Statistical heterogeneity refers to differences among results of studies combined in a meta-analysis beyond that expected by chance. On the one hand, excessive heterogeneity can diminish power to discover genetic signals; on the other, moderate heterogeneity can reveal important biological differences among studies. Given its double-edged nature, this thesis dissects heterogeneity in genetic association meta-analyses from three vantage points. First, a novel multi-variant statistic, M is proposed to detect genome-wide (systematic) heterogeneity patterns in genetic association meta-analyses. This was motivated by the limited availability of appropriate methodology to measure the impact of heterogeneity across genetic signals, since traditional metrics (Q, I² and T²) measure heterogeneity at individual variants. Second, given that meta-analyses comprising small numbers of studies typically report imprecise summary effect estimates; GWAS-derived empirical heterogeneity priors are used to improve precision in estimation of average genetic effects and heterogeneity in smaller meta-analyses (e.g. ≤ 10 studies). Third, a critical evaluation of the Han-Eskin random-effects model shows how it can identify small effect heterogeneous loci overlooked by traditional fixed and random-effects methods. This work draws attention to the existence of genome-wide heterogeneity patterns, to reveal systematic differences among the ascertainment criteria of participating studies in a meta-analysis of coronary disease (CAD) risk. Furthermore, simulation studies with the Han-Eskin random-effects model revealed inflated genetic signals at small effect loci when heterogeneity levels were high. However, it did reveal an additional CAD risk variant overlooked by traditional meta-analysis methods. We therefore recommend a holistic approach to exploring heterogeneity in meta-analyses which assesses heterogeneity of genetic effects both at individual variants with traditional statistics and across multiple genetic signals with the M statistic. Furthermore, it is critically important to review forest plots for small effect loci identified using the Han-Eskin random-effects model amidst moderate-to-high heterogeneity (I² ≥ 40%).

Introdução: A Estimulação Transcraniana por Corrente Contínua (ETCC) é uma intervenção não farmacológica com resultados discordantes quanto à sua eficácia para o tratamento do Episódio Depressivo Maior (EDM). Possivelmente devido a heterogeneidade dos estudos, as três meta-análises publicadas até agora sobre o assunto apontam para uma pequena vantagem a favor da técnica, dependendo da variável de desfecho analisada (melhora de depressão ou resposta). Todas essas meta-análises utilizaram dados agregados. Focamo-nos, no presente estudo, na realização de uma revisão sistemática da literatura e de uma meta-análise baseada em dados individuais de pacientes (MA-DIP) com Depressão Maior (DM) submetidos à ETCC na fase aguda. Este tipo de análise é mais preciso na avaliação da eficácia de uma intervenção e na obtenção dos preditores de respostas de tratamento, já que as características individuais de cada sujeito são consideradas ao invés das médias e frequências, como nas meta-análises de dados agregados. Objetivos: (1) avaliar a eficácia da ETCC na fase aguda da DM, (2) identificar preditores de resposta, remissão e melhora da depressão específicos e (3) avaliar a aceitabilidade da intervenção. Resultados: Os dados foram colecionados de 6 ensaios clínicos randomizados placebo controlados, totalizando 289 sujeitos. A ETCC ativa foi estatisticamente superior à estimulação simulada em relação à resposta (34% vs. 19%, respectivamente; RC=2,44, IC 95% = 1,38-4,32, NNT=7), remissão (23,1% vs. 12,7%, respectivamente; RC= 2,38, IC 95% = 1,22 - 4,64, NNT=9) e melhora da depressão (coeficiente = 0,35, IC 95% =0,12 0,57). Demonstrou-se que após ajustes para outros preditores e confundidores, depressão resistente a tratamento e doses mais altas de ETCC foram, respectivamente, inversa e diretamente associadas com a eficácia da ETCC. Conclusões: O tamanho de efeito do tratamento com ETCC foi comparável àqueles reportados, em outros estudos, para a Estimulação Magnética Transcraniana repetitiva (EMTr) e para o tratamento farmacológico (com antidepressivos tricíclicos e inibidores seletivos de recaptação de serotonina) na atenção primária para depressão. Os parâmetros mais importantes para otimização em ensaios clínicos futuros são a refratariedade da depressão e a dose da ETCC / Introduction: Transcranial direct current stimulation (tDCS) is a nonpharmacological intervention for depression. Randomised, sham-controlled clinical trials (RCTs) conducted hitherto have presented mixed results regarding its efficacy. Although recent meta-analyses suggest some efficacy when measuring depression symptoms using a continuous outcome, these meta-analyses were limited in their results as they used an aggregate data approach. We aimed therefore to perform an individual patient data (IPD) meta-analysis. In contrast to an aggregate data meta-analysis, an IPD approach uses the raw data of each participant within a study. IPD is more accurate in estimating the efficacy of an intervention since aggregate data meta-analyses present only summary estimates of efficacy. IPD meta-analysis is also superior to the aggregate data approach for obtaining predictors of treatment outcome, as the characteristics of each patient are assessed instead of the mean and frequency values obtained in the traditional aggregate data meta-analysis. Objectives (a) To provide precise estimates of tDCS efficacy based on continuous (depression improvement) and categorical (response and remission rates) outcomes, (b) to identify variables associated with tDCS efficacy and (c) to estimate the treatment acceptability. Results: Data were gathered from six randomised sham-controlled trials, enrolling 289 patients. Active tDCS was significantly superior to sham for response (34% v. 19%, respectively, odds ratio (OR) = 2.44, 95% CI 1.384.32, number needed to treat (NNT) = 7), remission (23.1% v. 12.7%, respectively, OR = 2.38, 95% CI 1.224.64, NNT = 9) and depression improvement ( coefficient 0.35, 95% CI 0.120.57). Mixed effects models showed that, after adjustment for other predictors and confounders, treatment-resistant depression and higher tDCS doses were, respectively, negatively and positively associated with tDCS efficacy. Conclusions: The effect size of tDCS treatment was comparable with those reported for repetitive transcranial magnetic stimulation (rTMS) and antidepressant drug treatment in primary care. The most important parameters for optimisation in future trials are depression refractoriness and tDCS dose

Background: Breast cancer is the most frequently diagnosed cancer and a leading cause of cancer death among females, both worldwide and in the UK. Although, UK incidence of breast cancer is rising, breast cancer mortality rates are falling, due largely to early detection and improved treatment. As a result there are more women living with a diagnosis of breast cancer than ever before. Due mainly to side-effects of adjuvant therapy, breast cancer patients may require diagnostic, therapeutic, supportive or palliative services many years post-diagnosis, which poses a major challenge to already stretched healthcare services. Accordingly, effective and inexpensive interventions that can alleviate treatment side-effects, improve health, quality of life and potentially reduce risk of early mortality are required for breast cancer patients. Awareness of the positive influence that physical activity can have on breast cancer development and outcome is an important determinant of physical activity levels. A higher level of physical activity before and after breast cancer diagnosis is related to a lower risk of all-cause and breast cancer-related mortality. Randomised controlled trials have reported beneficial effects of physical activity interventions on outcomes relating to health, quality of life and mortality risk among breast cancer survivors. Aims: The present project aimed to: 1) assess awareness of the role of physical activity on breast cancer risk and the sufficiency of physical activity undertaken in women attending the NHS breast screening programme (NHSBSP), 2) compare physical activity levels of women at different stages of breast cancer pathway, 3) investigate the effects of a low-cost six-month home-based physical activity intervention on physical activity, body mass, health-related quality of life (HRQoL), insulin resistance and blood lipid profiles of breast cancer survivors and 4) assess the effects of our home-based intervention on cardiorespiratory fitness in a subset of breast cancer survivors. Methods: A total of 309 volunteers (188 NHSBSP attendees, 41 breast cancer patients undergoing chemotherapy and 80 post-treatment breast cancer survivors) participated in the current project. Physical activity was assessed via the International Physical activity Questionnaires (IPAQ). In studies one and two, Body mass and body mass index (BMI) were assessed directly in chemotherapy patients and breast cancer survivors, and indirectly from self-reported values in NHSBSP attendees. While in study three, body fat percentage was measured via bioelectrical impedance analysis, HRQoL was assessed using the Functional Assessment of Cancer Therapy-Breast (FACT-B) questionnaire and fasting blood samples were taken to measure lipid, glucose and insulin concentrations at baseline and post-six month home-based physical activity intervention. In study four, a random subsample of 32 breast cancer survivors undertook an exercise tolerance test to establish peak oxygen uptake values. Results: A high proportion (70%) of NHSBSP attendees engaged in low-moderate levels of physical activity and performed low amounts of recreational physical activity. Attendees demonstrated high awareness (75%) of the role of physical activity in reducing breast cancer risk but those categorised as “low activity” were significantly unaware of insufficiency of activity (p<0.05). Chemotherapy patients and breast cancer survivors had significantly lower levels of total physical activity than NHSBSP attendees (p<0.001 and p<0.05, respectively). The randomised controlled trial revealed significant improvements in total physical activity, body mass (p<0.05), BMI (p<0.05) HRQoL (breast cancer subscale, p<0.01; trial outcome index, p<0.05) and total (p<0.01) and low-density lipoprotein (p<0.05) cholesterol concentrations in the intervention group compared to usual care, and significant improvements in cardiorespiratory fitness (p<0.05) in a subsample of breast cancer survivors allocated to intervention. Conclusions: Physical activity interventions that incorporate strategies aimed at increasing awareness of recommended physical activity guidelines may be required in populations at risk of breast cancer. A relatively large proportion of women at risk of breast cancer may not be sufficiently exposed to the potential benefits of physical activity on breast cancer outcomes. Post-treatment breast cancer patients may be more receptive to physical activity interventions as the negative effects of chemotherapy begin to resolve, and therefore, may benefit from physical activity interventions. Results suggest that a low-cost home-based physical activity intervention with counselling and telephone support can improve the health and HRQoL of breast cancer survivors, which may in turn potentially reduce risk of breast cancer and cardiovascular disease-related mortality. Given the encouraging results and its highly portable and feasible nature, our intervention represents a promising tool for use in health and community settings to benefit large numbers of breast cancer survivors. The current project supports the inclusion of physical activity promotion as an integral component for the management and care of breast cancer survivors.

Achtsamkeitsmeditation hat in den letzten Jahren sowohl in Wissenschaft und Heilkunde als auch in der breiten Bevölkerung eine bemerkenswerte Popularität erlangt. Es existieren sehr viele Studien über die Wirkung von Meditation, und die meisten davon zeichnen ein großes Potential dieser Praxis. Bei einer genaueren Betrachtung wird allerdings deutlich, dass in Anbetracht der großen Menge vorhandener Studien nur sehr wenige theoretisch begründete Wirkungsstudien vorliegen. An diesem Punkt möchte die vorliegende Dissertationsschrift ansetzen. Entsprechend besteht das übergeordnete Ziel der Arbeit darin, die Wirkungen und Wirkmechanismen von Achtsamkeitsmeditation umfassend zu explorieren. Zu diesem Zweck wurde in einem ersten Schritt zusammengefasst, welche Wirkungen der Achtsamkeitsmeditation bisher durch kontrollierte Studien untersucht wurden. In den 51 bis September 2011 veröffentlichten kontrollierten Studien an erwachsenen gesunden Probanden wurde eine mittlere Effektstärke von r = .26, CI95% = [.22;.29] beobachtet. Mittelgroße Effekte fanden sich in den Variablenkategorien negative Persönlichkeitseigenschaften, Stressempfinden, Empfinden von Empathie, Aufmerksamkeit, Ausmaß an habitueller Ängstlichkeit und dem Abschneiden in Intelligenztests. Um zu verstehen, wie diese (und möglicherweise weitere) Wirkungen der Meditation zusammenhängen, wurde in einem nächsten Schritt eine umfassende Theorie über die durch Achtsamkeitsmeditation zu erwartenden Wirkungen konstruiert. In diese Theorie gingen theoretische Überlegungen über die Wirkung von Meditation aus der westlichen Psychologie, Interviews mit Meditierenden auf verschiedenen Erfahrungsstufen und theoretische Erörterungen aus den historischen buddhistischen Schriften ein. Das Resultat ist ein Modell, das verschiedene Schritte meditativer Veränderung beschreibt: (1) spezifische Verhaltensmuster, (2) dadurch hervorgerufene Bewusstseinszustände, (3) die Verbesserung bestimmter Fähigkeiten, (4) die Stärkung von Gleichmut und das Erlangen von Einsichten als intendierte Hauptwirkungen von Meditation, (5) weitere proximale Wirkungen, die durch die Hauptwirkungen hervorgerufen werden. Daneben wurde eine Reihe von Wirkmechanismen identifiziert, die die verschiedenen Schritte des Veränderungsprozesses miteinander verbinden. In einem letzten Schritt wurde die aufgestellte Theorie einem ersten Test unterzogen, um deren Gültigkeit zu prüfen. Im Rahmen einer querschnittlichen Studie wurden die postulierten Schritte und deren Zusammenhänge durch den Vergleich von erfahrenen Meditierenden mit Nichtmeditierenden und von Meditierenden mit verschiedenen Erfahrungsniveaus weitgehend bestätigt. Zusammenfassend kann festgehalten werden, dass die entwickelte Theorie über die Wirkungen und Wirkmechanismen von Meditation das Potential besitzt, die grundlagen- und anwendungsbezogene Forschung zum Thema Meditation durch neuartige Hypothesen und Erklärungsansätze zu bereichern.

A few scholars have argued that the Internet is a valuable channel for social marketing, and that practitioners need to rethink how they engage with target audiences online. However, there is little evidence that online social marketing interventions can significantly influence behaviours, while there are few evidence-based guidelines to aid online intervention design. This thesis assesses the efficacy of online interventions suitable for social marketing applications, presents a model to integrate behavioural change research, and examines psychological principles that may aid the design of online behavioural change interventions.The primary research project used meta-analytical techniques to assess the impact of interventions targeting voluntary behaviours, and examined psychological design and adherence correlations. The study found that many online interventions demonstrated the capacity to help people achieve voluntary lifestyle changes. Compared to waitlist control conditions, the interventions demonstrated advantages, while compared to print materials they offered similar impacts, but with the advantages of lower costs and broader reach. A secondary research project surveyed users across an international public mobilization campaign and used structural equation modelling to assess the relationships between website credibility, active trust, and behavioural impacts. This study found that website credibility and active trust were factors in behavioural influence, while active trust mediated the effects of website credibility on behaviour. The two research projects demonstrated that online interventions can influence an individual’s offline behaviours. Effective interventions were primarily goal-orientated: they informed people about the consequences of their behaviour, encouraged them to set goals, offered skills-building support, and tracked their progress. People who received more exposure to interventions generally achieved greater behavioural outcomes. Many of these interventions could be incorporated into social marketing campaigns, and offer individually tailored support capable of scaling to massive public audiences. Communication theory was used to harmonize influence taxonomies and techniques; this proved to be an effective way to organize a diversity of persuasion, therapy, and behavioural change research. Additionally, website credibility and users’ active trust could offer a way to mitigate the negative impacts of online risks and competition.

近年來電腦科技進步、資料庫健全發展，使得處理大量資料的需求增加，因而發展出結合生物醫學與資訊統計兩大領域的生物資訊(Bio-informative)。這個新學門的特色在於資料量及資料變數的龐雜，但過多資料經常干擾資訊的篩選，甚至癱瘓資料分析，因此如何適當地縮減資料(Data Reduction)就變得必要。資料縮減常藉由維度縮減(Dimension Reduction)進行，其中常見的線性維度縮減方法首推主成份分析，屬於非監督式學習(Unsupervised Learning)的一種，而線性的監督式學習(Supervised Learning)方法則有SIR(Sliced Inverse Regression)、SAVE(Sliced Average Variance Estimate)及pHd(Principal Hessian Directions)。非監督式學習的主成份分析，主要在找出少數幾個維度而可以解釋代表自變數的變異程度，而監督式學習的SIR、SAVE及pHd則可以在縮減維度時，同時考量自變數跟應變數之間的關係，而找出可以解釋應變數的維度。 本研究為解決蛋白質質譜儀資料高維度的問題，將應用各種線性維度縮減方法，並分別使用CART(Classification and Regression Tree)、KNN(K-Nearest Neighbor)、SVM(Support Vector Machine)、ANN(Artificial Neural Network)四種分類器，比較各維度縮減方法的分錯率高低，以交叉驗證(Cross Validation)比較維度縮減方法的優劣。研究發現在四種維度縮減方法中，PCA及SIR在各種分類器下都有較為穩定的分錯率，表現較為一致，但SAVE及pHd較不理想。我們也發現在不同的分類器下，PCA跟SIR兩者有不同表現，正確率較高的分類器(SVM與ANN)與PCA結合，而正確率較低的分類器(CART與KNN)與SIR結合，會有較佳的結果。另外，我們也嘗試整合分析(Meta Analysis)，綜合幾種線性維度縮減方法，而提出邊際訓練效果法(Marginal Training Effect Method)與加權整合法(Meta Weighted Method)，其中發現邊際訓練效果法若可以挑選出有效的維度，可以在不同分類器下提高整體模型，而加權整合法則確保在不同分類器下，讓其分類模型具有較為穩定的準確率；並提出相關係數重疊法(Overlap Correlation Method)來解決需要決定維度大小的問題。

People with a diagnosis of schizophrenia have been found to perform poorly on tasks assessing semantic memory, and these impairments have been proposed to be related to certain symptoms, in particular Formal Thought Disorder (FTD). A systematic literature review and meta-analysis identified the need a) to determine whether semantic memory is a primary impairment in schizophrenia and not secondary to other cognitive impairments and b) what cognitive models could provide the best explanation for the impairment. With these aims, Studies One and Two compared the performance of a group of people with schizophrenia across a battery of semantic memory tests (Hodges, Salmon and Butters, 1992). In order to eliminate confounding variables, two clinical control groups were recruited for comparison, one with a probable degraded semantic memory arising from Alzheimer‘s Dementia (AD) and the other with a primary dysexecutive syndrome caused by acquired brain injury (ABI). From these comparisons, it was possible to profile the semantic memory impairment in schizophrenia with the conclusion that any deficits are task-specific. Unlike the AD group, the impairment did not seem to arise from a loss of stored knowledge but nor did a retrieval problem, in its simplest terms, offer the best explanation. Since the ABI group performed normally on the battery it is clear that a dysexecutive syndrome does not necessarily explain poor semantic memory performance. Qualitatively, the associations and categories formed by people with schizophrenia on tasks of semantic categorisation e.g. the Category Generation Test (CGT) (Green, Done, Anthony, McKenna and Ochocki, 2004) often resemble loosening of associations and psychotic speech. In order to understand more about the processes involved in the formation of these bizarre categories, I compared performance on the CGT of groups of people with schizophrenia, AD and ABI. I found that the people with AD performed fairly similarly to the people with schizophrenia in that they sorted cards in an idiosyncratic way but the ABI group performed normally, adhering to taxonomic categories. Although this result might suggest that the bizarre associations on the CGT in people with schizophrenia are caused by a deficit in semantic memory (and not a dysexecutive syndrome), further analysis found important differences between the AD and the schizophrenia group in the way the card sorts were formed. In addition, both these groups showed intact semantic memory knowledge of the items they mis-sorted, indicating that categorisation problems do not necessarily arise from a degraded memory store. The difficulties people with schizophrenia appear to have on tests of associations and categorisation (e.g. CGT) could arise from a disorganised semantic memory i.e. differences in the way in which concepts are interconnected. On the CGT, patients with schizophrenia were far more likely to sort items on the basis of thematic (situational) information suggesting a preference for thematic over taxonomic associations. To test this, participants were tested using a triadic comparison task which requires choosing whether an item is best associated with a taxonomic, thematic or perceptually related item. On this test patients performed comparably to controls suggesting that their semantic memory is organised normally and that the abnormalities in the way in which items are associated on some semantic memory tests, including the CGT, are task-specific. It has been proposed that one of the core problems in schizophrenia is that there is ―an aberrant assignment of salience‖ (Kapur 2003) to contextually inappropriate concepts due to a dysregulated dopamine system (Kapur 2003; Kapur et al 2005). It is possible that this could also explain the semantic memory impairments in schizophrenia i.e. certain less relevant concepts/ associations are chosen because they are experienced as more salient. To test this, a group of patients with schizophrenia were assessed using a test of semantic salience. Compared to controls, the patients made significantly more errors of salience including significantly more errors where large aberrant attributions of importance were given to items. The tendency to make errors on the salience test was highly correlated with errors on the CGT and also the semantic association tests, indicating a common underlying mechanism. Therefore, it can be concluded that the semantic memory impairments in schizophrenia are task-specific, not caused by a loss of semantic knowledge or a dysexecutive syndrome, but due to an aberrant assignment of salience to less relevant semantic concepts. More work is needed to understand the cognitive processes underlying this aberrant attribution process, and also the biological substrates involved.

The information service sector e.g. utilities, telecommunications and banking has grown rapidly in recent years and is a significant contributor to the Gross Domestic Product (GDP) of the world’s leading economies. Though, the information service sector has grown significantly, there have been relatively few attempts by researchers to explore this sector. The lack of research in this sector has motivated my PhD research that aims to explore the pre-established relationships between dependability, quality and customer satisfaction (RQ1) within the context of information service sector. Literature looking at the interrelationship between the dependability and quality (RQ2a), and their further impact on customer satisfaction (RQ2b) is also limited. With the understanding that Business to Business (B2B) and Business to Customer (B2C) businesses are different, exploring these relationships in these two different types of information firms will further add to existing literature. This thesis also attempts to investigate the relative significance of dependability and quality in both B2B and B2C information service firms (RQ3a and RQ3b). To address these issues, this PhD research follows a theory testing approach and uses multiple case studies to address the research questions. In total five cases from different B2B and B2C information service firms are being investigated. To explore the causality, the time series data set of over 24 to 60 months time and the ‘Path Analysis’ method has been used. For the generalization of the findings, Cumulative Meta Analysis method has been applied. The findings of this thesis indicate that dependability significantly affects customer satisfaction and an interrelationship exists between dependability and quality that further impacts customer satisfaction. The findings from B2C cases challenges the traditional priority afforded to relational aspect of quality by showing that dependability is the key driver of customer satisfaction. However, B2B cases findings shows that both dependability and quality are key drivers of customer satisfaction. Therefore, the findings of this thesis add considerably to literature in B2B and B2C information services context.

Gliome stellen mit 30-40 % die häufigsten intrakraniellen Tumoren dar. Darunter ist das Glioblastom, auch als Glioblastoma multiforme bezeichnet, mit ca. 50 % am stärksten vertreten und macht somit allein etwa 20 % aller Hirntumoren aus (Russell und Rubinstein 1989). Neben einer Metastasierung über den Liquor können diese Tumoren auch sehr selten als primäre Neubildung im Rückenmark vorkommen. Ebenso ist dort die Entwicklung eines sekundären Glioblastoms aus dem fortschreitenden Wachstum eines niedriggradigen Astrozytoms heraus möglich (Sure et al. 1997). Aufgrund ihres diffus infiltrativen Wachstums sowie der infausten Prognose wird das Glioblastom in der WHO-Klassifikation dem Tumorgrad IV zugeordnet (Louis et al. 2007). Ziel dieser vorliegenden medizinischen Dissertation ist es, mit Hilfe einer Meta-Analyse individueller Patientendaten einen integrativen Überblick über die bereits vorhandene Forschung dieser seltenen Erkrankung zu erlangen. Aus einer umfangreichen Literaturrecherche resultierten 82 geeignete Publikationen, sodass von insgesamt 157 Patienten unter anderem Daten zum Veröffentlichungszeitraum, Alter und Geschlecht des Patienten, Tumorbiologie und –lokalisation sowie Therapie und Outcome erhoben werden konnten. In der statistischen Auswertung wurde die Korrelation dieser Faktoren eingehend untersucht mit dem Ziel, signifikante Aspekte herauszufiltern, die das Überleben von Patienten mit spinalem Glioblastom positiv beeinflussen können. Trotz einer stetigen Verbesserung der bildgebenden Diagnostik durch Magnetresonanztomographie und Computertomographie sowie multimodaler Tumortherapie zeigte sich in der Auswertung der Ergebnisse, dass das mediane Überleben der Patienten mit spinalem Glioblastom bzw. Gliosarkom auf nur rund neun Monate zu beziffern ist. Dabei konnten sowohl das Alter bei Diagnosestellung als auch das Jahr der Veröffentlichung der jeweiligen Patientenfälle als statistisch signifikante und somit bedeutende prognostische Faktoren im Hinblick auf die Überlebenszeit der Patienten identifiziert werden.

Turnover is one of the most important phenomena for management scholars and practitioners. Yet, researchers and practitioners are often frustrated by their inability to accurately predict why individuals leave their jobs. This should be worrisome given that total replacement costs can exceed 100% of an employee’s salary (Cascio, 2006) and can represent up to 40% of a firm’s pre-tax income (Allen, 2008). Motivated by these concerns, the purpose of this study was to assess the predictive validity of commonly-investigated correlates and, by extension, conceptualizations of employee turnover using a large-scale database of scientific findings. Results indicate that job satisfaction, organizational commitment, and embeddedness (e.g., person-job fit, person-organization fit) may be the most valid proximal predictors of turnover intention. Results for a tripartite analysis of the potential empirical redundancy between job satisfaction and organizational commitment when predicting turnover intention align well with previous research on this topic and generally suggest that the two constructs may be empirically indistinguishable in the turnover context. Taken together, this study has important implications for the turnover and sensitivity analysis literatures. With regard to the sensitivity analysis literature, this study demonstrates the application of a sensitivity analysis for relative importance weights in the meta-analytic context. This new method takes into account variance around the meta-analytic mean effect size estimate when imputing relative importance weights and may be adapted to other correlation matrix-based techniques (i.e., structural equation modeling) that are often used to test theory.