Standard Score Classification (WJ-V)
| SS Range |
Classification |
Approx. Percentile |
Eligibility note |
| 131+ |
Very Superior |
98th+ |
— |
| 121–130 |
Superior |
92nd–97th |
— |
| 111–120 |
High Average |
76th–91st |
— |
| 90–110 |
Average |
25th–75th |
Within expected range |
| 80–89 |
Low Average |
9th–24th |
Context-dependent — see note below |
| 70–79 |
Low |
3rd–8th |
Typically qualifies as weakness |
| <70 |
Very Low |
<1st–2nd |
Significant weakness |
Low Average note (SS 80–89): Do not automatically label as a weakness. If the student's overall cognitive profile is also Low Average, a score of 85–89 may represent expected performance. Name it as a weakness only when it is meaningfully discrepant from cognitive ability OR ≥1.5 SD below the mean (SS ≤ 77).
RPI Interpretive Bands
| RPI Range |
Proficiency Level |
Instructional Implication (age/grade-level tasks) |
| 100/90 |
Very Advanced |
Extremely easy |
| 98/90 to 100/90 |
Advanced |
Very easy |
| 95/90 to 98/90 |
Average to Advanced |
Easy |
| 82/90 to 95/90 |
Average |
Manageable |
| 67/90 to 82/90 |
Limited to Average |
Difficult |
| 24/90 to 67/90 |
Limited |
Very difficult |
| 3/90 to 24/90 |
Very Limited |
Extremely difficult |
| 0/90 to 3/90 |
Extremely Limited |
Virtually impossible |
Bands come from W Difference ranges, so the RPI endpoints touch (82/90 ends one band and starts the next). Use the label printed on the Riverside Score report for a score that lands on a boundary. The same bands map to CALP levels 6 through 1 when CALP is reported.
Average SS + Low RPI — Why This Happens
The Core Reason
The SS and RPI answer genuinely different questions. The SS tells you the student's rank among peers — they outperformed X% of their norm group. The RPI tells you the student's functional proficiency at the difficulty level where peers are performing at 90%. A student can maintain their rank (average SS) while still finding those same tasks significantly harder than peers do (low RPI). Both facts are true simultaneously — they are not contradictions.
① Compensatory strategies
Older or higher-ability students who have developed workarounds (slowed self-monitoring, phonetic decoding on sight words, rereading, context guessing) may maintain their rank while their underlying automaticity remains genuinely limited. The SS reflects keeping up; the RPI reflects the cognitive cost of keeping up. Classic pattern in students with dyslexia who've had good intervention — they've closed the rank gap but not the fluency gap.
② Low end of the average band
An SS of 90–93 is technically classified as average but sits at the 25th–32nd percentile. The student's W score is meaningfully below the peer criterion, so the RPI can legitimately fall into the Limited to Average band (67/90 to 82/90). The classification says "average" but the proficiency picture is limited. Always report the percentile alongside the classification — it tells a more honest story.
③ Cluster composition masking
A cluster with one high and one low subtest can average to an average composite SS while the individual subtest RPIs reveal that one underlying skill is genuinely weak. The SS hides the split; the individual RPIs surface it. When a cluster SS looks average but the referral concerns don't fit, check individual subtest RPIs — especially in clusters that combine fluency and accuracy subtests.
④ Fluency-sensitive task gradients
On timed measures (Sentence Reading Fluency, Math Facts Fluency, Sentence Writing Fluency), small rank differences correspond to larger proficiency differences because the task penalizes non-automaticity more than rank ordering captures. A student at the 28th percentile on a fluency task may have an RPI well into the Limited range — they are doing the task, but not with the ease peers show.
📝 FIE Documentation Implication
When you see average SS / low RPI, the RPI is often the more honest picture of functional impact. The SS tells you the student isn't falling far behind peers on a normative ranking; the RPI tells you that performing those skills in real instructional contexts is significantly harder for this student than it is for peers. This is one of the stronger arguments for documenting adverse educational effect — the student may look average on a score table but experience the curriculum with meaningfully less proficiency and more cognitive effort. Document both the SS and the RPI in the narrative, and name explicitly what each one shows.
WJ V Nonverbal (NVCAI) — When and How to Use It
Same tests, different doorway
The WJ V Nonverbal is not a new battery. It is seven of the regular WJ V COG tests, with the same items, scoring, basal/ceiling rules, and norms. What changes is how the student learns the task: wordless video demonstrations and standardized gestures in Riverside Score, plus the allowable prompts specified for each test, in place of spoken directions. Riverside describes it as language-minimized, not language-free — the goal is to keep a student's English (or expressive language) from standing in for the ability being measured.
| Score |
Tests |
Time |
| NVCAI |
Matrices, Analysis-Synthesis, Spatial Relations, Block Rotation, Number-Pattern Matching, Letter-Pattern Matching, Visual Working Memory |
~45 min |
| NV Fluid Reasoning (Gf) |
Matrices, Analysis-Synthesis |
~15 min |
| NV Visual Processing (Gv) |
Spatial Relations, Block Rotation |
~14 min |
| NV Cognitive Processing Speed (Gs) |
Number-Pattern Matching, Letter-Pattern Matching |
~8 min |
Visual Working Memory counts toward the NVCAI only; it is a mixed visual working memory/visual processing task. Times are the Riverside Score test-set estimates. Reliability for the NVCAI was reported for ages 6 and up; Matrices and Visual Working Memory start at age 6.
When to consider it
When there is reason to think language demands would lower a cognitive score for reasons unrelated to the ability itself: emergent bilingual students, students with speech or language impairments or developmental language disorder, and students who are deaf or hard of hearing. For a student with no language barrier, the full WJ V COG gives a broader profile and stays the first choice.
Never mix verbal and nonverbal tests in the GIA
Do not give Matrices, Spatial Relations, and Number-Pattern Matching nonverbally and the Gc, Gwm, and Gl tests verbally, then let them combine into a GIA. If language is the barrier, verbal tests can pull the GIA down, and the GIA no longer estimates g for that student. In that situation the NVCAI is the better primary index — the concern is the GIA underestimating, not the NVCAI overestimating.
Fidelity: use only what the platform gives you
Give the video directions, standardized gestures, and the allowable prompts Riverside Score lists for each test — nothing more. Re-showing a demonstration beyond what is specified, pointing out features of the stimulus, or explaining in your own words moves the administration outside standardization, even when it is well meant. Detailed per-test guidance lives in Riverside Score.
Watch the retest effect
In Riverside's study, the second administration ran about 4–5 SS points higher than the first no matter which version came first (at least 2 days apart). If a student just took these tests with verbal directions, re-giving them nonverbally will likely run high from practice alone — do not read that gain as proof of language interference.
Note how the student handled the format
Video and gesture directions are new to some students. Record whether the student seemed to understand what to do and how comfortable they were with the format; uncertainty about the task can lower a score for reasons that are not ability. This belongs in Testing Behavior Observations.
What it does not measure
Gf, Gv, Gs, and visual working memory only — no Gc, Gl, Gr, Ga, or auditory working memory. Report it as an estimate of nonverbal reasoning and visual-spatial processing, not as an overall IQ, and place it within the full evaluation (language history, educational history, observations, other data).
📊 What the validation study found (Johnson et al., 2026)
- Verbal vs. nonverbal directions, 475 students K–12 (171 general education, 149 English learners, 155 with a language impairment; counterbalanced order). NVCAI means differed by less than 2 SS points within every group, verbal and nonverbal scores correlated .73–.84, and the version difference did not depend on group. The Gf and Gv clusters showed no version effect.
- Two small test-level differences, both favoring verbal directions: Number-Pattern Matching (about 2 SS points overall) and Visual Working Memory (about 1.3). Both are smaller than the test SEM. Letter-Pattern Matching was not given in this study because its nonverbal directions mirror Number-Pattern Matching.
- Against the KABC-II NU Nonverbal Index (108 students): r = .76, with nearly identical means.
- NVCAI reliability .96–.97 across ages, SEM about 2 SS points.
- Who was not in the sample: students with AU Level 3, ID, ED, TBI, or DD were excluded, so the comparability evidence does not yet cover those groups.
Read the tables, not just the prose, if you cite this bulletin: the paragraph on Number-Pattern Matching (p. 11) describes the difference as a nonverbal advantage and the group interaction as nonsignificant, while Tables 4–6 show verbal higher and a significant interaction (p = .009).
📝 FIE Documentation Language
[Student] was administered the WJ V Nonverbal, in which directions are presented through video demonstrations and standardized gestures rather than spoken language, because [his/her/their] [English language proficiency / receptive-expressive language skills / hearing status] could otherwise lower scores on tasks intended to measure reasoning and visual processing. This approach is consistent with the requirement that assessments be administered in the form most likely to yield accurate information on what the child knows and can do (34 CFR §300.304(c)(1)(ii)). [Student] [appeared to understand the demonstrated tasks / required the specified prompts on (test)]. The Nonverbal Cognitive Ability Index (NVCAI) estimates fluid reasoning, visual processing, processing speed, and visual working memory; it does not measure verbal knowledge or auditory memory and is interpreted alongside [his/her/their] language history and other evaluation data.
Johnson, E. S., Oh, H., Liu, H., & Roviso, N. (2026). WJ V Nonverbal: Development, validation, and interpretive guidance. Riverside Assessments, LLC.