(((JonathanWeisman)))Verified account

The lesson should have been “make sure you understand the basics of statistics before attempting to report on statistical findings.” The media’s desire to make the election as a whole a horserace is ridiculous and totally unsupported by (mountains of) data.

The media's desire to make the election a horserace is totally supported by mountains of television rating data.

You misspelled “make a point to learn the basics of statistical inference”.

Seriously? We should stop observing the situation, collecting data, analysing it, and reporting findings..... because on one occasion our analysis was inaccurate? GMAFB. I’d rather we just refined our processes to be more accurate if that’s ok with you.

No, use the data, not fear and spin/pressure from GOP strategists. All the data points exactly where it's been pointing. Minor deviations, and yet reporting has overblown it, seemingly into a horserace when it's not.

I see on your twitter feed that you believe Khashoggi was murdered in Turkey. Is this just your spin on the data collected? Or is it a reasonable conclusion based on reasonable analysis? I suggest the lattter. Don’t give me data. Experts CAN provide analysis AND predictions.

Not quite sure where you’re going with this. I’m saying use the data, so I’m agreeing with you. But there is no data showing GOP on verge of keeping House. Much of that is spin, hedging by reporters.

And yes, GOP pollsters/analysts cay have the opinion — analysis — that they think this is true, but reporters should report that there’s little to data backing that up instead of being spun/succumbing to fear and making it appear as if it’s a horserace when it’s not.

Amen. More discussion of the factors that could impact the outcome either way and how they are changing, less attempts to synthesize all of that into a prediction (esp when the prediction is more of a bet hedging).