Science Communication and the Problematic Impact of Descriptive Norms

In the empirical literature on descriptive norms, discussions of experimental results typically focus only on how descriptive norms can be used by policy-makers to bring about positive social change.

“in the empirical literature on descriptive norms, the focus in discussions of experimental results is typically only on how descriptive norms can be used by, for instance, policy-makers to bring about positive social change”

• cites Gerber and Rogers [2009]Not checked yet

  Not checked in the free run.

• cites Duguid and Thomas-Hunt [2015] sourceNot checked yet

  Not checked in the free run.

• cites Walton and Wilson [2018] sourceNot checked yet

  Not checked in the free run.

Meta-analyses of studies on descriptive norms have repeatedly confirmed the reality of the type of self-fulfilment effects discussed above.

“In contrast, meta-analyses of studies on descriptive norms repeatedly confirmed the reality of the type of selffulfilment effects discussed above (for example, Melnyk et al. [2019]).”

• cites Melnyk et al. [2019]Not checked yet

  Not checked in the free run.

Some detrimental effects of descriptive norms have been discussed in the context of public service announcements.

“Some detrimental effects of descriptive norms have been discussed in the context of 'public service announcements'”

• cites Cialdini [2003] sourceNot checked yet

  Not checked in the free run.

Data on descriptive norms and their impact on behaviour and cognition have been known for about 30 years.

“the data on descriptive norms and their impact on behaviour and cognition, too, have been known for about 30 years now (Cialdini et al. [1990]).”

• cites Cialdini et al. [1990] sourceNot checked yet

  Not checked in the free run.

Ethics committee approval requires social scientists to reflect on, and commit to avoiding or mitigating, risks of harm to study participants and the public.

“This requires them to reflect on, and commit to avoiding or mitigating, risks of harm to study participants and the public”

• cites Wassenaar and Mamotte [2012]Not checked yet

  Not checked in the free run.

Neither Fine nor Kourany consider whether the problem of self-fulfilment of scientific claims might be highly general and arise even for prima facie unproblematic claims.

“Neither Fine nor Kourany consider whether the problem of a self-fulfilment of scientific claims might in fact be highly general, potentially arising even for prima facie unproblematic claims”

Roberts et al. found that participants presented with descriptive norms of a group to which they did not belong subsequently tended to disapprove of norm-violating behaviour among members of that group.

“Roberts et al. found that study participants presented with descriptive norms of a group to which the participants themselves did not even belong still subsequently tended to disapprove of norm-violating behaviour among members of that group.”

• cites Roberts et al. [2019] sourceNot checked yet

  Not checked in the free run.

Henrich et al. found that behavioral scientists routinely publish broad claims about human psychology and behavior in top journals based on samples drawn entirely from WEIRD societies.

“Henrich et al. ([2010, p. 61]) found that 'behavioral scientists routinely publish broad claims about human psychology and behavior in the world's top journals based on samples drawn entirely from Western, Educated, Industrialized, Rich, and Democratic (WEIRD) societies.”

• cites Henrich et al. [2010] sourceNot checked yet

  Not checked in the free run.

Henrich et al. showed that these assumptions are not justified.

“Henrich et al. went on to show that these assumptions are not justified.”

• cites Henrich et al. [2010] sourceNot checked yet

  Not checked in the free run.

Authors failing to mention socioeconomic status tended to use more generics than those who did mention it.

“Strikingly, authors failing to mention, for instance, socioeconomic status tended to use more generics than those who did mention it (p.18373).”

• cites DeJesus et al. [2019] sourceNot checked yet

  Not checked in the free run.

There is evidence that social scientific results expressed with social generics tend to be viewed as more generalizable than findings expressed with non-generic language.

“There is evidence that social scientific results expressed with social generics (explicitly unquantified generalizations about kinds of people as such, not individual members or subsets of them) tend to be viewed as more generalizable than findings expressed with non-generic language (DeJesus et al. [2019]).”

• cites DeJesus et al. [2019] sourceNot checked yet

  Not checked in the free run.

Authors showed an overwhelming tendency to treat limited samples as supporting general conclusions by means of universalizing statements.

“Instead, authors showed an overwhelming tendency to treat limited samples as supporting general conclusions, by means of universalizing statements'.”

• cites DeJesus et al. [2019] sourceNot checked yet

  Not checked in the free run.

Scientists often used generics despite omitting information on sample features or having only small or unrepresentative samples as a basis.

“In fact, scientists often used generics despite omitting information on sample features (for example socioeconomic status) or having only small or unrepresentative samples (for example, Western, white, middle-class) as a basis.”

• cites DeJesus et al. [2019] sourceNot checked yet

  Not checked in the free run.

This meta-analysis and individual experimental studies suggest that descriptive norms often exert a significant causal influence on behaviour.

“This meta-analysis and several individual experimental studies that manipulated the scope of descriptive norms to track causal normbehaviour links (Kormos et al. [2015]; von Wagner et al. [2019]) suggest that descriptive norms do often exert a significant causal influence on behaviour.”

• cites Kormos et al. [2015] sourceNot checked yet

  Not checked in the free run.

• cites von Wagner et al. [2019] sourceNot checked yet

  Not checked in the free run.

There is reason to assume that many subjects tend to conform to descriptive norms in many different contexts.

“Given this, there is reason to assume that many subjects tend to conform to descriptive norms in many different contexts.”

• AssumptionThe authors state this as an assumption: In many but not necessarily all domains, descriptive norms are often but not necessarily always likely to have the effects discussed.

There are contexts and domains where descriptive norms have not been effective.

“There are contexts and domains where they have not been effective (Richter et al. [2018]; Paryavi et al. [2019]).”

• cites Richter et al. [2018]Not checked yet

  Not checked in the free run.

• cites Paryavi et al. [2019] sourceNot checked yet

  Not checked in the free run.

The ethical and epistemic problems posed by self-fulfilment effects are broader and more common than previously recognized.

“I argue that the ethical and epistemic problems that such self-fulfilment effects may pose are much broader and more common than hitherto appreciated.”

Staunton et al. found that students told most peers do not eat healthily reported lower intentions to eat healthily than controls.

“students told that most of their peers do not eat healthily (but think that people should do so) subsequently reported significantly lower intentions to eat healthily themselves compared to controls (Staunton et al. [2014]).”

• cites Staunton et al. [2014] sourceNot checked yet

  Not checked in the free run.

Because people conform to descriptive norms, science communication and scientific testimony can frequently produce significant social harm.

“Because of this tendency, science communication may frequently produce significant social harm.”

• Also stated in: § Abstract · plain

  “Due to this tendency, scientific testimony can produce significant social harm.”

One risk of public scientific testimony is that scientific claims cause cognitive or behavioural changes in audiences that help make those claims true.

“One such risk arises when scientific claims cause cognitive or behavioural changes in the audience that contribute to the self-fulfilment of these claims.”

Various other studies outside the laboratory report the same overall findings across a wide range of domains.

“Moreover, there are various other studies on descriptive norms specifically venturing outside the laboratory, reporting the same kind of overall findings from a wide range of different domains (see, for example, Cialdini et al. [1990]; Mollen et al. [2013]).”

• cites Cialdini et al. [1990] sourceNot checked yet

  Not checked in the free run.

• cites Mollen et al. [2013] sourceNot checked yet

  Not checked in the free run.

Gerber and Rogers found that descriptive norms emphasizing low expected voter turnout elicited less motivation to vote than messages emphasizing high turnout.

“presenting subjects with statements conveying descriptive norms emphasizing low (expected) voter turnout (in an election) elicited less motivation to vote than messages emphasizing high turnout (Gerber and Rogers [2009]).”

• cites Gerber and Rogers [2009]Not checked yet

  Not checked in the free run.

Sparkman and Walton found that US lunch-goers informed most people in the US consume high levels of meat ordered more meaty meals than controls.

“Similarly, US lunch-goers informed that most people in the US consume high levels of meat tended to order more meaty lunch meals than controls (Sparkman and Walton [2017])”

• cites Sparkman and Walton [2017] sourceNot checked yet

  Not checked in the free run.

A wide range of studies suggests that many people tend to conform to descriptive norms.

“A wide range of different studies suggests that many people tend to conform to descriptive norms.”

The potential harm of public scientific testimony is more closely linked to claims about negative features than to claims about positive features.

“the potential harm of PST is (as will become clearer below) more closely linked to claims about negative than positive features.”

• cites Peters [2022] sourceNot checked yet

  Not checked in the free run.

Many of the reviewed experiments were field studies with high ecological validity.

“Notice that many of the experiments just mentioned were field studies, that is, they were conducted in naturalistic contexts ensuring high ecological validity (see Schultz et al. [2007]; Goldstein et al. [2008]; Gerber and Rogers [2009]; Hallsworth et al. [2017]; Sparkman and Walton [2017]).”

• cites Schultz et al. [2007] sourceNot checked yet

  Not checked in the free run.

• cites Goldstein et al. [2008] sourceNot checked yet

  Not checked in the free run.

• cites Gerber and Rogers [2009]Not checked yet

  Not checked in the free run.

Mortensen et al. found that students told only a minority of peers engaged in water conservation increased their water use compared with controls.

“students told that only a minority of their peers engaged in water conservation subsequently increased the amount of water they used compared to controls (Mortensen et al. [2019]).”

• cites Mortensen et al. [2019] sourceNot checked yet

  Not checked in the free run.

Subjects frequently form perceptions of descriptive norms by inferring them from observing others, their comments, and the media, not only from direct summary information.

“Indeed, or other research found that subjects frequently form perceptions of descriptive norms by inferring these norms (often incorrectly) from simply observing others, their comments, and the media (for example, advertisements suggesting pervasiveness of smoking, or drinking; Borsari and Carey [2003]; Nan and Zhao [2016]; Liu and Shi [2019]).”

• cites Borsari and Carey [2003] sourceNot checked yet

  Not checked in the free run.

• cites Nan and Zhao [2016] sourceNot checked yet

  Not checked in the free run.

• cites Liu and Shi [2019] sourceNot checked yet

  Not checked in the free run.

Members of the public are likely to take public scientific testimony such as claims 1-10 to indicate descriptive norms.

“Since, in doing so, such generalizations convey descriptive norms, members of the public are likely to take PST, such as claims 1-10 to indicate descriptive norms.”

Rivis and Sheeran assessed 14 studies with a total sample of 5,810 and found strong evidence for the predictive validity of descriptive norms in intention formation.

“Rivis and Sheeran ([2003], p. 228) assessed fourteen studies (total sample size N 5 5,810) covering a wide range of behavioural domains, and they found 'strong evidence in support of the predictive validity of descriptive norms' in 'intention formation'.”

• cites Rivis and Sheeran [2003] sourceNot checked yet

  Not checked in the free run.

Because of their broad and often ambiguous scope, the ten claims are likely to be interpreted by the public as conveying what features are common among certain groups.

“But given their unconstrained scope, they are likely to be interpreted thus by the public.That is, they are likely to convey (perceptions of) what features are common among certain groups of people.”

These problems often arise because many people tend to conform to descriptive norms.

“Moreover, these problems are often due to a specific psychological phenomenon that has been neglected in the research on science communication: many people tend to conform to 'descriptive norms', that is, norms capturing (perceptions of ) what others commonly do, think, or feel.”

• Also stated in: § Abstract · hedged

  “(2) The reason for this is that self-fulfilment effects may often arise because of a common psychological phenomenon that has been neglected in the work on scientific testimony, namely, that many people tend to conform to 'descriptive norms'-norms capturing (perceptions of ) what others commonly do, think, or feel (Prentice [2007]).”

• cites Prentice [2007]Not checked yet

  Not checked in the free run.

DeJesus et al. analysed 1,149 psychology articles from 2015-16 and found that generics were ubiquitously used to convey results about a wide range of categories and constructs.

“DeJesus et al. ([2019], p. 18370) analysed 1,149 psychology articles published during 2015-16 and found that 'generics were ubiquitously used to convey results' in claims about a 'wide range of categories and constructs: People, women, children, adult people with schizophrenia', and so on.”

• cites DeJesus et al. [2019] sourceNot checked yet

  Not checked in the free run.

Public scientific testimony often carries ethical and epistemic risks.

“When scientists or science reporters communicate research results to the public, this often involves ethical and epistemic risks.”

Public scientific testimony, including the ten listed claims, is likely to convey descriptive norms.

“the key point to note for present purposes is that PST, including the ten claims listed above, is likely to convey descriptive norms.”

• Also stated in: § 4. Revisiting Public Scientific Testimony · plain

  “The data just mentioned are relevant for the normative theorizing on PST because, as noted, PST frequently communicates descriptive norms.”

The ethical and epistemic problems with self-fulfilment effects arise for a wide range of ordinary public scientific generalizations about negative features of individuals, groups, society, and social structures.

“(1) The ethical and epistemic problems with self-fulfilment effects of public scientific testimony are much broader and more common than hitherto appreciated, arising with respect to a wide range of ordinary public scientific generalizations about negative features of individuals, groups, society, and social structures.”

Meta-analyses of experiments on descriptive norms confirm the reality of the phenomenon.

“when it comes to the effects of the kind at issue here, meta-analyses of experiments on descriptive norms confirm the reality of the phenomenon.”

Generics are pervasive in many social scientific publications.

“Generics are pervasive in many social scientific publications.”

• cites DeJesus et al. [2019] sourceNot checked yet

  Not checked in the free run.

Duguid and Thomas-Hunt found that individuals informed the vast majority harbour stereotypes showed more stereotyping than those told few are biased.

“individuals informed that the vast majority of people harbour stereotypes that bias their social evaluations subsequently showed themselves more stereotyping in their own social evaluations than those told that only very few people are biased by their stereotypes (and controls;Duguid and Thomas-Hunt [2015]).”

• cites Duguid and Thomas-Hunt [2015] sourceNot checked yet

  Not checked in the free run.

Goldstein et al. found that hotel guests informed most people reuse towels were significantly more likely to reuse towels than guests given environmental reasons.

“In another study, hotel guests informed that most people reuse their hotel towels were subsequently significantly more likely to do so than guests given environmental reasons for that action (Goldstein et al. [2008]).”

• cites Goldstein et al. [2008] sourceNot checked yet

  Not checked in the free run.

Most of the public scientific testimony and descriptive norms mentioned above are unlikely to elicit the specific self-related anxiety found in stereotype threat, because they do not invoke under-performance.

“most of the PST and descriptive norms mentioned above (for example, 'men resist green behavior as unmanly', 'Americans eat too much processed meat') are unlikely to elicit the specific type of self-related anxiety found in stereotype threat, for they do not invoke any kind of under-performance to begin with.”

A likely self-fulfilment effect is decreased support for climate policies among conservatives.

“These effects include (1) reluctance towards green behaviour among men, (2) increased meat consumption among Americans, a (3) proliferation of falsehoods among social media users, (4) low US voter turnout,foot_10 (5) bias against women, (6) affective polarization among Republicans and Democrats,foot_11 (7) acceptance of fake news among 'smart' people, (8) decreased support for climate policies”

Even when recipients of public scientific testimony are not themselves mentioned in the testimony, the claims are still likely to incline them to expect members of the mentioned groups to act in ways conforming to the descriptive norm.

“even when it comes to recipients of PST who are not themselves mentioned in the PST, the claims are still likely to incline the recipients to expect members of groups who are mentioned in the PST to act in ways conforming to the descriptive norm.”

• cites Roberts et al. [2019] sourceNot checked yet

  Not checked in the free run.

A likely self-fulfilment effect is proliferation of falsehoods among social media users.

“These effects include (1) reluctance towards green behaviour among men, (2) increased meat consumption among Americans, a (3) proliferation of falsehoods among social media users, (4) low US voter turnout,foot_10 (5) bias against women, (6) affective polarization among Republicans and Democrats,foot_11 (7) acceptance of fake news among 'smart' people, (8) decreased support for climate policies”

Even if one doubts that the reviewed studies generalise to the ten claims, the preceding discussion directly supports that in some cases public scientific testimony leads to negative, descriptive-norm-related self-fulfilment effects.

“So even if one is sceptical about generalizing from the reviewed studies to the ten claims, the preceding discussion already directly supports the claim that in some cases PST leads to negative, descriptive norm related self-fulfilment effects.”

A likely self-fulfilment effect is affective polarization among Republicans and Democrats.

“These effects include (1) reluctance towards green behaviour among men, (2) increased meat consumption among Americans, a (3) proliferation of falsehoods among social media users, (4) low US voter turnout,foot_10 (5) bias against women, (6) affective polarization among Republicans and Democrats,foot_11 (7) acceptance of fake news among 'smart' people, (8) decreased support for climate policies”

In stereotype-threat cases, the underlying psychological mechanism is distinct from the mechanism highlighted in this paper.

“However, in these cases, the underlying psychological mechanism is distinct from the one that I highlighted here.”

• cites Bicchieri [2006]Not checked yet

  Not checked in the free run.

Some kinds of descriptive norms, such as dynamic norms or don't-versus do-norms, have been found to be more powerful than others.

“And some kinds of descriptive norms (for example, dynamic norms, or don't-versus do-norms) have been found to be more powerful than others (Sparkman and Walton [2017]; Bergquist and Nilsson [2019]; Mortensen et al. [2019]).”

• cites Sparkman and Walton [2017] sourceNot checked yet

  Not checked in the free run.

• cites Bergquist and Nilsson [2019] sourceNot checked yet

  Not checked in the free run.

• cites Mortensen et al. [2019] sourceNot checked yet

  Not checked in the free run.

The ten claims indicate to audiences that the described cognitive or behavioural features are common among people.

“But they also explicitly or indirectly (for example, claim 10) indicate to an audience that the cognitive and/or behavioural features described in the claims are common among people.”

The empirical basis of the argument introduced in this paper is more robust than that of Fine's and Kourany's arguments.

“Moreover, the empirical basis of the argument I introduce above is more robust than that of Fine's and Kourany's arguments.”

Scientists are not under an obligation to consider all possible harms of their professional actions.

“However, scientists are not under an obligation to consider all possible harms of their professional actions.”

• cites Douglas [2009]Not checked yet

  Not checked in the free run.

There is an inductive basis for assuming that ethical and epistemic problems are likely to arise with respect to a wide range of public scientific testimony.

“Since, as noted, various different studies found the same phenomenon with different property ascriptions in diverse domains, there is an inductive basis for assuming that the ethical (behaviour-related) and epistemic (truth-, belief-, or knowledge-related) problems that this raises are likely to arise with respect to a wide range of PST.”

The study participants were in much the same situation as people receiving public scientific testimony in everyday life.

“The study participants were thus in much the same situation as people are when they receive PST in everyday life via TV, online, newspapers, and so on.”

Some public scientific testimony might trigger stereotype threat and thereby cause pernicious self-fulfilment effects.

“As Fine and Kourany rightly emphasize, some PST might trigger stereotype threat and so cause pernicious self-fulfilment effects.”

• cites Fine [2012]Not checked yet

  Not checked in the free run.

• cites Kourany [2016] sourceNot checked yet

  Not checked in the free run.

A likely self-fulfilment effect is reluctance towards green behaviour among men.

“These effects include (1) reluctance towards green behaviour among men, (2) increased meat consumption among Americans, a (3) proliferation of falsehoods among social media users, (4) low US voter turnout,foot_10 (5) bias against women, (6) affective polarization among Republicans and Democrats,foot_11 (7) acceptance of fake news among 'smart' people, (8) decreased support for climate policies”

The descriptive-norm effects were in many cases harmful.

“Importantly, the descriptive-norm effects were in many cases harmful.”

Many experiments on descriptive norms can themselves be seen as instances of public scientific testimony because they involve public claims by scientific or authoritative sources about negative features.

“many of the experiments on descriptive norms discussed earlier can in fact be viewed as involving actual instances of PST on their own: they involve public (descriptive-norms conveying) claims by scientific or otherwise authoritative sources about negative features”

• cites Schultz et al. [2007] sourceNot checked yet

  Not checked in the free run.

• cites Goldstein et al. [2008] sourceNot checked yet

  Not checked in the free run.

• cites Gerber and Rogers [2009]Not checked yet

  Not checked in the free run.

A likely self-fulfilment effect is bias against women.

“These effects include (1) reluctance towards green behaviour among men, (2) increased meat consumption among Americans, a (3) proliferation of falsehoods among social media users, (4) low US voter turnout,foot_10 (5) bias against women, (6) affective polarization among Republicans and Democrats,foot_11 (7) acceptance of fake news among 'smart' people, (8) decreased support for climate policies”

A likely self-fulfilment effect is increased meat consumption among Americans.

“These effects include (1) reluctance towards green behaviour among men, (2) increased meat consumption among Americans, a (3) proliferation of falsehoods among social media users, (4) low US voter turnout,foot_10 (5) bias against women, (6) affective polarization among Republicans and Democrats,foot_11 (7) acceptance of fake news among 'smart' people, (8) decreased support for climate policies”

A likely self-fulfilment effect is low US voter turnout.

“These effects include (1) reluctance towards green behaviour among men, (2) increased meat consumption among Americans, a (3) proliferation of falsehoods among social media users, (4) low US voter turnout,foot_10 (5) bias against women, (6) affective polarization among Republicans and Democrats,foot_11 (7) acceptance of fake news among 'smart' people, (8) decreased support for climate policies”

Even very modest self-fulfilment effects in individual cases might accumulate to significant overall social harm.

“This means that even only very modest selffulfilment effects in individual cases might still accumulate to significant overall social harm.”

Given the data on descriptive norms, there is reason to believe the ten claims are likely to contribute to self-fulfilment effects among the public receiving them and belonging to the demographic groups concerned.

“Since that is so, given the data on the directive impact of descriptive norms just reviewed, there is reason to believe that these ten claims are also likely to contribute to self-fulfilment effects among the public receiving these claims and belonging to the demographic groups concerned.”

• AssumptionRests on an assumption: The public receiving the ten claims includes members of the demographic groups concerned.

A likely self-fulfilment effect is acceptance of fake news among 'smart' people.

“These effects include (1) reluctance towards green behaviour among men, (2) increased meat consumption among Americans, a (3) proliferation of falsehoods among social media users, (4) low US voter turnout,foot_10 (5) bias against women, (6) affective polarization among Republicans and Democrats,foot_11 (7) acceptance of fake news among 'smart' people, (8) decreased support for climate policies”

Some of the reviewed studies can be viewed as actual cases of public scientific testimony communicating descriptive norms.

“In fact, some of the studies just mentioned can be viewed as involving actual cases of PST doing precisely that.”

If a particular empirical claim has been experimentally supported with, for example, a WEIRD sample only, then even though using a generic might not necessarily involve making a false claim, it would still be inaccurate in that it leaves the social reference class of the claim more indeterminate than the experimental data warrant.

“Because if a particular empirical claim has been experimentally supported with, for example, a WEIRD sample only, then even though using a generic might not necessarily involve making a false claim, it would still be inaccurate in that this leaves the social reference class of the claim more indeterminate than the experimental data warrant.”

There is reason to believe that key aspects of the problem Henrich et al. highlighted still persist today.

“While ten years have passed since Henrich et al.'s article, there is reason to believe that key aspects of the problem that the article highlighted still persist today.”

• cites Simons et al. [2017] sourceNot checked yet

  Not checked in the free run.

The easier it is for subjects to generalize social scientific claims to most members of groups, the higher the likelihood that they will construe these claims as conveying descriptive norms.

“The guiding thought is that the easier it is for subjects to generalize social scientific claims to most members of groups, the higher the likelihood that they will construe these claims as conveying descriptive norms.”

Norms phrased as 'don't' are more effective at inducing conformity than norms phrased as 'do'.

“there is evidence that don't-descriptive norms are in fact more powerful in eliciting conformity than do-descriptive norms (seeBergquist and Nilsson [2019]).”

• cites Bergquist and Nilsson [2019] sourceNot checked yet

  Not checked in the free run.

A large amount of current PST contains overgeneralizations.

“currently much PST involves over-generalizations (Simons et al. [2017];DeJesus et al. [2019])”

• cites Simons et al. [2017] sourceNot checked yet

  Not checked in the free run.

• cites DeJesus et al. [2019] sourceNot checked yet

  Not checked in the free run.

Claims from social science that describe negative traits of people, groups, societies, or social structures can bring about mental and behavioural shifts that help make those claims come true.

“social scientific claims about negative features of individuals, groups, society, or social structures can cause cognitive and behavioural changes that contribute to the self-fulfilment of these claims.”

Some relevant effects may also come from natural science generalizations, yet the social psychological mechanisms behind them are probably different.

“Some of the effects relevant here might also arise from natural scientific generalizations (for example, that endocrine disrupters cause infertility).However, the underlying social psychological processes are likely to be distinct; see also(Turnwald et al. [2019]).”

• cites Turnwald et al. [2019] sourceNot checked yet

  Not checked in the free run.

The PST generality constraint does not remove every possible harmful self-fulfilling effect of PST.

“following the PST generality constraint does not eliminate all possible detrimental self-fulfilment effects of PST.”

The preceding sections suggest that these effects are likely to arise in a wide range of cases.

“Still, the preceding sections suggest that these effects are likely to arise in a wide range of cases.”

The public may be very diverse in its individuals and groups.

“it might be highly heterogeneous with respect to the individuals and groups it contains; see(O'Connor and Joffe [2014]).”

• cites O'Connor and Joffe [2014]Not checked yet

  Not checked in the free run.

The impact of PST partly depends on the identity of its audience.

“the effects of PST depend partly on who the PST audience is (for example, different political orientation might lead to different trust assignments and information uptake;Nisbet et al. [2015]).”

• cites Nisbet et al. [2015] sourceNot checked yet

  Not checked in the free run.

For several of the ten earlier claims, ethical, epistemic, or feasibility worries probably do not outweigh the advantages of restating them with explicit quantifiers and reference to specific samples or narrow populations.

“With respect to several of the ten claims discussed earlier, there are unlikely to be ethical, epistemic, or feasibility concerns overriding the benefits of rephrasing these claims with explicit quantifications and relativization (including past tense reference) to specific samples or narrow target populations.”

PST can be received and spread by people other than its intended audience.

“it may be picked up and disseminated beyond those for whom it is intended (for example, by other individuals who believe they understand the study but lack the expertise to understand the caveats and limitations).”

The public communication of information that policymakers need to implement a mask mandate need not necessarily reduce mask wearing among the public.

“If so, the public communication of that information need not necessarily reduce mask wearing among the public.”

This ambiguity or openness to different interpretations is itself in tension with scientific accuracy.

“This ambiguity or openness to different interpretations itself is in tension with scientific accuracy.”

Even if PST uses explicitly quantified claims, it may still frequently concern a majority of people.

“even when it involves explicitly quantified claims, PST might often pertain to a majority of individuals.”

These problematic self-fulfilling effects probably occur often.

“I provided reasons to believe that this is likely to happen frequently”

A wider descriptive norm produces a stronger impact on people's intention to act.

“the broader the scope of the descriptive norm (for example, 50% versus 80%), the stronger the effect on intention to act(Kormos et al. [2015];von Wagner et al. [2019]).”

• cites Kormos et al. [2015] sourceNot checked yet

  Not checked in the free run.

• cites von Wagner et al. [2019] sourceNot checked yet

  Not checked in the free run.

It inaccurately suggests generalizability or 'projectability' where the data may not support it.

“It inaccurately suggests generalizability or 'projectability' (Munton [2019]) where the data may not support it.”

• cites Munton [2019] sourceNot checked yet

  Not checked in the free run.

When PST concerns a majority of people, the danger of problems linked to descriptive norms remains.

“In these cases, the risk of problematic effects related to descriptive norms will persist.”

The triggers and immediate social settings that prompt conformity to norms vary with the content of the descriptive norm.

“Just as in the studies discussed above, they will differ depending on the content of the descriptive norm”

There could be cases where the negative consequences of PST-based self-fulfilment effects are outweighed by the benefits of the public scientific testimony.

“there could be cases where the negative consequences of PST-based self-fulfilment effects are outweighed by the PST's benefits.”

These problematic self-fulfilling effects are connected to the widespread human inclination to follow descriptive norms.

“such problematic self-fulfilment effects are tied to many people's tendency to conform to descriptive norms.”

Sometimes the possible effects of PST are difficult to predict.

“Thus, in some cases, the potential effects of PST are not easy to determine.”

Taken together, the observations of Henrich et al., Simons et al., and DeJesus et al. suggest that many generics and other broad generalizations currently found in public scientific testimony in freely accessible scientific articles can be replaced with claims involving quantifiers other than 'all' or 'most' in result reports.

“Taken together, Henrich et al.'s, Simons et al.'s, and DeJesus et al.'s observations suggest that many generics and other broad generalizations currently found in PST in scientific articles that are often freely accessible online can be replaced with claims involving quantifiers other than 'all' or 'most' in the result reports.”

• cites Henrich et al. [2010] sourceNot checked yet

  Not checked in the free run.

• cites Simons et al. [2017] sourceNot checked yet

  Not checked in the free run.

• cites DeJesus et al. [2019] sourceNot checked yet

  Not checked in the free run.

Broad, generic generalizations formed on the basis of only small samples are not necessarily inaccurate.

“Notice that broad, generic generalizations formed on the basis of only small samples are not necessarily inaccurate.”

Since generics may be interpreted as referring to 'some', 'many', 'most', or 'all' Ps, generics in public scientific testimony involve an ambiguity that may lead different people to different interpretations depending on their expertise concerning P.

“Since they may be interpreted as referring to 'some', 'many', 'most', or 'all' Ps, generics in PST involve an ambiguity that may lead different people (for example, scientific experts versus non-experts) to different interpretations depending on their expertise concerning P.”

In some situations, the pull to follow descriptive norms may outweigh people's motivated reasoning that favors their political in-group.

“Recent studies suggest, however, that people's tendency to follow descriptive norms might in some cases even be stronger than their motivated cognition in favour of (for example) their political in-group:Pryor et al. ([2019], p. 1) found that 'counter to self-categorization theory's prediction', informing 'participants that an action was unpopular amongst people they did not [politically] identify”

• cites Pryor et al. [2019]Not checked yet

  Not checked in the free run.