What Is PB Max in Audiology and How Is It Measured?

PB Max is the highest word recognition score a person can achieve on a phonemically balanced word list when the volume is set to the most favorable level. In clinical audiology, it represents the ceiling of your ability to understand speech under ideal listening conditions, and it is one of the most informative numbers on an audiogram. The score is expressed as a percentage of words correctly repeated, and comparing it against the degree of hearing loss on a standard tone test reveals whether your difficulty understanding speech is proportional to how much sound you are missing or whether something else is going on.

The Performance-Intensity Function

To understand PB Max, picture a simple graph. The horizontal axis is intensity, the volume at which words are played to you. The vertical axis is the percentage of words you get right. For most people, the curve starts low on the left (words are too quiet to catch), rises steeply as volume increases, and then flattens out into a plateau once the speech is loud enough. The score at that plateau is your PB Max. It does not keep climbing forever because once all the speech information is audible, turning the volume up further does not help.

In a textbook-ideal test, the audiologist would map out the entire curve by presenting word lists at several different intensities, plotting the results, and reading off the peak. That gives the most precise estimate. In practice, time constraints mean most clinics try to hit the plateau in one shot by choosing a single presentation level thought to land near the peak. Research has shown that for people with primarily cochlear hearing losses up to about 50 dB, testing at a level selected by an adaptive tracking procedure or at 95 dB SPL can serve as a reasonable single-level estimate of PB Max.

Computer simulations of one commonly used adaptive approach suggest that at least 50 words should be presented to reliably estimate the level at which PB Max occurs before measuring the actual recognition score at that level, and that a 20 dB intensity difference between paired presentations works well for zeroing in on the peak.

What “Phonemically Balanced” Means and Why It Matters

The “PB” in PB Max stands for phonemically balanced. A phonemically balanced word list is designed so that the speech sounds in it appear in roughly the same proportions as they do in everyday conversation. If the language you speak uses the “s” sound twice as often as the “th” sound, the word list should reflect that ratio. The idea is to make the test a fair sample of real-world listening rather than something skewed toward easy or hard sounds.

Building these lists from scratch for a language that does not already have one is a substantial research project. A Tamil-language effort, for example, started with 500 two-syllable words, had 40 native speakers rate each word for familiarity, and then passed the survivors through expert review by audiologists, speech pathologists, linguists, and psychologists. Only 270 of the original 500 words made the final cut, and those were divided into four lists of 25 words each.

An Arabic-language project took a different approach, working with monosyllabic consonant-nucleus-consonant words and constructing eight lists of 26 words each, arranged so that each consonant appeared once at the beginning of a word and once at the end within each list.

⁴

The most widely used English-language lists, such as the CID W-22 and the NU-6, have been around for decades. They use monosyllabic words like “ship,” “ranch,” and “growth” and come in sets of 50 or 25. Whether you are tested with 25 or 50 words affects how precisely the score can be interpreted statistically, a point covered further below.

Choosing the Right Presentation Level

One of the less obvious decisions in measuring PB Max is how loud to play the words. The goal is to find the level where you perform best, but there are several shortcuts clinics use, and they do not all produce the same score for every type of hearing loss.

A study comparing five common presentation-level methods found that for people with mild, gently sloping hearing loss, all five methods produced essentially the same result, so the choice did not matter much. For people with moderately severe to severe losses, though, only one method consistently reached the true maximum: presenting the words at 5 dB below the uncomfortable loudness level. The other four methods underestimated PB Max for that group.

⁵

This has real consequences. If the words are played too softly for someone with a significant hearing loss, the score underestimates their actual speech understanding ability. That could lead a clinician to recommend more aggressive intervention than necessary or to overstate the severity of the problem. In large-scale studies, PB Max scores are often measured at 30 dB above the speech reception threshold or at the most comfortable listening level.

⁶

Live Voice Versus Recorded Delivery

There are two ways to deliver the words: an audiologist reads them aloud through a microphone while monitoring a VU meter to keep the volume consistent (called monitored live voice, or MLV), or the words come from a standardized recording. The recording method is more repeatable and is generally considered the clinical gold standard, but live voice is faster and still common in busy clinics.

Research has consistently found that live voice introduces more variability, both in the scores patients receive and in how long the test takes. One study found significantly more variability in administration time across testers when using live voice compared to recorded presentations.

⁷

A more recent study confirmed this pattern and added an important detail: a statistically meaningful difference in word recognition scores between groups with normal hearing and groups with hearing loss was detected only when the recorded method was used, not when using live voice. In other words, the recorded approach was more sensitive to real differences in hearing ability. Live voice delivery was faster, but the extra variability it introduced muddied the clinical picture.

⁸

If your audiologist uses live voice, the result is still clinically useful, but it comes with wider error bars. For tracking changes over time or making high-stakes decisions like cochlear implant candidacy, recorded materials are strongly preferred.

When a PB Max Score Is Considered Disproportionately Poor

The real diagnostic power of PB Max lies in comparing it to the degree of hearing loss measured by the pure-tone audiogram. There is a well-established statistical relationship between the two: as the pure-tone average gets worse, PB Max tends to decrease, but within a predictable range. In a large study of nearly 2,000 subjects, researchers used nonlinear quantile regression to define the lower boundary of expected scores at each level of hearing loss. If your PB Max falls below the 95th-percentile confidence limit for your pure-tone average, the score is considered disproportionately poor.

⁶

A disproportionately poor PB Max suggests that something beyond a simple loss of loudness is degrading speech understanding. The most common culprits include damage to the auditory nerve (as seen in conditions like vestibular schwannoma, previously called acoustic neuroma), auditory processing disorders, and certain inner-ear pathologies that distort the neural signal even when enough sound reaches the cochlea.

One phenomenon that audiologists look for is called rollover: PB Max is reached at a moderate intensity, but the score actually drops when the volume is increased further. This was historically considered a hallmark of nerve-related (retrocochlear) lesions. However, research on patients with confirmed acoustic neuromas found that the rollover index was not actually characteristic of those tumors, which complicates the old diagnostic rule of thumb.

⁹

The takeaway for patients is that a PB Max score that seems too low for your hearing loss is a flag, not a diagnosis. It prompts further investigation, often with imaging or additional specialized tests, rather than pointing to a single cause on its own.

How Reliable Is a Single Score?

Because PB Max is based on a relatively small number of words, each individual score comes with some inherent statistical uncertainty. If you scored 80% on a 25-word list today and 72% next week, did your hearing actually change, or is that just normal test-to-test variability? To answer that, audiologists use critical difference tables, which tell them how large a change between two scores has to be before it is considered statistically meaningful rather than random noise.

A computer simulation study that re-derived one of the most widely used critical difference tables found that the original table was most accurate for list sizes of 25 words. The updated table differed from the original in about a quarter of its entries, though no individual range shifted by more than one word in either direction.

¹⁰

In practical terms, a 25-word list scored in 4-percentage-point increments (each word is worth 4%) has fairly wide confidence intervals. A score of 80% on a 25-word list might represent a true score anywhere from the low 60s to the mid-90s. This is why clinicians who need tighter precision will use 50-word lists or test at multiple levels, and why comparing scores across visits requires consulting the critical difference values rather than just eyeballing the numbers.

How Cognitive Function Interacts With PB Max

PB Max is presented as a test of auditory ability, but recognizing words is not a purely ear-level task. Your brain has to hold the sound in memory, match it against stored vocabulary, and produce a response, all of which draw on cognitive resources. This means that a lower PB Max does not always reflect worse hearing in a straightforward way, especially in older adults.

A cross-sectional study of older adults with hearing loss found that participants with lower speech recognition scores showed more pronounced impairments in memory and executive function, while attention and language abilities were comparatively less affected. The researchers noted that cognitive domains like processing speed and working memory directly influence speech discrimination, so some of the decline in scores likely reflects cognitive limitations layered on top of auditory ones.

¹¹

This overlap creates an interpretive challenge. An older adult with mild dementia and moderate hearing loss may score poorly on a PB Max test, but the score underestimates their auditory capacity because cognitive decline is absorbing some of the blame. Experienced audiologists factor this in by considering the patient’s overall cognitive status, pacing the test appropriately, and sometimes using closed-set formats (where the patient picks from a short list of options) instead of the open-set repetition that standard PB word lists require.

PB Max in Hearing Aid and Cochlear Implant Decisions

PB Max is not just a diagnostic number. It plays a direct role in deciding what kind of amplification or intervention to recommend. Unaided PB Max scores provide useful information for hearing aid fitting and for counseling patients and their families about the realistic benefits of hearing aids.

¹²

If your unaided PB Max is high, say above 80%, hearing aids that simply restore audibility will likely work well for you. Most of your speech-understanding machinery is intact; you just need the sound delivered at the right levels. If your unaided PB Max is moderate, around 50 to 70%, hearing aids will help but may not restore fully natural conversation, especially in noisy environments. In that range, counseling about realistic expectations becomes especially important so that patients are not disappointed when their expensive devices do not make speech crystal clear in a crowded restaurant.

At the low end of the scale, a very poor PB Max, particularly when measured with best-aided conditions, is one of the criteria audiologists use to evaluate cochlear implant candidacy. The logic is straightforward: if hearing aids are making speech loud enough but your recognition score remains very low, the bottleneck is not volume. A cochlear implant bypasses the damaged sensory cells and stimulates the auditory nerve directly, offering a different path to speech understanding. Current guidelines in many countries use aided word recognition scores in the range of 60% or below on sentence-level materials as part of the candidacy evaluation, though the exact cutoffs and test materials vary by program.

Why Scores From Different Clinics May Not Be Directly Comparable

A PB Max score of 72% from one clinic and 80% from another does not necessarily mean your hearing changed. Several procedural variables can shift the number without reflecting any real auditory difference. The word list used matters: lists with more common, everyday words tend to yield higher scores than lists packed with less familiar terms. Whether the test was delivered live or from a recording matters, as discussed earlier. The specific presentation level matters, especially for more severe losses. Even the talker on the recording can make a difference, since some voices are inherently clearer than others.

This is why audiologists pay attention to testing conditions when comparing results over time. The most informative comparisons come from tests done at the same clinic, with the same materials, at the same presentation level, by the same method. When those variables are held constant, even small changes in score can be clinically meaningful, provided they exceed the critical difference threshold for the list size used.

Patients sometimes arrive at a new audiologist’s office with a PB Max score from a previous provider and wonder why the new test gives a different number. The answer is usually not that their hearing improved or worsened overnight. It is that the testing conditions differed in one or more of these procedural ways. A good audiologist will note the testing conditions alongside the score so that future comparisons are apples to apples.

The Language Problem in Global Audiology

Most of the foundational PB word lists were developed for English, and specifically for American English. Clinics testing patients in other languages face a genuine gap. You cannot simply translate an English PB list into another language and expect it to remain phonemically balanced, because different languages have entirely different inventories of speech sounds in different frequency distributions.

This is why research groups around the world have invested in building language-specific PB materials from the ground up, as the Tamil and Arabic projects illustrate. Each effort involves selecting words that reflect the phonemic structure of that language, testing them for familiarity with native speakers, validating them with expert panels, and then verifying that the resulting lists produce stable, repeatable scores across listeners. The Tamil project pruned 500 candidate words down to 270 across multiple rounds of review.

For languages that lack validated PB materials, clinicians sometimes resort to improvised word lists or sentence-based tests, but the scores from those tests do not map neatly onto the normative data developed for standardized PB lists. This is an active area of research in audiology, and the number of languages with properly validated word recognition materials grows slowly each year. If you are tested in a language for which standardized PB lists exist, the results are interpretable against established norms. If not, the score is still informative in relative terms but harder to benchmark against population data.