Revisiting the Correct Information Unit: Measuring Informativeness in Unstructured Conversations in People With Aphasia

Abstract

Purpose The purpose of this study was to determine whether the correct information unit (CIU) can be reliably applied to unstructured conversational discourse in people with aphasia (PWA). The CIU was developed by Nicholas and Brookshire (1993) to measure word-level informativeness in structured monologue-level discourse and is widely used by clinicians and researchers for this purpose. A case study ( Oelschlaeger & Thorne, 1999 ) investigating the use of the CIU in conversation has suggested potential issues with interrater reliability (IRR), which has discouraged application of the CIU to this discourse context. However, no further research has been conducted to replicate or extend this finding. Given a clinical and research need for reliable linguistic measures appropriate for use in unstructured conversation, revisiting the reliability, stability, and suitability of the CIU is indicated. Method The CIU protocol developed by Nicholas and Brookshire (1993) was modified according to the needs of conversational discourse, resulting in the CIU in conversation (CIUconv) protocol. Two speech-language pathology graduate student research assistants completed training on use of the CIUconv with the 1st author. Sixteen conversations held by 8 PWA (i.e., 2 conversations each) were used as language samples to determine IRR of percent CIU (%CIU) in conversation through use of the CIUconv. Test–retest stability of %CIU as applied per this protocol was then assessed across the 2 conversations collected for each PWA. Results Use of the CIUconv resulted in excellent IRR of %CIU for each research assistant and the 1st author. Likewise, test–retest stability for the measure was excellent. Results were evaluated at both the group and individual levels. Conclusions %CIU demonstrated excellent interrater and test–retest reliability when applied to unstructured conversation using the CIUconv procedure, which was developed to account for expected linguistic characteristics of conversation. These findings suggest that %CIU may be a feasible, reliable measure of informativeness in unstructured conversation in PWA when the CIUconv is used by trained raters.

Spoken Discourse Assessment and Analysis in Aphasia: An International Survey of Current Practices

“…Another limitation in using discourse measures in outcome sets is that studies often report insufficient details around how language samples were collected and analyzed, which creates replication and reproducibility challenges. With few exceptions (e.g., Leaman & Edmonds, 2019;Roberts & Post, 2018), researchers have rarely (a) reported on how the raters responsible for transcribing, coding, and/or analyzing language samples were trained, or (b) made rater training documents freely available (e.g., see Thompson, n.d.).…”

Spoken Discourse Assessment and Analysis in Aphasia: An International Survey of Current Practices

“…Another limitation in using discourse measures in outcome sets is that studies often report insufficient details around how language samples were collected and analyzed, which creates replication and reproducibility challenges. With few exceptions (e.g., Leaman & Edmonds, 2019;Roberts & Post, 2018), researchers have rarely (a) reported on how the raters responsible for transcribing, coding, and/or analyzing language samples were trained, or (b) made rater training documents freely available (e.g., see Thompson, n.d.). Furthermore, in addition to a large number of spoken discourse measures being reported in the aphasia literature, only infrequently (e.g., Boyle, 2014Boyle, , 2015Brookshire & Nicholas, 1994;Pritchard et al, 2018) have researchers intentionally studied the psychometric properties of specific discourse measures or developed robust normative data (Dietz & Boyle, 2018;Pritchard et al, 2017;Wallace et al, 2018).…”

Informative and semantic language features of people with dementia displayed during reminiscence therapy

“…CIU analysis may also be a feasible, reliable measure of informativeness in unstructured conversations. 21 Therefore, understanding linguistic aspects through CIU analysis will contribute to understanding the comprehensive communication abilities of PWD.…”

“…One study reported that CIUs contribute to the overall meaningfulness and lexical informativeness of a given conversation. CIU analysis may also be a feasible, reliable measure of informativeness in unstructured conversations 21 . Therefore, understanding linguistic aspects through CIU analysis will contribute to understanding the comprehensive communication abilities of PWD.…”