Hacker News Reader

Odds at online bookmakers and betting exchanges still rate a "No" vote as around twice a likely as a "Yes" vote.

http://www.oddschecker.com/politics/british-politics/scottis...

As well they should. This one poll gives the appearance of a trend towards Yes and a dramatic close finish, but Yes have always been behind by a fairly consistent margin [0]. Just so happens the odd poll (~2/50 now) finds a Yes majority by crook or by hook (e.g. SNP commission a poll) and it makes a nice news piece. I did a quick analysis of that polling data should anyone be interested [1].

[0] http://ukpollingreport.co.uk/scottish-independence-referendu...

[1] http://blm.io/blog/scottish-independence-polls/

Phenomenal work, well done. I was slightly disappointed that the (beautiful) blog post was not authored in R Markdown. ;-)

One comment, your analysis only takes into account data up to July of this year. It is entirely plausible that the recent debates could have affected the polls dramatically (there’s precedence for this, prime example being the Kennedy–Nixon debate), something that the BBC data is consistent with.

Reply on news.ycombinator.com