Given how much liberals bitch about "dog whistling" (i.e. claiming every innocuous thing non-liberals say that they disagree with is racist etc), it is instructive to look at experiments about it:
"The dog-whistle hypothesis, or racial priming theory, rests on the assumption that white Americans in the post-Civil Rights era maintain conflicting and ambivalent attitudes toward racial issues, particularly those pertaining to African Americans... Whites have shown increasing support for racial egalitarianism since the Civil Rights era, including rejection of discrimination and commitment to abstract norms of racial equality ...
In such an environment, racial priming theory (Mendelberg 2001; Valentino et al. 2002) suggests that elite appeals can harness whites’ underlying racial dispositions and prejudice to influence their policy opinions, but only when the racial content of the message remains outside conscious awareness... According to the theory, racial cues must therefore be subtle and implicit, or they will be rejected as racist...
We explored whether dog-whistle effects may be moderated by political ideology. Given scholars’ claims that veiled racial appeals have been a key tactic in Republican Party strategy over the past century (Haney-López 2014; Lamis 1990), we might expect that conservatives are more responsive to appeals reminiscent of familiar partisan rhetoric. Alternatively, we might expect dog-whistle effects to be stronger among political liberals, a group particularly likely to have internalized norms of egalitarianism and colorblindness, thus meeting the assumptions of racial priming theory...
Among participants as whole, we found no evidence that the implicit appeal increases the strength of participants’ racial prejudice on their evaluation of welfare programs, contrary to what racial priming theory would predict...
Instead, we found a three-way interaction between experimental condition, racial resentment, and political ideology suggesting that the dog-whistle effect was driven by liberal participants. Liberals high in racial resentment were less supportive of welfare after reading the implicit racial appeal compared to a message with no racial appeal. Results for explicit messages, however, were more ambiguous... we also did not find evidence that white liberals reject explicit racial appeals"
Beyond welfare, the paper also looks at gun control. While the authors also note that "appeals for stricter firearm regulations historically have not linked these regulations to negative stereotypes of racial minorities", they find once again that "liberals high in racial resentment were more likely to voice support for increased restrictions on gun ownership after reading the implicit racial appeal compared to a message with no racial appeal"; significantly, this happened with the explicit appeal too. Meanwhile, conservatives didn't change their support for it, and moderates decreased their support for it. In other words, conservatives do not just support gun rights for white people. So much for that liberal canard. Indeed, it's liberals that support gun control for minorities - even when this is explicitly racist.
The authors propose, as a likely explanation for why liberals are the ones receptive to "dog whistling":
"norms of colorblindness in liberal political culture impair the development of structural explanations of negative outcomes among black Americans, leaving liberals ill equipped to rebuff messages linking outcomes such as welfare utilization or criminal justice involvement to negative racial stereotypes. While liberal news sources are more likely than conservative news sources to frame responsibility for negative outcomes as resulting from social causes rather than personal problems or inadequacies (Kim, Carvalho, and Davis 2010), strong norms of colorblindness in liberal political culture mean negative outcomes among black Americans as a group are rarely discussed. Rather than tracing these outcomes to structural factors like concentrated poverty or lack of access to dominant institutions, discussion of these issues is often self-censored (Bonilla-Silva and Ashe 2014). Therefore, social and structural explanations of disproportionate incidence of negative outcomes like welfare usage and crime may not be readily accessible to liberals, leaving them vulnerable to rhetorical attempts to link their views about the proper political response to these issues to latent racial stereotypes."
In other words, since liberals are unwilling and/or unable to consider a broad range of explanations for poor outcomes of "minorities" (or even discuss the fact that such outcomes exist), they become more receptive to "dog whistling". This is the price of (self-imposed) ignorance.
Truly, the lady doth protest too much, methinks
As Leonydus Johnson observes: "If racist dog whistles are meant to be heard only by racists, then it is quite interesting how progressives seem to be the only people who can hear them."
More succinctly, "If you keep hearing dog whistles, you must be the dog"
It is significant that even the liberals did not "reject explicit racial appeals".
Yet, it must be noted that the study measured racial resentment using the Symbolic Racism Scale, which is very problematic: those cognizant of history, who believe blacks have agency and generally believe that black are equal would register as high in "racial resentment". Indeed, to be high in "racial resentment" is basically to not be a grievance monger, which would also correlate strongly with pro-welfare attitudes. So unthinkingly accepting the claims that these people are racist is simplistic. The study also supposedly measured "explicit racial prejudice", but there was almost no information on how this was measured, with just one example question: "It is a bad idea for blacks and whites to marry one another."
Of course, claiming that everything you disagree with is a "dog whistle" is a great way to place it outside the Overton Window. The article references Reagan on welfare reform:
Today I’d like to speak to you about a gathering crisis in our society: It’s a family crisis. To some it’s hidden, concealed behind tenement walls or lost in the forgotten streets of our inner cities
Besides revealing that, regardless of the effects on the "intended base", liberals see poor and black people as the same (revealingly repeated more than 2 decades later with Biden equating black kids and poor kids - somehow they don't get called out for this racist belief), if talking about welfare reform is a "dog whistle" and thus "racist", it mean you never need to reform welfare.
Unfortunately, the study only looks at white people, limiting the conclusions about "dog whistles" and racism (for example, if minorities also respond to "dog whistles", that suggests that these may not really be dog whistles).
There're some other studies which purport to find that "dog whistling" is a thing, but they're problematic.
"The Effects of Dog-Whistle Politics on Political Violence" (which I note is not peer-reviewed) finds that the volume of tweets by elected officials with many white nationalist followers correlated with hate crimes. Yet, tweets coded by research assistants as having "explicit anti-minority content" did not correlate with hate crimes; you'd expect "explicit anti-minority content" (however that is coded: for all we know, saying covid came from China would count as "explicit anti-minority content") to be even more effective at motivating white nationalists than dog whistles, but they have no effect. That suggests that this is a spurious correlation - it is more likely that both the volume of tweets from the politicians and the hate crimes were motivated by some other events.
Meanwhile, Race and policing in the 2016 presidential election: Black lives matter, the police, and dog whistle politics finds that "claims of supporting the police were connected to votes for Trump mainly among those with high levels of racial resentment" - but given the problematic way in which "racial resentment" is defined, the conclusion is suspect. Plus, the fact that the study found that there were BLM supporters with high levels of "racial resentment" suggests that there're issues with this measurement (either that or that many "BLM supporters" are engaging in preference falsification). Neither conclusion is flattering for liberals. The study also complicates the claim that "support for the police" is "racist", since "Support for the police was only associated with vote choice among those with high racial resentment. To be clear, many of those low in racial resentment also felt warmly toward the police, but these views were not connected to vote choice". To say nothing of the usual issues with observational studies.
