Which of these is NOT a guideline for constructing a frequency distribution?

When constructing a frequency distribution, one of the key guidelines is that intervals should not overlap. Each data point should belong to one and only one interval to maintain clarity and ensure accurate counting. Overlapping intervals can lead to confusion regarding which interval a data point belongs to, ultimately skewing the results and making the distribution less useful for analysis.

Choosing logical intervals is essential because it helps in effectively grouping data in a way that reflects the underlying distribution. This helps in identifying patterns and trends. Counting cases in each interval is a fundamental requirement. It allows for the construction of a frequency distribution that accurately represents the data's distribution.

Including all data is also a crucial guideline as every piece of data in the dataset has a role in understanding the overall distribution. Omitting data could result in an inaccurate representation.

Thus, the choice indicating that intervals must overlap is the only statement that does not conform to the guidelines for creating an effective frequency distribution.