Athena

The government is us; we are the government, you and I. - Theodore Roosevelt

[Athena @BunnySql]

@DuquesneLight your new outage map is still horrible. We'd have to click on every single instance under a zip code to find our specifics. You need better labeling. This is basic BI stuff here.

[Athena @BunnySql]

And what is a peppermint tooth? That was the answer for words that come before tooth... Even googling, there is no such thing as a peppermint tooth.

Come on, do better or don't bother.

[Athena @BunnySql]

@LinkedIn When the only hint Queens gives you is the answer, the puzzle is not solvable without a hint. This has happened several times.

[Athena @BunnySql]

@loralof @VillMom @Angry_Staffer @laurapastine Fucking idiot

[Athena @BunnySql]

@edanniels0893 @The_OG_Menace @JoJoFromJerz Because you clearly are a hateful idiot.

[Athena @BunnySql]

@HomoConNYC @Angry_Staffer Hey dumbass, it's not a "Boogeyman". HIS people are the ones who wrote and published it.

It's dumbasses like you who are going to cry when these policies actually cause you damage. You asked for it, so when it happens, take it like the dumbass you are.

[Athena @BunnySql]

@PJ_Weber @Brian9665781821 @RonFilipkowski Or they'll be glad for the horror caused. Gotta get the others after all. Be damned the consequences because they think as white Christians they are safe and they are right.

They are neither, but they will definitely blame everyone else first.

[Tristan Snell @TristanSnell]

Do NOT panic during this first part of the evening.

Rural counties report first because they have fewer votes to count -- tends to result in Republicans coming out to early leads.

Democratic count tends to come in later in the evening.

Keep Kamala and carry on a la.

[Athena @BunnySql]

@mattcrossette @JoJoFromJerz Yeah, duh. Smart people, you know, anyone who has more than a 3rd grade intelligence level, doesn't have to use their hands at all.

[Athena @BunnySql]

@Nicoabi420 @BarackObama

[Athena @BunnySql]

@CyrusWJones @PAStateDept Keep listening to those voices in your head.

[Athena @BunnySql]

@mattcrossette @JoJoFromJerz Yeah, able to count to 10 without taking his shoes off or wearing sandals. Can't say the same for you.

[Athena @BunnySql]

@chicagojimmy @AKBrews My liver already hates me.

[Mark Greene @RemakingManhood]

Men. Times up. We know exactly what this election is about. It’s obvious now. It’s a war on women and women are dying. VOTE for all the women in your life. Vote for all the daughters who haven’t been born yet. VOTE #HarrisWalz2024

[Athena @BunnySql]

@AP_from_NY @AOC Bots can't vote, and idiots shouldn't. Either way that leaves you out.

[Athena @BunnySql]

@AKBrews I wish I could. Votes by mail weeks ago.

[Mama Maya @rockstarronan]

Taylor Swift wrote a song about my son, Ronan, after he died of cancer. She put me on as a co-writer, and donated the proceeds directly to pediatric cancer -no strings, no hidden motives, just pure compassion and integrity.

In stark contrast, the Trump family engaged in significant fundraising efforts for children with cancer through the Eric Trump Foundation. However, reports later revealed that funds were used to pay Trump-owned properties for hosting fundraising events, raising concerns about conflicts of interest and self-dealing. What could have been a mission solely dedicated to supporting children in need became tainted by financial practices that appeared to benefit the Trump Organization, casting a shadow over the charitable intentions initially claimed.

This election is a no-brainer. It’s a choice between compassion and corruption, between someone like Kamala Harris, who has spent her career fighting for justice, and Trump, who has exploited the vulnerable at every turn. Kamala has consistently shown a commitment to supporting communities and ensuring people’s rights are protected, while Trump and his family have even stooped to taking funds meant for children with cancer.

If integrity, empathy, and basic decency matter, then this choice couldn’t be clearer. Supporting Kamala and her values means standing up for the kind of future where compassion isn’t just talk—it’s action. We owe it to ourselves and to the next generation to reject self-interest and stand on the right side of history.

[Prof Jeffrey S Morris @jsm2334]

Nice article discussing relevant statistical principles regarding polling. https://t.co/rLmqonBtxf

Bottom line is that a poll has variability and bias.

The variability is easily quantifiable and is the reported +/- % and is a function of the sample size and realized proportions.

The bias is much trickier, amounting to the difference between the demographics of poll responders and demographics of the target population, which is those that will actually cast a vote, i.e. voting turn-out.

This is inherently unknown and the only way to adjust for it is through voter turnout models making subjective (and hopefully empirically informed) assumptions about the demographics of those who will actually vote.

A naive summary of % voting for each candidate assumes this bias is zero and presumes the responders to the poll are representative of the target population of voters, and is naive.

Thus, pollsters use turn-out ("likely voter") models to predict the voter turnout demographics, including party, age, race, sex, income, urban/suburban/rural, recall of previous voting history, etc., and then they attempt to correct the bias of the polled sample by reweighting the results based on the posited turnout demographics.

This is mathematically simple, just amounting to a simple weighted average of the poll results for certain demographic groups (BTW same concept as is used to compute "age adjusted" mortality when reporting death incidence data)

The problem is that different pollsters use different models, and we really don't know which model is correct, and every election has unique issues that make turnout different from previous elections.

Here is what I wish the poll-aggregation websites would do:
1. Aggregate raw poll results split by demographics in a database
2. Specify various different turnout models making assumptions about the demographics of voters, some based on past turnout, and others based on prospective new patterns based on issues of the day (e.g. increased female turnout in this election)
3. Compute results for each proposed turnout model.

Results could be reported a number of different ways:
1. Report results for each turnout model
2. Give a weight/probability to each turnout model based on expected likelihood, and compute results based on that weighted average (called model averaging in Statistics)
3. Provide a dashboard where the user could specify their own weights/probabilities to each turnout model, and then get the computed results based on them.

Also, if this were done, they could have a secondary hierarchical model to produce likelihoods of each of the potential turn-out models based on accruing voting data as results roll in on election day, and use this to update projected results in real time.

This has the potential to be more accurate than the current election day projection approaches.

I wish I would have thought of that a month ago, but maybe someone will build something like that for future elections.

@NateSilver538 @RCPolling #Election2024 #PresidentialElection2024 #electionpolls

[TrumpsTaxes (@trumpstaxes.com on bsky) @TrumpsTaxes]

Omg they did a male version of the ‘secret vote’ ad and it’s even better than the female version.

[Athena @BunnySql]

@GhostofPrickles @jonathanstea I am so sorry. And I'm sorry that idiots like this try to minimize the issue and bring it all back for you.