Measure of Cluster Proximity (MCP)

The Measure of Cluster Proximity (MCP) compares a two-consonant target cluster with its actual production. Each cluster receives a score from 0 to 100 that reflects how many target consonants were produced and whether each was correct, substituted, deleted, or separated by an epenthetic vowel.

Use MCP to distinguish productions such as complete cluster deletion, cluster reduction, substitutions, vowel epenthesis, and a correct two-member production.

Before you begin

Check the syllabification and phone alignment of the IPA Target and IPA Actual words you want to analyze. Assign a speaker to each record so that the result is included in the correct participant section.

The stock query searches IPA Target words for exactly two consonants in one of the four supported cluster contexts. Clusters containing three or more target consonants are not scored.

Parameters

The analysis opens at the Parameters step. The principal controls are:

Table 1. MCP parameters
Parameter Default Effect
Report Title Measure of Cluster Proximity Sets the title of the generated report.
Combine results for all selected participants Off Combines selected participants instead of producing separate participant results.
Cluster types All four selected Includes or excludes the supported cluster contexts. Report columns and context-specific aggregate tables follow this selection.
Diacritic Options Compare diacritics Controls whether all or selected diacritics are ignored when a target and actual phone are classified as correct or substituted.
Aligned Phones Empty IPA Target and IPA Actual Phonex matchers Optionally restricts the aligned cluster productions included in the result.

The remaining expandable controls are the standard group, word, tier-data, and participant filters. See Common Query Parameters and Aligned Phones.

Figure 1. Measure of Cluster Proximity (MCP) Parameters
Measure of Cluster Proximity (MCP) Parameters

Cluster contexts

All four contexts are selected by default:

  • OOi — word-initial onset cluster.
  • OOm — word-medial onset cluster.
  • COm — word-medial coda-plus-onset sequence.
  • CCf — word-final coda cluster.

Scoring

The Pattern column describes the phone alignment. Each symbol represents one aligned position:

  • C — the target and actual phones are correct.
  • S — the target phone is substituted.
  • D — the target phone is deleted.
  • V — an epenthetic vowel occurs between the consonants.

The pattern determines the proximity value. Scores use 6.25-point increments where needed to place epenthesis patterns between their corresponding non-epenthesis patterns.

Table 2. Proximity values
Pattern Proximity Pattern Proximity
DD0CC100
DS, SD12.5SS75
SVD, DVS18.75CS, SC87.5
CD, DC25SVS37.5
CVD, DVC31.25CVS, SVC50
CVC62.5
Note:
If a result displays the fallback pattern XX, review its phone alignment. The analysis could not classify that alignment and assigns it a value of 0.

Report output

The report contains the following tables:

Table 3. MCP report tables
TableContents
Summary One row per participant, including role, cluster count, overall proximity, and the count and proximity for each selected context.
Participant Summary One row per session for that participant, with the same overall and context-specific measures.
Aggregate (all clusters) Counts each distinct IPA Target, IPA Actual, pattern, and proximity combination across the participant's sessions.
Aggregate (OOi/OOm/COm/CCf) Repeats the aggregate inventory for each selected context. A table reports No results when that context has no clusters.
Listing Lists every scored cluster with its session, speaker, record, target and actual cluster and word, context, pattern, and proximity.