A study can provide strong evidence of discrimination without telling us which country or ethnic group is the most racist. That ought to be an obvious limit. It is easy to lose sight of it when reliable local findings become part of a much wider political argument.
Matched-application studies are a useful example. Researchers send employers comparable applications with different names and measure the responses. Where the name is the relevant difference, a gap in callbacks is evidence of unequal treatment in the setting being studied.
That matters. It shouldn't be dismissed because the finding is uncomfortable.
What was actually compared?
A study of employers in one country can't rank employers in countries it didn't examine. A study of one kind of ethnic disadvantage can't by itself establish the prevalence of every other kind.
The same caution applies to claims about the Western world as a whole. Research conducted in Western labour markets can establish a problem there. It doesn't establish that the problem is uniquely Western, or that it is less serious everywhere else.
Discrimination involving caste, sect, region or ethnicity elsewhere needs evidence of its own. We shouldn't assume it is absent because it wasn't included in the study we happen to be discussing. Nor should we assume researchers have never investigated it.
Surveys answer a different question
The World Values Survey asks, among other things, about willingness to have neighbours of a different race. Cross-country answers can challenge assumptions about where prejudice exists. They don't measure the same behaviour as an employer deciding whom to interview.
Self-reported attitudes can be influenced by wording, translation and what respondents think is acceptable to admit. Employer studies have limits too. Neither source becomes useless because it can't answer every question, but they can't simply be combined into a definitive league table of racism.
I want the conclusion to fit the evidence. A demonstrated hiring disadvantage is a reason to address that disadvantage. It isn't sufficient evidence for a global judgement about a people. Asking where a study looked and what it measured strengthens the argument rather than dismissing the research.