Thread by @Peoples_Pundit

1. It is patently false to claim online polls are easier to "rig" than other data collection modes. Data sources and methodology need sunlight, not the data collection mode, itself.

This is why all pollsters should comply w/ @AAPOR Transparency Initiative.

2. FYI, data collection modes are online, landline, cellphone, etc. Data source is the source of the population being surveyed, i.e. voter files, online panel, etc.

AAPOR's TI requires you disclose this information at the time of release, and more. Here's the kicker...

3. Our online panels are sourced by voter files, meaning they are in fact the same source, verifiably, and adding an online component simply gives a pollster another mode by which to interview the same population. Not all respondents participate at equal rates across all modes.

4. Maybe the real problem is in our understanding of what is an online poll. The Drudge Report's website poll isn't an "online" poll. It's not at all scientific. Online panel polling isn't remotely close to the same thing, though clearly many big names don't know the difference.

5. Also FTR, our online panels vote at higher rates than the entire voter file, itself. Naturally, I can't speak to all online panels, but our panel and our partners' panels have shown this to be true. It's not an industry secret.

https://twitter.com/Peoples_Pundit/status/1059432131653439493

6. It's worth repeating how much even tiny disparities in participation between modes can have an impact. Rather than trashing something they don't understand, anchors on certain networks should concern themselves with their own network's Donnelly+7s.

How'd that work out?