DaveNewcastle
Established Member
Exactly!. . . . the omission of a "None" button will lead to highly false results as respondents have to put something and introduces the coin toss problem and a massive potential shift within the results, as coin toss theory rarely holds, resulting in a large skew of data and wasting a lot of your time later.
Ie. A respondent wants to put, "None of the Above" or "Change Nothing" but without being provided that button, will randomly stab any buttons to complete the survey to carry out their moral obligation to do so and/or gain a sense of satisfaction for helping you by completing the questionnaire; this will happen a lot more than you think, resulting in a massive bias and utterly useless results, that WILL be shot down by anyone reading the report that has any sense at all between their ears.
The results WILL be meaningless for that very reason.
A little more intellectual rigour is required to render the results of an open ended survey meaningful.
A broader cross-section of respondents would help too, but one particular sub-section would seem to be necessary to provide useful outcomes: i.e. on-board surveys of the First Class passengers on a few long-distance TOCs whose preferred style of travel is ostensibly being analysed.
Without asking the users, and identifying their/our responses, then we're not revealing the opinions of the people whose market choices would be affected by the proposals.
But as Nym says, the biggest distortion of any survey dataset is the lack of a 'none of the above' option. Basic stuff.