The concept emergent phenomena tells that the behavior of the phenomena is complex. Simple course and effect relationships cannot explain it. Statistical analysis does not give meaningful results.
Typical rules governing an emergent phenomena contain non-linear relationships, feedbacks and time lags. Simple linear extrapolation does not work.
Think, for example, a kettle of water on a stove and the relationship of the input power and the temperature of the air just above the water level. Nothing happens, when you put the power on. After a while the temperature starts to rise but this rise ends though the power is still on.
If you put a cover on the kettle, the rise ends earlier but the final temperature remains the same.
Models are the way to understand the behavior of emergent phenomenon. To create and validate a model you need to have accurate and relevant data. When you run the model all of the variables of the model must match the observations all the time. If not, you need to correct the model.