Considering the Calvinist position on limited atonement, Millard Erickson, a “moderate Calvinist,” writes the following,
“We conclude that the hypothesis of universal atonement is able to account for a larger segment of the biblical witness with less distortion than is the hypothesis of limited atonement.” 
This is an interesting observation confirming that coherence and incoherence are indicators of valid and invalid interpretation. Erickson calls this “distortion.” Note that he is also affirming the interpretive criteria of comprehensiveness or explanatory power and explanatory scope when he talks about universal atonement being able to “account for a larger segment of the biblical witness.”
Note Erickson’s hermeneutic. It includes the principle that our theological conclusions must “make sense” in relation to other portions of “the biblical witness.” Furthermore, the biblical scope of that coherence is an important consideration. What we propose to be true of one text must “account for a larger segment of the biblical witness.” And it should not do so with “less distortion.” Erickson obviously considers it important to identify this “distortion” or incoherence, and he feels he can do so reliably. “Distortion” should not be explained away or summarily dismissed as unimportant. Logical and moral coherence across the breadth of Scripture is guiding Erickson in determining what is most likely the biblical truth regarding the scope of the atonement. That which causes the least amount of “distortion” given the whole testimony of Scripture is the better interpretation.
I just wonder why Erickson, as a Calvinist in all other respects aside from the atonement issue, doesn’t apply these same hermeneutical principles to his other Calvinist doctrinal conclusions and beliefs. It’s odd that he does not see that this same principle of “less distortion” applies to his other doctrines of “total inability,” unconditional election” and “irresistible grace.”
 Millard J. Erickson, Christian Theology, (Grand Rapids: Baker, 1985), 835.