Testing and evaluations bring up many emotions in parents and children alike. Used well, a comprehensive evaluation provides a measure of the child’s progress and a road map for the way forward. But how are you supposed to untangle the web of jargon and questions surrounding your child’s testing? Let’s discuss…
Criterion-Referenced vs. Norm-Referenced
Criterion-referenced tests measure whether or not the test-taker knows certain material or can perform a certain task. For young children, these tests often look like checklists (does your child know the following words: dog, duck, bear, sheep, etc.). The test-taker’s score is compared to a correct answer key, not to other test-takers. Criterion-referenced assessments let us know what the child is doing based on his own knowledge, not in relation to his peers or other test-takers. Scores from these kinds of tests might let us know things like, “[Child] is demonstrating 15/25 skills on this Auditory Development Checklist” or “[Child] is performing all of the tasks in the 12-18 month age range” or “[Child] correctly identified 85% of the sight words on this list.”
Norm-referenced tests, on the other hand, compare the child’s performance to an average sample of the population, and compare the child to others who have taken this test. Scores from these kinds of tests sound more like, “[Child] performed in the 50th percentile on language tasks when compared to age-matched peers.” (Don’t worry if this sounds like confusing jargon — keep reading, I’ll explain it in a minute!)
Questions to Ask about Testing
What does this test assess? What exactly are we testing here? There are more commercially-available tests on the market than I can count (or than you would like to read about), but before any professional gives your child an assessment, they owe it to you (and the child, if he/she is old enough to understand, given an explanation at their developmental level) to give a thorough explanation of what the test is about. Does it measure language (grammar, vocabulary, pragmatics/social language), reading/writing skills, auditory abilities, overall development (gross and fine motor, self-help skills, cognitive development), etc.? Is the test focused on receptive (understanding) or expressive (using) language abilities? For the data we gather from a test to be meaningful, first we have to be sure what exactly we’re trying to measure.
A sample explanation to the parent(s) would look like this:
“Today, we are going to do some testing to do a test called the EVT-2, the Expressive Vocabulary Test, second edition. It’s a test to measure Leah’s expressive vocabulary skills. That means that this test will look at how many words she uses on her own in her speech. I’ll show her a picture, and ask her to label it. The score will let us know how the number of words she knows and uses compares to other children her age with typical hearing and development.”
A sample explanation to the child would look like this:
“Today, we’re going to look at some pictures in this book. I will show you the picture and you can tell me the name of what you see. If you don’t know, it’s okay to guess. Just do your best.”
Who is the sample group used to norm this test? For a norm-referenced test (see above), the standards for scoring are set based on measures of a sample population. This sample group is supposed to be representative of the larger population of people who will take the test in terms of geographic location, race/ethnicity, age, ability, etc. Most tests are normed to try to be based on as “average” a test-taker as possible to eliminate bias. Some tests are normed specifically for children with hearing loss, but I would avoid those. Why? Well, quite frankly, I don’t want the children in my practice to be doing well for a child with hearing loss. I want them to be doing well. Period. Our children with hearing loss are not going to be attending school just with other children with hearing loss, and they’re certainly not going to be competing for college admission or for jobs in the real world only against other people with hearing loss, so I want to know how their skills stack up against the general population. After all, age-appropriate speech, language, and listening skills are our goal.
What about bilingual children? Some tests are normed for children who speak more than one oral language, most often English and Spanish. Others were not normed on children who speak two languages, so we must interpret the scores with caution. A general rule of thumb is: if the child was not represented in the sample group, you must take their scores with a grain of salt, because in those cases we are comparing apples and oranges, not apples to apples. You can still administer these standardized tests to bilingual children, but you may gain more information from the raw score (does the number of items correct increase over time?) and an item analysis (see below) than you do from standard scoring.
What will we do with the information gained from this test? What’s the point? How is the information we discover here going to affect my child’s therapy goals or educational program? How will we use this data to keep moving him forward?
Does this test account for dialectical differences? As I discussed in this article, there are actually many different ways to speak English (or any other language), that are just as rule-governed and “correct” as the standard dialect. For a child from a home where African-American (Vernacular) English (AA(V)E) is the norm, saying “the girl hair” instead of the Standard American English (SAE) “the girl’s hair” is an equally correct possessive form, and the child should not receive points off for that response. Evaluators need a firm grasp of overall child language development as well as language development in special populations and dialects to ensure that they are not mistakenly under- or over-scoring children for perceived “mistakes” that are actually perfectly acceptable in their community. Now that our field has grown in our understanding and research of dialects (learning that they aren’t, in fact,
The majority of norm-referenced assessments are scored around the “Normal Curve” or “Bell Curve.” Shaped like a bell, fattest in the middle, it represents that fact that, for almost any human behavior, the vast majority of people perform in the average range, with fewer and fewer people in both the most-delayed and most-advanced groups. A score of 100 is the mean (more on this below), and then we take fifteen point chunks away from the mean in either direction to be our standard deviation (amount away from average).
Most evaluations yield several different kinds of scores and information about the test-taker’s performance. Some might be:
Raw Score: this is the number of questions the test-taker got right. Think of it as the number of “points earned.” For children who are bilingual or very delayed (so delayed that a standard score cannot be determined), comparing changes in raw score over time may be another way to measure performance.
Standard Score: This is the score of how the child did compared to the sample population. The mean score, let’s think of this as “the most average child in the world,” would score 100. But, we know that that child does not exist. For all eight-year-olds, for example, there is a range of what is considered normal. If you took a class of third-graders, about 50% would be reading at a third grade level, 25% below that, and 25% above. Thus, standard scores on tests have an “average range” of 85-115. CAUTION: I have heard professionals tell parents that children with hearing loss scoring an 85 on standardized assessments means they’re perfectly ready to compete in the mainstream and scoring “on par with hearing peers.” This is 100% absolutely not the case. Barely scraping by is a hard, hard way to go through school. “Just okay” is not okay, and don’t let anyone tell you differently. Scoring an 85 means you’re barely scraping the bottom range of what is considered normal compared to your hearing peers… and that’s in an ideal, sterile testing environment. Add in a noisy classroom or a teacher who speaks too softly, and you have no wiggle room before you fall out of “normal” and into “delayed.” Barring other disabilities, learning disorders, or cognitive issues, I want children in my practice to be scoring 100-115+ on these assessments. This gives them the wiggle room they need to compete. Their skills are good enough that they have a few points “to spare” to make up for the many less-than-ideal listening situations that we know exist in the real world.
Standard Deviation: As explained above, for a normal curve, this represents fifteen point chunks away from the mean in either direction. Scoring one standard deviation away from the mean +/- creates that 85-115 “average range.” We call this scoring WNL, Within Normal Limits. Anything greater than one standard deviation above or below 100 represents exceptional performance.
Range: Each standard score usually comes with a range of a few points above and below that also represent the child’s performance. For example, his SS might be 92 with a “confidence interval” of +/- 5 points, meaning 87-99. This accounts for the fact that all of us perform differently from day to day. Maybe the child is just recovering from a cold, or maybe he’s really “on” today. Either way, we know that behavior changes, so this range represents our best guess as to where he would fall on any given day. Make note of this, too, in light of what’s described above. If the confidence interval on your child’s score puts him sometimes below that 85-115 range, even if his score on testing day was, say, an 89, then there’s work to be done!
Percentile: A percentile score tells us where the child falls in relation to his peers. For example, a child in the 99th percentile means that 99% of other children his age will score below him.
Age- and Grade-Level Equivalents. Many tests will also translate the child’s performance score into an age- or grade-level equivalent. Caution: While these may be somewhat helpful, it’s important to take them with a grain of salt. As discussed above, there is a wide range of what is considered “normal development” for children at each age and grade. Additionally,
What to Consider When Testing Children with Hearing Loss
As discussed above, I think it is important to assess children with hearing loss compared to their hearing peers, not compared to children with hearing loss. Here are some other things to think about regarding children with hearing loss and assessments:
Test the child under his best conditions. Are the child’s hearing devices working? Are they working up to the gold standard? If not, it’s not a fair testing situation. It’s also best to assess children when they’re happy, healthy, well-rested and well-fed. We all do our best under those conditions. Additionally, when I assess children with hearing loss, I do not attempt to use the “hand cue” or visual distraction techniques common in AVT to prevent the use of speechreading. I want the child to have all information possible (even though most children raised with AVT are so auditory they don’t tend to search for the speaker’s mouth anyway). I also want to sit on the child’s best aided side, if there’s a difference between ears, and conduct the test in an acoustically-friendly environment. If the child uses and FM system, use it for the test, too.
Consider chronological age versus hearing age. While we want to make quick progress to close the gap between chronological age and hearing age, it’s also important to look at test scores in light of how long the child has had good access to sound AND access to good listening and spoken language intervention. A four-year-old child who has had undetected hearing loss for the first three years of her life may very well score well below her chronological age on her first round of testing. This does not necessarily indicate some great cognitive problem, just a lack of time, experience, and access. Chronological age versus hearing age should not be an excuse for low scores or poor progress, but rather an explanation that helps us better understand where the child is presently and how we need to work to catch up.
What to Look for in Your Child’s Testing
Growth over time. Is your child making more than one month’s progress per month? That’s what is needed to close the language gap and help her keep pace with her hearing peers. Most tests are not valid if administered too frequently, but comprehensive testing every six months should reveal more than six months’ progress. If this isn’t happening, it’s time for a change! See THIS article for tips.
Are there gaps or holes? Does your child excel at receptive (understanding language) tasks but lag behind in expressive (using language) abilities? While it is normal that receptive language develops before expressive language (for example, you have to understand the word “dog” before you begin to use it on your own), too big of a gap, or a gap that persists over time, is a red flag.
Item analysis. A good evaluator will not only give you the scores, but also an analysis of the items your child answered correctly and those that she missed. Many tests are actually comprised of sub-tests, with each one measuring a different area (e.g. one measuring understanding of basic concepts, one measuring the child’s ability to follow auditory directions, one measuring the child’s vocabulary, etc.). An analysis of these subtests can let you know specific areas where your child can improve. Item analysis will let you know weak and strong points and help you and your therapy team design goals to patch up these holes!
A comprehensive evaluation. A comprehensive evaluation does not make a determination about your child based on just one test. Instead, we want to look at all aspects of the child’s development… and from all angles! Standardized tests should cover all areas of speech, language, listening, and academic skills in depth. Checklists of listening and social skills completed out by parents and teachers add to the picture. A portfolio of the student’s academic work can fill in the picture of how the child is using her skills in the classroom. All of these aspects are important and should be considered when making determinations about child goals or placement.
Results-based goals. If the child missed something on the test, I want a concrete goal written to address it. It’s our own fault as parents and professionals if we see a child make the same mistake twice without trying something different to fix it!
Though Auditory Verbal Therapy is all about parent coaching and helping parents learn to be their children’s first and best teachers, testing is one time when the therapist needs to take the front-row seat. To preserve the validity of the test, we cannot give the child any hints, clues, or help to complete the tasks. After all, we want to know what the child knows on his own, not what he knows with mom’s help, or, in situations extreme parental interference, what mom herself knows. Even praise during a test has to be carefully constructed. It’s okay to say neutral comments like, “Good working!” or “I like how you’re concentrating” or “Thanks for being patient!” but giving evaluative feedback, like, “You got it!” or “That one wasn’t right” throws off the validity of the test. As hard as it is, in testing situations, we have to stay quiet.
This doesn’t mean that parents are out of the picture during testing, though. As I mentioned above, a good professional makes sure that both parents and child understand the purpose and procedure of the testing and explains the results thoroughly in a way that is parent-friendly and easy to understand. In addition, while I can’t have parents help their child on the test, I do like to ask questions afterward, like, “Do you think he would have known that at home?” or “Do you think his performance today is equal to what you see in real life, better, or worse?” These parental observations don’t influence the final score, but they do give perspective on the child as a whole, and help us evaluate his performance in context. Remember, children don’t develop in a vacuum!
We often think of basic audiological testing as a “hearing test,” the audiogram graph you’re probably familiar with seeing (read HERE for more information on how to interpret your child’s audiogram), but there are actually a variety of other tests that your audiologist may perform. In terms of listening and spoken language development, the most important of these are the speech perception tests, which measure the child’s ability to understand speech with and without his devices in quiet and in noise. There are many different tests in this category. Some measure the child’s ability to point to the picture he hears from a set of four, others test the ability to repeat words and sounds, others the ability to hear and repeat sentences in noise. If your audiologist is not adding these tests to your child’s regular appointments, ask for them! Hearing tones on the audiogram is great, but it doesn’t tell us how the child is actually hearing what’s most important — speech!
Other Types of Testing
Sometimes, a child can benefit from calling in someone outside the realm of listening and spoken language. Often an occupational therapist, developmental neuropsychologist, reading specialist, or other professional can provide an added dimension that helps us understand the child’s needs, especially if the child has additional disabilities outside of the hearing loss.