Jump to content

Talk:Opinion polling for the next Italian general election

Page contents not supported in other languages.
From Wikipedia, the free encyclopedia

local regression

[edit]

The local regression is strange. It clearly gets skewed around that big time gap in data (ca 1 month, June to July 2024). It looks like the regression is performed over a set number of data points, irrespective of when those data points were taken. The resulting trend line is misleading. 82.4.208.61 (talk) 19:31, 21 November 2024 (UTC)[reply]

to be clearer: all parties are essentially flat at the moment, whereas the local regression shows them on the rise or falling. The June-July gap is around the european elections. Just before and just after elections, it is common to see big jumps. The regression should be adjusted to better reflect that, IMO. This could be done by averaging over a set time-range, rather than just a number-of-data-points range. 82.4.208.61 (talk) 19:34, 21 November 2024 (UTC)[reply]