Aren’t incumbency and fundraising downstream of ideology and thus should not be adjusted for if you want to estimate the causal effect of ideology on electoral success?
Yes, adjusting for incumbency and fundraising will induce post-treatment bias and will give you an estimate of the causal effect of moderation that does NOT go through incumbency or fundraising. Moderation is also measured imperfectly with error, so there will also be attenuation bias towards zero.
Not at all. This is about how much each candidate raised individually. And as the system works today, fundraising ability is part of how “good” a candidate is, so controlling for fundraising is methodically wrong. It’s saying “moderate candidates would do worse if they couldn’t have raised as much money as they did”, which may be true but it’s imagining a different system than the one currently in place, and in fact they did raise this money. In more technical terms, one cannot adjust for a variable that is affected by ideology if one wants to estimate the casual effect of ideology on electoral success
All candidates do worse when there is less money to spend, not just moderates. I think you are suggesting that candidate moderation causes donations to increase, so controlling for donations erases some of the effects of moderation. Do you have any evidence that this is the case? In general election races, do moderate candidates get more donations than other general elections with progressive candidates? Hypothetically, would progressive candidates in those same races get less money than the moderate candidates did? Or are other variables more important, like how competitive the race is, whether the candidate is an incumbent with a base of support, if the candidate has been endorsed by other popular politicians, how many out-of-state donations there are, etc.? Primary donations certainly could be based on progressive vs moderate politics, but it is not clear to me that the same is true in the general election.
You also suggest that candidate moderation led to incumbency in the 2024 election, which would be difficult to prove given the research referenced in the previous post that candidate moderation has not led to increased electability for Democrats.
"one cannot adjust for a variable that is affected by ideology if one wants to estimate the casual effect of ideology on electoral success"
Yes, you absolutely can, at least in a statistical sense. In a regression model, you can include ideology, fundraising, and also an interaction term, or ideology x fundraising, and determine whether any or all are significant variables.
I don't know whether that can be done with any of the data discussed here, but in the hypothetical case, "adjusting for a variable" does not mean "throwing out" that variable. It means estimating whether a variable has any explanatory power on its own, or whether all of the "information" it contains, in this case the ability it has to explain electoral success, is simply a product of it being correlated with another variable (a "confounding" variable, like fundraising). In this case he's demonstrated that fundraising alone can explain the observed correlation, at least in the small sample sizes seen here.
Put another way, if you "control" for the effect of fundraising and the correlation with ideology disappears, then there never was a correlation with ideology.
Diet correlates with weight. If you "control" for the effect of calorie consumption and the correlation with weight disappears, then there was never a correlation with diet.
But yes, in fact, calorie consumption IS the primary driver of weight, and when attempting to test effectiveness of certain diets, the research is extremely weak on being able to show effectiveness of ANY approach other than calorie restriction.
Even advice like "eat more protein" or "eat more fiber," the logic in both cases is to reduce overall calories consumed by eating foods that lead one to feel more full and thus consume less.
The "Democrats" are going all in on a strategy because its supports the status quo fascism they've always wanted and revered. They have no interest whatsoever in any modicum of progressive change. They and their supporters have made out like the bandits they are, precisely by sucking up to the most greedy, repressive forces available at every given instant They want to prevent any change in that. The strategy of subverting, besmirching , or sabotaging any progressive candidate or message is all they are ever interested in, and the fact that it has led to out and out fascism would be counted as a feature, not a bug.
You think that more left candidates would be able to raise similar amounts of money if the party apparatus got behind them? Enough so that the fundraising advantage would largely dissipate?
The point here is that "moderates being able to raise more money" is _in part_ a function of strategy at the high level of the democratic party. While strategic shifts and priority changes at the party level wouldn't 100% redirect the funds, these types of grand strategic visions and post-mortem assessments need to recognize the endogeneity of their choices.
You think that more left candidates would be able to raise similar amounts of money if the party apparatus got behind them? Enough so that the fundraising advantage would larger dissipate?
I don't know about "enough", but the idea is it's not like moderates just randomly enter a race amid others and then successfully attract money in a free marketplace of donations, but rather that the party apparatus can function as a significant gatekeeper/director of funds in early stages (and later) -- so it's not as clearcut as fundraising success correlating with electability.
I'n a researcher with 30 years of experience. You are proving nothing. You could try constructing focus groups and providing them with sample platforms. But PAC funding tells you what a PAC thinks the candidate represents. Progressives want to "believe" this to be true.
I agree that he didn't prove it is more advantageous to move left or be progressive, but I do believe he conclusively proved that the NYT case for the moderate advantage was completely ungrounded. If you disagree, can you explain how the analysis presented in his second figure doesn't prove that?
The type of broken brain it takes to think that Democrats need more moderates after the last few generations of American decline is insane.
Aren’t incumbency and fundraising downstream of ideology and thus should not be adjusted for if you want to estimate the causal effect of ideology on electoral success?
Yes, adjusting for incumbency and fundraising will induce post-treatment bias and will give you an estimate of the causal effect of moderation that does NOT go through incumbency or fundraising. Moderation is also measured imperfectly with error, so there will also be attenuation bias towards zero.
Are you saying Citizens Untied has zero effect on the electoral process?
Not at all. This is about how much each candidate raised individually. And as the system works today, fundraising ability is part of how “good” a candidate is, so controlling for fundraising is methodically wrong. It’s saying “moderate candidates would do worse if they couldn’t have raised as much money as they did”, which may be true but it’s imagining a different system than the one currently in place, and in fact they did raise this money. In more technical terms, one cannot adjust for a variable that is affected by ideology if one wants to estimate the casual effect of ideology on electoral success
All candidates do worse when there is less money to spend, not just moderates. I think you are suggesting that candidate moderation causes donations to increase, so controlling for donations erases some of the effects of moderation. Do you have any evidence that this is the case? In general election races, do moderate candidates get more donations than other general elections with progressive candidates? Hypothetically, would progressive candidates in those same races get less money than the moderate candidates did? Or are other variables more important, like how competitive the race is, whether the candidate is an incumbent with a base of support, if the candidate has been endorsed by other popular politicians, how many out-of-state donations there are, etc.? Primary donations certainly could be based on progressive vs moderate politics, but it is not clear to me that the same is true in the general election.
You also suggest that candidate moderation led to incumbency in the 2024 election, which would be difficult to prove given the research referenced in the previous post that candidate moderation has not led to increased electability for Democrats.
"one cannot adjust for a variable that is affected by ideology if one wants to estimate the casual effect of ideology on electoral success"
Yes, you absolutely can, at least in a statistical sense. In a regression model, you can include ideology, fundraising, and also an interaction term, or ideology x fundraising, and determine whether any or all are significant variables.
I don't know whether that can be done with any of the data discussed here, but in the hypothetical case, "adjusting for a variable" does not mean "throwing out" that variable. It means estimating whether a variable has any explanatory power on its own, or whether all of the "information" it contains, in this case the ability it has to explain electoral success, is simply a product of it being correlated with another variable (a "confounding" variable, like fundraising). In this case he's demonstrated that fundraising alone can explain the observed correlation, at least in the small sample sizes seen here.
Put another way, if you "control" for the effect of fundraising and the correlation with ideology disappears, then there never was a correlation with ideology.
Diet correlates with weight. If you "control" for the effect of calorie consumption and the correlation with weight disappears, then there was never a correlation with diet.
What do you mean by "diet?"
But yes, in fact, calorie consumption IS the primary driver of weight, and when attempting to test effectiveness of certain diets, the research is extremely weak on being able to show effectiveness of ANY approach other than calorie restriction.
Even advice like "eat more protein" or "eat more fiber," the logic in both cases is to reduce overall calories consumed by eating foods that lead one to feel more full and thus consume less.
The comment was meant to be read sarcastically, but evidently my ability to perform snark is severely lacking. Probably for the best.
I was parodying your claim that if an effect goes away after controlling for a mediator variable, the correlation was not causal.
It would be like saying diets don't work because they don't correlate with weight after controlling for calories consumed. Absurd!
If you haven’t figured out that the New York Times has a problem, you have not been paying attention.
"Moderation helps you win."
"If you control for being a winner it doesn't."
The "Democrats" are going all in on a strategy because its supports the status quo fascism they've always wanted and revered. They have no interest whatsoever in any modicum of progressive change. They and their supporters have made out like the bandits they are, precisely by sucking up to the most greedy, repressive forces available at every given instant They want to prevent any change in that. The strategy of subverting, besmirching , or sabotaging any progressive candidate or message is all they are ever interested in, and the fact that it has led to out and out fascism would be counted as a feature, not a bug.
If moderates can raise more money… and money helps candidates win…isn’t that a reason to nominate more moderates?
You think that more left candidates would be able to raise similar amounts of money if the party apparatus got behind them? Enough so that the fundraising advantage would largely dissipate?
The point here is that "moderates being able to raise more money" is _in part_ a function of strategy at the high level of the democratic party. While strategic shifts and priority changes at the party level wouldn't 100% redirect the funds, these types of grand strategic visions and post-mortem assessments need to recognize the endogeneity of their choices.
You think that more left candidates would be able to raise similar amounts of money if the party apparatus got behind them? Enough so that the fundraising advantage would larger dissipate?
I don't know about "enough", but the idea is it's not like moderates just randomly enter a race amid others and then successfully attract money in a free marketplace of donations, but rather that the party apparatus can function as a significant gatekeeper/director of funds in early stages (and later) -- so it's not as clearcut as fundraising success correlating with electability.
Sadly I don’t think swing voters care much about corruption.
Could be a nice unifying message if they are already primed to dislike how things have been going lately.
Great article.
I'n a researcher with 30 years of experience. You are proving nothing. You could try constructing focus groups and providing them with sample platforms. But PAC funding tells you what a PAC thinks the candidate represents. Progressives want to "believe" this to be true.
I agree that he didn't prove it is more advantageous to move left or be progressive, but I do believe he conclusively proved that the NYT case for the moderate advantage was completely ungrounded. If you disagree, can you explain how the analysis presented in his second figure doesn't prove that?