Updated: Oct 25, 2020

UPDATE: In the final week our recommendation is to donate to flip the Texas State House here. Donations will be used on highly efficient digital ads where the spend can be increased easily as more funds come in. This is our top recommendation right now.


Since Justice Ruth Bader Ginsburg’s passing, more than $100 million dollars have been donated to Democratic candidates. The energy and urgency around this election is clear. However, is the money going where it can make the biggest impact?

Republican candidates have out-raised Democratic candidates in the critical states of Florida and Texas. Women and minority candidates are underfunded. Click here to go straight to our ActBlue page which allocates funds to the 10 state house races where we believe your money will have the biggest impact. You can also explore the data in our interactive tool here.

Here’s how we determined which races to fund:

At Data 2 the People we have been working on 2020 elections up and down the ballot, and in the course of our work we have collected data on many different races. We want to help people sort through the data to find the races that need the money most. We’re excited to share our work: aggregated election data from across states and races to find candidates we believe are most in need of funds.

To start, we’ve focused on historical election outcome data and current fundraising data for the 2020 cycle. We’ve identified 61 priority state legislature races and have highlighted five of our top races below. Our goal is to be transparent about our process so that we all can make clearer decisions, and so that Democrats can win big in November, up and down the ballot!

1. First we selected states.

2. Second, we analyzed the funding data so we could understand where to look.

3. Third, we selected candidates within states, for each level of government.

4. Finally, we ranked the list of candidates.

We will update this analysis as we collect additional fundraising data and welcome your comments and ideas!

Go here to donate easily to our priority state legislature races via ActBlue.


Step One | State Selection

There are some states that are more important than others in 2020.

This first phase of analysis focuses on 14 states (Colorado, Florida, Indiana, Iowa, Kansas, Michigan, Minnesota, Missouri, Montana, Nevada, North Carolina, Pennsylvania, Texas, and Wisconsin) selected based on three categories of election importance:

  1. Presidential Swing States (selected based on FiveThirtyEight forecasting)

  2. Senate Swing States (selected based on FiveThirtyEight forecasting)

  3. State Legislatures with the power to redistrict (selected based on Princeton Election Consortium analysis)

Some states fall into more than one category or have additional important features. For example, in Missouri, Kansas, and Indiana the Republicans hold a supermajority at the state legislative level, and in Missouri there’s also a ballot initiative, Amendment 3, which would allow for more partisan gerrymandering in the state.

For now, we have limited our analysis to states with more straightforward election formats. However, in the future, we will plan to add more states (such as Maine with its ranked choice voting, or Arizona with its multi-candidate system) and more races, providing a more complete picture of the 2020 election.

Step Two | Analysis of High-level Funding Differences by Party & State

With a narrowed pool of states and race types, we wanted to understand funding levels across race types to see, at a high-level, where there might be over-funded and under-funded races. The federal Senate races in these states are more well funded ($167M raised by Democrats vs. $153M raised by Republicans), whereas at the state legislature level Democratic state house candidates have raised less than Republican candidates ($98M by Democrats vs. $108M by Republicans).

For U.S. Senate races, we see that Democratic candidates have raised more money than their Republican opponents in many of the states. The numbers are big—on average $17M for the Democratic candidate and $15M for the Republican candidate. Many of the races do not have a Democratic candidate currently in the seat, and because it often takes more dollars to take a seat from a Republican incumbent, and because that incumbent may have reserves from previous races, it makes sense that the Democratic candidate has raised more to compete (for example, in 2018 the most extreme case was $80M spent by Beto vs $40M by Ted Cruz in TX).

In 2020, Texas is a striking example where the Republican has raised much more money than the Democratic candidate—though polling indicates that this race is less close than other Senate contests. In Michigan, the Democratic incumbent is being out-raised slightly and in Montana, the Democratic candidate has been slightly out-raised by the Republican incumbent.

The Center for Responsible Politics website here has data on all Senate races

In Colorado, Iowa, and Montana the funding is close and the Democratic candidate is challenging an incumbent, so these candidates likely need more funding.

Now turning to the State legislatures where we see Democratic candidates outspent in aggregate, we also see interesting results by state:

In a few states, Republican candidates have out-raised Democratic candidates by a lot: Florida, Texas, Indiana (the most Republican leaning of our list). In other states Democratic candidates are ahead.

The most troubling gaps are found in Florida and Texas, because both are key states where the state legislature will draw new district lines after the census. If Republicans are in charge, there is a high potential for gerrymandering, which would change the balance of power for a decade.

Two high level goals for State legislatures are flipping chambers to be Democratic controlled and breaking Republican supermajorities (states where Republicans hold more than ⅔ of seats and therefore do not not need to work with Democrats to pass certain legislation).

Individual races matter for understanding where we can flip a chamber or prevent a super majority. We look at the difference in the vote share in 2018 compared to the funding difference in 2018 to see how they related that cycle:

There are indeed some races where the Democratic candidate vote difference was very small (close to the horizontal line) and the Republican raised more money. We also see that for competitive races in our focus states there are a few outliers where the Republican candidate spent vastly more money (TX 23, TX 25, and PA 28).

In 2020 we definitely see some raises that are outliers again, and then others where the funding is close. This gives us a framework to look into which races to focus on!

Step Three | Select Specific Candidates

At this point, after RBG’s passing, we suspect both Republicans and Democrats will be pouring money into federal Senate races. So, we recommend giving to Senate candidates even if the Democratic candidate has more money. In particular, Colorado, Iowa, and Montana would be good races to give to, because the Democratic and Republican candidates have similar funding, but the Democratic candidate is challenging an incumbent Republican so may need more.

Aggregating current fundraising data and historical election results, we used three criteria to select potential state legislative races based on 2018 data:

  1. The seat is currently held by a Democrat, but the 2018 margin of victory was less than 10 percentage points and the Democratic candidate had to outspend their opponent by a factor of 2 or more

  2. [The converse of Criteria 1] The seat is currently held by a Republican, but the 2018 margin of victory was less than 10 percentage points and the Republican candidate had to outspend their opponent by a factor of 2 or more

  3. The seat is currently held by a Republican but the race was decided by a very small number of votes (less than 2,000) and the Democratic candidate was outspent

To do this analysis, some states from our initial list were removed for this first phase because they require additional data processing due to idiosyncrasies in their election processes (i.e., fundraising reporting differences, or runoff elections, or multiple candidates can win for a given position)

This analysis produced 61 candidates, all of which can be found here. The breakdown of number of races by state is here:

The most candidates are in Texas and Florida, because, as noted above, the funding gaps are biggest there. Some of the states are indeed less important for the state legislature itself because there is already a Democratic majority, but we include them because we believe there can be a “trickle up” effect where if a local candidate turns out Democratic voters those voters will vote for Biden or the competitive Senate race in that state.

We then further limited candidates based on 2020 fundraising data using the following rules:

  • If the Democratic candidate is running as an incumbent in 2020 and they have currently raised 50% more than their opponent, we’ve excluded them

  • If the Republican candidate is running as an incumbent in 2020 but the Democratic candidate has currently raised 100% more than the Republican, we’ve excluded them

72% of the candidates we have identified are female and 21% are under represented minorities. We did not collect data on gender and race for the races we did not identify. That said, we suspect that this indicates that women and underrepresented minorities are more likely to be underfunded.

Step Four | Rank Races

To prioritize among these candidates, we’ve ranked using two criteria. We initially looked at the votes per dollar for candidates in 2018 and forecasted that, but the results were confusing because of some idiosyncratic data, so we simplified.

First we categorize each race into three categories depending on 2018 vote outcome:

  • “Toss Up” - where the vote margin was within 5-points on either side.

  • “Slight Rep” - where the Republican won in 2018 by between 5- and 10-points

  • “Slight Dem” - where the Democrat won in 2018 by between 5- and 10-points

Second, we sort by the ratio of Republican to Democratic candidate 2020 funding.

You can view the data in our interactive tool here.

The sorted list is here.

State Legislature Races Highlights:

Texas State House District 54

Democratic Candidate: Keke Williams

Republican Candidate: Brad Buckley (Incumbent)

Why we like Keke: Texas needs to flip 9 seats in the legislature to break the Republican majority and ensure fair redistricting after the 2020 census. In 2018, the Democratic candidate for this seat lost by less than 4,000 votes, but was out-raised by more than 15 to 1. Today, Keke has narrowed the funding gap to 3 to 1 but is still behind with only $66k compared to the $180k raised by Brad Buckley, the Republican incumbent. Keke is an army veteran who is endorsed by the AFL-CIO, the Sierra Club, and Texas Democrats with Disabilities. Brad Buckley the Republican incumbent has an F rating from Equality Texas and a score of 60/100 on Environment Texas’s 2019 scorecard. Learn more about Keke. Donate to Keke.

Florida State House District 115

Democratic Candidate: Franccesca Cesti-Browne

Republican Candidate: Vance Aloupis (Incumbent)

Why we like Franccesca: Florida is a critical swing state for the Presidential. It also has a Republican trifecta at the state level (meaning that Republicans hold a majority of seats in the State legislature and also there is a Republican Governor). Democrats need to flip 14 seats in the state House to win a majority in that chamber. District 115‒a district that Hilary Clinton won in 2016, but is currently held by a Republican‒is ready to be flipped Blue. The 2018 race was decided by 579 votes, despite the Republican candidate (now incumbent), Vance Aloupis, out fundraising his opponent by 4x. Today, Franccesca has narrowed that fundraising gap to 3 to 1 and secured endorsements from Emily’s List, the NRDC, and Run for Something. Given the tiny margin of victory in 2018 and Clinton’s success in 2016, we think that with additional support Franccesca can win.

Learn more about Franccesca. Donate to Franccesca.

Iowa State House District 47

Democratic Candidate: Shelly Stotts

Republican Candidate: Phil Thompson (Incumbent)

Why we like Shelly: Iowa is four seats away from a Democratic State House and Shelly is one of the best possible candidates to make that happen. In 2018, the Democratic candidate lost their race by just under 900 votes, even though they raised only 38% as much money as their opponent, which shows the district could be promising for Democrats. Additionally, while the district voted for Trump in 2016, Obama garnered 51% of the vote in 2012. In 2020, Shelly has raised quite a bit less than her Republican counterpart, but the race overall has raised much less, meaning even just a small amount of cash could help Shelly to clinch the seat. Shelly spent 35 years as a classroom teacher. Phil previously worked at the National Rifle Association Institute for Legislative Action. He has a 0/100 on the AFL-CIO 2019 legislative scorecard. Learn more about Shelly. Donate to Shelly.

Indiana State House District 89

Democratic Candidate: Mitch Gore

Republican Candidate: Cindy Kirchhofer (Incumbent)

Why we like Mitch: Indiana is just one seat away from breaking a Republican supermajority and we bring you our top choice for breaking that supermajority. Mitch, a Captain with the Marion County Sheriff's Office, has been endorsed by a number of officials and organizations, including Run for Something, which picks young, promising progressive candidates like Mitch. Mitch’s opponent, Republican Cindy Kirchhofer, won her reelection in 2018 by just 218 votes, even though she had raised more than 5 times as much money as her then-competitor. This year, while Mitch is still slightly behind in fundraising as compared to Cindy, he only has to make up a gap of $13,500 (as of publication). Mitch’s fundraising position is vastly better than in past races, and further donations to help close this gap could help Mitch flip this long-held Republican seat blue. Republicans currently control Indiana entirely, and with plans to redistrict in 2021, this is a very high leverage seat. Learn more about Mitch. Donate to Mitch.

Kansas State House District 2

Democratic Candidate: Lynn Grant

Republican Candidate: Kenneth Collins (Incumbent)

Why we like Lynn: Kansas is just one seat away from breaking a Republican supermajority and we bring you a top choice for breaking that supermajority. Lynn is running in Kansas District 2, which was decided by just 72 votes in 2018. Currently, Lynn’s competitor has out-raised her, but only by $3,000 at the time of analysis, and the gap is dwindling. This race could tighten very easily. If elected, Lynn would advocate for Medicaid expansion and fight against food insecurity, which is a growing problem in Kansas. A win for Lynn could make for a fairer, more balanced legislature for Kansas. Learn more about Lynn. Donate to Lynn.

Next Steps

We plan to add US Congressional races and Arizona and Maine, which we were not able to include so far.

We also will share back results from all of the races we are highlighting! Sometimes giving to down-ballot candidates feels hard to track, so we want to make it easy for our audience to share in the successes they’ve contributed to building by sharing the outcomes of these elections. Stay tuned.

Please reach out with questions and comments. And please donate!

Authors: Ben Tanen, Rebecca Grunberg, Sarah McGowan, Carolyn Henderson, Elena Grewal

1,754 views0 comments

Updated: Jun 8, 2020

Dr. Elena Grewal, Alex Orenstein

We are appalled by how black people are treated by the police in the US and by how deeply racism is ingrained in our country. Black lives matter. At Data 2 the People we are using data to help elect candidates for public office aligned with our values - candidates who are going to support the BLM movement.

As President Obama shared, local elections are a mechanism for change, writing “the elected officials who matter most in reforming police departments and the criminal justice system work at the state and local levels. It’s mayors and county executives that appoint most police chiefs and negotiate collective bargaining agreements with police unions...

- Barack Obama, Instagram, June 1, 2020

Taking this advice, we look into four questions:

  1. Which are the worst cities for Black people with respect to police violence?

  2. What type of elected officials can we support in those cities to make an impact?

  3. What are the election dynamics?

  4. What can we do to support those elections with data?

1. What are the worst cities for Black people with respect to police violence?

First we needed to actually find the data. This is difficult, a symptom of the problem that perpetuates the problem. As Bryan Stevenson, author and protagonist in Just Mercy, writesno one in this country can tell you how many people were killed by the police last year, because we don’t require that data. People have been trying for two decades to mandate the disclosure of that kind of information.”

We found a dataset compiled by the Washington Post. The Post started tracking more than a dozen details on every fatal shooting in the U.S. since 2015. They continue to enhance this data as more facts emerge about individual cases, seeking help from the public. We also found a database of police killings as part of the Mapping Police Violence project created by We the Protesters. The Washington Post had 5 more months of data, through June 1 of this year, so that is what we use. We did notice some differences between the datasets prior to this year that we would like to look into further; at a high level, the data matched.

We then mapped the data by metro area:

(Click on the map to interact with it + view the legend)

We look at the raw number, not the percentage of the population, because we want to understand where the impact of changing the local officials will be biggest.

The specific cities in the US with the highest number of police shootings of Black people by the police since 2015 are the following:

The 15 cities above represent about ~20% of all police killings of black people by gunshot. While there are other forms of police brutality not captured here, we believe this is a useful proxy for police violence.

2. What type of elected officials can make change?

District Attorneys -- District attorneys are the “chief prosecutor” typically at a county level (depends on the state) and lead a staff of prosecutors. The District attorney is usually the only member of their staff who is elected. The DA’s office decides whether to prosecute police officers. Color of Change has created a database of prosecutors here. DAs can take money from police unions, creating a conflict of interest. To find DA candidates whom we want to support, we can look for those who didn’t take money from policy unions.

Mayors -- In some places mayors hire and fire police chiefs, in others they appoint police chiefs, and in other places they control the city council as well. Mayors also have some power over the police budget depending on the place (e.g. can veto the budget). They are also leaders of the city so are important moral leaders with influence beyond their formal scope of authority.

Sheriffs -- Many states have elected sheriffs, who can serve in a role similar to police chief. The job depends on where they are. In Harris county, the sheriff is the police force for the unincorporated part (2 million people). They are also often responsible for the county jail conditions.

State Attorneys -- Some states have “state attorneys” instead of “district attorneys.” The function is similar. They could overnight stop charging thousands of cases and can decide on what sentence to ask for.

Attorney General -- The state attorney. Our understanding is that for police killings the attorney general at the state level does not typically have much impact.

City Council -- Typically determines the budget for the police can reallocate funds. They also can start new programs that could make a big difference.

Judges -- There are many elected judges. They make sentencing determinations which is key for determining what happens to police and Black people. Sometimes there is an elected “chief judge” who sets policies.

Ballot initiatives / Propositions -- Not an elected official, but another way to change laws and the police. For example, passing a constitutional amendment giving returning citizens the right to vote after a felony conviction in Florida.

We welcome additional input on this list!

3. What are the election dynamics?

We started by looking at one location as an example. Houston, Texas is a part of Harris County, and so we took a look at the election results from 2018.

In Houston, there were 1.2M ballots cast across all elected positions (~50% of registered voters, ~25% of the overall population). On the ballot, there were 76 judges up for election, and another 9 justices of the peace. Despite this seemingly large opportunity to make a difference, 10% of voters didn’t cast a vote for a judge at all. More importantly, regarding those who actually did vote for a judge, we wonder how many voters really knew who would be best to vote for, for the competitive judges races (of which there were many).

This gives a sense of just how many elected officials there are who can influence outcomes for the criminal justice system and the importance of working on those races so the best candidate wins and so voters know why it matters. We’re in the process of gathering additional data on races for the above offices to calculate margins and target candidates.

Out of the 15 cities we identified, 80% are led by a Democratic mayor and 40% are led by a Black mayor. Party affiliation is only one signal here

4. What can we do with data?

Our mission at Data 2 the People is to help elect candidates who will support making the world better, and BLM is a critical movement in achieving that mission.

There are multiple ways we can use data to help candidates for these races:

1. Raise more money:

Candidates spend a lot of time fundraising (for some races up to 30-40 hours per week). We can use data to identify opportunities to increase fundraising efficiency.

2. Know who to reach out to:

Most small races don’t have the budget to reach every voter personally, so data can help the campaign to determine who will be best to focus on.

3. Test what works:

So much money is spent without a clear sense of the return on the investment. Testing and experimentation can be the new norm for how a campaign is run.

In addition to helping candidates win, we can help to educate the public (and to aid our candidate selection) by looking into data on the candidates and their records, along with data on the issues in their jurisdictions.

After the Election:

Electing those who support the BLM movement will not necessarily create change without a strong agenda and the ability to navigate the pushback that will come from attempting to create change. That said, the right leaders in place will help.

Call to Action:

We welcome data help. We are a community of like-minded data experts. You can find our values here.

This is the most important election in US history, and we want to do everything we can to help.

455 views0 comments

Updated: Apr 22, 2020

Dr. Elena Grewal

Follow us on Twitter

With the proliferation of COVID-19 forecasts, the average person and government official alike is likely feeling overwhelmed. Is everyone already infected? How many deaths will we really see from this? Will we need to shelter in place for two weeks, or for six months? What will happen to my business or to my children’s schools? Predictions of what lies ahead have, arguably, never been given more attention, nor have the stakes been higher.

As data scientists involved in modeling this pandemic, we know how forecasts help us navigate the unknown. Many of us look to them for a shred of certainty in the midst of so much that is out of our control. But, forecasts are not as certain as we could hope. Even amongst experts, the accuracy of different forecasts is hotly debated. However, nearly all of these forecasts are missing something critical -- and lives depend on this missing consideration.

What’s missing is this: the goal of a COVID-19 forecast is not to be accurate. It’s to save lives.

What we need to realize is that in this dynamic, unprecedented situation, all forecasts have short shelf lives. We must think of forecasts not as inevitable futures, but as guideposts that inform us about the imperative to act and care for one another today.

With this new goal in mind, we’d like to offer 3 practical tips to help us make the most of forecasts.

  1. Beware of any forecast that tells you an exact number. Look for a range to prepare your expectations. If you are making a forecast, share that range.

  2. Consider the ethics of a too high or too low forecast. If you are seeing lots of different forecasts make sure you pay attention to the high numbers too. Forecasters: do not discount high predictions.

  3. Account for how people will react to the forecast. As soon as people’s actions change, many predicted scenarios become less likely. A wise forecaster sets expectations that the outcome will change based on how people respond.

Let’s dive into each of these.

First, know that all data related to COVID-19 are uncertain. The number of COVID-19 cases is subject to how many tests are available and how those tests are administered. Are tests given to only the very sick? To anyone who wants one? Even data on the number of deaths, while likely more accurate, has uncertainty. Was everyone who died from the virus tested? Were there delays in the testing results?(see this article for more) It is impossible to have certainty in a forecast given this shaky data foundation.

All statistical models, even with rock solid data, produce probabilities, not certainties. This uncertainty is typically communicated by statisticians with a “confidence interval” which estimates an upper and lower bound for the forecast.

The problem is that many COVID-19 forecasts being shared and actively used by the public and by policy makers are not using confidence intervals. Any time data is expressed as an exact number, it gives a false sense of certainty.

So projecting 300 hospital admissions in Washington, D.C., on May 14 (as the Penn CHIME model does) is not helpful. Saying “9 days till the projected peak” (as the IHME model cited by the White House does) is similarly irresponsible.

Screenshot of the Penn CHIME model on April 20th where the forecasts show exact numbers:

I believe the forecasters sharing exact numbers mean well; perhaps they think it’s simpler for people to interpret an exact number rather than a range, or perhaps they aren’t able to estimate the uncertainty. Yet scientists are also educators, and as such, they should share the range of potential outcomes. If they do not, ask for that range. Compare multiple sources of models. The discrepancy between forecasts is not a scary thing; it’s a realistic reflection that there are a range of potential outcomes.

Recommendation for the forecaster: Present uncertainty with a range of numbers such as a “worst case” and a “best case,” which is clear language that the average person can understand.

Second, make sure to pay attention to the high estimates. It may feel good to look at the low end of a range, but we argue there is a greater risk in not paying attention to the higher estimates. In addition, low estimates often are created assuming that we will take preventive measures.

We should ask ourselves:

If we act on a forecast that is too low, what will happen? We may not prioritize changing behavior, changing policy, or investing in research. We will not be adequately prepared to prevent fatalities. Associated economic suffering in the long term will be enormous as well. If a forecast is too low, significantly higher fatalities are likely.

If we act on a forecast that is too high, what will happen? Governments may over-prepare or mis-allocate resources. Given the limited preparation taken by most governments so far, over-preparing seems to be a lower risk. The economic impact of over-preparing could be greater in the short-term (arguably not as high in the long term though, because the economic impact will last much longer if the virus spreads widely). We can compensate for a loss of a job. We can’t compensate for loss of life.

As an example of how this can go wrong, on March 2nd, scientists in the UK predicted that there could be an astonishing 500k deaths from COVID-19. Yet, the forecasts were said to be unlikely even by the scientists themselves. The result: the UK waited too long to act. Of course, this isn’t an occasion to cherry pick high numbers. The model choices need to be reasonable. The point is that it is not a good idea to downplay high numbers just because they “seem too high.”

Recommendation for the forecaster: Ensure that the high numbers are not overlooked, using language such as “the number could be as high as X”.

Finally, let’s address the broader point of the impact of sharing a forecast on the outcome. The prophet’s dilemma warns that telling people a prediction can cause a reaction that either contradicts or fulfills that prediction--also known as self-defeating or self-fulfilling prophecies.

This became especially salient in 2016, when Nate Silver and most political forecasters predicted Hilary Clinton would win, and she didn’t. Research showed that because of widely-shared forecasts of her win many individuals may not have voted thinking their vote was not needed, which then cost Clinton the election. In this case, the forecast became a self-defeating prophecy: it led to its own inaccuracy.

The same is true for COVID-19 forecasts. Any projection that is widely-shared can potentially change— for better or for worse— the outcome, depending on how people respond to it. When we see a forecast we should make sure we look into how the forecast has accounted for the impact of behaviors changing. If no changes are accounted for, we should look at a forecast as “what will happen if we continue behaving the same way we have behaved before.”

Recommendation for the forecaster: Make it clear what assumptions you make about human behavior changing when producing the forecast.

Thus, the goal of producing or reading a forecast shouldn’t be to find the most accurate prediction, but to save lives. What ends up happening will be highly dependent on actions we take, and forecasts directly influence this.

As consumers of forecasts, we can hold our statisticians accountable. The role of the scientist and statistician during this pandemic is not simply to share numbers but to take responsibility for the outcome of sharing those numbers. At times like these, where we are in a crisis and faced with immense consequences of decisions, this is even more urgent. Together, we can make the high forecasts wrong.

Thanks for reading, and if there is something we missed here, please let us know in the comments!

#covid19 #forecasting

649 views0 comments