audiometry

Further testing


This topic introduces you to basic information about assessments that are done after initial screening audiometry. This includes both further testing and interpretation of those results. In order to know whether further assessment is required, the audiogram as well as the client’s history need to be taken into account. This leads to making decisions about what further testing is required, the significance of the results, how to correctly report these results, and identifying the need for referral.

After an initial history and audiogram it is time to look at the information you have gathered and decide whether the client can be discharged, or to look at options for improving hearing and whether a referral is required.
Each client has different needs, depending on their age, general health, severity of loss and situation in life. Some needs are physical, but we also need to look at emotional, psychological, sociological and financial needs. Remember, a person’s hearing loss causes problems not just for themselves but also for their families.

Activity
The next time you see a client, write down all the needs of your client in each of the five categories cited above and next to each need a way to overcome it. Eg.: Psychological – client worried about cosmetic appearance of aid – show client actual size and colour of hearing aid, discuss small hearing aid (if audiogram allows).

Identifying need for further testing

Once you have done the initial testing (pure tone audiometry and bone conduction thresholds) you will need to look at your results and consider whether further testing is required.

The simplest type of audiogram is acquired by measuring with air and bone conduction, using only pure tones. If all the thresholds fall within the ‘normal’ range, ie.<20dBHL, there is no need for further assessment.

If, however, any of the thresholds exceed 20dBHL, or there is a difference between the ears of more than 10dB at any two frequencies, more testing is required.

If there is an asymmetry between the ears or the air and unmasked bone thresholds differ by more than 10 dB, masking will have to be applied. There are two types of masking: bone conduction masking and air conduction masking.

Bone conduction masking

You will need to apply bone conduction masking if the bone conduction is 15dB or more better than the air conduction threshold of either ear.


Air conduction masking

If there is a 40dB (or greater) air-bone gap at any frequency, then air conduction masking is required at those frequencies.


Speech masking

If the average air conduction thresholds in one ear are significantly worse than the bone conduction thresholds in the other ear, you may have to do speech masking. The details of these calculations are set out below
If any of the thresholds fall outside of the normal range, speech testing needs to be performed.

Tympanometry

It is good practise to perform tympanometry on everyone, however, not all clinics do this. If you do not routinely perform tympanometry, you should definitely perform it on those people who show some conductive element in their hearing loss, that is a conductive hearing loss or a mixed hearing loss.

Tympanometry should always be performed when testing children.


Post-basic assessment: masking


Why do we need to Mask?


When we hear sounds, most are heard by going through the normal pathway of the ear, that is, by air conduction. However, some of the sound is also passed through the bones of the skull, which is called bone conduction. This is the reason that our own voice sounds different when we hear it recorded. We hear our own voice mainly by bone conduction, which usually gives it a richer, deeper sound. When we hear ourselves on a recording we are hearing our voice by air conduction which makes it sound thin and higher in pitch. When we hear loud sounds, there may be a mixture of air and bone conducted sound.

When we are testing someone’s hearing, we need to be sure of what is recorded as air conducted sound and what is bone conducted sound, and in which ear it was heard. Masking is a technique which allows us to do just that.
masking1.png
Interaural attenuation
When we apply an air conducted sound to one ear, we hear it only in that ear, until the sound is so loud that some of it is transmitted through the bone and into the other ear (remember, the cochlea is encased in a bony structure). In some people this occurs at relatively soft sounds, in others the sound needs to be very loud before it ‘crosses over’. This depends on the size and shape of the head. To be safe, we assume that it only takes 40dB for sound for it to cross over to the other side and is heard in the other ear (also known as the non-test ear). This is called the interaural attenuation. If the interaural attenuation is exceeded, sound will spill to the other cochlea. The amount of this sound is called the cross over.
maskingb.png



With air conducted sounds the interaural attenuation varies from 40-60dB. When using headphones, we always assume 40 dB interaural attenuation, as this is the worst case scenario. When using insert earphones, we can assume 60dB interaural attenuation, as inserts reduce the amount of bone conducted sound.

When using the bone conductor the interaural attenuation is 0dB, that is because the sound is applied directly to the bone. So a sound presented through a bone conductor will be heard equally in both ears (if the hearing is the same in both ears), regardless of where the conductor is placed on the head.


Bone conduction masking


When using the bone conductor, we obtain the cochlear threshold of the better ear, regardless of on which side the bone conductor is situated. Therefore, if there is a significant difference of hearing level (15dB or more) between the unmasked bone conduction threshold and the air conduction thresholds in either ear, then we want to establish whether the discrepancy is caused due to a conductive problem (middle ear) or a sensorineural problem (cochlear or retro cochlear). To do this we need to isolate the worse ear. That means we need to keep the better ear busy, so it’s not interfering in our investigation of the worse ear. While we are keeping the better ear busy, we then test the worse ear (or test ear).

We do this by introducing a masking noise into the better ear.

• Masking noise is always presented via a headphone or insert phone, that is, by air conduction.

• The masking noise must be suited to the signal, ie. Narrow band noise for pure tones and speech noise for speech signal.

• The masking noise needs to be audible by the better ear, that is, it must be introduced at least at air conduction threshold level.

• It must be loud enough to take care of the cross-heard signal.

• It must not exceed the interaural attenuation, otherwise the masking will spill into the worse ear and affect threshold. This is called overmasking.

• The earphone on the test ear side must be off the ear (usually up near the temple) as occlusion may otherwise occur and skew test results.

maskingc.png

Air conduction masking


In the case of there being a large gap between the AC thresholds and the unmasked BC thresholds, you will need to mask the AC thresholds. The worst case scenario for cross-over occurring using AC is 40dB. Therefore if there is a 40dB gap, you must mask. The principle is the same as BC cross-over, except that with AC testing the levels have to be higher before they are transmitted by the bone (remember in the case of BC testing, the sound is applied directly to the bone, whereas in AC testing they are not). In the case of a dead ear, for example, you may be getting thresholds although the ear has no hearing. It will mimic the hearing ear, but at levels of 40dB or more below. This is called a shadow effect. When masking is introduced, the thresholds quickly drop down to the real levels.

maskingd.png

When do you need to mask?

Bone conduction- always, except when AC and BC thresholds are within 10dB of each other.

Air conduction- if the AC thresholds are 40dB worse than the unmasked bone conduction thresholds.

How much masking is needed?
For both AC and BC testing:

masking level = Air Conduction threshold (Non-Test-Ear) + 10dB

How do you mask?

In order to obtain masked bone conduction (BC) thresholds, the bone conductor must be placed behind the ear we want to establish the thresholds from ie. The test ear (TE). We apply masking to the other (better) ear by placing the headphone on it. This is called the non-test ear (NTE). We DO NOT place the headphone over the test ear but next to or above it, as this may cause an inaccurate result due to occlusion. The arrangement looks a little lopsided and may take some practice to achieve.

In the case of AC masking, the headphones are placed as usual, with pure tones applied to the TE and masking to the NTE.

Continuous masking noise is then introduced through the headphone to the NTE, 10dB above the air conduction (AC) threshold and pure tones are introduced through the bone conductor on the test ear. When threshold has been established in the TE, the masking noise in the NTE is increased by 10 dB and the tone in the TE is repeated at threshold level. If the client responds, turn up the masking level by a further 10 dB and present the tone again, at threshold level. If the client responds again, the threshold is established. This procedure is called establishing a masking plateau. A plateau of 30 dB must be reached before the threshold is considered valid, ie. that none of the signal has ‘crossed over’ from the NTE. (Remember that we started with 10dB to begin with, we then increase twice more by 10dB to get a 30dB masking plateau).

If the client does not respond once the masking noise has been turned up, the signal in the TE is increased at 5 dB increments until the client responds again. Then a new plateau should be attempted. You must increase the signal on the TE, until a 30 dB plateau is achieved, ie. The client must respond three times to the same level in the TE while the masking noise is increased in the NTE.

maskingscenario1.png

maskingscenario2.png

Overmasking

Sometimes when we need to mask, there is the risk of overmasking. This is when the masking level gets to a point where it is so loud that it spills back into the test ear. This doesn’t happen very often, but is possible especially when there is a bilateral conductive hearing loss. That is, a lot of masking is needed to be audible via the airconductive pathway, but because the other ear has good bone conduction thresholds, the masking is heard by the test ear. This means that the test ear now has trouble hearing the signal. In this case, it is not possible to establish a masking plateau. Every time the masking is turned up, the signal tone cannot be heard. It needs increasing, until it is heard again, but then the masking noise will blot it out again. In this case, try to get a 20dB masking plateau, but also write in the margin that a full masking plateau was not established due to overmasking, or ‘Naunton’s masking dilemma’. You are at risk of overmasking when:

overmasking.png

maskinge.png

Activity 1


Below you will find a number of terms associated with masking.
In your journal write a definition or description for each term.
You can check your answers against the terms in the Moodle course GLOSSARY

With air conducted sounds the interaural attenuation varies from 40-60dB. When using headphones, we always assume 40 dB interaural attenuation, as this is the worst case scenario. When using insert earphones, we can assume 60dB interaural attenuation, as inserts reduce the amount of bone conducted sound.

When using the bone conductor the interaural attenuation is 0dB, that is because the sound is applied directly to the bone. So a sound presented through a bone conductor will be heard equally in both ears (if the hearing is the same in both ears), regardless of where the conductor is placed on the head.

maskingactivity.png

Activity 2


Practice masking at your clinical visits whenever possible. Masking takes a while to get the hang of, so the more practice you get, the sooner you will feel it coming to you naturally. Get your workplace supervisor to check your technique.

Follow these steps for bone conduction masking:
1. Place the bone conductor on the mastoid bone behind the test ear
2. Place the headphone over the other ear (NTE) and to the front of or above the test ear.
3. Mask the NTE by starting at AC threshold + 10dB
4. Test the bone conduction threshold going up in 5dB steps
5. Add in 10dB of masking every time you get a response
6. When you have done this three times at the same threshold, you can mark it in the audiogram

For air conduction:
1. Put masking into the better ear (NTE) at threshold + 10dB
2. Retest worse ear (TE) in 5 dB steps
3. When you have a response, add in masking in 10dB steps until you have done this three times, with a response.


Speech testing

Speech testing is one of the basic test battery components. It helps us to assess the speech intelligibility our client has. Speech intelligibility is mostly, but not always linked to the audiogram. For example, a client who has a typical presbycusis loss with good low frequencies and poor high frequency hearing is expected to do well on words such as CAN but may have trouble with words such as THIN. There, are however, some cases where the speech intelligibility is markedly better or worse than expected. In Australia we use phonetically balanced word lists (PB word lists). These have been designed to carry a proportional balance of phonemes as used in the English language.

When performing speech testing we like to test at different levels, so that we can graph a speech curve. To graph a speech curve we should present at least 6 different speech lists (of 10 words each) at different levels. We start 30dB above the three frequency average (3FA) which consists of 500Hz, 1000Hz and 2000Hz, and then increase the level by 10dB until the client scores their best score, and then below the starting level until the client scores very poorly.

Typical speech curve

speechcurve.png

Most Comfortable Level (MCL)
Some clinicians like to do speech testing at the client’s most comfortable level (MCL). There are points for and against this.

The aim of rehabilitation in hearing is to get the client to hear to the best of their ability. The variation in level within speech is about 30dB. This means that when you present speech at 30dB above threshold, there is every chance that your client should hear everything in the speech spectrum between 500-2000 Hz. In this technique we like to do a speech curve, that is, we get the client to perform at different levels, so we can see where their best performance is, and this can be taken into account when setting the hearing aid.

When using MCL in speech testing it is usually not done as part of a curve, ie only one or two lists are presented at one level. This requires less time, but also yields less information. Furthermore, the MCL of a client may be less than 30dB above 3FAHL and the client is therefore then not given the chance to achieve their best score. However, the argument for using MCL is that it is probably the level at which the client will end up using the hearing aid, even if they hear better at a higher level, and is therefore a better indication of the final result you can expect for your client.

There is no right or wrong in the different techniques, just different philosophies. What method do you use in your clinic? Talk to the other clinicians and ask why they use a particular technique. Perhaps you would like to use a combination of both.

Reliability and validity

The important thing to remember about speech testing is that it only gives us a rough indication as to how a person is performing. A 5% or 10% improvement between levels is not significant when you are only presenting 10 words at a time. Speech discrimination results have problems with validity and can be impacted by issues such as native language (ie those people whose first language is not English, and even those whose first language is an English from elsewhere such as American, Canadian, Scottish or Caribbean), as well as the pronunciation of the speech presenter (if done by clinician). There are also issues of variability, which are dependent on changing levels if the speech is presented live voice, compared to pre-recorded material which has been equalised. Also female vs male voices, having a cold and the quality of the test set up can all affect speech discrimination scores and their reliability. As well as all that, because of time restraints, we tend to only present a limited amount of speech material, which renders small changes in scores insignificant when calculated statistically.

Activity 3


Some lists are scored by words correct, others by phonemes. Can you find an example of each type of material? Which do you think is the more sensitive material?

Speech tests can be presented through voice recorder/CD or live. List the advantages and disadvantages of each method.

Check your answers here

Speech masking

Masking is also required when testing speech, if there is danger of cross-over occurring. In the case of speech testing, we are testing at levels above threshold. This is called supra-threshold testing. We are therefore more prone to risk cross-over signal. In speech testing we use the average hearing loss, that is the average of thresholds at 500Hz, 1000Hz and 2000Hz and add 30dB as our starting presentation level.

Speech presentation level = average AC (500Hz, 1000Hz, 2000Hz) + 30dB
Also known as 3FAHL

We then need to look at whether this presentation level is likely to spill into the other ear. The spill or cross-over will be heard in the non-test cochlea, so we need to look at the average bone conduction threshold of the non-test ear. We deduct 40dB (for air conducted interaural attenuation) from the presentation level and compare it to the average bone conduction in the other ear. If it is larger, the presentation signal will be heard by the non-test cochlea, and we will need to mask.

If Presentation Level (PL) – 40dB (interaural attenuation) > average BC of NTE

To establish whether masking is needed, we also need to establish the minimum level at which a speech score can be obtained. This is called the minimum discrimination score level (MDSL). The reason we need to take into account the MDSL is because Audiometers are calibrated with a pure tone. The speech signal is a broadband signal, and as such has more energy at any given level than a pure tone of the same level. Therefore it is possible and indeed often the case, that at the average threshold level, where a person ought to be just hearing a speech signal, they might hear it better than expected, because there is more energy in the signal. This would cause your speech discrimination score to be better than expected.

The final formula looks like this:
When to mask.
maskingformula1.png
How much masking do we need?

Well, we need to start with the 3FAHL of the non-test ear (NTE), the ear we want to present the masking to. Then we need to add enough masking to overcome the cross-over. The cross-over we calculated above, it is the difference between the presentation level etc and the average BC of the NTE

maskingformula2.png

When masking for speech we use speech masking noise, which is a broad band noise with similar frequency weighting to speech. On top of this we need to add a ‘buffer’ called the effective speech masking level (ESML). The ESML, again has to do with the fact that the audiometer has been calibrated with a pure tone. Because the masking noise is calibrated to pure tone levels, but has to mask a speech (broadband) signal, it may not be effective. That is, ideally a 20dB masking noise should completely mask a 20dB speech signal, but this may not be the case. You may need extra masking, to ensure that the speech is really masked. An explanation of how to establish ESML and MDSL for your audiometer will follow below.

maskingformula3.png

Overmasking

With all masking techniques there is a chance of the masking noise spilling back into the test ear. This is called overmasking. This sometimes happens when there is a large bilateral conductive loss. It can happen more easily in speech testing because we are testing at supra threshold levels, ie. Levels that are higher than hearing threshold. The principle is the same, though, as for pure tone overmasking:
overmaskingformula.png

This is called Naunton’s masking dilemma. If this occurs, you must write this on the audiogram or in your report.


Calculating MDSL and ESML

These levels only need to be established once, providing there are no changes at the annual calibration of your audiometer.

MDSL – the level at which a group of normal hearers begin to score (4%) with the speech material used, ie. The minimum discrimination score level.

ESML – the level at which that speech material is effectively masked

A minimum of ten people with hearing within normal limits is needed to establish these levels effectively. It will need to be done separately for each audiometer and different speech materials.

To calculate MDSL present speech material at decreasing levels until the subject is scoring only one or two phonemes. Take the lowest level at which any of the people tested scored (we want to mask for the worst case scenario) as your MDSL.

N.B. It is important to do this for different speech materials used as PB word list MDSL will be higher than a Boothroyd list MDSL because a whole word is needed for a minimum score rather than just a phoneme.

To calculate ESML take the level at which a subject is scoring 100%. Then introduce speech masking noise into the same ear until the score is reduced to less than 5 % (ie the speech is ‘masked out’). The difference between the level of speech presented and the masking level is the ESML. That is, if the speech material was 40 dB and the masking level 50dB, your ESML is 10dB.
NB: Some older audiometers do not allow masking to be presented ipsilaterally, ie. Into the same ear as the signal. If this is the case with your audiometer, assume that the ESML = 0dB. However, it is important for you to understand the principle of ESML and why it is taken into account.

Once you have established the MDSL and ESML for your audiometer(s), it is a good idea to write them down on a strip of paper and tape them on or near the audiometer. This way anyone testing can immediately see what your levels are.


Activity 4


Test 10 people with normal hearing and establish the MDSL and ESML for the audiometer you are using. Use the tables below to help you record your results.

mdslactivity.png

Practical

Test speech masking using the following steps:
1. Find the average threshold of 500,1kand 2k (3FAHL) for the ear you want to apply masking to
2. Add 30dB to the 3FAHL
3. Take away 40dB (for interaural attenuation) Take away (or add! if the MDSL is negative, (-(-)=+)) MDSL value
4. Compare to the average BC thresholds in the other (test ear)
5. If there is a gap you need to mask
6. To calculate level for masking, add gap to 3FAHL to masking (non-test) ear
7. Add ESML value
8. Present masking noise at this level in masking ear and present speech to test ear at 30dB above average threshold.
9. If you test at more than one level, the masking level will change accordingly, ie if you present speech at presentation level + 10dB, then the masking level also has to be increased by 10dB.


Tympanometry

Tympanometry is a test which looks at the middle ear and function of the eardrum. It is part of a larger test battery known as impedance testing which also includes acoustic reflex testing. For the purpose of this course we will only look at screening tympanometry and screening reflexes.

The Tympanometer


The tympanometer is a machine that lets us measure middle ear function. There are three measurements involved with tympanometry:
1. tympanic membrane compliance
2. middle ear pressure
3. middle ear volume

Tympanic membrane compliance tells us how mobile the eardrum is. If there is fluid in the middle ear, then the tympanic membrane will not move as readily, for example.

Middle ear pressure should be the same as the outside pressure, ie. 0 mmH2O. The Eustachian tube which connects the middle ear to the pharynx opens a little each time we swallow or yawn and thus equalises any differences in pressure that may occur, eg when we drive down from a steep mountain, our ears ‘pop’. If there is an infection in the throat which causes swelling around the opening of the Eustachian tube, air cannot get into the middle ear. Since the middle ear is lined with skin which breathes, soon a vacuum is created which gives us a negative middle ear pressure. Sometimes when we blow our nose, extra air is forced into the middle ear and thus a positive middle ear pressure is created. As far as pathology is concerned, we will concentrate on negative middle ear pressure.

Please do not confuse negative middle ear pressure or even fluid in the ear with a middle ear infection (otitis media). Fluid in the ear occurs when prolonged negative middle ear pressure draws fluid from the skin into the middle ear cavity. This in itself is harmless. An infection only occurs if bacteria travel up the Eustachian tube and then cause an infection of the skin lining the middle ear. When fluid is present this may occur more readily because of the warm, moist environment. If there is an infection, it is always accompanied by pain and fever, and if pus forms, eventually the tympanic membrane may rupture. Fluid in the ear alone is often reabsorbed with no pain or infection. Sometimes if it remains in the cavity for extended periods of time, fluid will turn to ‘glue’, a sticky substance which needs to be removed by a surgeon.

Middle ear volume is in itself relatively arbitrary, as it depends on how far the probe is pushed into the ear canal. If however, the middle ear volume is significantly larger or smaller than the normal range, it may indicate a perforated eardrum in the case of large volume, or a middle ear growth (cholesteatoma) if it is very small.

The probe

In order to make tympanometry measurements we need to insert a probe into the ear canal. Therefore you must look into the ear canal before you start, first of all to look at the ear drum, but also to check that there is no wax or other debris present. Next you need to select an eartip that will fit the client’s canal snugly. If there is leakage, the test will not work.

The probe of the tympanometer is made up of three parts.

1. a tube that delivers pressure
2. a tube that delivers sound
3. a microphone outlet

probe.png

The tympanometer measures the compliance of the eardrum by putting pressure into the earcanal and also producing a sound. As the pressure increases, the eardrum becomes less compliant (it moves less) and the sound mostly bounces off. The microphone measures how much sound comes back from the ear drum. Then we reduce the pressure, and in a normal ear, the eardrum will be at its most compliant and the sound is conducted well and does not bounce back. Then we reduce the pressure and again the eardrum becomes taut. We graph this movement in a chart called a tympanogram.
tympanogram1.png


Screening tympanometers usually do all this automatically, but you will still need to read it and put the results in your report. If your tympanometer is not automatic you will need to note down the cc values at –200mmH2O, at the peak of the curve (not necessarily 0mmH2O!!!) and at +200mmH2O. The height of the curve is measured from +200mmH2O to the peak.


The different types of tympanograms

As an audiometrist you will only need to report the values of each of these three measurements without having to report on abnormality, but it is useful for you to have an understanding of the results.

Tympanograms come in three versions, Type A, Type B and Type C.

Type A denotes a tympanogram with normal middle ear pressure ie 0 mmH2O (+/- 100mmH2O).


tympanograma.png















Type B denotes a tympanogram with no peak. This occurs when there is fluid or glue in the middle ear cavity which prevents the eardrum from moving. In conjunction with a high volume, a type B can also indicate a perforated eardrum. A type B tympanogram should always be referred to a doctor.

tympb.png
















A type C tympanogram has a negative peak and is commonly seen when the client has a cold. For it to qualify as a type C the peak must be <-100mmH2O. NB.:In children this is extended to -150mmH2O.

tympc.png







Screening reflexes

As part of the impedance test battery we perform acoustic reflex testing. Audiologists tend to do ‘the works’, that is, diagnostic reflex testing which can give us clues as to the integrity of the acoustic nerve. Reflex testing assesses a reflex arc of nerves that run between the ears. When a loud sound is presented, a tiny muscle that is attached to the stapes contracts, moving the stapes away from the oval window. It is designed to reduce the amount of signal going into the ear. This reflex works in both ears simultaneously, regardless of which ear the sound is presented to. This means that there are two ways of measuring the reflex, contralaterally and ipsilaterally.

In the contralateral test the sound is presented to one ear and the reflex measured in the other. In the ipsilateral test the sound is presented to the same ear where the reflex is measured. Reflex testing is done with pure tones at different frequencies. In screening tympanometry, screening reflexes are often part of the test setup and essentially confirm the tympanogram result. Eg, if we get a type B tympanogram, we do not expect to get any reflexes. We tend to also only use ipsilateral reflexes in a screening setup. The most common reflexes tested are 500Hz and 1000Hz, although some also test 2kHz and 4kHz. In a screen test, tones are presented only at one level and recorded as present or absent. (In the diagnostic test, a threshold is obtained).





Activity 5


Get your supervisor to show you how to do a tympanogram. It is important to choose the correct probe tip size for each ear, as a seal must be established for the test to run properly.

Do a tympanogram on yourself and record the volumes at –200mmH2O, the peak, and at –200mmH2O. Then calculate the height of the peak.

Do tympanograms on different people at your clinical visits and record your findings.

Check how the tympanometer at your clinic is set up for reflexes. At what level does it present and at what frequencies? If your tympanometer is not set up for automatic screening reflexes, what is the protocol regarding screening reflex testing at your clinic?

What are the three main components of a tympanometer (impedance bridge)? Illustrate your answer with the help of diagrams.


The significance of various test results

When you look at your test results you must keep in mind the reliability and validity of those results.

Reliability
Reliabilty is related to how repeatable a test result is. Providing that your test equipment and operating technique are intact, objective tests (such as tympanometry) are very reliable. However, most of the tests you will do in the clinic are subjective tests, such as audiometry and speech tests. These involve an element you cannot control fully, ie the client. Things to take into consideration your client’s:
1. hearing loss
2. lucidity (alertness and ability to comprehend)
3. language skills (if they are not proficient in English)
4. motivation for test (does not want to do test or financial reward for HL)
5. mental ability (mental handicap)
6. physical ability (dexterity)
7. age (very young or old)
8. understanding of test protocol

Any of the above items may contribute to variations in test reliability. In order to increase reliability, we need to be aware that variations may occur and how to minimise them.

Activity 6


Choose a client you are seeing and test their speech discrimination each time they come to see you. When you have done 3-4 tests, compare them. Are they exactly the same? What might have caused differences?


Validity

Validity is a measure of how significant your test results are. This generally means how statistically significant, but also the validity of the test stimulus. That is, whether your audiometer has been calibrated properly and within the last 12 months, so you can be sure that the pure tones are at the correct frequency and level. Another example is the speech test. In order for the test to have statistic significance you must use a minimum number of items. If you use only one list of 10 items, your test is not statistically relevant. Speech tests, particularly, are subject to difficulties in validity if they are cut short. This happens often in clinics, unfortunately, because of time restraints or client fatigue. Speech lists should also be phonetically balanced to reflect the phonetic spread of the language you are testing. Different speech tests are designed for specific purposes, so you must be sure you are using the correct speech material for the test you wish to perform. If your client is of a non-English speaking background, you must take into account that the speech test will not reflect their true ability. Similarly, if the words you are using are unfamiliar to your client you may not get valid results.



Activity 7


What speech material do you usually use in your clinic? Find out and write down what factors have been built into the list to increase validity of the test results. How many words/lists do you need to present for the result to be statistically significant?






ANSWERS

Activity 3
AB (Arthur Boothroyd) word lists are marked by the phoneme, PB (Phonetically Balanced) word lists are marked by the word. The AB lists are more sensitive, because if you get part of the word right you get credit for it. If you get part of the word right in the PB word list, you get no credit for it

Presenting via recording means that the voice is always the same, ie no colds or emotion play a part, it is level equalised so that the levels are always even, it is always clear. The disadvantages are that it is slow compared to live speech, the examiner’s voice may be more familiar, it is faster.