Based on data gathered by echo-sounder buoys attached to drifting fish-aggregating devices (dFADs) across tropical oceans, we applied a machine learning protocol to examine the temporal trends of tuna-school associations with drifting objects both in comparison to previous studies, and in the context of the ‘ecological trap’ theory. Using a binary output, metrics typically used in the literature were adapted to account for the fact that the entire tuna aggregation under the dFAD was considered. The median time it took tuna to colonize the dFADs for the first time varied between 25 and 43 d, depending on the ocean, and the longest soak and colonization times were registered in the Pacific Ocean. The continuous residence times of tuna schools were generally shorter than continuous absence times (median values: 5-7 and 9-11 d, respectively), in line with the results found by previous studies. Using a regression output, 2 novel metrics, namely aggregation time and disaggregation time, were estimated to obtain further insight into the symmetry of the aggregation process. Across all oceans, the time it took for tuna aggregations to depart from individual dFADs was not significantly longer than the time it took for the aggregations to form. This does not align with what would be expected if the association were ‘strong and long-lasting’ as proposed by one of the aspects of the ‘ecological trap’ theory. The value of these results in the context of the reasons driving the aggregation process is discussed, and further analyses to enrich and make use of this data source are proposed.