Which level of measurement would best define data collected by a survey with only yes or no responses?

The best level of measurement for data collected by a survey with only yes or no responses is nominal. Nominal data is used for labeling variables without any quantitative value. In this case, the responses of "yes" and "no" categorize the participants' views or behaviors but do not imply any order or ranking. Since the responses are merely different categories without any numerical implications, they fall under the nominal scale.

In contrast, ordinal data involve categories that have a meaningful order or ranking but do not quantify the difference between them. Interval data provide a scale where the difference between values can be meaningfully measured, including a zero point, which is not applicable here. While nominative is not a widely recognized level of measurement in this context, nominal is the appropriate term for surveys that classify responses without implying any ranking or order.