What imparts on a pollster a pro-D or pro-R house effect? The samples they use?
Sampling techniques can be one factor. Another is the weighting that they use. No sample is going to be truly representative of the general population in terms of age, race, gender, income, education, etc... So the pollster has to weight their results for what they believe the actual electorate will look like in the current election. Each pollster's weighting is their "secret sauce" and will greatly affect how predictive their results are.
That makes sense, and correlates with some knowledge I've picked up over the years.
For instance - the reason some polls were off in 2016 was because they didn't weigh for educational status, which proved to be one of the big fault lines for the election.
