Reading-level testing is often used to place students within a reading series, usually for instructional purposes.  While this might sound like diagnostic testing and a good idea, in reality, it is not.  The reason is that such testing uses leveled passages on a diverse population of students to determine a "level" based on only a single measure.  For example, if students are English language learners and leveled vocabulary words are embedded in the passages, there will be a bias against these students even though they may be able to understand complex sentence patterns and topics.  Likewise, the nature of the passages will affect student achievement.  Student interest, passage length, paragraph length, and type of literature are all specific variables used to determine a general "reading level" that teachers or parents often apply to all reading materials.  Therefore, for a true diagnosis, a profile is needed which can only be obtained by looking at multiple reading measures.  Only a full diagnostic battery can provide this profile and thus insight into student achievement and knowledge.  

For instance, at the minimum, a diagnostic battery that looks at decoding, vocabulary, and comprehension would provide a more complete picture.  By using multiple measures, you can scientifically isolate skills and eliminate confounds.  Successful reading and answering of questions on a "reading level" test only gives you a positive when a student has all skills working together.  Thus, any single deficit skill will work to bring the results down significantly.  Furthermore, the teacher has no idea why the student did poorly, only seeing that he or she performed below his or her grade level in reading.  Giving the student books at this low grade level isn’t necessarily going to help him or her remediate.  The student needs to work on the specific skills that may be low.