It really should. I'm a PhD student (which does not overly qualify me anyway) but the more I look into works with quantitative measuring methods, the more I have a hard time trusting polls and statistics in general. Even in academic papers, or at least the ones I reviewed, more often than not the data is massaged in some way and leaves a lot to be desired.

It's best to immediately get suspicious if a polling company is owned by some parent firm with a clear conflict of interest.