Commenting on the need for textual criticism (a careful analysis of the text) of the Hebrew Bible, Tov explains:

The biblical text has been transmitted in many ancient and medieval sources which are known to us from modern editions in different languages: We now have manuscripts (MSS) in Hebrew and other languages from the Middle Ages and ancient times as well as fragments of leather and papyrus scrolls two thousand years old or more. … All of these textual witnesses differ from each other to a greater or lesser extent. Since no textual source contains what could be called “the” biblical text, a serous involvement of biblical studies clearly necessitates the study of all sources, including the difference between them…¹

Not only are there differences in all ancient manuscripts, but also in our Bibles that have been published in modern times:

Textual differences are also reflected in modern editions of the traditional text of the Hebrew Bible, the so-called Masoretic Text…, since these editions are based on different manuscripts…Similar discrepancies between the various ancient witnesses are even reflected in the modern translations.

One would not have expected differences between the printed editions of the Hebrew Bible, for if a fully unified textual tradition would have been possible at any one given period, it would certainly seem to be so after the invention of printing. Such is not the case, however, since all the editions  of the Hebrew Bible, which actually are editions of [the Masoretic Text], go back to different medieval manuscripts of that tradition, or combinations of such manuscripts…, so that the editions also necessarily differ from each other. Moreover, these editions reflect not only the various medieval manuscripts, but also the personal views of the different editors. Furthermore, each edition contains a certain number of printing errors. Therefore, there does not exist any one edition which agrees in all of its details with another… Some editions even differ from each other in their subsequent printings…without even informing the readers.²

Differences in the different versions of the text go from structural differences in chapter and verse divisions to more significant variations in letters and whole words (again, this is in the basic Masoretic Text of the Hebrew Bible). Tov notes that these differences go back to the differences that can be seen in the medieval manuscripts that are behind the modern versions.  He then goes on to give some reasons behind these differences. The first reason for variations is that they are caused by textual “corruptions” or “mistakes” made in the transmission of the text.

Most of the texts–ancient and modern–which have been transmitted from one generation to the next have been corrupted in one way or another. For modern compositions the process of textual transmission from the writing of the autograph until its final printing is relatively short, so that the possibilities of its corruption are limited.8 In ancient texts, however, such as the Hebrew Bible, these corruptions (the technical term for various forms of “mistakes”) are found more frequently because of the difficult physical conditions of the copying and the length of the process of transmission, usually extending until the period of printing in recent centuries. The number of factors which could have created corruptions is large: the transition from the “early’ Hebrew to Assyrian (“square”) script…, unclear handwriting, unevenness in the surface of the material (leather or papyrus) on which the text was written, graphically similar letters which were often confused…, the lack of vocalization…, and unclear boundaries between words in early texts…, etc.

The above corruptions to the text are accidental. Tov discusses another reason for textual variations, but these are not accidental.

A second phenomenon pertains to corrections and changes inserted in the biblical text. In contradistinction to mistakes, which are not controllable, the insertion of corrections and changes derives from a conscious effort to change the text in minor and major details, including the insertion of novel ideas. Such tampering with the text is evidenced in all textual witnesses…, including [the Masoretic Text]. Tradition ascribes to the soferim, “scribes,” 8, 11, or 18 such “corrections” in [the Masoretic Text] itself…, but even if these transmitted corrections are questionable, many other similar ones are evidenced elsewhere…
Corruptions as well as various forms of scribal intervention (changes, corrections, etc.) are thus evidenced in all textual witnesses of the Hebrew Bible, including the group of texts now called the (medieval) Masoretic Text as well as in its predecessors, the proto-Masoretic texts. Those who are unaware of the details of textual criticism may think that one should not expect any corruptions in or any other sacred text, since these texts were meticulously written and transmitted. Indeed, the scrupulous approach of the soferim and Masoretes is manifest in their counting of all the letters and words of [the Masoretic Text]… Therefore, it is seemingly unlikely that they would have corrupted the text or even corrected it. Yet, in spite of their precision, even the manuscripts which were written and vocalized by the Masoretes contain corruptions, changes, and erasures. More importantly, the Masoretes, and before them the soferim, acted in a relatively late stage of the development of the biblical text, and before they had put their meticulous principles into practice, the text already contained corruptions and had been tampered with during that earlier period when scribes did not as yet treat the text with such reverence.  Therefore, paradoxically, the soferim and Masoretes carefully preserved a text that was already corrupted…
The preceding analysis has surmised that [the Masoretic Text], too, contains occasional errors. In our analysis of the witnesses of the biblical test no exception is made in this regard for [the Masoretic Text], because that text, like all other texts, may have been corrupted in the course of the scribal transmission.³

These types of textual corruption or even found in many of the very oldest copies of the biblical texts that we have — the versions found among the Dead Sea Scrolls.  These biblical texts from Qumran are sometimes more like the Septuagint and older Greek texts than the Masoretic texts (see, e.g., Exod., Sam./Kings, minor prophets); sometimes have multiple versions in Hebrew, Greek, and Aramaic which do not match each other (see, e.g., Lev., Deut., Josh., Sam./Kings, Ezek.); and sometimes there are longer and shorter versions of the same text (see, e.g., Judg., Jer., Psalms). The Psalter at Qumran is not complete and includes extra psalms not found in our Bible.

In summary, Tov explains:

It has become clear from the preceding paragraphs that one of the postulates of biblical research is that the text preserved in the various representatives (manuscripts, editions) of what is commonly called the Masoretic Text, does not reflect the “original text” of the biblical books  in many details. ((Ibid., 11))

1. Tov, 2; bold emphasis mine
2. Ibid., 3
3. Ibid., 9; italics in original, but bold is mine