Participants should prepare the data by placing the noun cards in piles based on their root-initial consonant - one pile for <b>, one for <n>, one for <t>, etc. For now, any words starting with a vowel should be set aside. 

When the initial sorting is done, each pile should be subdivided according to the vowel in the first syllable of the root. For example, roots starting with <bV> would be further sorted into the piles <ba>, <bo>, <bi>, etc. 

Begin the SAME-DIFFERENT Test. Select a consonant to evaluate and, as before, ask one participant to read through a pile of cards while everyone else listens. 

For example, while a participant reads aloud all word roots beginning <ba...>, the facilitator and other participants should listen together and decide if the sound written as <b> always sounds the same before the phoneme/a/. Does the language only have one [b] sound, or does it have 2, for example an implosive /ɓ/ and an egressive /b/? If a contrast exists, is there any indication in the transcriptions that some participants have already differentiated between these sounds, such as writing <b> for some words and <bh> for others? If so, combine the piles of words beginning <bha...> and <ba...>, then use the SAME-DIFFERENT Test to sort the words based on the sound of the initial consonant. 

Continue this process until all root-initial sounds have been evaluated using the SAME-DIFFERENT Test. Pay special attention to any set of consonant sounds that are phonetically close, such as <b> and <bh> in the example above. Put data for similar sounding consonants together and check whether everyone can hear the contrast. Ideally, everyone should come to the same conclusion for each individual word. 

It should be sufficient to check each consonant with only one vowel. For example, to continue the illustration above, any phonemic contrasts in the language relating to <b> should be established after checking the pile of root words that begin <ba...>. However, for a more thorough study, participants could check each consonant with every possible vowel combination (i.e. <bo...>, <bi...>, <be...>, etc.). 

Keep track of which vowels can combine with each consonant in the syllable-initial position of root words.

If the SAME-DIFFERENT Test shows that 2 different consonant phonemes have been written with the same symbol up to this point,* participants will now need to choose 2 symbols, one for each phoneme. 

As before, it is ideal for orthography choices to be...

a) As close as possible to the language community's intuition,

b) Non-contradictory with the writing system(s) taught in local schools, and 

c) Proven to work well for related languages. 

The group may also discover that 2 allophones, such as [d] and [r], have been written with 2 different letters instead of only one. This can happen as a result of following the writing system of another language (for example, English or French) where sounds like [d] and [r] are contrastive. Where 2 different letters have been used to represent a single phoneme, facilitators should use checking sessions and other exercises to make sure that native speakers can write these letters consistently in the proper environments (for example, <dV> and <VrV>). It is important to value the language community's preference while also challenging them by bringing awareness to potential issues. 

*For example, if the language has contrast between implosive /ɓ/ and egressive /b/ but both have been written as <b>

When all needed orthography decisions have been made, update cards based on the letters chosen. Next, chart the language's consonant phonemes. 

Proof of contrast can be presented in a table that is arranged either by place of articulation or in alphabetical order. The goal of the table is to show contrast using ‘minimal syllables’ in the root-initial position. By using this type data for the table, any gaps in distribution will stand out, highlighting possible cases of complementary distribution. If the consonants are not in complementary distribution, this proves that they must be contrastive. 

Words containing closed syllables in the root should be sorted by their root-final consonant.  

Further divide the resulting piles by the preceding vowel. For example, the <...t> pile would be split into piles for <...at>, <...ot>, <...it>, etc. During this sorting process, it is important to be aware that allophonic variations of vowel phonemes may surface before certain consonants. However, participants will normally group all allophones of the same vowel together intuitively.

Proceed with the SAME-DIFFERENT Test for root-final consonants, working systematically through each <...VC> pile. Make note of which vowels may come before each consonant.   

Make a separate chart for root-final consonants. This will highlight any generalizations about consonants in the root-final position as compared to root-initial.  Having separate charts will reveal cases of contrast, limited distribution, and complementary distribution. 

Use the SAME-DIFFERENT test, as before, to evaluate consonants in all relevant root-medial positions. 

Evaluating the C2  position of a consonant sequence may help prove contrast for consonants with limited distribution. For example, in many Bantu languages, prenasalized obstruents can occur at the beginning  of nouns (as a noun class marker) but not at the beginning of verbs. In the C2 position, however, prenasalized consonants occur frequently in both nouns and verbs. So while contrast may be difficult to show based on the evaluation of C1 alone, it can be proven with a set of contrastive consonants in the C2 position followed by the same vowel. 

Chart results in a new table and compare it with the previous ones, noting occurrences of both limited distribution and complementary distribution. Observe any gaps in consonant distribution and determine whether they are systematic or accidental.