No. Just as a clock that is not working shows the time more accurately than the most accurate clock twice a day, non-scientific research sometimes gives accurate results. However, the results are far more likely to be inaccurate this way.
The most common forms of non-scientific research are telephonically conducted by television programs and self-selection research on the Internet. These two ways have two disadvantages.
First, their samples represent self-selection, or rather, they are arbitrary in relation to the one who implements them. Such surveys tend to attract individuals who are biased toward the subject of the survey, rather than providing a representative sample.
Second, such surveys rarely collect additional information (such as age and gender) that would allow for some judgment about the sample.