In the aldol/cross aldol reaction, of $α,β$-unsaturated carbonyl compound, under what circumstances does a 1,2-addition i.e. aldol condensation (carbanion formed directly attacking carbon of carbonyl group) and 1,4-addition i.e. Michael condensation (carbanion formed attacking $β$ carbon)?
These are the following points I considered:
Stability of resulting product.
The amount of conjugation in the compound. What I feel, extended conjugation ($γ,δ$-unsaturated double bond as well) would always lead to 1,4-addition.
The group attached to the $β$ carbon. For example, an electronic withdrawing group would increase the probability of 1,4-addition and an electronic donating group would reduce the probability, favoring a 1,2-addition.
The steric crowding around the carbonyl or the unsaturated carbon system.
Are these arguments valid?
I have added a few examples (I have just written the final product and not the mechanism as my question lies in the very step after generating carbanion). In examples 1, 2, and 4, aldol condensation (1,2) has taken place but in example 3, Michael condensation (1,4) has taken place.
Is there a way we can account for this?