Weinert & Miller (1996) suggest that English wh-clefts are a heterogeneous class in that they can have varied degrees of structural integration. Many such constructions depart structurally from the canonical wh-cleft which consists of a wh-clause, the copula and a focus constituent, and in which all the three elements are brought together into a fully integrated utterance. In the types of wh-clefts displaying looser structure, their lack of syntactic integration has so far been related to such linguistic features as (a) omission of the copula, (b) non-canonical copular complementation, e.g. independent main clauses instead of standard infinitival phrases appearing in the focus constituent, (c) lack of a clearly identifiable copular complement, (d) the focusing effect of the wh-clause extending over several clauses (Weinert & Miller 1996; Koops & Ross-Hagebaum 2008; Hopper & Thompson 2008; Callies 2012). Although the disintegrating effect of these features has been observed, the extent of the phenomenon in modern English has not been properly established and other non-integration features have not been investigated. Therefore, the purpose of this study is to use corpus data to identify and examine such non-integration features and to investigate the extent – expressed quantitatively – to which these features are found in wh-clefts with the verb do in the wh-clause. The article also points out the formulaic status of those wh-clauses which become disconnected from their focus phrases.