LESSON 97

Speech Test Materials in Turkish

Name: Speech Test Materials in Turkish
Uploaded: 2026-11-04T03:30:00+03:00
Duration: 10 min 49 s
Description: Speech tests depend on language, so Turkish needs materials developed specifically for it. In this lesson we look at how the Turkish language shapes test design, the one-syllable and three-syllable word lists used in Türkiye, closed-set testing for children, and sentence tests such as the Turkish Matrix Sentence Test, the Turkish HINT and QuickSIN-TR.

Speech tests depend on language, so Turkish needs materials developed specifically for it. In this lesson we look at how the Turkish language shapes test design, the one-syllable and three-syllable word lists used in Türkiye, closed-set testing for children, and sentence tests such as the Turkish Matrix Sentence Test, the Turkish HINT and QuickSIN-TR.

Watch on YouTube

Follow along

Hello, and welcome to the AIHEAR Audiology Academy.
In this lesson, we look at the word lists and sentence tests used for speech audiometry in Türkiye.
We will see how the structure of Turkish shapes test design and what each material is chosen for.
Unlike pure tone tests, speech tests depend on language.
Translating an English list into Turkish is not enough; the two languages differ in how often sounds occur and in word structure.
A word list must represent the sound structure of the tested language in a balanced way.
Words should be familiar in everyday language, to reduce the risk of scoring a word wrong just because the patient does not know it.
To test the same patient again, we need several equivalent lists, that is, lists of similar difficulty.
Finally, to interpret the results, we need norms collected from normal-hearing speakers of Turkish.
So the test should be given in the patient's first language and judged against the norms of that language.
In Turkish, stress usually falls on the last syllable of a word.
So spondees, two-syllable words with equal stress on both syllables as in English, hardly occur naturally in Turkish.
Turkish is an agglutinative language; most words in everyday speech have two or more syllables.
Rules such as vowel harmony affect which sounds can appear together within a word.
In recent one-syllable lists, words with a consonant, vowel, consonant structure are preferred.
In Türkiye, work on developing one-syllable word lists goes back about fifty years.
Over this time, many lists that take phoneme balance into account have been prepared.
For example, the adult test published by Mungan Durankaya and colleagues in 2014 has three lists of fifty words each.
In these lists, words were chosen from the most frequent one-syllable words in Turkish and screened for familiarity.
The study found no significant difference between the lists, showing that the lists were equivalent.
One-syllable lists are used to measure the word recognition score and the performance curve.
But different centers may use different lists, so the list used must always be stated in the report.
For the speech recognition threshold, or SRT, lists from several centers are used in Türkiye.
A study published in 2023 compared the Hacettepe, Marmara and Başkent university lists for word and phoneme structure.
All three lists use three-syllable words; the Hacettepe list was reported as the most widely used, with the richest word variety.
The Marmara list was compiled from materials for children and may be more sensitive to high-frequency losses.
The Başkent list consists mostly of nouns.
The scarcity of equally stressed two-syllable words in Turkish is given as one reason for using three-syllable words in threshold testing.
The strength of Turkish word lists is that they are built on the sound structure of Turkish and use familiar words.
These lists have been used in clinics for years and are part of routine testing in many centers.
According to review studies, there is no national standard list of multisyllable words for threshold testing.
The lists' curves of correct responses by intensity, their psychometric curves, have often not been published.
Using the same recorded version across clinics is also not yet common.
There is also a continuing need for standard, recorded lists suitable for children.
For children, the test material should suit the child's age, language level and way of responding.
For school age, a thesis study developed four one-syllable lists of twenty five words each for children aged seven to twelve.
In younger children, closed set responses by pointing to pictures or objects are used instead of open set repetition.
In this approach, the child points to the word they heard among a few pictures; speaking skills do not affect the result.
In Turkish children with implants, the MTP test, with one, two and multisyllable words, was used in three, six and twelve choice forms.
For speech in noise, the Turkish Hearing in Noise Test, or HINT, has a children's version with norms for ages six to twelve.
In a closed set test, the pictures are first introduced to the child and named together.
Pictures the child does not know are left out; the goal is to measure hearing, not vocabulary.
Words are presented as recordings, or without facial cues in a way that prevents lipreading.
The child touches or points to the picture of the word they heard.
Scoring accounts for chance: in a four-choice test, even random pointing gets about twenty five percent correct.
This is why closed set scores are not directly compared with open set scores.
Sentences contain context, so they are closer to everyday speech than word lists, and most are given in noise.
The Turkish Matrix Sentence Test, or TURMatrix, was published in 2015.
It was developed jointly by the matrix test team in Oldenburg and Turkish researchers, and it is comparable with other matrix tests.
The Turkish Hearing in Noise Test, or HINT, is also in use.
In 2023, a children's version of this test with twelve lists of ten sentences was published.
In 2025, QuickSIN-TR, a Turkish adaptation of the Quick Speech-in-Noise test with nine equivalent and two practice lists, was published.
At the heart of a matrix test is a table made up of five word classes.
In TURMatrix, each class, name, number, adjective, object and verb, contains ten familiar Turkish words.
Each test sentence is built in a fixed pattern from one word picked at random from each column.
The sentences are grammatically correct but semantically unpredictable, so the patient cannot use context to guess words.
This structure allows many equivalent lists to be generated, so the same patient can be tested again and again.
Responses can be given open set, by repeating, or closed set, by choosing the words on a screen.
The graph shows schematically how the percentage of words correct changes with signal-to-noise ratio in normal-hearing Turkish adults.
The ratio at which half the words are understood is the SRT50; with fixed ratios, its mean was about minus eight point three decibels.
The curve is very steep: when the signal-to-noise ratio changes by one decibel, understanding changes by about fourteen percent.
In adaptive testing with open set responses, the mean SRT was about minus seven point two decibels.
With closed set responses, this value was about minus seven point nine decibels.
Because the curve is steep, even a few decibels of worsening means a clear difficulty understanding speech in daily life.
When working with matrix tests, pay attention to two points.
First, the patient's score improves over the first lists as they learn the word set; practice lists are given for this learning effect.
Second, results of the Turkish version of a test are compared only with Turkish norms, never with norms of another language.
Stating the version, the response format and the noise type in the report also makes comparison easier.
To verify pure tone thresholds, three-syllable word lists are used for the SRT.
For the speech discrimination score, one-syllable word lists are preferred.
For young children, closed set tests with pointing to pictures or objects are suitable.
To measure understanding in noise, sentence tests such as the matrix test or the Turkish HINT are used.
In hearing aid counseling, sentence tests in noise turn the patient's everyday complaint into a number.
For monitoring over time, the same test, the same version and the same conditions should be used every time.
Speech tests depend on language; they are given in the patient's first language, with that language's norms.
In Türkiye, common SRT lists use three-syllable words, and word recognition lists use one-syllable words.
For young children, closed set tests with picture pointing are used, and chance level is taken into account.
For understanding in noise, there are tests such as the Turkish Matrix Sentence Test, the Turkish HINT and QuickSIN-TR.
Test yourself: Why do SRT lists in Turkish mostly use three-syllable words?
The correct answer is B: Turkish has few equally stressed two-syllable words.
Because stress usually falls on the last syllable, English spondees have hardly any natural equivalent in Turkish.
In the next lesson, we will measure word recognition at several levels to learn about the performance-intensity function and rollover.
Subscribe to the channel so you never miss a new lesson. See you soon!

Keywords

Click a keyword to search all lessons for that term.

Prerequisites

LESSON 96 Word recognition score