Challenges  with Zero Raw Scores

Some norm-derived scores associated with 0 raw scores can be spuriously inflated for young examinees. This is due to the often low variability of the norming sample at very young ages. When most of the very young examinees in the norming sample are only able to answer 0, 1, or 2 questions correctly, the resulting sample variance is very small; the scores don’t “spread” as much as they do at older ages. This low variance translates into very small SD-upper values and SD-low values, which in turn translates to norm-derived scores that may look “too high” for the examinee’s age.

Note that the variance for a given test is going to differ based on the test administered, and the age/grade group you are comparing your examinee's performance to as part of running a report and deriving scores. 

In summary- 

So, an examinee might score 0 on a particular test but end up with an inflated standard score, e.g. a score over 100. In this instance, we do not want users to mistakenly interpret that the child has above-average ability when, in fact, we have very little information about his or her actual ability. The inflated standard score in this case is an artifact of the statistical process of norming.