D. People who expressed no party preference during the in-person interviews shortly before Country X's last election did not outnumber the people who expressed a preference for the Conservative Party.
This doesn't strengthen our argument. If we knew the opposite i.e. people who expressed no preference actually outnumbered the people who preferred C, it might lend some support to our argument. It would seem then that people who prefer R are many but they do not speak up in in person interviews. So telephonic interviews may be better.
Say if 100 people were interviewed and 40 favoured C while 60 favoured R. The poll will indicate that R will win.
But what if out of 60 of R, 30 do not turn up to vote while all Cs turn up to vote? Then actually C will win.
~ Ellaborating
KarishmaB reply From my understanding.
When C outnumbered NoneAcutal Poll Result
Case 1: Total People=100, C=35, R=55, None=10
Survey C = 35, R= 25, None = 5
In this case, there is no need of telephonic survey, we just have to increase the survey user base to arrive at the required result. This can be achieved by either In-person or Telephonic interviews.
When Data set is changed ~
No party preference outnumber the C supporter.Acutal Poll Result
Case 2: Total People=100, C=35, R=55, None=10
Survey C = 35,R= 5, None = 25
Now, for C to win election some person who are actually supporting None in the survey lied in the survey and are actually supporting C. Therefore to determine actual number, In-person survey is inaccurate and telephonic survey might be helpful to determine the accurate poll results.