Talk:Opinion polling for the 2022 Italian general election

Strange artifacts in graph

Does anyone have the raw data and a data analysis file in something like R? Some of these points look really weird and I can’t tell if the problem is the data or the analysis.

Maybe the points have been Winsorized or truncated somehow? Closed Limelike Curves (talk) 18:20, 3 December 2021 (UTC)

The data is the exact same that is in the table (literally, the numbers are copy-pasted into the Excel file that get's recognized by R to automatically create the chart). Which points do you see as weird? Impru20^talk 18:36, 3 December 2021 (UTC)

Hi, again, do you happen to have a copy of the spreadsheet (or the RData)? It's hard to work with the table as currently formatted. Closed Limelike Curves (talk) 20:20, 17 September 2022 (UTC)

I see; do you have a copy of the spreadsheet?

The reason I think it looks weird is that all the numbers look almost like they’re drawn from a uniform distribution, which doesn’t make any sense… Closed Limelike Curves (talk) 20:09, 26 December 2021 (UTC)

I agree this is very strange. I don't know whether the issue is in the original data or in the graph, but the plot does not look like legitimate polling data. As well as the constrained uniform-looking spreads in some areas (especially e.g. 2018/2019 for Forza Italia), there's also the extremely gradual changing lines of outlying data, like in Lega for 2020 and Forza in 2021 (which looks more like a moving average or something - no way separate polls would give that similar results in a row like that).

Also the data just seems to have an unusual amount of consistency and lack of outliers (look at FdI for example). Definitely something going on here. Yaldi5 (talk) 12:01, 21 April 2022 (UTC)

There’s a few possibilities here: 1. Fraud—some of these polls were completely fabricated. This isn’t as uncommon as you’d think, which is why you have to make sure to stick to regular pollsters. 2. Extreme pollster herding; it’s possible that Italian pollsters are just unwilling to stick their necks out, so they suppress or reweight outliers to be more in line with consensus. While possible, this would have to be the most extreme case of herding I have ever seen, suggesting a lack of ethics (like suppression of polls) rather than just overly-aggressive weighting. 3. Misinterpreting sources—maybe some of these “Polls” aren’t polls, they’re moving averages or model-based regression lines.

It could also be something else I’m not thinking of. Closed Limelike Curves (talk) 16:32, 21 July 2022 (UTC)

Italia Sovrana e Popolare (ISP)

Should a column for Italian Sovrana e Popolare (ISP) be added? Its voting intentions are currently being estimated by two pollsters (BiDiMedia and Termometro Politico) around 1.0 - 1.7 %. The levels are similar to those of IC and UP, which are included in the table. Tomjoad82 (talk) 08:06, 29 August 2022 (UTC)

https://en.wikipedia.org/wiki/User:Yakme removed the update with on the ground that "just two polling institutes out of >10 is not enough to have a separate mostly-empty column for ISP". I don't necessarily disagree (that's why I asked here first). What would you say a sufficient number of polling institutes is, 4? At the moment, the minimum in this page is 4 (UP). Tomjoad82 (talk) 08:24, 31 August 2022 (UTC)

We (actually mostly myself) keep up-to-date a summary table of all parties covered by pollsters: Talk:Opinion polling for the next Italian general election/Pollsters table. I just finished updating it. Usually the criterion was that when a majority of pollsters polled a party, then it was inserted. In this case it looks like UP is polled by 5 pollsters out of 14, and ISP 2/14. In principle we should also remove the UP column, but 5 pollsters is a reasonable amount. On the other hand, just two pollsters is the same as minor fringe parties like Alternative for Italy, should we also add that one then? I believe the current status is good enough for now, let's see what happens in the following weeks with polls (waiting for the first Ipsos and Demos&Pi polls after the lists were closed). Yakme (talk) 10:17, 31 August 2022 (UTC)

IP's edit warring

Can the IP explain how this is done for "political motives" when it is only for readability purposes and all polls are still there and easy to access with just a click? From mobile, they are not collapsed at all and it is very frustrating. Also who are those "bad actors" and "paid political actors"? For someone pretending to be an admin and claiming to know the "regs", they sure do not seem to know rules like WP:NOPERSONALATTACKS, WP:EDITWARRING, and WP:CONSENSUS. Davide King (talk) 21:00, 31 August 2022 (UTC)

@Davide King I already asked for page protection. Hopefully in a few hours these IPs will be barred from editing this page. Never mind, problem solved. Yakme (talk) 21:04, 31 August 2022 (UTC)

Semi-protected edit request on 31 August 2022

Just looked at this page for the first time and noticed that someone has hidden polls done before August 5th, this isn't allowed as all polls in 2022 must be visible and NOT collapsed by default. Apparently someone is vandalizing the entry by repeatedly hiding those polls. 84.222.57.48 (talk) 21:11, 31 August 2022 (UTC)

 Not done This is intentional, in order to have a smoother and more readable article. It is done with all older polls, in order to limit the length of the tables. However hiding by default is not equivalent to removing: you can click on "show" and the polls you are interested in will still be there. Yakme (talk) 21:25, 31 August 2022 (UTC)

More Polling aggregation

There are more polling aggregations that should be added such as, Simulatore Elezioni GEHSC Free Weed. CHRpedia (talk) 13:47, 1 September 2022 (UTC)