Oct 15, 2026 · 37 min · 10 segments
In this episode of UCL Uncovering Politics, our new co-host **Dr Julia de Romémont** explores how elected policymakers approach decisions about public service delivery, and whether their preferences…
Julia de RomémontHostFor those of you who are very, more interested in kind of research approaches, right, like, uh, we tend to defer in, like from paper to paper or from project to project how strong our expectations are about what we might find, and what Ellie just described is where you have, like, no strong priors about what might, you might observe, and so you have competing hypotheses, and you try to discriminate between which one fits more with the data, and that's also a way to kinda see, like, okay, we might think about all of these ways it could work.
Sometimes some people have more strong priors, and we have more reason to believe it's definitely gonna go this way, and then they'll just have one hypothesis.
Ch- following up on that, so how do you go about actually testing those hypotheses? You run what is called a survey experiment.
So one of the co-authors on this paper, Nathan Lea, the other co-author is Paolo Belardinelli, run, runs Civic Pulse-
... which is an NGO based in the United States which offers a really fantastic resource, which is also quite rare to come by-
... which is that essentially they have worked for a long time with a sample of bureaucrats, local bureaucrats, and also at, at various levels in the US, and as such, can give you access to what is at least hopefully a relatively representative sample-
... of bureaucrats from across the United States, which is very difficult to get hold of.
In lots of countries, you have to essentially sort of field your own survey.
So that means reaching out to civil servants, to bureaucrats, and asking them to complete your survey-
... which not to get too much into the weeds here, but can have pretty sort of negative effects on your response rates-
... meaning that the degree to which the respondents who respond to your survey are kind of said to be representative of the, the wider bureaucracy.
For those of you who are very, more interested in kind of research approaches, right, like, uh, we tend to defer in, like from paper to paper or from project to project how strong our expectations are about what we might find, and what Ellie just described is where you have, like, no strong priors about what might, you might observe, and so you have competing hypotheses, and you try to discriminate between which one fits more with the data, and that's also a way to kinda see, like, okay, we might think about all of these ways it could work.
Sometimes some people have more strong priors, and we have more reason to believe it's definitely gonna go this way, and then they'll just have one hypothesis.
Ch- following up on that, so how do you go about actually testing those hypotheses? You run what is called a survey experiment.
So one of the co-authors on this paper, Nathan Lea, the other co-author is Paolo Belardinelli, run, runs Civic Pulse-
... which is an NGO based in the United States which offers a really fantastic resource, which is also quite rare to come by-
... which is that essentially they have worked for a long time with a sample of bureaucrats, local bureaucrats, and also at, at various levels in the US, and as such, can give you access to what is at least hopefully a relatively representative sample-
... of bureaucrats from across the United States, which is very difficult to get hold of.
In lots of countries, you have to essentially sort of field your own survey.
So that means reaching out to civil servants, to bureaucrats, and asking them to complete your survey-
... which not to get too much into the weeds here, but can have pretty sort of negative effects on your response rates-
... meaning that the degree to which the respondents who respond to your survey are kind of said to be representative of the, the wider bureaucracy.
The rest of this transcript — segmented and speaker-labeled, so you land on the exact moment something was said
Search every transcript — by keyword, by phrase, or by meaning, across every show Radar indexes
Trends — what is surging across podcasts, measured against its own baseline
Alerts — when a name you follow appears in a newly indexed episode
No account is needed to search Radar.
