1. Does this mean that a voter voted for Warnock but not Ossoff?
  2. I am confused since I would think they appeared on the same ballot and the tallies for candidates would be made at the same time the ballots are counted.

Can someone please clarify?


5 Answers 5


The special election that Warnock ran in and the regular election that Ossoff ran in happened at the same time, and voters voted in each at the same time, but they were separate races.

As of now, the afternoon of January 6, Warnock's lead over Loeffler is significantly larger than Ossoff's lead over Perdue. The difference is enough that many news organizations are confident enough to call Warnock as the winner in his race but not yet confident enough to call Ossoff as the winner than his.

Although most voters probably voted for two candidates from the same party, there were evidently some voters who voted for Warnock but did not vote for Ossoff, and some voters who voted for Perdue but not for Loeffler.

Furthermore, as Bobson points out in an excellent comment:

There's also some voters who just don't vote in one or the other race, either accidentally or because they don't have a preference. Or cases where one vote wasn't valid for some other reason (not properly marked, voted for both, etc). As of right now, the reported totals are only off by 96, out of 4.4 million total votes.

  • 11
    @user2716511 There's also some voters who just don't vote in one or the other race, either accidentally or because they don't have a preference. Or cases where one vote wasn't valid for some other reason (not properly marked, voted for both, etc). As of right now, the reported totals are only off by 96, out of 4.4 million total votes.
    – Bobson
    Commented Jan 6, 2021 at 19:40
  • 3
    @Bobson That explains it. The total votes (as last shown on the NYT page) in the Ossof/Perdue race is 4,408,642. In the Warnock/Leoffler race it is 4,408,738. Interesting how voters perceive the two contests! Commented Jan 6, 2021 at 19:44
  • 4
    Further, the "live" vote-counts provided on many news sites like the NYT aren't actually live, there's some delay involved in reporting across many counties, basically what you're seeing is the vote totals being reported in batches whenever local officials get around to it. There's a lot of steps involved and things don't always make sense at first glance in the middle of the operation.
    – Ryan_L
    Commented Jan 6, 2021 at 19:49
  • 4
    Perdue was also an incumbent senator who's been in office for six years and was seeking re-relection; Loeffler was appointed to the seat just over a year ago, so if nothing else she has much less of a record and less experience campaigning. Commented Jan 6, 2021 at 19:55
  • 1
    @user2716511 Warnock's opponent was apparently viewed more unfavorably than Ossof's, which probably is a significant factor in the difference.
    – phoog
    Commented Jan 6, 2021 at 22:30

Races are called before all the votes are counted.

If Amy has 10,000 votes, and Bob has 5,000, and there are 1,000 more to count, you can safely say that Amy is the winner. If, on the same ballot, Chris has 7,600 votes and Danny has 7,400 votes, then the media is not going to call the second race. The 1,000 remaining votes could tip the balance towards Danny in a way that they couldn't for Bob.

The exact breakpoint at which a race becomes "callable" is slightly subjective. If Emily has 7,999 votes and Fred has 7,001, the the remaining 1,000 votes could technically let Fred win. But the odds of that happening are exceedingly slim (requiring Fred to win every remaining vote), so most media outlets would probably feel safe calling the race for Emily.

Media outlets use known voter patterns and the known results to estimate what the remaining votes are going to be. This allows them to call races before they are technically decided, but leaves them open to error if the remaining votes sufficiently defy expectations. The most famous example of this is probably "Dewey beats Truman", where a newspaper went so far as to publish a newspaper with that headline before learning that Truman had actually overtaken his opponent.

  • 8
    On the Emily-Fred example, it also helps if the 1000 votes are from an Emily friendly area. If they were from an extremely Fred friendly area, maybe they wouldn't call it just yet.
    – Damila
    Commented Jan 6, 2021 at 20:36
  • 3
    Re the Chicago Tribune's infamous headline: To be fair, there are a number of factors which make such events highly unusual. For a start, they had to write that headline much further in advance compared to when races are usually called today. The article was largely based on pre-election polling, some of which was several months old, and exit polls in their modern form basically didn't exist in 1948. Finally, the entire media establishment had been printing similar headlines for months (e.g. describing Dewey as "our next president" etc.).
    – Kevin
    Commented Jan 6, 2021 at 21:28
  • 3
    An important point for those who are not used to the fine details is that the media calling an election is just an opinion and in no way legally binding. It is the electoral board overseeing the elections the one that gives the actual, valid results, and it is way slower (although they provide provisional results fairly quickly, but they wait until at least the first counting of the votes is over).
    – SJuan76
    Commented Jan 6, 2021 at 21:35
  • 1
    The races are technically decided when the polls close. Not making this distinction is what has fueled the chaos of the last few months. Commented Jan 7, 2021 at 2:14
  • 3
    According to fivethirtyeight reporting, another issue is that media outlets seem reluctant to call an result which may be subject to a mandatory recount. This applies under Georgia's state electoral procedures. Warnock's estimated majority was above the cutoff and was expected to increase as the final vote counts were announced, but Ossof's was below the cutoff until later in the counting process.
    – alephzero
    Commented Jan 7, 2021 at 2:46

Does this mean that a voter voted for Warnock but not Ossoff?

tl;dr: Yes. At least 19K voters did that.

There was a third race on the Georgia runoff ballot, which, in conjunction with the final (yet still unofficial) numbers, provides a little more clarity:

Georgia Jan 5, 2021 Runoff Totals

The reason Warnock's race was called before Ossoff's, was simply due to the fact that his lead was slightly larger right from the beginning, and therefore his lead became insurmountable earlier.

Analyzing this data we can see that Ossoff won by approximately 37K votes, whereas Warnock won by approximately 75K votes. (About a 38K differential.) Note that Warnock had 19K more total votes than Ossoff, which lines up mathematically with a 38K differential (because every 1 vote that shifts from Rep to Dem becomes a 2 vote differential.) It appears to be the case that most people voted for both Rep, or both Dem, but it's likely that more than 19K people voted for Warnock (Dem) and also Perdue (Rep). One possible theory for this is that Black Republicans may have been more likely to split their vote in favor of Warnock. One study (from 2014) showed that 12% of Black adults in Georgia identified as Republican.

Of note, is that in the the third race on the ballot, the Republican candidate won. When asked about this in a news conference on Jan 6, 2021, Gabriel Sterling, a top Georgia election official, stated:

It means there’s people who did not vote for Senators Perdue and Loeffler, who did vote for that, which makes me think there’s probably Republicans who chose not to vote for the two incumbent senators.

  • 1
    Most of this answer is great, but the speculation about Black Republicans is almost certainly wrong since ballot splitting was more common in whiter precincts Commented Jan 7, 2021 at 21:48
  • 2
    @NoahSnyder that's interesting, but it only would discount the theory if Black Republicans were dispersed evenly throughout the precincts, which is extremely unlikely to be the case. It's more likely that the majority of Black Republicans live in right leaning precincts, which would be predominately white. If true then it fits nicely with that data. (Speculation only, I'm sure we could find data on that too...)
    – TTT
    Commented Jan 7, 2021 at 22:45
  • 1
    There's literally no evidence towards your speculation and the evidence we do have cuts against it (though I agree it doesn't conclusively refute it). The simplest explanation is that incumbents typically have an advantage while appointees don't. An alternate explanation is that Warnock's advertising was more effective, which is suggested by the fact that Ossoff did better than Warnock in the Chattanooga media market (and other non-GA media markets). Commented Jan 8, 2021 at 0:15
  • @NoahSnyder definitely no evidence for it! I will do some more research on that point.to see if we can more accurately categorize approximately what 20K voters did. Occam's Razor would indeed point to incumbent advantage.
    – TTT
    Commented Jan 8, 2021 at 6:16

Arcanist has the right idea, but gives a race that is 100% certain. Let's consider about where we were yesterday morning.

With about 90,000 votes outstanding, Warnock had a lead of about 50,000 votes, while Ossoff had a lead of about 15,000 votes (these are approximate for the sake of argument). Since most of the vote remaining to be counted was in heavily Democratic areas of the state, it was unlikely Loeffler could make up her deficit against Warnock. But Purdue still had a path to overtake Ossoff for much of the day. Ossoff would ultimately win by about 35,000 votes.

The principle most press use is that you have to have a decent chance of picking up enough votes to close the gap. That was why Fox News, and then the Associated Press, called Arizona early for Joe Biden

At [the point AP called the race], an estimated 80% of the vote was in, and Biden was up by a margin of 136,000 votes.

And later

In making race calls, the AP has been traditionally known for its caution. It is sometimes deliberately slower than other networks, because it likes to say it isn't making projections but, rather, calls based on solid math. Editors at the AP stress that they make a call only when they have determined that a candidate has no remaining path.

As to your other question

Does this mean that a voter voted for Warnock but not Ossoff?

It does indeed appear that way. Current numbers (with 98% shown reporting) show Ossoff up by 35,615, while Warnock is up by 73,404. Vote totals are within 100 votes for both races, so there's no clear undervoting (where someone voted in one race but not the other).


Aren’t they both on the same ballot?

(from the question title)

Indeed they are on the same ballot, i.e. there is just one physical piece of paper on which the two Senate votes are cast. An image of that ballot can be found online (from Pickens County, Georgia; an absentee/provisional/emergency ballot) and is reproduced below.

Sample absentee/provisional/emergency ballot for the Georgia runoff elections on 05/01/2021 from Pickens County, Georgia. Taken from the official county website and thus assumed to be public domain.

Although these votes are on the same ballot they are in different boxes and thus these boxes can be marked differently. Some voters will have engaged in split-ticket voting by voting for Perdue and Warnock (and some, although fewer, will have voted for Ossoff and Loeffler). There is a third independent race (public service commissioner) which again is to be marked separately so you can probably find all eight mathematically possible variations among the over four million ballots cast in the runoff.

(In addition, the totals show that some voters chose not to vote for public service commissioner – i.e. left that box blank – which means there might even be all 27 ways to fill out the ballot found somewhere in the results if one includes blanks.)

You must log in to answer this question.

Not the answer you're looking for? Browse other questions tagged .