There are cases where small differences are critically important, and the larger context might obscure these differences. In these scenarios, truncating the Y-axis can be helpful for emphasizing the variations in the data that matter. For instance, if you're tracking minute changes in a vital medical reading, it's essential to be able to see those fluctuations clearly.
Scientifically speaking, the representation of data should always prioritize clarity and accuracy. It's neither universally right nor wrong to truncate the Y-axis. Instead, the decision should be based on the specific use case, the intended audience, and the importance of the message the data is meant to convey.
The best practice is to always be transparent about how data is visualized. If a Y-axis is truncated, it should be evident to viewers, and the reasons for doing so should be justifiable based on the data's context and importance.