Matched guise effects can be robust to speech style

Abstract

When investigating how listeners evaluate the social meaning of variability in speech, researchers using the Matched Guise Technique (MGT) must decide whether to use read speech or conversational speech stimuli. An MGT experiment comparing social evaluation of /ɪŋ/ ∼ /ɪn/ variation in read and conversational speech styles found no evidence that the social evaluation of this variation differed across frame utterance styles. This suggests that use of read speech stimuli can be an appropriate methodological choice in MGT research.

Emerging from below the social radar: Incipient evaluation in the North West of England

“…in teaching (Baratta ). It has also been shown recently that matched‐guise techniques are not confounded by speech style, with subjects reacting similarly to the use of non‐standard forms in conversational speech as in elicited speech (Tamminga ). This is an important finding and one that allows us to generalise results beyond this specific experimental context with a reasonable degree of confidence.…”

6 Social Meaning and Variation in Perception: Beijingers’ Attitudes Towards Beijing Mandarin

“…The meaning of -in/-ing variation in English has been studied extensively, showing that talkers are associated with different personal attributes (e.g. intelligence) depending on whether they use the -in or -ing variant (Campbell-Kibler, 2006Tamminga, 2017). More recently, perceptions of gender and sexuality in language use have also been researched using digitally manipulated guises (Campbell-Kibler, 2011;Levon, 2006;Levon & Fox, 2014;Pharao et al, 2014).…”

Production and perception of the Pin-Penmerger

“…Participants from across the United States were recruited via Amazon Mechanical Turk, an online platform where people do tasks for pay over the Internet. In the past several years, Mechanical Turk and similar platforms have been used for a variety of linguistic research (e.g., Degen & Goodman, 2014;Kanwal, Smith, Culbertson & Kirby, 2017;Kim, Wyschogrod, Reddy & Stanford, 2016;Tamminga, 2017), and have been shown to produce results similar to those of laboratory studies (Berinsky, Huber & Lenz, 2012;Enochson & Culbertson, 2015;Tamminga, 2017;Wang, Huang, Yao & Chan, 2015). Participation was limited to users in the United States, and data from participants who reported having grown up outside of the United States or not speaking English as a native language were discarded (n=12).…”