I posted earlier that 'professional survey takers' are not necessarily a bad thing. As long as they're who they say there are and are reading and answering your questions fully, you have nothing to worry about.
When you do have to worry is if the panelist is not who they say there are. For example you have a retired single male posing as the mother of a family of 4. Fortunately, the online panel companies have done a good job in the last 2 years to make sure that panelists are who they say they are. It isn't fool proof, but it is much more reliable than some think (see the ARF study which I refered to in my previous post).
The other major issue is whether your survey takers are answering your questions fully. Are they just skimming through the survey to get it done as quickly as possible?
Here are two very easy ways you can make sure that respondents are reading the questions you ask (and are answering accordingly):
a) for one of your questions, simply tell the respondent to 'Answer 'Slightly Agree' in the question below'. If someone hasn't answered 'Slightly Agree' to that question, pull their entire response set from your data - you can't be sure they read any of the questions.
b) ask the same question twice. It has to be an obvious question so that respondents couldn't reasonably answer in different ways. (Age, number of children, etc). Again, if it's not consistently answered, pull that respondent's entire data as you can't rely on their responses.