This thesis is a sociolinguistic study into lexical regional variation in New Zealand English. The study utilises qualitative responses provided in an online survey and analyses them using quantitative methods. Classification and Regression Trees (CARTs) are employed with fixed and random effects to determine statistically significant trends in the data. This study focuses on testing a number of social and geographical factors such as age, gender, mobility, and region in order to determine the influences over any given individual’s preferred lexical choice. Anecdotally, many people in New Zealand, whether linguistically trained or not, are aware that there is a difference between the way people from the Southland region speak in comparison to people who live in other areas of New Zealand. However, Bauer and Bauer (2000) propose a three-way split pertaining to the regional speech in New Zealand, suggesting a Northern New Zealand English as well as the well documented Southland New Zealand English. This means that there would also be a standard central New Zealand English spoken throughout the majority of the country. This thesis seeks to determine whether the lexical regional trends seen in the study of Bauer and Bauer’s participants are continued on into today. The online survey analysed in this thesis was available between 2012 and 2017 and contained 36 questions pertaining to the words individuals use to describe both everyday items, childhood games and daily interactions. For example: • What do you call the public place you go to watch a film? • If you missed school without permission, for example to go to the park or into town, what word would you use to describe it? • A person who didn’t have any friends was called what? Of the 36 survey questions, seven are analysed in depth in this study, using data from 2400 New Zealand-born respondents. Three questions regard everyday items, and tend to show strong influences of time, as well as hints of regional variation. The remaining four questions relate to childhood games, and show clearer regional patterns, as well as influences of time. Although not every question demonstrates a three-way split between the northern, central and southern areas of New Zealand, there is evidence that the northernmost regions behave differently to the rest of the country, in addition to the predicted divergent behaviour of the southern regions. The preferences differ for each question, with single lexical items being the majority variant for most respondents in some cases, and up to six lexical items being somewhat equally preferred terms in other questions. Overall, although change over time is significant in predicting lexical choice, there is also strong quantitative evidence to argue for regional variation country-wide. These results have implications for identifying and documenting regional variation in New Zealand English as it emerges in the vocabulary of New Zealanders.