Research on the dialects of English spoken within the United States shows variation regarding lexical, morphological, syntactic, and phonological features. Previous research has tended to focus on one linguistic variable at a time with variation. To incorporate multiple variables in the same analysis, this thesis uses a latent class analysis to perform a cluster analysis on results from the Harvard Dialect Survey (2003) in order to investigate what phonetic variables from the Harvard Dialect Survey are most closely associated with each dialect. This thesis also looks at how closely the latent class analysis results correspond to the Atlas of North America (Labov, Ash & Boberg, 2005b) and how well the results correspond to Joshua Katz's heat maps (Business Insider, 2013; Byrne, 2013; Huffington Post, 2013; The Atlantic, 2013). The results from the Harvard Dialect Survey generally parallel the findings of the Linguistic Atlas of North American English, providing support for six basic dialects of American English. The variables with the highest probability of occurring in the North dialect are ‘pajamas: /æ/’, ‘coupon: /ju:/’, ‘Monday, Friday: /e:/’ ‘Florida: /ɔ/’, and ‘caramel: 2 syllables’. For the South dialect, the top variables are ‘handkerchief: /ɪ/’, ‘lawyer: /ɒ/’, ‘pajamas: /ɑ/’, and ‘poem’ as 2 syllables. The top variables in the West dialect include ‘pajamas: /ɑ/’, ‘Florida: /ɔ/’, ‘Monday, Friday: /e:/’, ‘handkerchief: /ɪ/’, and ‘lawyer: /ɔj/’. For the New England dialect, they are ‘Monday, Friday: /e:/’, ‘route: /ru:t/’, ‘caramel: 3 syllables’, ‘mayonnaise: /ejɑ/’, and ‘lawyer: /ɔj/’. The top variables for the Midland dialect are ‘pajamas: /æ/’, ‘coupon: /u:/’, ‘Monday, Friday: /e:/’, ‘Florida: /ɔ/’, and ‘lawyer: /ɔj/’ and for New York City and the Mid-Atlantic States, they are ‘handkerchief: /ɪ/’, ‘Monday, Friday: /e:/’, ‘pajamas: /ɑ/’, ‘been: /ɪ/’, ‘route: /ru:t/’, ‘lawyer: /ɔj/’, and ‘coupon: /u:/’. One major discrepancy between the results from the latent class analysis and the linguistic atlas is the region of the low back merger. In the latent class analysis, the North dialect has a low probability of the ‘cot/caught’ low back vowel distinction, whereas the linguistic atlas found this to be a salent variable of the North dialect. In conclusion, these results show that the latent class analysis corresponds with current research, as well as adding additional information with multiple variables.

Half of the students who begin college do not complete a degree or certificate. The odds of completing a degree are decreased if a student has a low socio-economic status (SES), is the first in a family to attend college (first-generation), attends multiple institutions, stops out multiple times, reduces credit loads over time, performs poorly in major-specific coursework, has competing family obligations, and experiences financial difficulties. Stopping out of college does not always indicate that a student is no longer interested in pursuing an education; it can be an indication of a barrier or several barriers faced. Institutions can benefit themselves and students by utilizing person-centered statistical methods to re-engage students they have lost, particularly those near the end of their degree plan. Using demographic, academic, and financial variables, this study applied latent class analysis (LCA) to explore subgroups of seniors who have stopped out of a public four-year Tier One research intuition before graduating with a four-year degree. The findings indicated a six-class model was the best fitting model. Similar to previous research, academic and financial variables were key determinants of the latent classes. This paper demonstrates how the results of an LCA can assist institutions in the decisions around intervention strategies and resource allocations.

Emerging adulthood is often-overlooked in current gastrointestinal (GI) health research; however, epidemiological evidence suggests that GI disorders are increasing in this population. The purpose of this study was to first define common GI symptom subgroups within emerging adults and then to characterize these group differences with key biopsychosocial factors encompassing diet, depression and anxiety symptoms, as well as physical and social functioning related to quality of life. A total of 956 emerging adults from a southeastern US university were surveyed on GI symptoms, psychosocial factors, and demographics. Latent class analysis uncovered three statistically significant GI symptom patterns within the sample identified by the degree of severity: Normal (n=649), Mild (n=257), and Moderate (n=50). This study demonstrated that significant impairment in GI functioning emerges at much earlier ages that are commonly assumed. In addition, these GI symptom levels were associated with important biopsychosocial factors. Assessing GI functioning in emerging adults may provide important insights into understanding the development of FGIDs.

Latent class modelling is one method used in the evaluation of diagnostic tests when there is no gold standard test that is perfectly accurate. The technique demonstrates maximum likelihood estimates of the prevalence of a disease or a condition and the error rates of diagnostic tests or observers. This study reports the effect of departures from the latent class model assumption of independent misclassifications between observers or tests conditional on the true state of the individual being tested. It is found that estimates become biased in the presence of dependence. Most commonly the prevalence of the disease is overestimated when the true prevalence is at less than 50% and the error rates of dependent observers are underestimated. If there are also independent observers in the group, their error rates are overestimated. The most dangerous scenario in which to use latent class methods int he evaluation of tests is when the true prevalence is low and the false positive rate is high. This is common to many screening situations. / Thesis / Master of Science (MS)

Intimate partner violence (IPV) is a prevalent potentially traumatic experience that increases risk for posttraumatic stress symptoms (PTSS). However, there is still considerable heterogeneity in PTSS among women exposed to IPV. Research on IPV has examined the ways in which different characteristics of IPV exposure have separately related to risk for PTSS, specifically the type (physical, psychological, economic, sexual), frequency (number of incidents), severity (minor, severe), and mode of violence (in-person, online). However, it may be important to examine how the integration of these characteristics of IPV differ across ���������������������� ���� ������ ���� ������������ �������������������� �������������� ���������� The current study integrated these characteristics to assess classes of IPV and the relevant associations between concurrent and future PTSS. 264 women between the ages of 18-24 (Mage=20.41, SD=2.99) were recruited as part of a greater longitudinal study examining the relationship between PTSS and co-occurring psychopathology following exposure to IPV and/or sexual assault in the past year. Four classes of IPV across four characteristics of IPV (type, severity, frequency, and mode) were identified with latent class analysis (LCA). (1) history of both mild and severe psychological, physical, and sexual IPV in person and online, (2) history of mild and severe psychological IPV and mild sexual IPV occurring in person and online, (3) history of mild psychological IPV occurring in person and online, (4) past history of one type of IPV occurring in person. Class membership and concurrent and future PTSS were found to be associated with class membership.

No / Almost all studies in health research control or investigate socioeconomic position (SEP) as exposure or confounder. Different measures of SEP capture different aspects of the underlying construct, so efficient methodologies to combine them are needed. SEP and ethnicity are strongly associated, however not all measures of SEP may be appropriate for all ethnic groups. Methods We used latent class analysis (LCA) to define subgroups of women with similar SEP profiles using 19 measures of SEP. Data from 11,326 women were used, from eight different ethnic groups but with the majority from White British (40%) or Pakistani (45%) s, who were recruited during pregnancy to the Born in Bradford birth cohort study. Results Five distinct SEP subclasses were identified in the LCA: (i) "Least socioeconomically deprived and most educated" (20%); (ii) "Employed and not materially deprived" (19%); (iii) "Employed and no access to money" (16%); (iv) "Benefits and not materially deprived" (29%) and (v) "Most economically deprived" (16%). Based on the magnitude of the point estimates, the strongest associations were that compared to White British women, Pakistani and Bangladeshi women were more likely to belong to groups: (iv) "benefits and not materially deprived" (relative risk ratio (95% CI): 5.24 (4.44, 6.19) and 3.44 (2.37, 5.00), respectively) or (v) most deprived group (2.36 (1.96, 2.84) and 3.35 (2.21, 5.06) respectively) compared to the least deprived class. White Other women were more than twice as likely to be in the (iv) "benefits and not materially deprived group" compared to White British women and all ethnic groups, other than the Mixed group, were less likely to be in the (iii) "employed and not materially deprived" group than White British women. Conclusions LCA allows different aspects of an individual’s SEP to be considered in one multidimensional indicator, which can then be integrated in epidemiological analyses. Ethnicity is strongly associated with these identified subgroups. Findings from this study suggest a careful use of SEP measures in health research, especially when looking at different ethnic groups. Further replication of these findings is needed in other populations.

Background: Cannabis use (CU) and disorders (CUD) are highly prevalent among adolescents and young adults. We aim to identify clinically meaningful latent classes of users of cannabis and other illegal substances with distinct problem profiles. Methods: N= 3021 community subjects aged 14–24 at baseline were followed-up over a period ranging up to 10 years. Substance use (SU) and disorders (SUD) were assessed with the DSM-IV/M-CIDI. Latent class analysis (LCA) was conducted with a subset of N= 1089 subjects with repeated illegal SU. The variables entered in the LCA were CU-related problems, CUD, other SUD, and other mental disorders. Results: Four latent classes were identified: “Unproblematic CU” (class 1: 59.2%), “Primary alcohol use disorders” (class 2: 14.4%), “Delinquent cannabis/alcohol DSM-IV-abuse” (class 3: 17.9%), “CUD with multiple problems” (class 4: 8.5%). Range and level of CU-related problems were highest in classes 3 and 4. Comorbidity with other mental disorders was highest in classes 2 and 4. The probability of alcohol disorders and unmet treatment needs was considerable in classes 2–4. Conclusion: While the majority of subjects with repeated illegal SU did not experience notable problems over the 10-year period, a large minority (40.8%) experienced problematic outcomes, distinguished by clinically meaningful profiles. The data underline the need for specifically tailored interventions for adolescents with problematic CU and highlight the potentially important role of alcohol and other mental disorders.

Background: Cannabis use (CU) and disorders (CUD) are highly prevalent among adolescents and young adults. We aim to identify clinically meaningful latent classes of users of cannabis and other illegal substances with distinct problem profiles. Methods: N= 3021 community subjects aged 14–24 at baseline were followed-up over a period ranging up to 10 years. Substance use (SU) and disorders (SUD) were assessed with the DSM-IV/M-CIDI. Latent class analysis (LCA) was conducted with a subset of N= 1089 subjects with repeated illegal SU. The variables entered in the LCA were CU-related problems, CUD, other SUD, and other mental disorders. Results: Four latent classes were identified: “Unproblematic CU” (class 1: 59.2%), “Primary alcohol use disorders” (class 2: 14.4%), “Delinquent cannabis/alcohol DSM-IV-abuse” (class 3: 17.9%), “CUD with multiple problems” (class 4: 8.5%). Range and level of CU-related problems were highest in classes 3 and 4. Comorbidity with other mental disorders was highest in classes 2 and 4. The probability of alcohol disorders and unmet treatment needs was considerable in classes 2–4. Conclusion: While the majority of subjects with repeated illegal SU did not experience notable problems over the 10-year period, a large minority (40.8%) experienced problematic outcomes, distinguished by clinically meaningful profiles. The data underline the need for specifically tailored interventions for adolescents with problematic CU and highlight the potentially important role of alcohol and other mental disorders.

This study uses latent class analysis and latent transition analysis to model and compare patterns of offending over time for males and females by: (1) identifying qualitative dimensions of offending; (2) modeling how patterns of offending change over time; and (3) exploring factors related to patterns of offending. This is a secondary analysis of data from the Edinburgh Study of Youth Transition and Crime, a longitudinal study consisting of a cohort of about 4,000 young people from secondary schools in the City of Edinburgh who responded to questionnaires administered between 1988 and 2001, when they were about 12, 13, 14, and 15 years old. Previous studies of offending have used trajectory modeling to explore the course of offending from onset to termination, but the models are generally based on a count of types of offences aggregated across individuals over time, making it difficult to determine whether individuals exhibit more versatility or specialization in offending or switch offences from one point in time to another. In addition, most of the studies on patterns of offending have focused primarily on males. An understanding of patterns of offending over time for both males and females is important for the design and selection of developmentally appropriate prevention/treatment strategies. The present study adds to the literature by (1) further exploring the small and understudied literature on offence transitions; (2) examining more closely the development of female offending separately from and in relation to male offending; and (3) exploring a range of factors (criminogenic and non-criminogenic) related to the development of offending for both males and females. While the same number of qualitative dimensions (latent classes) characterised male and female offending in this study, there were some structural differences. There was also evidence of shifts in the qualitative dimensions for males and females over time. Finally, those factors classified as criminogenic were more likely to differentiate among the latent classes than those classified as non-criminogenic.

Objectives We first sought to assess if bicyclist typology was associated with health. Second, we investigated whether bicyclist typology was related to health through physical activity and commute bicycling. Finally, we sought to develop profiles of disposition toward commute bicycling following proposed changes to a specific destination and the significance of pertinent covariates. Methods Data from the 2014 Georgia State University-Bicycling Survey were used. We first estimated the adjusted odds of worse health-related quality of life by bicyclist typology. A mediation model was then used to estimate the relative total and direct effects of bicyclist typology on health-related quality of life and relative indirect effects through physical activity and commute bicycling. A finite mixture modeling approach was used to identify latent classes of disposition toward whether proposed changes to a specific destination would increase likelihood of commute bicycling. The manual 3-Step protocol was used to assess the effect of covariates on the probability of latent class membership. Results Respondents who had never bicycled, were not motivated to commute bicycle, and who required greater bicycle facilities to feel comfortable commute bicycling had higher odds of worse health-related quality of life. Physical activity and, to a lesser extent, commute bicycling status mediated the effect of bicyclist typology on health-related quality of life. The seven-class solution was decided on as the “best” model for disposition toward whether proposed destination improvements would increase the likelihood of commute bicycling. Several covariates were identified that impact the probability of latent class assignment. Conclusions Initial evidence of a health disparity by bicyclist typology was revealed. Physical activity appears to serve as the primary means through which bicyclist typology has an effect on health. Urban environments that make physical activity, including commute bicycling, more comfortable for a larger proportion of the population may be a potential important health intervention. Understanding the patterns of disposition toward whether proposed destination improvements would increase the likelihood of commute bicycling may assist in targeting and prioritizing commute bicycling-related interventions toward subpopulations of interest.