A merger of /e/ and /ae/ in pre-lateral contexts has been reported among Australian English (AusE) speakers in Melbourne (Loakes et al. 2017). It appears to be completely entrenched for some speakers, but still in progress for others, with highly gradient individual production patterns (Author 1 2019). There are no previous investigations of the merger in spontaneous speech. We present an acoustic phonetic analysis from 12 AusE speakers (7F, 5M) of short front vowels /ɪ, e, ae/ in pre-alveolar stop (/t, d/) and pre-lateral contexts in a wordlist task (n=216 in /hVt/ and /hVd/ contexts; n=111 in /hVl/); and sociolinguistic interviews (n=2744 in pre-alveolar contexts; n=591 in pre-lateral contexts). Data was automatically segmented, with F1 and F2 taken at a selected point between 0-0.2 in each segmented vowel (following Author 1 2019). Findings show robust acoustic differences between /e/ and /ae/ preceding /t, d/ for all speakers across both data types. However, individual differences emerge for pre-lateral /e/ and /ae/, with highly variable and gradient production patterns. In wordlist data, this ranges from speakers with clear merger behaviour to those maintaining categorical distinctions. Differences also emerge between the data types, e.g. merger behaviour by one participant shows clear lowering of /e/ preceding /l/ in the wordlist, but striking retraction of /el/ beyond /ael/ in their interview. Some differences interact with compression of the vowel space in natural speech, but certain speakers also produce more vocalised laterals in /el/ than /ael/ in the interview, potentially influencing F2 variability. These findings illustrate the value of incorporating diverse data types, highlighting the “pivotal role of speech style” (Docherty et al. 2018: 1). Participants also undertook a perception task and a wordlist with simultaneous ultrasound tongue imaging. Future work will triangulate acoustic, articulatory and perceptual patterns to further elucidate the mechanisms underpinning the merger-in-progress.