As Ganesan points out, we have to do here with distinct works.  The oldest one passing under this title was usually known simply as the Atharvaśiras.  I have made a study of this short work, many years ago, and have been quite delinquent in finishing and getting into print.  In fact I have returned to it since last summer and it will indeed see the light of day very soon.  For the moment, let me say that its textual history is quite complex and quite interesting.  It exists in several recensions that differ markedly.  Something called Atharvaśiras is mentioned already in the Mahābhārata, but the archetype of the text actually transmitted in surviving manuscripts is probably no older than the middle of the first millennium CE, composed in western India, and often included in collections of Upaniṣads.  It also gets called Atharvaśirṣa.

A late development in this history is the appearance of a new recension in Maharashtra, where an extremely corrupt and abbreviated form of the text is transmitted under the name Rudrātharvaśīrṣa or Śivātharvaśīrṣa along with four other wholly distinct, imitative works likewise styled -atharvaśīrṣa but dedicated to Gaṇapati, Nāyāyaṇa, Devī, and Sūrya, which together with Śiva are referred to as the pañcāyatana.

The Gaṇapatyatharvaśīrṣa alone, I think, remains popular today as a devotional text among Smārta Brahmins in Maharashtra, and usually when you mention the name 'Atharvaśirṣa' that is what they have in mind.  It is quite different from the (Śiva-/Rudra-)Atharvaśīrṣa, which itself is a highly divergent form of the old Atharvaśiras.  Note, by the way, that the word śiva is wholly absent in the old Atharvaśiras, and the deity is known as Rudra.  Even in the second section, which consists of a litany of names of God, the name Śiva does not appear except as a later addition.

I am just now finishing editing and translating Nārāyaṇa's commentary on the text; the edition and translation and study of the text are already complete, so the thing should finally appear soon.

Tim