Can Weighting Compensate for Sampling Issues in Internet Surveys?