of the weight they
were given by the authors. We first identified reason mentions, i.e. specific passages of
a text mentioning a reason. As long as a passage explicitly stated a reason, we counted
it as a reason mention, no matter its length. After collecting all reason mentions, we
clustered them into thematic areas. Within each thematic area, we grouped reason
mentions into reason types according to the principle or value they appealed to. Next,
we categorized each reason mention according to whether the author claimed it
counted for or against the VFI.
Given that different authors use different definitions of the VFI, and not all authors use
the term “value-free”, we employed two minimal criteria to identify it. First, whenever
the term “value-free ideal” was explicitly employed, we took it as an instance of the VFI.
Second, when the term “value-free ideal” was not explicitly employed, we took any ideal
implying at least that non-epistemic values should not influence epistemic phases of
science (data analysis, inference, theory assessment, etc.) as an instance of the VFI. For
example, we counted as an instance of the VFI the view discussed by Inmaculada de
Melo-Martin and Kristen Intemann according to which “scientists need not, and
should not, endorse non-epistemic values related to their research, as doing so may
bias their assessment of what the evidence is” (de Melo-Martin & Intemann, 2012, p.
60), even though they do not use the term “value-free ideal”.
Figure 1 PRISMA flow chart of the selection process
Study selection and characteristics
The database search yielded 530 articles, duplicates excluded. Of these, 15 were eligible
for inclusion. We added 3 more articles retrieved via other methods. The total number
of included articles is 18 (see table 1).
In this section, we present our findings by describing the definitions of the VFI
addressed in the reviewed articles, their general stance towards it, and the reasons for
and against it that they consider, sorted by thematic area.
Table 1 Reviewed articles arranged by year, including journal, number of reasons
mentions, and general stance towards the value-free ideal (VFI). N.B. Reason mentions
in an article do not necessarily represent the view of its authors (they may be the report
of someone else’s view). Moreover, though they may play a role in the reasoning behind
the general stance assumed by the reviewed authors, epistemic reasons were not coded,
and therefore not included in the count of reason mentions. This is why, for instance,
de Melo-Martin & Intemann (2016) have only reasons for the VFI, but nonetheless do
not endorse it.
Definitions of the VFI, and general stances
While all the definitions of the VFI used in the reviewed articles match our minimal
criteria, some of them are stated in a weaker form, e.g. “whereas a hypothesis may
Number of reason
Philosophy of Science
Behavior and Philosophy
Philosophy of Science
Science and Engineering Ethics
Public Affairs Quarterly
Science and Engineering Ethics
de Melo-Martin &
Perspectives in Biology and Medicine
European Journal for Philosophy of
Studies in History and Philosophy of
European Journal for Philosophy of
de Melo-Martin &
Philosophy of Science
Perspectives on Science
Resnik & Elliott,
Accountability in Research Policies and
Studies in History and Philosophy of
Resnik & Elliott,
Studies in History and Philosophy of
Studies in History and Philosophy of
contingently be held for valuational reasons, it is epistemically preferable if it is held
for evidential reasons” (Hudson, 2016, p.168), and some in a stronger form, e.g.
“science is not entitled to formulate value-judgments in which any non-cognitive
values are taken into account” (Lekka-Kowalik, 2010, pp. 33-34)
. The former
definition proposes the VFI as an ideal to strive for, while admitting that sometimes
non-epistemic values do affect scientific reasoning. By contrast, the latter maintains
that non-epistemic values cannot ever be considered.
3 articles (de Melo-Martin & Intemann, 2012; Koertge, 2000, 2004) do not employ the
phrase “value-free ideal”, nor similar ones, but discuss ideals of science matching our
minimal criteria for the VFI. Similarly, Rottschaefer (2003) talks of “value-neutrality”,
but defines it in a way matching our criteria. Therefore, we consider these 4 articles on
a par with those explicitly addressing the VFI.
11 of the articles reviewed do not endorse the VFI, 7 do. After criticizing and
acknowledging the shortcomings of its traditional formulations, 3 of the articles not
endorsing the VFI propose what we will call “value-limiting ideals” (VLIs) which do not
reject all non-epistemic values in science but only some and in certain contexts. The
three proposed versions of VLIs in the reviewed articles are:
• Non-epistemic values should not be used in place of evidence (direct role), but
only to determine the standards for sufficient evidence (indirect role) (Douglas,
• The justification of findings should not be based on non-epistemic values that
are incompatible with the values of the intended public. Instead, scientists
should endorse a “value-apt ideal”, according to which the communication of
their findings should be guided by values shared with the intended public (John,
• Whenever complete value-neutrality is not possible, scientists can be influenced
by non-epistemic values as they make this influence explicit—i.e. they are
transparent about it (Resnik & Elliott, 2016).
The complete list the various definitions of the VFI and VLI used in the reviewed
articles is available in the online supplementary material.
In the literature the terms “cognitive” and “non-epistemic” are sometimes used interchangeably. See
for instance Douglas, 2013 and Rooney, 1992.
Thematic areas, reason types, and alleged implications
We identified three thematic areas – societal, political, and professional– reflecting the
groups affected by the endorsement of the VFI: society, politics, and the scientific
community itself. We grouped different reason mentions in 10 reasons types, each of
which appeal to a different principle to argue for or against the VFI. 2 are reasons
against the VFI, 4 are reasons for the VFI, and 4 have been used to argue both ways.
Table 2 displays reason types, including for each of them a synthetic description, the
number of articles mentioning them, and the bibliographic references.
Table 2 Overview of the reasons for and against the value-free ideal (VFI). The table includes: thematic areas (societal, political, and
professional); reason types with a brief explication; alleged implications (for or against the VFI) with the number of articles
mentioning them; bibliographic references.
Abandoning the VFI would impair public trust, because the main reason
why the public trusts scientists is that their claims are deemed objective and
independent of personal preferences.
(Bright, 2018, p. 2228, 2243), (Douglas, 2008, p. 5), (Hudson, 2016,
p. 180, 182), (John, 2015b, p. 9), (Koertge, 2000, p. 54)
Holding the VFI even though science cannot be value-free would be
detrimental for public trust. It is better to acknowledge the full range of
values involved in scientific research in a transparent way, rather than
maintaining a false appearance of complete value-neutrality.
(Douglas, 2008 p. 7; Resnik & Elliott, 2016 pp. 35-36)
Holding the VFI could discourage scientists pursuing their responsibilities
towards society. Since they are moral agents who receive public funding,
scientists should actively tackle urgent issues and consider the broad
societal consequences of their research.
(Bueter, 2015, p. 22), (Drenth, 2006, p. 15), (Hudson, 2016, p. 187),
(Lekka-Kowalik, 2010, p. 39), (Resnik & Elliott, 2016, p. 36, pp. 36-
37), (Rottschaefer, 2003, p. 244)
Keeping scientific claims value-free preserves the personal autonomy of
members of the public. Otherwise, the general public would rely on
knowledge based on values they may not share.
(Betz, 2013, p. 207), (John, 2015a, p. 80), (John, 2019, p.69)
Holding the VFI would make science less discriminatory, because non-
epistemic values could distort scientific findings, making them sexist or
(Rottschaefer, 2003, p. 227)
Holding the VFI would reinforce discrimination, because it would make it
easier for undesirable values to go unnoticed. We should better actively
inject politically progressive values in science instead.
(Koertge, 2000, pp. 49-50), (Koertge, 2004, p. 870),
(Rottschaefer, 2003, p. 228)
By holding the VFI, and thus avoiding value judgments, scientists can
prioritize giving useful information to policymakers. By contrast, including
political values within science would rule out some possible solutions to
public problems based on principle.
(Bright, 2018, p. 2228), (de Melo-Martin & Intemann, 2016, p. 503),
(John, 2015b, p. 5), (Koertge, 2004, p. 874), (Resnik & Elliott, 2016,
Holding the VFI would make scientific claims useless to policymakers.
Attempting to strip scientific claims of any value assumption would make
their scope less broad, thus reducing the possible applications of scientific
knowledge to policy.
(Ahn, 2020, p. 63), (de Melo-Martin & Intemann, 2012, p. 67),
(Douglas, 2008, p. 7), (John, 2015b, p. 9, p. 12)
Abandoning the VFI would infringe basic democratic principles. As
scientific knowledge is used to determine the political route of democratic
societies, it should not be influenced by the value-judgments of unelected
(Betz, 2013, p. 207), (de Melo-Martin & Intemann, 2016, p. 503),
(John, 2015a, p. 93), (John, 2015b, p. 5)
The rejection of the VFI jeopardizes research integrity, as it is an essential
feature of both the ethics and the practice of science to be guided by
epistemic values only.
(Ahn, 2020, p. 57), (Douglas, 2008, p. 6), (Resnik & Elliott, 2016, p.
35), (Resnik & Elliott, 2019, p. 2)
Holding the VFI could cause communication breakdowns between
scientists: information about the values that underlie judgments are crucial
for successful communication between scientists.
(de Melo-Martin & Intemann, 2012, p. 67)
Endorsing the VFI allows ethical consideration to be minimized in science
curricula. Were we to abandon the VFI, then ethics would need a more
prominent position, and this would mean less attention for learning the
(Hudson, 2016, p. 188)
Autonomy of scientists
Holding the VFI would preserve the autonomy of scientists from political
pressure. In fact, were political values to guide scientific research, scientists
would censor themselves whenever their findings are in contrast with those
(Douglas, 2008, p. 6)
Holding the VFI would breach the autonomy of scientists. Were scientists
to exclude any value-judgment in their work, they would not be free to act
ethically, ending up being ‘minds to be hired’.
(Douglas, 2008, pp. 12-13), (Hudson, 2016, p. 187),
(Lekka-Kowalik, 2010, p. 38)
Under this label we include reasons based on the consequences of endorsing or not
endorsing the VFI for society in general.
Among these reasons, the ones concerned with public trust are the most represented,
as they are mentioned by 6 articles. It is mostly argued that holding the VFI increases
the trustworthiness of science in the eyes of the public, but on two occasions it is argued
that openly acknowledging the presence of values in science would be more beneficial
than rejecting those values. In particular, one article evaluates public trust both as a
reason for and against the VFI. Another well-represented reason type is social
responsibility, which includes general claims against the VFI based on the
responsibilities that scientists have towards society. While the source of these
responsibilities may be different, all of these reasons against the VFI imply that
scientists should actively acknowledge and address non-epistemic values, rather than
misleadingly presenting their work as value-free.
Under this label we include reasons based on the consequences of endorsing or not the
VFI for politics. Although there may be some overlap with societal reasons, the two can
be distinguished insofar as political reasons focus specifically on society’s use of science
through political bodies.
Most of the political reasons are concerned with how the VFI impacts public
intervention: the process of policy-making and its effectiveness. The reasons
considered in the reviewed literature were split almost equally between benefits for and
challenges to policy-making. Some other political reasons concerned the relation
between the VFI and democratic principles, which determine policies in democratic
societies. These principles were seen as a reason for endorsing the VFI, because the
VFI is seen as preventing scientists from assuming technocratic power. In democratic
societies policy-makers use scientific knowledge to make decisions on behalf of
citizens. They alone, because democratically elected, are allowed to make value-
judgments. Therefore, the argument goes, it would be inappropriate if scientific
knowledge were already influenced by value-judgments made by unelected people, i.e.
Under this label we include reasons based on the consequences of endorsing or not the
VFI for the professional life of researchers themselves.
Most of these reasons support the view that abandoning the VFI would jeopardize
research integrity, which comprises both ethical and epistemic integrity. Therefore,
these reasons have also an epistemic component. Nonetheless, they were included,
because falling short of epistemic standards is considered detrimental not (only)
because it would weaken scientific knowledge, but because it would infringe ethical-
professional standards. Another well-represented professional reason type is the one
concerned with the autonomy of scientists. On the one hand, it is feared that imposing
the VFI would make scientists incapable of autonomous moral judgements. On the
other, it is feared that were political values to play a role within science, scientists could
refrain from making claims not aligned with those values, regardless of the evidence.
In both cases, scientists would lose the autonomy and freedom to practice their
This review identifies non-epistemic reasons for and against the VFI considered in the
literature. While it cannot be used to settle the academic debate, it offers to RI scholars
an overview of the possible advantages and disadvantages – for society, politics, and
the scientific community – of publicly endorsing the ideal that science should be
“value-free”. We believe this overview be useful for drafting codes of conduct for RI, as
further discussed below.
In this section, we contrast our findings with empirical data and the broader academic
literature in order to highlight their relevance to RI and codes of conduct while
suggesting possible ways to evaluate them.
Implications for research integrity
Codes of conduct for RI, in offering normative guidance to researchers, must per
definition make some assumptions about what ideal scientific research looks like. How
they deal with non-epistemic values is a crucial part of that ideal. Should codes
communicate the norm that scientists should conduct their research in a way that is as
value-free as possible? Or should codes acknowledge the importance of societal and
political values? Which position is assumed by current codes of conduct? Here it is
interesting to note that in the reviewed articles preserving RI is mentioned only as a
reason in favor of the VFI (Ahn, 2020; Douglas, 2008; Resnik & Elliott, 2016, 2019).
However, it should be mentioned that the distinction between “epistemic” and “non-
epistemic” values is not one that codes of conduct operate with. These typically refer to
both epistemic as well as ethical-professional integrity. For instance, the European
Code of Conduct for Research Integrity speaks of the importance of “respect” (ESF-
ALLEA, 2017), and the U.S. National Academies refers to “fairness” as a guiding
principle for scientists (National Academies of Sciences, Engineering, and Medicine
(U.S.), 2017). None of these could pass as a purely epistemic value. Rather, they may
be read as moral values, or moral virtues
Moreover, the importance of virtues for RI has been recently highlighted by the work
of Robert Pennock (Pennock, 2015; Pennock & O’Rourke, 2017). Pennock’s virtue-
ethics approach for the ethics of science by no means denies the importance of
searching for truths in science. Quite the contrary, it implies that both epistemic
soundness and scientists’ virtues are necessary in order for science to flourish
(Pennock, 2019). Thus, from this perspective, the sharp distinction between epistemic
and non-epistemic (in this case, moral) virtues underpinning the VFI does not hold.
This is in line with the work of those who have criticized the distinction, and claimed
that epistemic and non-epistemic values lie on a spectrum (Longino, 1996; Rooney,
. According to both perspectives, whether specific values are appropriate
or inappropriate in science is not determined by their being epistemic or not.
Does this mean that codes of conduct for RI implicitly reject the VFI? Although they
seem incompatible at least with some formulations of it, some of them hint at the VFI
in passing. For example, the abovementioned European code states that “research
draws on the work of the community of researchers and ideally develops independently
of pressure from commissioning parties and from ideological, economic or political
While virtues and values do not – strictly speaking – coincide, they are very much related. For instance,
Rik Peels and colleagues claim that some of the principles of codes of conduct such as honesty “can be
understood generally as a value—some good-making property that actions, studies, people, events, or
instruments can have—or as a virtue, that is, as a moral or intellectual character trait of researchers or
perhaps even teams and organizations.” (Peels et al., 2019, p. 4). In addition to this, the scope of the
values in science debate has been so broad, that Justin Biddle has proposed to talk about “non-epistemic
factors” rather than values (Biddle, 2013). Given these considerations, the virtues discussed by Pennock
can be considered non-epistemic factors. We thank an anonymous reviewer for pointing to this
See Lacey (2017) and Steel (2010) for defenses of this distinction.
interests.”(ESF-ALLEA, 2017, p. 3). It is not difficult to see in this passage the core of
the VFI: even if the influence of political or societal values cannot be avoided, in an
ideal situation research should not be influenced by them. Moreover, this brief passage
matches the expectations of Europeans that economic interests and political agendas
should not unduly influence scientific research (European Commission, 2021). Thus, a
distinction is made between two different types of non-epistemic value: moral-
professional values to be pursued, and economic and political values, which ideally
should be avoided.
More research would be needed to identify which precise ideal or ideals underlie
current codes of conduct—the VFI or other ideal. Nonetheless, it is still an open
question to what extent codes of conduct should openly endorse the role that societal
and/or political values can play in scientific research. What long-term impact would
that have on the integrity of research? For now, we can suggest that the benefits and
downsides of the VFI identified in this article will help future discussion on to what
extent the VFI should be endorsed by future codes.
Implications for public trust
To foster public trust in science is often considered one of the main aims of policies
that foster RI. It is believed that maintaining RI and conveying this state of affairs to
society is likely to increase public trust. However, one can separately raise the question
on how scientific institutions should convey this state of affairs. Would an official
endorsement of the VFI increase public trust in science as claimed by some authors, or
would it decrease it as argued by others? As the relations between society, politics, and
science have increasingly received attention by researchers and institutions, a wide
array of relevant data is now available. These data can be a good starting point for
scientific institutions to evaluate whether endorsing the VFI is a useful measure to
improve public trust.
To start with, does the public currently trust science? And if so, for what reasons? While
scientists tend to believe the public does not trust them (Ceci, 2015; Nature, 2015),
recent surveys return a reassuring picture. Both in the UK and the USA, scientists enjoy
high levels of public trust, following increasing trends (Ipsos MORI, 2020; Pew
Research Center, 2020). Likewise, most Europeans maintain positive views of science
and its impact on society, and attribute to scientists the same qualities (intelligence,
honesty, reliability…) they think they should have (European Commission, 2021).
However, this image changes when it comes to sensitive topics. For instance, despite
their general positive attitude towards scientists and science, 50% of Europeans believe
that scientists cannot be trusted about controversial and technological issues, because
economic and political interests are involved (European Commission, 2021). These
apparently contradictory data show how trust in science is a multidimensional concept:
it is topic-dependent, and moreover, the public tends to have high levels of trust in the
principles and methods of science, but not in scientific institutions (Achterberg et al.,
2017; Aupers, 2012; Huber et al., 2019; Miller, 2004). This seems to support the view
that people trust science when its methods and principles are seen as value-free, but
tend not to trust scientists and scientific institutions when perceived as guided by non-
epistemic interests. At the same time, most Europeans agree that researchers should
take account of the needs of everyone within society, especially in the developing of
new technologies (European Commission, 2021). This is in line with the reasoning of
those who reject the VFI by appealing to the social responsibility of scientists to address
socially significant research.
From this discussion, there does not seem to be a straightforward answer to the
question: does the VFI improve public trust in science? At this point, it is worth
reminding that while fostering public trust may seem a worth pursuing goal in itself,
trust in science should always be contingent (MacCoun, 2015): people should trust
scientists as long as they are trustworthy, i.e., they are reliable and unbiased (Carrier,
2017). To misleadingly present research as value-free and unbiased as a strategy to
increase public trust does not seem a morally acceptable option. However, supporters
of radical transparency as a way to deal with the presence of values in science need a
stronger case. In fact, there is currently not enough evidence to conclude that radical
transparency would remedy the loss of public trust, at least when this is lost due to low
replicability of research (Wingen et al., 2020). Quite the opposite, a preliminary study
suggests that disclosure of value-commitments by researchers has null or negative
effect to the trust of people (Elliott et al., 2017).
Implications for scientific expertise and science-based policy
One of the most powerful aspects of research is that the knowledge it produces can
have an impact on the whole of society. Whether acting as experts or simply
contributing to public knowledge by means of scientific publications, researchers
inform the decision-making process of political bodies around the world. And these
decisions, in turn, have a direct impact on the lives of people. This is a crucial point
where research and society meet and where endorsing or not the VFI could make a real
difference. How should scientific experts behave and communicate? Should they select
research questions according to the (non-epistemic) needs of society? Should they
merely endeavor to communicate the science, or should they tailor their messages for
the needs of society?
This issue is closely related to questions about public trust. For instance, when
questions are general and abstract, interviewees tend to report high levels of trust,
while when they are about specific politicized topics, distrust emerges (MacCoun,
2015). Particularly striking is the fact that in the USA public trust in science has been
growing only among Democrats, while Republicans’ trust in science has stayed stable
and low (Pew Research Center, 2020). People seem to fear that science could be driven
by some political agenda that they do not endorse. Indeed, making sure that scientists,
especially when they are acting in capacity of experts, are honest and reliable, and are
not skewing their findings in favor of one or another party is a crucial component of
RI. Authors appealing to public intervention and democratic principles to argue for the
VFI give voice to this worry. According to them, holding the VFI is a promising way
towards effective and democratic policy-making. However, this view seems to contrast
with 61% of European citizens agreeing that non-scientists should be involved in
research as a way to promote the values of society (European Commission, 2021).
Furthermore, it does not address the scenario where value-free science is unattainable.
A possible way to maintain democratic principles in this scenario is to make sure that
only democratically endorsed values are allowed. If values cannot be removed from
science, to include only democratically shared values seems a reasonable second-best
alternative to maintain RI within a democratic society. How to practically implement
this is far from uncontroversial. Philip Kitcher’s ideal of well-ordered science—
probably the most influential VLI aimed at incorporating democratic values into
science (Kitcher, 2001)—has been criticized mainly for the lack of details that would
make its implementation impossible. For instance, it has been recently argued that it
offers “no identifiable ideal research agenda as a benchmark against which we can
assess the research agenda of real-world science” (Philippi, 2020, p. 374). More
generally, incorporating democratic values in science may come at a cost for
researchers, who may lose autonomy in their choices (Schroeder, 2017).
Charting the disagreement on the VFI
While the reviewed reasons point to the importance of the consequences of endorsing
the VFI or alternative VLIs, where does the weight of opinion in the reviewed literature
stand on this issue? The results in Table 1 clearly indicated that the majority of
reviewed sources reject the VFI (7 versus 11). However, further details suggest
alternative possible interpretations of where the weight of opinion lies.
The first important nuance to bring into this discussion is that those who endorse VLIs
define VLIs in very similar ways to the VFI. For instance, compare Betz’s definition of
[…] the justification of scientific findings should not be based on non-
epistemic (e.g. moral or political) values (Betz, 2013, p. 1)
with Douglas’s proposed ban on the direct role for non-epistemic values:
[…] values should only be used to weigh the importance of uncertainty,
by considering the consequences of error. This role is in contrast to a
more direct role for values, namely that values could be taken as reasons
in themselves to accept or reject an empirical claim (Douglas, 2008, p. 8)
So, while Table 1 lists Douglas 2008 as “not endorsing” the VFI and Betz 2013 as
“endorsing”, one can legitimately ask the question just how much these two resources
differ on the VFI or “VFI-like” VLIs. Both agree that non-epistemic values should not
be used as direct support for accepting or rejecting empirical claims. Hence, one may
consider the meaningful distinction to be that between those who endorse a VLI and
those who do not (instead of those who do and do not endorse the VFI). Then a very
different picture emerges of where the weight of opinion lies: it appears that a majority
endorses VLIs (10 endorsing versus 8 not endorsing). For the interpretative details
behind this categorization, we refer readers to the online supplementary material.
Moreover, even this analysis may overstate the level of disagreement, since many
authors holding different normative positions on the VLIs are nonetheless driven by
similar concerns. Only 4 articles (Ahn, 2020; Bueter, 2015; de Melo-Martin &
Intemann, 2016; Rottschaefer, 2003) of those not endorsing the VFI explicitly claim
that non-epistemic values are epistemically beneficial. These articles claim that at least
in some fields value-judgments are fundamental to accomplish the epistemic aims of
science (Ahn, 2020; de Melo-Martin & Intemann, 2016; Rottschaefer, 2003), and that
pluralism of values would benefit inquiry by providing more perspectives (Bueter,
2015). In this perspective, the weight of opinion shifts even more in favor of some of
the key concerns underlying the VFI: virtually 14 sources endorse not using non-
epistemic values for epistemic ends, whereas 4 do.
This suggests caution when the VFI is rejected. Sometimes it can simply entail the
descriptive claim science is not and cannot be a “view from nowhere” and that is
somehow always influenced by values. In this sense, the rejection of the VFI can be
tantamount to saying it is a fiction. However, such a rejection may, ultimately, not
entail a normatively different stance on the exclusion of certain types of (non-
epistemic) values, for instance concerning the role that political or societal values
should play in the process of scientific research. One may still strongly reject, for
instance, the extreme view that scientific research is only acceptable when it conforms
to a particular political ideology. In this sense, those authors who reject the VFI could
still in principle agree that the VFI is a “useful fiction” in helping to prevent such
approaches to scientific research.
Another source of confusion is that similar concerns (about non-epistemic values) lead
to different conclusions about the VFI. Some authors view maintaining public trust as
a reason for endorsing the VFI; for others, the exact same reason is used to reject the
VFI, in the assumption that science cannot be value-free and that claiming the opposite
would be a form of window-dressing if not blatant deception. Both lines of reasoning
have plausibility, and here empirical research may provide some progress. In the
preceding discussion we used available data on the relations between society, politics,
and science to suggest possible ways to evaluate those reasons; however, it would be
helpful if dedicated studies on the impact of the VFI could be designed and carried out.
In charting where the major non-epistemic reasons for and against the VFI, the present
review is limited by the interpretative decisions inherent to the method of a reasons-
based review. The last section discussed how one can draw different lessons from the
results according to one’s interpretative framework. In this section we would like to
point to some other limitations.
The first is related to the necessity of reading the full text of the articles to extract non-
epistemic reason mentions. This means that some relevant reasons may have been
present in articles excluded on basis of title and abstract. We minimized this risk by
keeping a high number of articles (65) until the last phase of selection, and by retrieving
records via other sources.
Another limitation lies in our consideration of journal articles only. In fact, influential
contributions in this area have appeared in monographs and collections. Nonetheless,
for purposes of this review, namely to offer to scholars working on RI an overview of
what is currently debated among philosophers of science, we did not consider this to
be a fatal limitation, for several reasons. First, were there any major non-epistemic
arguments included in books, we would expect these to be discussed in the journal
literature too. Second, not including a non-epistemic reason presented in a book that
is never mentioned in recent journal articles would be a good (though non-conclusive)
basis to consider that reason less relevant for charting patterns of consensus in the
relevant scholarly community. Third, the relatively low number of non-epistemic
reasons found in the literature, together with our knowledge of the non-journal
literature let us think that we have not missed any major non-epistemic reason.
Finally, questions regarding the value-free ideal of science of course have a long
history, going back at least to Max Weber’s work (Weber, 1949). However, between the
90s and the early 2000s, seminal contributions in social epistemology and feminist
philosophy of science (e.g. Douglas, 2000; Kitcher, 2001; Longino, 1990; Solomon,
2001) revived and reshaped the debate on values in science in such a way that including
articles predating them would return a skewed image of what is currently discussed in
philosophy of science.
Even if science is in fact not value free, does the endorsement of the VFI have beneficial
consequences for scientific research? Or do the benefits of rejecting the ideal outweigh
the benefits of endorsing it? In bringing attention to these questions, our review
charted some of the main non-epistemic reasons for endorsing and rejecting the VFI.
While many authors reject the VFI, many do nonetheless hold that political and societal
values should not determine whether an empirical claim is accepted or rejected. The
review also pointed to the difficulty of taking an unqualified stance on the VFI. The
dangers that non-epistemic values pose to the integrity of research are contingent not
only on the specific phase of inquiry (research, dissemination…) and the specific
science, but also on the kind of values at stake.
To the extent the VFI is a “useful fiction” for the scientific community, it should be
possible to empirically test some of the envisioned consequences for society, politics,
and the professional ethics of research. This would clear up discussion in the
philosophical debate, and at the same time would better inform future decision on
whether or not to include the VFI in future codes of conduct. For instance, if empirical
research confirmed that presenting science as value-free has a positive effect on the
trust of the public, would it be a reason to present science as such? We believe the RI
community can find answers to this and similar questions by engaging critically with
the reasons here identified and more generally by engaging with scholars involved in
the values in science debate.
Our findings and discussion suggest two future research lines where collaboration
between scholars working on values in science and those working on RI could prove
fruitful. The first is the empirical testing of the non-epistemic effects of endorsing the
VFI, in particular as regards public trust and RI. The second would be to investigate
systematically to what extent current codes of conduct for RI endorse the VFI or other
VLIs. Do they commit to any specific philosophical position or do they manage to do
without it? Answers to that question, together with a clearer picture of the effects of
endorsing the VFI (and related VLIs) would help show how codes of conduct for RI
could be redesigned and made more effective.
Supplementary Information The online version contains supplementary material
available at https://doi.org/10.1007/s11948- 022-00427-9.
Author Contributions JA performed the literature search, data analysis, and
drafted the article. KD and HD conceived the original idea for the research, revised the
literature search, the data analysis, and the draft. All authors contributed to the
definition of the research question and the design of the research, and approved the
final version of the article.
Funding This research is part of a project funded by the FWO, Research Foundation
– Flanders (Grant Number: G0D6920N).
Achterberg, P., de Koster, W., & van der Waal, J. (2017). A science confidence gap:
Education, trust in scientific methods, and trust in scientific institutions in the
United States, 2014. Public Understanding of Science, 26(6), 704–720.
Ahn, S. (2020). How non-epistemic values can be epistemically beneficial in scientific
classification. Studies in History and Philosophy of Science Part A, 84, 57–65.
Assarut, R., & Eiamkanchanalai, S. (2015). Consumption Values, Personal
Characteristics and Behavioral Intentions in Mobile Shopping Adoption.
Market-Trziste, 27(1), 21–41.
Aupers, S. (2012). ‘Trust no one’: Modernization, paranoia and conspiracy culture.
European Journal of Communication, 27(1), 22–34.
Betz, G. (2013). In defence of the value free ideal. European Journal for Philosophy of
Science, 3(2), 207–220. https://doi.org/10.1007/s13194-012-0062-x
Biddle, J. (2013). State of the field: Transient underdetermination and values in
science. Studies in History and Philosophy of Science, 44(1), 124–133.
Bright, L. K. (2018). Du Bois’ democratic defence of the value free ideal. Synthese,
195(5), 2227–2245. https://doi.org/10.1007/s11229-017-1333-z
Bueter, A. (2015). The Irreducibility of Value-Freedom to Theory Assessment. Studies
in History and Philosophy of Science, 49, 18–26.
Carrier, M. (2017). Facing the Credibility Crisis of Science: On the Ambivalent Role of
Pluralism in Establishing Relevance and Reliability. Perspectives on Science,
25(4), 439–464. https://doi.org/10.1162/POSC_a_00249
Ceci, C. (2015). Take concepts of chemistry out of the classroom. Nature, 522(7554),
de Melo-Martin, I., & Intemann, K. (2012). Interpreting Evidence: Why Values Can
Matter As Much As Science. Perspectives in Biology and Medicine, 55(1), 59–
de Melo-Martin, I., & Intemann, K. (2016). The Risk of Using Inductive Risk to
Challenge the Value-Free Ideal. Philosophy of Science, 83(4), 500–520.
Douglas, H. (2000). Inductive Risk and Values in Science. Philosophy of Science,
Douglas, H. (2008). The Role of Values in Expert Reasoning. Public Affairs Quarterly,
Douglas, H. (2009). Science, Policy, and the Value-Free Ideal. University of Pittsburgh
Douglas, H. (2013). The Value of Cognitive Values. Philosophy of Science, 80(5), 796–
Drenth, P. J. D. (2006). Responsible Conduct in Research. Science and Engineering
Ethics, 12(1 Special Issue), 13–21.
Elliott, K. C., McCright, A. M., Allen, S., & Dietz, T. (2017). Values in environmental
research: Citizens’ views of scientists who acknowledge values. PLOS ONE,
12(10), e0186049. https://doi.org/10.1371/journal.pone.0186049
ESF-ALLEA, (European Science Foundation and All European Academies). (2017).
The European Code of Conduct for Research Integrity. https://allea.org/wp-
European Commission. (2021). European citizens’ knowledge and attitudes towards
science and technology. https://doi.org/10.2775/303708
Huber, B., Barnidge, M., Gil de Zuniga, H., & Liu, J. (2019). Fostering public trust in
science: The role of social media. Public Understanding of Science, 28(7), 759–
Hudson, R. (2016). Why We Should Not Reject the Value-Free Ideal of Science.
Perspectives on Science, 24(2), 167–191.
Ipsos MORI. (2020). Veracity Index 2020. https://www.ipsos.com/ipsos-mori/en-
John, S. (2015a). Inductive Risk and the Contexts of Communication. Synthese, 192(1),
John, S. (2015b). The example of the IPCC does not vindicate the Value Free Ideal: A
reply to Gregor Betz. European Journal for Philosophy of Science, 5(1), 1–13.
John, S. (2019). Science, Truth and Dictatorship: Wishful Thinking or Wishful
Speaking? Studies in History and Philosophy of Science, 78, 64–72.
Kitcher, P. (2001). Science, Truth, and Democracy (1st ed.). Oxford University
PressNew York. https://doi.org/10.1093/0195145836.001.0001
Koertge, N. (2000). Science, values, and the value of science. Philosophy of Science,
67(3), S45–S57. https://doi.org/10.1086/392808
Koertge, N. (2004). How Might We Put Gender Politics into Science? Philosophy of
Science, 71(5), 868–879. https://doi.org/10.1086/421417
Kuhn, T. S. (1977). Objectivity, Value Judgment, and Theory Choice. In The essential
tension: Selected studies in scientific tradition and change (11. print., pp. 225–
239). Univ. of Chicago Press.
Lacey, H. (2017). Distinguishing Between Cognitive and Social Values. In Current
Controversies in Values and Science. Routledge.
Lekka-Kowalik, A. (2010). Why Science Cannot Be Value-Free: Understanding the
Rationality and Responsibility of Science. Science and Engineering Ethics,
16(1), 33–41. https://doi.org/10.1007/s11948-009-9128-3
Longino, H. E. (1990). Science as Social Knowledge: Values and Objectivity in
Scientific Inquiry. In Science as Social Knowledge. Princeton University Press.
Longino, H. E. (1996). Cognitive and Non-Cognitive Values in Science: Rethinking the
Dichotomy. In L. H. Nelson & J. Nelson (Eds.), Feminism, Science, and the
Philosophy of Science (pp. 39–58). Springer Netherlands.
Ludwig, D. (2016). Ontological Choices and the Value-Free Ideal. Erkenntnis: An
International Journal of Analytic Philosophy, 81(6), 1253–1272.
MacCoun, R. J. (2015). The Epistemic Contract: Fostering an Appropriate Level of
Public Trust in Experts. In B. H. Bornstein & A. J. Tomkins (Eds.), Motivating
Cooperation and Compliance with Authority: The Role of Institutional Trust
(pp. 191–214). Springer International Publishing. https://doi.org/10.1007/978-
Miller, J. D. (2004). Public Understanding of, and Attitudes toward, Scientific
Research: What We Know and What We Need to Know. Public Understanding
of Science, 13(3), 273–294. https://doi.org/10.1177/0963662504044908
National Academies of Sciences, Engineering, and Medicine (U.S.) (Ed.). (2017).
Fostering integrity in research. The National Academies Press.
Nature. (2015). Misplaced faith. Nature, 522(7554), 6–6.
Page, M. J., McKenzie, J. E., Bossuyt, P. M., Boutron, I., Hoffmann, T. C., Mulrow, C.
D., Shamseer, L., Tetzlaff, J. M., Akl, E. A., Brennan, S. E., Chou, R., Glanville,
J., Grimshaw, J. M., Hróbjartsson, A., Lalu, M. M., Li, T., Loder, E. W., Mayo-
Wilson, E., McDonald, S., … Moher, D. (2021). The PRISMA 2020 statement:
An updated guideline for reporting systematic reviews. BMJ, 372, n71.
Peels, R., de Ridder, J., Haven, T., & Bouter, L. (2019). Value pluralism in research
integrity. Research Integrity and Peer Review, 4(1), 18.
Pennock, R. T. (2015). Fostering a Culture of Scientific Integrity: Legalistic vs.
Scientific Virtue-Based Approaches | American Association for the
Advancement of Science. Professional Ethics Report, 2(28), 1–3.
Pennock, R. T. (2019). An instinct for truth: Curiosity and the moral character of
science. The MIT Press.
Pennock, R. T., & O’Rourke, M. (2017). Developing a Scientific Virtue-Based Approach
to Science Ethics Training. Science and Engineering Ethics, 23(1), 243–262.
Pew Research Center. (2020). Trust in Medical Scientists Has Grown in U.S., but
Mainly Among Democrats.
Philippi, C. L. (2020). Well-Ordered Science’s Basic Problem. Philosophy of Science,
87(2), 365–375. https://doi.org/10.1086/707539
Resnik, D. B., & Elliott, K. C. (2016). The Ethical Challenges of Socially Responsible
Science. Accountability in Research Policies and Quality Assurance, 23(1), 31–
Resnik, D. B., & Elliott, K. C. (2019). Value-Entanglement and the Integrity of Scientific
Research. Studies in History and Philosophy of Science, 75, 1–11.
Rooney, P. (1992). On Values in Science: Is the Epistemic/Non-Epistemic Distinction
Useful? PSA: Proceedings of the Biennial Meeting of the Philosophy of Science
Association, 1992, 13–22.
Rooney, P. (2017). The Borderlands Between Epistemic and Non-Epistemic Values. In
Current Controversies in Values and Science. Routledge.
Rottschaefer, W. A. (2003). Assessing the Role of Non-Epistemic Feminist Values in
Scientific Inquiry. Behavior and Philosophy, 31, 225–249.
Rudner, R. (1953). The Scientist Qua Scientist Makes Value Judgments. Philosophy of
Science, 20(1), 1–6.
Schroeder, S. A. (2017). Using Democratic Values in Science: An Objection and
(Partial) Response. Philosophy of Science, 84(5), 1044–1054.
Smith, R. (2005). Medical Journals Are an Extension of the Marketing Arm of
Pharmaceutical Companies. PLOS Medicine, 2(5), e138.
Sofaer, N., & Strech, D. (2012). The Need for Systematic Reviews of Reasons. Bioethics,
26(6), 315–328. https://doi.org/10.1111/j.1467-8519.2011.01858.x
Solomon, M. (2001). Social empiricism. MIT Press.
Steel, D. (2010). Epistemic Values and the Argument from Inductive Risk*. Philosophy
of Science, 77(1), 14–34. https://doi.org/10.1086/650206
Strech, D., & Sofaer, N. (2012). How to write a systematic review of reasons. Journal
of Medical Ethics, 38(2), 121–126. https://doi.org/10.1136/medethics-2011-
Weber, M. (1949). The Methodology of the Social Sciences (E. A. Shils & H. A. Finch,
Trans.; 1st ed.). The Free Press.
Wingen, T., Berkessel, J. B., & Englich, B. (2020). No Replication, No Trust? How Low
Replicability Influences Trust in Psychology. Social Psychological and
Personality Science, 11(4), 454–463.