Under the hood, Brexit Polls are unreliable

After the fiasco of the Opinion polls during the General elections, it appears British Pollsters continue to have problems with their Brexit Opinion Polls. 

I chose 3 polls conducted during the last one week for my analysis. They are YouGov, Opinium and ORB. All 3 of them are online polls and have been weighted to ensure fair representation of all demographic groups.

1. All 3 polls show a narrow range between 43-45%, in itself this is quite reassuring. However, we saw similarly low variances during the General elections. At the end, every one of the polling firms missed the Tory vote share by a big number. Secondly, the ORB poll appears to have extrapolated for Don't Knows and is projecting a large % for 'Leave'

2.  Analysis by Age : The General election polls threw inconsistent data at Age, SEC and Region level. These inconsistencies continue in these polls. For example, the gap between Opinium and the rest in Scotland is preposterous (19% points). Amongst voters in London, the gap between YouGov and ORB is quite large (11%) . While the gap does narrow at England+Wales level, the large gaps under the hood across the three polling firms raise serious concerns about sampling and reliability of these surveys.

3. SEC analysis:There is a substantial gap amongst less affluent voters between the two polls. 

When one looks at Age wise analysis, all the 3 polls show very different numbers amongst voters in the 18-34 age band. 

While all the polls appear to be predicting a close race, there are substantial gaps between polls when one looks at region/age and SEC level. These gaps between various polls also existed during the General elections and cannot be easily explained by margin of error. These gaps in my view put a question mark to the overall remain/leave numbers put out by pollsters. I would strongly recommend users to use this polling data with great caution.