data for politics #22: What happened in the NY-14?

By Kevin Morris (@ktnmorris)

In June, the NY-14 became perhaps the most famous congressional district in the country. On June 26th, 28 year-old Alexandria Ocasio-Cortez bested Joe Crowley in the district’s federal primary. Ocasio-Cortez ran on an unabashedly progressive platform, calling for Medicare for all, the abolishment of ICE, and a federal jobs guarantee. She has been heralded as the face of the insurgent left, demonstrating the potency of progressive policy among the Democratic base.

So what the hell happened on September 13?

New York State held its state primary on September 13th, featuring incumbent Andrew Cuomo facing off against progressive education activist Cynthia Nixon. Successes such as Ocasio-Cortez’s and Ayanna Pressley in the MA-07 stirred the hopes of many New York progressives that Nixon could bring a serious challenge to the Cuomo regime. In the end, she fell flat, outperforming Zephyr Teachout’s run against Cuomo in 2014 by just two percent.

Somewhat surprisingly, Nixon seriously underperformed in parts of New York City that had shown an appetite in June for progressive candidates. In the NY-14, for instance, Nixon won just 32 percent, compared to Ocasio-Cortez’s 57 percent. In the NY-09, where challenger Adem Bunkeddeko won 48 percent of the vote in the federal primary in June, Nixon won just 36 percent of the vote.

Some hot-takes have implied that it was just a matter of turnout. Dave Weigel, for instance, argued that Crowley lost by not turning out Cuomo-machine-motivated voters.

That story, however, doesn’t carry much water. Although there was a relationship between increased turnout and lower support for Cynthia Nixon, the correlation doesn’t explain the downward shift in support for the progressive candidate; even precincts where turnout barely increased supported Cynthia Nixon far less than they supported Ocasio-Cortez:

Some of this is likely due to the racial demographics of the district. Although black voters seemed willing to take a chance on a Latina from the area, it appears they were far less likely to take a chance on Cynthia Nixon. In Queens, support for Ocasio-Cortez and support for Nixon were highly correlated, but that relationship broke down in the Bronx:


This corroborates what we saw across the city, where Cynthia Nixon underperformed worst relative to Jumaane Williams in predominantly black communities:

cynthia_jumaane (1).png

Comparing these two elections is a little bit tricky because of that turnout issue. Not only did the share of registered voters who turned out change, but the composition also likely changed. Even in a district where turnout remained the same, it’s unlikely that exactly the same voters participated. We won’t be able to tease that out till next month.

But when we focus only on races that happened last week, it becomes even clearer that the hopes of progressive candidates didn’t rise and fall with the Cuomo machine. Last week, Zellnor Myrie and Alessandra Biaggi defeated members of the conservative Independent Democratic Caucus. Alessandra (whose senate district overlaps with Ocasio-Cortez’s congressional district) and Myrie ran progressive campaigns and were effective at getting folks to turnout. Surprisingly, even where neighborhoods supported these progressive candidates for the state senate, they didn’t support Cynthia Nixon:

This is all a bit puzzling - why were neighborhoods that were so happy to support progressive challengers in the federal primary and for state senate unwilling to cast their ballots for Cynthia Nixon? Certainly some of it can be attributed to the Cuomo machine turning out voters, but that’s far from the whole story there was a split even in races where turnout couldn’t have been a factor. It points to the complexity of running city and statewide campaigns versus races covering a smaller geography. It points to the potency of the identity politics of the candidate, and the sway that Cuomo has over voters. We should, however, resist the easy narratives. Clearly, Cynthia Nixon was not simply interchangeable with other progressive candidates for NYC voters. Her defeat, happily, doesn’t point to the impotence of progressive policies in the neighborhoods that didn’t vote for her.

Subscribe to our mailing list

* indicates required

Data for Politics #21: The Day After

By Kevin Morris (@ktnmorris)

Well, machine politics are alive and well in New York State, as yesterday’s primary election showed. In a race described by the New York Times Editorial Board as “dirty politics, nearly as sleazy as it gets,” incumbent Andrew Cuomo falsely accusing his opponent, Cynthia Nixon, of antisemitism. He also “enticed” builders of the new Gov. Mario M. Cuomo Bridge to rush its opening, placing politics above safety. These stories were just from the last week of the campaign, yet he sailed on to win the Democratic primary for governor. Jumaane Williams lost his bid for Lieutenant Governor, as did Zephyr Teachout for attorney general. Blegh.

We’ll be puzzling over the results for weeks to come - and that’s before we even get information about who actually voted from the state next month.

Today, we’re doing a preliminary dive into the results in New York City. Of course, the race was competitive across the state (Cynthia Nixon won 36 percent of the vote outside of New York City, compared to 33 percent in the city) but New York City makes gathering the precinct-level data the day after the election far easier than the rest of the state. To that end, we’re going to take a look at how support for Cynthia Nixon, Jumaane Williams, and Zephyr Teachout were correlated with a few demographic factors at the precinct level. Of course, the voter file doesn’t yet include data on who actually voted. In the charts and maps that come below, we’re determining the average characteristics of each precinct by looking at the demographics of all actively registered Democrats.

First, a couple maps to situate ourselves.



These maps aren’t particularly surprising, but that doesn’t mean that they aren’t frustrating. The progressive slate of candidates generally did very well in whiter, wealthier neighborhoods like downtown Brooklyn, much of Manhattan, and gentrifying parts of Queens such as Long Island City and Astoria. With the exception of Williams (more on that below), the candidates failed to win much support in majority-Black neighborhoods in Brooklyn and the Bronx.

Thanks to the hyper-segregated nature of New York City, these neighborhoods in the outer boroughs are also lower income and face significantly more economic instability. These are precisely the neighborhoods that would likely have benefited most from Nixon’s ambitious housing affordability plan, or her Medicare for All proposal. These neighborhoods have long been big supporters of Andrew Cuomo - in fact, the Bronx was Cuomo’s strongest county in his 2014 race against Zephyr Teachout. Nixon likely knew that winning over these communities would be tough, choosing to launch her campaign in Brownsville, Brooklyn. In the end, it wasn’t enough.

But there are some reasons for hope following yesterday’s race. Based on an examination of where Nixon underperformed relative to Jumaane Williams, it looks like communities in Central Brooklyn and the Bronx aren’t necessarily opposed to progressive candidates:


Cynthia Nixon underperformed most in areas with large black populations. There could be a couple of different reasons for this: Jumaane Williams is a black man, and it is possible that communities there supported him largely for that reason. However, it’s also more likely that these communities actually preferred his politics, as did the rest of the city. Andrew Cuomo has long enjoyed the support minority communities, and it’s not surprising that that support did not evaporate. On this down-ballot race, though, support for the progressive candidate didn’t drop off. This should give us hope that after the Reign of Cuomo a progressive candidate has a strong chance of running a winning campaign in these neighborhoods.

There is also reason for encouragement when we look at the relationship between a precinct’s average age and the rate at which they supported the progressive candidates. Much like we saw when we looked at the city’s federal primaries, younger precincts were much more likely to cast their votes for Cynthia Nixon. Before we dive in to the age analysis, though, a big caveat: these are the average age of registered voters, not the folks that actually cast a ballot. We won’t be able to calculate the average of people who actually participated for about a month, at which point we’ll update these charts.


Despite the fact that Black neighborhoods in New York City supported Andrew Cuomo, there’s a strong relationship between age and support for Cynthia Nixon when we limit our scope to plurality-Black precincts:


Younger precincts were similarly more supportive of Jumaane Williams and Zephyr Teachout:


It’s a frustrating morning here in New York. We’ll likely be stuck for four more years with a governor with a penchant for corruption, misinformation, and often outright lying. Much is now riding on how Tish James acts as the presumptive attorney general. She has been good for the city, but cozied up to Cuomo, his donors, and IDC members. It’s unclear how effective she’ll be at holding Cuomo accountable.

There are a few bright spots: 5 of the 6 members of the IDC in NYC (and six of eight statewide) facing primaries lost their reelection bids, making a truly Democratic State Senate significantly more likely. Below, the results of all six races are combined. Where folks turned out, progressive candidates triumphed:

Elsewhere, incumbent Martin Dilan lost his seat to progressive provocateur Julia Salazar:


As we keep on seeing, there’s hope that when younger voters turn out at higher rates and make up a larger share of the electorate, New York State might finally start electing the progressive leaders it deserves. Not this year, but maybe someday. As for me, I think my bike will be seeing more and more use as the subway, once an engineering wonder of the world, continues to crumble. Good luck out there.

Subscribe to our mailing list

* indicates required

data for politics #20: A #MeToo Effect? Attitudes About Gender Equality and Workplace Harassment

By Meredith Conroy (@sidney_b), Alexander Agadjanian (@A_agadjanian) and Hayley Cohen (@HayleyCohen)

Over the weekend the news broke that CBS Chairman and CEO, Les Moonves, would step down from his position, following allegations of sexual harassment and assault by several women, spanning many years. The flood of sexual harassment allegations against mostly powerful men in business, media, politics, and entertainment have largely come to be associated with the #MeToo movement, which is a movement that encourages women to publicly acknowledge or express their experiences with sexual harassment. The origin of “Me too” stems from the work of activist Tarana Burke who has worked with young women of color survivors for years; the hashtag was popularized online by actor Alyssa Milano last October.  


As attention to the movement has grown, news media interest in the effects of the movement on attitudes about workplace harassment has heightened. For example, in April, Glamour and GQ polled over 1,000 men between the ages of 18 and 55 to get a sense of how men are reacting, while Vox surveyed women’s reactions. In August, Huffpo released their survey which shows that political partisanship shapes opinions of the movement, with Democrats (men and women) more supportive than Republicans.

Below, we present some evidence of a #MeToo effect by looking at panel data from the Voter Study Group, which asked respondents about gender equality and sexual harassment in the workplace in 2016 and 2018.

The Voter Study Group (VSG) is a collaborative that partners with YouGov to survey the same group of respondents overtime, thereby more precisely capturing change over time than surveys administered to random samples. While the VSG does make their raw data publicly available, the data for 2018 is not yet accessible. Therefore, we rely on their weighted crosstab reports for 2016 and 2018 to present change in opinions about gender equality and harassment in the workplace. Once the 2018 data are available we can look more closely at group-based differences to see whether certain populations are more or less responsive to #MeToo.

The figure shows the percent of respondents who agree with statements about workplace gender equality and harassment. The questions asked respondents whether they strongly agree, agree, disagree, or strongly disagree with each statement. We collapsed the agree and strongly agree categories for the figures below; percentages for  respondents who said “disagree” or “strongly disagree” are not shown. Although these questions do not invoke the #MeToo movement explicitly or capture whether respondents are familiar with it, they do probe relevant attitudes are measured before and after the movement’s prominence, and therefore are a useful proxy for opinion change on these topics.


According to the crosstab reports provided by VSG, Americans’ broader views of women’s professional lives have changed little or not at all since 2016 with one exception (Footnote: The margin of error for the 2016 and 2018 survey is 2.2 and 1.8). Among those who agree that “sexual harassment against women is no longer a problem in the US,” we see a six point shift; in 2016, 23 percent of respondents agreed that sexual harassment was not a problem, but in 2018, only 17 percent agreed with this sentiment. This change signals an increase in awareness of sexual harassment in the US.

But this increased awareness of sexual harassment against women in the workplace seems to be just that--an awareness that harassment exists. Increased awareness does not correspond with more Americans agreeing that women miss out on jobs because of discrimination (Q4), or changes in attitudes about whether women who complain about harassment cause more problems than they solve (Q6). Although the data suggests sexual harassment is more widely acknowledged as a problem, there seems to be no change in attitudes toward the women who speak out about the problem.

Taking this data in context with the other studies we mentioned further illuminates #MeToo’s reach. Specifically, for many men, #MeToo is not synonymous with sexual harassment in the workplace. In the GQ survey of over 1,000 men ages 18-55, 41% said they had never heard of the #MeToo movement. But the 2018 VSG study, conducted during the same month, shows that 75.8% of men disagree that sexual harassment is no longer a problem. While the VSG survey includes all men 18 and over and the GQ study includes only men 18-55, the vast distance between these two figures suggests that there are men who recognize the problem of sexual harassment without having heard of the #MeToo movement.

The #MeToo movement, and the discussion of #MeToo in the media may have led to the increased awareness of sexual harassment against women in the workplace. However, this data suggests that the movement may not be changing Americans’ views on other key issues that are connected to harassment: discrimination in the workplace, gender equality, and attitudes about women who speak out against harassment. Although the #MeToo movement may be talked about as a watershed moment for sex equality, its reach is not universal and many attitudes about equality, opportunity, and harassment aren’t yet moved.

Meredith Conroy (@sidney_b) is a senior adviser to Data for Progress. Meredith has a PhD in political science, and is a professor at California State University, San Bernardino campus.

Hayley Cohen (@HayleyCohen) is a graduate of Tufts University and former Lake Research Partners Project Director.

Alexander Agadjanian (@A_agadjanian) is a research associate at the MIT Election and Data Science Lab.


The 2016 and 2018 Voter Study Group Surveys were conducted by Democracy Fund and YouGov. The 2016 Voter Study Group Survey was administered online to 8,000 adults (age 18 and over) who had internet access between November 29 and December 29, 2016. The margin of error for this survey is plus or minus 2.2 percent.

The 2018 Voter Study Group Survey was administered online to 6,005 adults (age 18 and over) who had internet access between April 5 and May 14, 2018. The margin of error for this survey is plus or minus 1.8 percent. Respondents were first interviewed in 2011 as part of the 2012 Cooperative Campaign Analysis Project (CCAP). Respondents to the 2016 VSG survey could have previously been interviewed 2 or 3 times: December 2011, November 2012, and July 2016. Respondents to the 2018 VSG survey could have previously participated in 5 surveys: December 2011, November 2012, and July 2016, December 2016, and June 2017.

See the code here.

Subscribe to our mailing list

* indicates required