2019 年 94 巻 p. 13-22
In 2016, media outlets began conducting dual-frame random digit dialing
（RDD） surveys that sample mobile phone numbers in addition to the traditional
targets with landline phone numbers. The resulting greater outreach to hitherto underrepresented youths and mobile-only users has improved the credibility
of polls. Still, new challenges have been identified with surveys targeting
mobile phone users. Examples include a predominant share of males among
respondents and the difficulty in ensuring validity of survey caused by inconsistency
with methods used for surveys that target only landline phone users.
Unfortunately, there are no promising alternatives for opinion polling in case
RDD surveys grow difficult. Worse still, failures in predicting election results
and mistrust of the media have dampened confidence in polls. Going forward,
pollsters must gain people’s trust regarding opinion polls by properly explaining
how they are supposed to be conducted.