Inflation is high, and President Biden’s approval is low. News stories are focused on the rising prices at the pump, upsetting voters. But don’t blame the media for the poor performance of the president’s party. Laurel Harbridge-Yong finds that increasing gas prices hurt presidential approval, regardless of media coverage. Eric Merkley finds that media coverage of inflation—and the economy more generally—is more favorable for Democratic presidents than Republicans. The media is hyping short-term negative changes in inflation, but that is normal. And the results for Biden, like other presidents, will be negative.

Guests: Laurel Harbridge-Yong, Northwestern University; Eric Merkley, University of Toronto

Studies: Presidential Approval and Gas Prices” and “Partisan Bias in Economic News Content

Transcript

Matt Grossmann: Inflation hurts presidents, especially gas prices, and it’s not the media’s fault. This week on The Science of Politics, for the Niskanen Center, I’m Matt Grossmann. Inflation is touching 30 year highs, and President Biden’s approval remains low. News stories are focused on the rising prices at the pump upsetting voters. But don’t blame the media for the poor performance of the president’s party.

It turns out inflation hurting the president is a long running pattern, and the media has actually been harder on prior Republican presidents. This week, I talked to Laurel Harbridge-Yong of Northwestern University about her chapter with John Krosnik and Jeffrey Wooldridge presidential approval on gas prices. She finds that increasing gas prices especially hurt presidents, and that pattern does not depend on media coverage. Incessant coverage is not necessary for translating inflation to voters.

I also talked to Eric Merkley of the University of Toronto about his American politics research article, Partisan Bias in Economic News Content. He finds that media coverage of inflation and the economy more generally is actually more favorable for Democratic presidents than Republicans. The media usually hypes short term negative changes, and it’s not just for Biden. Harbridge-Yong explains the long running relationship between gas prices and presidential approval.

Laurel Harbridge-Yong: This is a project where we’re looking at the relationship between gas prices and presidential approval. When we look at data from January, 1976 through July, 2007, we find that even when taking into account other measures of the economy such as employment and inflationary pressures in other areas of the economy, that we still find an effect of gas prices on presidential approval.

We’re also controlling for other factors like rally events that might boost presidential approval and scandals that might decrease presidential approval. What we find is that an increase in gas prices is correlated with a decline in presidential approval, and more importantly, this relationship is not impacted by the amount of media coverage on gas prices at the time suggesting it’s working through people’s own evaluations of what gas prices are doing. For instance, in one of the models we find that a 10 cent increase in gas prices is associated with a .60 percentage point decline in presidential approval.

Matt Grossmann: The relationship is longstanding and not dependent on partisanship, even with increasing polarization.

Laurel Harbridge-Yong: The benefit of looking at 30 years or so of data, which we do here and at the monthly level, which gives us a reasonable amount of power to look for a relationship here, is that we’re not finding a relationship that’s just driven by one president or one time period. We’re looking to see, is this a pattern that’s generally holding over this 30 year period? But the analysis doesn’t break things down by time period, and I think that is certainly a limitation. You could think that things have changed over time.

I think the thing I was thinking most about when you raise this question was whether the increased polarization among the public and the extreme partisan polarization over presidential approval ratings might be changing this a bit. People have noticed that beginning with Bush and then accelerating with President Obama and President Trump and now Biden as well, that presidential approval is highly polarized by party. Co-partisans of the president approve at very high levels regardless of what the president does or what the economy does. Opposing partisans do not approve of the president regardless of what he does or what the economy does, and then maybe it’s independence who are moving the most.

That certainly, particularly on the opposing partisans, there may be a floor effect of thinking about how far approval could really fall. I was looking at Gallup’s data for Biden, and when he first came into office, I think approval among Republicans was 12%. It’s now down to six, so it has dropped, but there wasn’t a lot of room for it to fall. Likewise, the in-party protection side of it, Democrats have dropped as well, but not a lot, from 98% to 90%. It’s really the independents who are the biggest movers in the Gallup data. I think they show that independent move from around 61% approval to 37% approval. That’s a much bigger swing.

When you think about how partisan polarization has played in, it’s certainly possible that there’s a weaker relationship between gas prices or other measures of the economy and presidential approval in more recent years. It’s not something that we looked at, but it’s certainly a possibility with the data and the limitation of what the current analyses can show.

Matt Grossmann: Merkley finds that the media overall is nicer toward Democratic presidents in economic coverage.

Eric Merkley: In that article, I conduct a sentiment analysis of top circulating US print media, so from 1985 to 2013. In the coverage of inflation and unemployment, those two issues specifically. I find generally that the tone of coverage is more positive under Democratic presidents after controlling for objective measures of economic performance. More crucially, and I think the more important contribution here, is that I find that in the short run, the news media becomes more negative with increases in the unemployment and inflation rates, but again, only under Republican presidents.

The big takeaway is there seems to be some systematic pro-Democratic bias, in at least the news media that I sampled in this article, and especially in its responsiveness to negative economic information. Then I have a bunch of supplementary analyses that reviewers push me on to look at the effects of this in economic news more broadly and compare across different media formats, outlet endorsements. I also look at the volume of content as well, and I find similar results as well. So it holds up fairly nicely.

Matt Grossmann: In recent decades, there’s been a lot more economic coverage of unemployment. But the news mostly focuses on the most negative conditions.

Eric Merkley: It’s certainly true that one of them is a top tier issue and the other one not so much, at least over the periods that I studied. Unemployment gets far, far more coverage, and the unemployment rate generally is predictive of a much broader set of economic news that the inflation rate probably isn’t.

Unemployment is certainly a top tier issue. Inflation covered much less readily. Through this period of study, it was a period of consistently low inflation with a couple exceptions, but certainly nothing compared to the 1970s. That’s just a artifact of when the data became available and all that. So there are differences there.

There are also differences in negativities. There’s a lot of scholarly research on negative media bias, that the tendency for journalists to accentuate the negative across the board. A lot of really good systematic evidence of this. I don’t look at this directly for the reason that… The Lexicoder Sentiment Dictionary, the tool I use to evaluate the tone of news articles. It doesn’t have a true neutral point. It’s a relative measure of relative positivity and negativity.

There’s some work by Stuart Soroka in journal politics back in 2012 looking at… Well, the media’s only really responsive to negative changes in economic conditions, so increases in the unemployment rate, for instance, and his focus was on the unemployment rate for that. But he’s also found, by matching Lexicoder scores to human coders to identify where the neutral point is in that scale, he’s able to show that negative news is twice as likely to be covered as positive news on the whole in the US context. Maybe it might be softer in other countries. But there’s definitely a negative news effect. My interest here was, “Okay, we know that there’s this focus on negativity. Does it matter more under Republicans than Democrats?” That was the starting off point for my research.

Matt Grossmann: Merkley finds that on inflation and unemployment, Democrats got a lot more positive coverage and it’s less responsive to real economic data.

Eric Merkley: The differences are sizeable. I would say they’re probably not overwhelmingly so, but subjective interpretations of how big the effects are, it’s going to vary person to person. That’s my interpretation of it. Effectively, the overall effect of having a Democratic president on the tone ranges between .6 and .7 standard deviations for unemployment and inflation coverage. I take that as a reasonably large effect. Not a huge one, but meaningful.

In terms of responsiveness, I find that something like a half point increase in the unemployment rate in the short run would lead to about a standard deviation decrease in tone under Republican administrations, and then there’s no significant effect for Democrats. That’s sizeable. It would be relatively comparable to the effect of having a recession, but, and this a key point to understand about this design that is only in the short run. That effect vanishes after that time point. That’s how the model’s designed. This focus on responsiveness, it’s not a long-lasting effect. It only exists for that quarter. Then, for the next change in the unemployment rate, the media does something different. That’s an important thing to keep in mind.

Matt Grossmann: Both papers use historical data, so let’s dig in to each project. Harbridge-Yong looks over a long time period providing a new analysis of what happened to George W. Bush.

Laurel Harbridge-Yong: I certainly don’t think that the conventional narrative of a boost following 9/11 followed by a regression to the mean combined with people’s unhappiness with Katrina, with the Iraq war, so for…

… with people’s unhappiness with Katrina, with the Iraq War, so forth, all kind of contributing to a decline. But I also think that as we point out in the figure, if you just look at the simple bivariate relationship, there’s also a very strong relationship in that figure between his approval rating and gas prices.

And so I think the benefit of the broader analysis, which moves beyond one president and looks over a 30 year period, is to say, “Is there anything to the story that gas prices might have been contributing to Bush’s approval?” And I think obviously our results don’t separate by president, but I think they’re suggestive that they were part of the story. They weren’t all of the story. In the model we control for both the boost that President Bush would’ve gotten from 9/11. So we have a variable that is an indicator for 9/11 that captures September, October and November of 2001.

We have presidential approval in the prior month, which shows a negative effect, which essentially a regression to the [inaudible 00:11:05]. So presidents who get a boost in approval, then likely decline. Presidents who have a low approval for a scandal or something, likely move back up. And we have measures of the Iraq War and other things that are capturing these other conventional measures. And so our results are saying that even accounting for these other factors, we’re still finding in effective gas prices. So I think the answer is it’s a bit of both. Certainly those things matter and our results suggest they do in terms of where there are significant effects of those control variables, but that on balance, at least across presidencies, gas prices also matter.

Matt Grossmann: She’s also not sure if Biden is being hurt by gas prices alone.

Laurel Harbridge-Yong: I think the place where I would say maybe there’s a little bit of caution in applying what we found is that right now, at least from my understanding, it’s very hard to separate the broader inflationary trends from gas prices. They’ve been moving together from all of the evidence that I’ve been able to see. And so in our paper, we tried argue that there was a unique or additional effect from gas prices overall. And moreover, that this was happening through people’s own recognition of gas prices, not necessarily through media coverage, which would suggest this more sociotropic country-wide wellbeing as opposed to pocketbook wellbeing.

And I think right now the challenge is that these are both moving together. And a lot of the same factors that may be driving inflation at a macroeconomic level are also the same factors that are driving the higher gas prices. And so it’s a lot easier to separate these things if you have… General inflation is not moving up that much, but gas prices are rising because of OPEC decisions or other things like that. I think the fact that they’re a visible indicator to people, I would say probably means they’re still part of the story of Biden’s decline in approval.

Matt Grossmann: She finds that the media does track real gas price change, but that isn’t necessary.

Laurel Harbridge-Yong: Certainly it’s reasonable to think that media coverage of inflation has been increasing and rightly so in terms of covering what’s happening on the ground. I think they’re certainly covering the broader inflation part, this bigger story of supply chain issues that’s been facing us first several months now. And these are certainly telling voters more about how the economy is doing as a whole. So again, pointing to these sociotropic measures.

But again, if we think about what our results were that gas prices mattered and they mattered even without the amount of media coverage suggesting it was people’s own recognition of what gas prices were is that even if the media weren’t talking a lot about inflation right now, our evidence would suggest that people would still be recognizing that gas prices are rising and may still be holding the President accountable for that. So, it’s unlikely that the media is solely to blame for why Biden’s approval is dropping as inflation and gas prices are rising.

Matt Grossmann: The media isn’t always the necessary intermediary.

Laurel Harbridge-Yong: Based on my reading of the past literature, that in general when people have thought about the role of the media in helping people figure out how to use economic indicators in approval of the president is the media is doing one or both of two things. One is they’re providing information, letting people know whether an indicator has gone up or down, if it’s in a bad place or not. And the other is just that by talking about it, they might be priming people to use that indicator in their evaluations of the President. And in the case of gas prices, people don’t really need the media to tell them what the change in gas prices is. So there’s past work that we cite in the paper by [Stephen Salabahare 00:14:48] and colleagues that found that people are significantly more accurate in their perceptions of gas prices than they are of the unemployment rate.

And this accuracy is also related to how often you drive a car, which makes a lot of sense that if you drive a lot and you’re filling your tank up a lot, you’re a lot more attuned to what the gas price is than if you drive more sporadically. And so I think because they’re visible, it really becomes just a question of, is the media providing this priming effect? And in our findings, there was not a significant relationship of media coverage when we interacted with those gas prices.

And so it suggested that that wasn’t happening. Could that maybe be different at a different point in time if the media talks about it differently? Certainly, but just over the time period that we were looking at, and even with a lot of different ways of measuring the extent of media coverage, we couldn’t find a significant relationship there in terms of that interactive effect between gas prices and media coverage.

Matt Grossmann: That’s likely because gas prices are the most visible form of inflation.

Laurel Harbridge-Yong: The main reason I think that gas prices don’t need media coverage is because they’re visible to people, that people get gas regularly. You see it as you drive by the gas station what they’re advertising is the price. You fill up regularly. You know, “Okay, last time I filled up my tank, it cost $40. Now it costs $45.” And so, it’s a repetitive thing that people do and it’s where it’s the one item they’re buying. And so, I don’t know any evidence on this, but you could certain imagine that people right now might see that their grocery bill has gone up, but they don’t necessarily know if the price of milk, versus eggs, versus bread, versus meat, where those things are moving because there’s a lot in their cart and some people may be tracking things specifically, but a lot of people may not. And so gas prices, it’s usually the only thing that you’re buying when you’re stopping at the gas station. So it’s visible, it’s repetitive in terms of doing it, and it’s a single item ticket that you’re buying at once.

Matt Grossmann: [Merkely 00:16:51] agrees that the media may not be as important on some parts of inflation.

Eric Merkley: There is this broader debate about to what degree does the news media actually influence people’s perceptions of the economy? That is because people do have some sort of real world ability to evaluate these changes and certain things like gas prices are very obtrusive. They affect people in their daily lives. And so, media framing, media agenda setting, all that just may not be as important on some dimensions of the economy than others. So I think it’s a very interesting question.

And it’s very hard to causely identify the effect of the media on economic perceptions, considering there is an [inaudible 00:17:36] argument that others have made, and I think there’s compelling, theoretical reasons to expect it, that the media follows the public to some degree as well. So I think it’s super interesting. There might be variation in that obtrusiveness, people’s ability to learn about the economy in their real life. And so, the ability of the media to frame the issue might be considerably less on those issues and gas prices may well be one of those.

Matt Grossmann: He compares Republicans and Democrats at similar levels of economic performance.

Eric Merkley: That’s the counterfactual that we have to engage in when we think about media bias is if this is a Republican president, and I say it as a normal order Republican president, given the time period of my analysis, what would tone be with these conditions? And my analysis suggests that it would be more negative. I think since the rise of Trump, which happened all after the time period I studied, I wonder if there’s some differences for other reasons, like how quickly does the media get knocked off of inflation stories when Trump does something outrageous and they move on to something else? So I wonder how much this extends into the new era, but nonetheless, that’s the counterfactual. So for me, it’s not surprising that such a increase in inflation is highly, highly novel.

It’s highly, highly negative and it taps into these pathologies in journalist reporting that it’s not surprising that there’s a media storm that comes out of this to some degree, taps into all those dimensions. And so, we have to think about the counterfactual. Ultimately partisans are going to always think that the media’s being mean to their side at some level, but we really do have to think through the counterfactual, and that’s why you can’t look at one event and say, “Well, that’s bias.” You have to look over time or find other… some creative research design that allows you to credibly examine that you won’t be able to examine the counterfactual directly, but to sort of get that in that direction.

Matt Grossmann: Bias was concentrated in newspapers, endorsing Democrats.

Eric Merkley: So I look at how the bias varies based on newspaper endorsements. I find that overall bias is stronger among Democratic endorsing papers, but I think an interesting thing is we’ve seen newspaper endorsements turn so dramatically against the Republican…

Newspaper endorsements turned so dramatically against the Republican party now in the Trump era, it seems we lose all leverage over this sort of question using endorsements. So just an interesting thing to note for future research.

Matt Grossmann: Democrats might do better overall in the economy, but this analysis is not picking that up.

Eric Merkley: The changes that I see are not as vulnerable, those sorts of patterns are not as vulnerable to persistent differences in the levels because they operate independently, is often the case. So, that element of the design is less susceptible to those issues, which is why I think that is the main contribution of this paper, because I couldn’t control for everything and so there is certainly a pattern that economies are better under democratic presidents. And so maybe it’s not fully accounted for, but for the short run analyses, I don’t think as the design is as vulnerable to that sort of problem.

Matt Grossmann: He finds that print media does predict public views and broad cast news is less responsive to economic conditions.

Eric Merkley: I focus on print media and print also includes newswires, so the AP, AP’s in my sample. I see them as agenda setters for the broader media environment, therefore deserving of attention, and there’s some evidence in the many, many supplementary analyses that are attached to this article, I find that print and AP is much more strongly predictive of people’s economic perceptions than broadcast media. So there’s some evidence that it’s these outlets their the most important for what I’m interested in, that is influencing economic of perceptions.

Other points that it really didn’t seem like broadcast news was particularly responsive to the unemployment conditions. Now there’s some methodological questions here because [inaudible 00:21:57] does well over large amounts of text where you can average cross a whole article for positive negative words. Broadcast transcripts are shorter, a lot shorter, so there’s a lot more noise in that measure. So I don’t use broadcast for that reason in the main text, but I do look at it in the supplement and find that broadcast news is more positive towards Democrats, controlling for all the things that controlled for. I don’t find any responsiveness bias, but that’s because I don’t find any responsiveness at all to short run changes in the economy.

Matt Grossmann: The media focuses on the real levels of inflation, but also reacts to big negative changes.

Eric Merkley: So it’s important to separate out the short run in the long run here. So I find both effects in my paper, that is there’s short asymmetric, responsiveness, but there’s also kind of an overall bias in the press towards democratic presidents. So, there’s both of those components. And similarly, in terms of how the media responds to the economy, there’s also short run and long run components. So there’s short run changes at a particular time point to an immediate change in inflation and unemployment. And the best evidence suggests that they really only focus on the short run negative. But in the long run, at least for inflation in my paper, it seems like the media is responsive to the level of inflation as well. So Stuart [inaudible 00:23:26] doesn’t find that to be the case for unemployment, but for inflation there does seem to be a long run effect of inflation in the models that I estimate.

So, that means that there is going to be some responsiveness of the economy to the overall level of inflation. So, if inflation does improve over the long run tone should probably match that eventually, just, it’s not going to be responsive to that short run improvement, but it’ll respond over time. So yeah, I do think if this is a story about inflation and it is a story about gas prices, you could see some rebounding. But again, our ability to know what exactly it is behind decreases in Biden’s approval is pretty limited at this stage.

Matt Grossmann: Both projects are returning to longstanding political science questions. Harvard Young says that other work points to gas price effects being about real pocketbook concerns.

Laurel Harbridge-Yong: The recent piece in American politics research that you reference by Kim and Yang, so they find that constituencies with longer average driving times to work are more likely to hold the president accountable for gas prices. And suggest that this is further evidence of people using pocketbook voting, rather those kind of sociotropic or broader economic evaluation. And I think this piece actually does provide a nice linkage between the chapter that we’ve just been discussing, which showed an over all relationship between gas prices and presidential approval. And because of the lack of significance of the media finding also pointed to pocketbook considerations in some of the more recent work. And so the, the Kim and Yang piece takes advantage of the fact that this overall relationship between gas prices and presidential approval isn’t fixed across different groups within the country. So suggesting that people with longer commute times are going to be more sensitive to this.

And so in the recent work, which is a forthcoming paper in LSQ or Legislative Studies Quarterly with Sarah Anderson, Daniel Butler in August [inaudible 00:25:15] is we look at the factors that affect state legislators support for the gas tax. And again, here we find that legislators whose constituencies have longer driving commute times are more likely to oppose an increase to the gas tax, more likely to favor a lower gas tax. And by contrast, those who have longer commutes on public transit as opposed to driving are more supportive. And so we kind of frame this paper in terms of thinking about the implications for legislator’s willingness to support investment in public transit and kind of the double dividends for carbon emissions. So, support for a higher gas tax can reduce driving. If more people take public transit, it reduces driving.

And so when I kind of think about these two different pieces of work, obviously an important distinction is that gas tax policy is clearly in the domain of legislators. So this is something that’s set at the federal level for the federal tax, the state level for the state gas tax. And it has important implications for decarbonization policy, for funding of transportation projects and others, but that the cost can be concentrated among some constituents. And the distinction there then so gas taxes are in the control of legislators, gas prices overall are much less directly tied to the policy decisions of legislators or the president. But I think the Kim and Yang piece highlights that again, the impact of higher gas taxes isn’t felt the same across the country. And so when we think about kind of who might be most responsive to these increases, again, this isn’t in the chapter that we’re talking about here with Craws, Nick and Woodridge, but I think the Kim and Yang piece kind of is a nice extension of that, which says that the impact of gas prices on presidential approval may also matter more where people are having to drive more to work.

Matt Grossmann: This is a return to previous views of how the economy matters in politics.

Laurel Harbridge-Yong: When people first started kind of thinking about the factors that went into presidential approval, there was a bit of assumption that it would be kind of pocketbook considerations, that it would be how individuals were doing economically. And what the research kind of consistently showed in that kind of 1970s, 1980s or so, was that it actually wasn’t pocketbook considerations, it was this broader kind of sociotropic set of views or kind of the wellbeing of the country as a whole. And I think part of that was that people don’t necessarily hold the president accountable for how they’re doing. I think that the kind of American ethos, so kind of individualism and stuff does suggest that many people view that the government is not necessarily responsible for their own economic wellbeing. That there’s a lot of kind of personal choice there, random chance, everything else, but that it’s not just up to the government.

And so it was more about if the economy as a whole is doing well, the president would be rewarded, if the economy as a whole is doing poorly, the president would be penalized. And so I think the evidence on inflation and unemployment certainly trended in that direction. People also tended to need media coverage to kind of know what those national numbers were. You might know that some people were laid off at a plant in your town, but you don’t necessarily know if that’s indicative of a broader trend across the country without kind of some greater attention from the media and kind of information there about what the trends are.

I think in the current environment, inflation has become pretty salient to people given how quickly it’s risen. But I think people probably still need kind of a sense of what it’s doing across the country and not just like, oh, things seem more expensive here. But as we said in the chapter, given the lack of relationship with media attention for gas prices, it was suggestive that this was working through more of a pocketbook relationship. It doesn’t rule out that people were extrapolating from what they saw as gas prices themselves to how this was affecting the country as a whole. But it moves things maybe in the at least keeping pocketbook on the table, where a lot of previous analyses kind of took pocketbook off the table. Certainly the Kim and Yang piece as well suggests that there is an element here that certainly pocketbook that if you’re thinking nationally about it, it shouldn’t matter whether you personally drive a lot and have to fill up your tank of gas a lot or not. And so that certainly suggests that there’s a pocketbook element going on.

Matt Grossmann: Inflation hurts presidents, even accounting for other good economic news.

Laurel Harbridge-Yong: In our model and I think any good model of presidential approval, you’re including kind of multiple indicators. So unemployment has been looking better and better over the course of the Biden administration, so that is accounted for the type of model that we have. And then when we think about kind of the inflationary side, I mean, you’re right economists may have differing views on how problematic it is for the country.

Have differing views on how problematic it is for the country that inflation is going up, but I think from a political perspective, inflation is a challenging thing for a president because the costs that people are facing are going up much higher than their wages are going up. And so what they’re seeing is that their money isn’t going as far, and then obviously the media coverage of it and everything else in terms of kind of hyping this as a problem is certainly going to then potentially hurt presidential approval.

But what the real economic consequences of that are and what it means for the health of the economy is a bigger story for economists to kind of tell to media organizations. But I think that again goes to kind of why, for these kind of broader measures of inflation and unemployment, the media does play a really important role because people’s context and kind of understanding of what a particular number means is explained to them by the media.

And so I do think you’re probably right. If the media talked about inflation differently, it might have a slightly different effect, but again, just given what the raw findings of the paper are and past work as well, they’re not looking at content analysis of what media coverage is saying. Historical record shows that when inflation goes up, presidential approval goes down. So the content of media coverage may not be that important.

Matt Grossmann: And Merkley responds to the old issue of media bias. He says his work might be indicative of broader media bias in favor of Democrats.

Eric Merkley: I can’t think of a reason, at least at the moment, that the patterns observed here would not be indicative of the broader media environment. I don’t think there’s anything about unemployment and inflation together that would mean that, okay, well, we observe pro-Democratic bias here, but it wouldn’t exist on other issues. Maybe there’s a reason out there for that. It could be about that economic conditions generally are more positive for Democrats, so it could be that, but generally speaking, I think the findings will extend, but it’s hard to say I do.

I do replicate this work with a larger tone measure based on two million news articles about the economy as a whole, so not just about unemployment and not just about inflation. So a completely different set of news article with a much larger set of newspapers, and I find the exact same thing. So that makes me a bit at least confident that it extends to coverage of the economy more broadly.

But there’s lots of other issues out there and I think the reason why I focused on these issues is because they are performance issues where there are observable metrics of performance that people generally agree, quibbles about inflation aside, generally agree what’s a positive outcome, what’s a negative outcome. There are all sorts of issues that are directional in nature and this sort of design, so directional meaning there’s an ideological valence to the issue, say like abortion or gun rights, and this sort of design would be completely unable because there’s no… It would be unable to tease out media bias.

Matt Grossmann: But there’s much more research to be done on media bias.

Eric Merkley: I think there’s just a lot of unanswered questions about how much influence does the media have on people’s perceptions of the economy and how that varies over time and across different dimensions of the media. And it could be that the story is very, very complex, and I tend to agree with it.

And this research ultimately, it has its limitations. Presidents are systematically different in performance and you can’t control for everything in any given model. And so there’s lots of unanswered questions still about my design and there’s lots of other academic articles on media bias that find different effects. A lot fine no effects. So this is very much still an open question about to what degree the media is biased against Republicans, how it varies across issues, how it may vary over time, that there might be a case for increasing bias against Republicans, but not because of any sort of journalist bias, but because of education polarization in the United States making the base of news consumers systematically more Democratic in their orientation. And so if journalists and editors are responsive to consumer preferences, they might move to the left accordingly.

So there’s lots of unanswered questions about this field of research and I don’t think my article here is by any means, any stretch of the imagination, kind of an open and shut this is the reality. And there’s so many other biases in the media that are probably much more substantively important. That we have a very commercialized media system means that we have content that is kind of systematically skewed towards novelty and conflict and sensationalism, and all this stuff almost certainly dwarfs the effect of any sort of liberal media bias if it exists.

And there are other, just to plug another article, I have this work with Alan Jacobs, Tim Hicks and Scott Matthews looking at class bias in American news media. And we find that because the news media is responsive to these economic indicators like the unemployment rates and GDP, but these indicators are no longer representative of economic gains for everyone, but more now targeted towards the rich, that there is a class bias in the news media that emerges unintentionally from that sort of process.

Matt Grossmann: There’s a lot more to learn. The Science of Politics is available biweekly from the Niskanen Center and part of the Democracy Group Network. I’m your host, Matt Grossmann. If you liked this discussion, you should check out our previous episodes, How to Change Americans’ Views of Inequality, Will a Good Economy Save Trump, Why Rising Inequality Doesn’t Stimulate Political Action, How the Media Economy Drives Local News, and How the News Media and Social Media Shape American Voters. Thanks to Laurel [inaudible 00:36:11] and Eric Merkley for joining me. Please check out Presidential Approval and Gas Prices and Partisan Bias and Economic News Content, and then listen in next time.

Photo credit: iStock