The weird world, and equally weird measurement models: Reactive indicators and the validity revolution