No full-text available
To read the full-text of this research,
you can request a copy directly from the authors.
Log-likelihood-based Pseudo-R2 in Logistic Regression: Deriving Sample-sensitive Benchmarks
Abstract
The literature proposes numerous so-called pseudo-R2 measures for evaluating “goodness of fit” in regression models with categorical dependent variables. Unlike ordinary least square-R2, log-likelihood-based pseudo-R2s do not represent the proportion of explained variance but rather the improvement in model likelihood over a null model. The multitude of available pseudo-R2 measures and the absence of benchmarks often lead to confusing interpretations and unclear reporting. Drawing on a meta-analysis of 274 published logistic regression models as well as simulated data, this study investigates fundamental differences of distinct pseudo-R2 measures, focusing on their dependence on basic study design characteristics. Results indicate that almost all pseudo-R2s are influenced to some extent by sample size, number of predictor variables, and number of categories of the dependent variable and its distribution asymmetry. Hence, an interpretation by goodness-of-fit benchmark values must explicitly consider these characteristics. The authors derive a set of goodness-of-fit benchmark values with respect to ranges of sample size and distribution of observations for this measure. This study raises awareness of fundamental differences in characteristics of pseudo-R2s and the need for greater precision in reporting these measures.
... To test the association between age and whether a request would belong to a certain genre, logistic regression models were calculated, and the adjustment of McFadden's pseudo-R² proposed by Horowitz (1982), R² MFH , was used to assess the model fit. Hemmert et al. (2018) found this pseudo-R² to be most stable across different sample sizes and asymmetric distributions of the binary outcome variable, which is a desirable feature in this analysis, where the requests for a certain genre are often just a small percentage of the total requests (Table 2). They also propose and .28, ...
... They also propose and .28, if the number of successes is above 38% and below 62% (Hemmert et al., 2018). The likelihood for boys and girls to request certain genres was estimated using 2 × 2 contingency tables of the requesting child's sex and whether the request was of the genre in question. ...
... All of these models suggest a very minor effect size, according to Hemmert et al. (2018); even though Figure 3 shows that the genre distribution within the request changes with age, there seem to be only minor effects of age on the likelihood to request a certain genre. It should also be noted that, in the cases of a capella, German songwriter, and theme song, either most or all of the requests were categorized as children's music. ...
In research on the development of musical preferences, children are often asked either to evaluate musical pieces that have been previously selected by adult experimenters, or to give self-reports on preferences for different musical styles. This study involved analyzing a sample of 1,412 freely and publicly expressed music requests from children aged between 4 and 11 years taken from a German radio program, with regard to age-and sex-specific differences. The music was categorized into genres, as listed on Spotify, and examined using methods of music information retrieval provided by the Spotify Developer application programming interface. Results showed that, at younger ages, the requests were generally more evenly distributed across different genres. Regarding single genres, we observed small positive relationships between age and the likelihood of requesting the genres pop or electro and small negative relationships between age and the likelihood of requesting A capella, German songwriter, indie, theme songs, or children's music. Furthermore, we found that boys requested significantly more rock or hip-hop, whereas girls had a higher tendency to ask for pop. Finally, age-(but no sex-) related differences regarding the Spotify features of valence and liveness of the requested music were found, which were related to the preference for children's music at younger ages. The results thus suggest that previous findings regarding differences between boys and girls and an increasing formation of distinct genre preferences during infancy also apply to single song requests made on a radio show.
... The core findings highlight how 50% of SAFE studies use the pseudo R 2 as a goodness of fit measure as the coefficient of determination R 2 cannot be applied to nonlinear categorical dependent models as a measure for goodness of fit [40][41][42]. However, Hemmert et al. [43] and Williams [44,45] argue that reporting unknown pseudo R 2 is meaningless given the plethora of existing measures and their definitional differences. However, only Mac an Bhaird [25] and Guercio et al. [35] acknowledge that they employ McFadden's Pseudo R 2 and none of the literature comments on these measures as a goodness of fit. ...
... However, AIC and BIC appear in only 7.1% of the SAFE literature. Table 2 shows common inference tests-to see if the model is significant-proposed in the econometrics literature for nonlinear binary dependent variables models such as the Wald test [45,48,59] and the likelihood ratio test [43,48,60]. However, Table 2 shows that the SAFE literature is scant on diagnostic testing for joint significance of variables with the Wald test present in 25% of studies and the Likelihood ratio test present in just 14.3% of literature with no discussion on the results of these tests. ...
Research on small and medium-sized enterprises (SMEs) access to bank finance is vital for the euro area economy. SMEs heavily represent the European business sector, employing around 100 million people and accounting for more than half of the Gross Domestic Product. Research studies in the field often rely on the ECB/EC Survey on the Access to Finance of Enterprises (SAFE). Many studies employ probit or logit models with categorical dependent variables derived from SAFE. The research findings show that hardly any study employs the simpler linear probability model (LPM), with a dominant lack of research providing evidence that justifies the model selection process and suitability. However, it is well known that different econometrics models can lack consistency and frequently yield different results. Yet, the literature has no consensus on the best econometric approach. In addition, there is a lack of robustness tests in the literature to ensure model validity, underlining the need for a comprehensive review of the methodological framework that dominates SAFE data use. This paper addresses the identified research gap by introducing a robust methodological framework that helps researchers identify and choose an appropriate categorical model when using SAFE data. The study adds significant value to the extant literature by identifying four criteria that need to be considered when selecting the appropriate model among three common binary dependent models: LPM, probit and logit models. The findings show that the probit model was appropriate is all cases but that the LPM should not be disregarded, as it can be used in two cases: when considering the interaction between monetary policy and debt to assets and monetary policy and innovation. The use of the LPM is justified as a less complex econometric model, allowing for clearer communication of the results. This innovative, robust approach to choosing the appropriate econometric categorical dependent model when employing SAFE data contributes to support policy effectively.
... The use of "nnet" package (Ripley, 2022) was to multinomial logistic regression and odd ratios analyses. The log-likelihood-based pseudo-R² reported in multinomial regression models cannot demonstrate the proportion of explained variance on the dependent variable as the ordinal R² outcomes can (Fagerland & Hosmer, 2012;Hemmert et al., 2016). Thus, the use of "generalhoslem" package (Jay, 2019) was to assess the goodness-of-fit of the multinomial logistic models in this study. ...
... Table 3 shows the result of multinomial logistic regression for natural ventilation (NV) usage patterns. The reported Nagelkerke's pseudo-R² of 0.29 represented an improvement in the likelihood of this model over a null model (Hemmert et al., 2016). The goodness-of-fit test of this model failed to reject the null hypothesis (p-value > 0.05) as desired, and thus, there was no evidence of a poor model fit in this case. ...
Natural ventilation is a default conditioning strategy in the Brazilian residential sector, while fans and air conditioners are complementary strategies. However, climate change and the increasing air conditioning penetration in this sector threaten the prevalence of natural ventilation and the potential wind-driven (breeze) performance on households’ thermal comfort. A questionnaire survey launched across Brazil assessed multiple aspects of natural ventilation at home: perceptions, usage patterns and motivations behind its use or avoidance. Data analysis methods were multinomial logistic regression and contingency tables of categorical data. The findings indicated that households’ preference for a conditioning strategy related to income and energy-saving concerns (economic aspects). The frequency of use of natural ventilation showed a decreasing trend towards the higher income level and preference for air conditioning. In contrast, the frequency of use of natural ventilation tended to increase as households considered it more positively. Moreover, participants who preferred to use natural ventilation at home expressed less dissatisfaction with the oscillation and unpredictability of the breeze from natural ventilation. The survey outcomes highlight the benefits of a favourable scenario for natural ventilation at home, potentially impacting households’ preferences and routines.
... The performance of the logit model was analysed in terms of its predictive capacity and goodness of fit (Pseudo-R 2 ). In particular, the McFadden Pseudo-R 2 was used because it reflects the criterion being minimised in the logistic regression estimation as well as the variance accounted for by the logistic regression model [50]. The IBM SPSS Statistics software, version 28, was used to perform the logit model and to analyse goodness of fit. ...
... In terms of the model's goodness of fit (Table 5), the McFadden Pseudo-R 2 value is 0.449. Considering that a McFadden Pseudo-R 2 value ranging from 0.200 to 0.400 indicates a 'good' model fit and >0.400 an 'excellent' model fit [50], the model performed can be said to have an 'excellent' goodness of fit. Moreover, in terms of predictive capacity (86.5 %), results again confirm this goodness of fit. ...
New mobility technologies related to autonomous, connected, and shared vehicles have prompted the entry of new players into the automotive industry, which has influenced the industry’s traditional configuration of regional status. Under the global value chain (GVC) approach, this research proposes a new framework for defining a ‘core–periphery’ spatial model of the automotive industry. Under that model, based mainly on the key
variables of domestic firms linked to new mobility technologies, analysis is made of the comparative status of regions of the European automotive industry traditionally regarded as peripheral (Portugal) and semi-peripheral (Spain). Results indicate that domestic firms located in each of those two regions do not differ in terms of
decision-making power, first-level supply positioning, added value of activities, and technological innovation.
This implies that the two regions now share the same status within the new (autonomous, connected, and shared) mobility value chain. This has relevant implications for public policies throughout the European automotive industry. Policies should focus on innovation in new mobility technologies and on the creation of an ecosystem
adequate to develop strong domestic capabilities around these new mobility technologies, in order to ensure more favourable regional status in the spatial model of this competitive industry.
... McFadden porque refleja el criterio que se minimiza en la estimación de la regresión logística, así como la varianza explicada por el modelo de regresión logística (Hemmert et al., 2016). Se utilizó el software IBM SPSS Statistics, versión 28, para realizar el modelo logístico y analizar la bondad del ajuste. ...
... En la Tabla 9 se presentan los estadísticos descriptivos de las variables de las dos muestras. (Hemmert et al., 2016), puede decirse que el modelo realizado tiene una "buena" bondad de ajuste. Además, en términos de capacidad predictiva (81,1%), los resultados confirman de nuevo una 'buena' bondad de ajuste. ...
El desarrollo de modelos de negocio digitales está afectando a la cadena de valor tradicional de la movilidad, lo que implica nuevos retos y cambios a futuro para los fabricantes de automóviles. Además, las nuevas tecnologías de movilidad han propiciado la entrada de nuevos actores en la industria, lo que ha influido en la división espacial tradicional del trabajo y en la configuración del estatus regional de la industria.
Adoptando el enfoque de Cadena de Valor Global, la tesis analiza la evolución, orientación a corto plazo y visión de futuro de la adopción de modelos de negocio digitales de movilidad por parte de los fabricantes de automóviles. Asimismo, en esta tesis se propone un nuevo marco para definir un modelo espacial "centro-semiperiferia- periferia" de la industria automovilística. Bajo este modelo, basado principalmente en las variables clave de las empresas nacionales vinculadas a las nuevas tecnologías de la movilidad, se analiza el estatus comparativo de las regiones de la industria europea del automóvil tradicionalmente consideradas periféricas y semiperiféricas.
Sin embargo, la digitalización actual no solo ha cambiado la Cadena de Valor Global, sino que también ha redefinido el panorama empresarial, donde la gestión de datos es esencial. Herramientas como la Inteligencia Artificial y el Big Data abren nuevas posibilidades, especialmente desafiantes en entornos B2B por la incertidumbre tecnológica y la competencia. A su vez, estas circunstancias provocan que el emprendimiento sea especialmente difícil y que los emprendedores se encuentren en constante alerta. En esta línea, la tesis explora estas oportunidades de emprendimiento y proponen acciones para mejorar las perspectivas en los servicios B2B en la nueva economía digital.
Los resultados obtenidos muestran que las empresas están adoptando modelos de negocio centrados en plataformas digitales y servicios de conectividad en la industria de la movilidad. La recolección y gestión de datos, junto con la interconectividad, son aspectos clave para el desarrollo futuro tanto de la industria como de estas empresas. La digitalización está transformando las relaciones en la cadena de valor, lo que podría implicar que los fabricantes de automóviles cedan parte de su control a nuevos actores,
como son los socios tecnológicos y de servicios, lo que implicará una mayor distribución del poder a lo largo de la cadena de valor.
Además, los resultados sugieren que las empresas nacionales de las regiones analizadas no difieren en términos de poder de decisión, posicionamiento de primer nivel, valor añadido o innovación tecnológica. Por lo tanto, puede concluirse que, actualmente, comparten el mismo estatus dentro de la nueva cadena de valor de la movilidad. Esto tiene implicaciones relevantes para las políticas públicas en toda la industria europea del automóvil. Las políticas deberían centrarse en la innovación en nuevas tecnologías de movilidad y en la creación de un ecosistema adecuado para desarrollar industrias nacionales alrededor de estas, con el fin de garantizar un estatus regional más favorable en esta industria competitiva.
Los resultados también ofrecen una nueva perspectiva tanto de la oportunidad empresarial como de la alerta emprendedora, destacando que los modelos de negocio más prometedores en B2B son aquellos relacionados con la inteligencia artificial, la gestión de datos y la sostenibilidad. Por lo tanto, se recomienda que las empresas inviertan en estas áreas y limiten otras como el alquiler de vehículos o la personalización de servicios. Además, los emprendedores deben adoptar la alerta emprendedora, fomentando la innovación, la creatividad y la adaptabilidad; la gestión del riesgo, la creación de redes y el desarrollo de habilidades empresariales son competencias altamente valoradas, mientras que los entes públicos deberían actuar indirectamente a través de medidas legales y fiscales.
... where ( ) and ( ) indicate the likelihood of the target prediction mo and baseline model (null model), respectively. Similar to R-squared used in linear reg sion to calculate the proportion of explained variance, pseudo-R-squared measures degree of improvement in the model likelihood over a null model, which is a simple b line model containing no predictor variables (Hemmert et al. 2018). ...
... where L M target and L(M baseline ) indicate the likelihood of the target prediction model and baseline model (null model), respectively. Similar to R-squared used in linear regression to calculate the proportion of explained variance, pseudo-R-squared measures the degree of improvement in the model likelihood over a null model, which is a simple baseline model containing no predictor variables (Hemmert et al. 2018). ...
Survival models have become popular for credit risk estimation. Most current credit risk survival models use an underlying linear model. This is beneficial in terms of interpretability but is restrictive for real-life applications since it cannot discover hidden nonlinearities and interactions within the data. This study uses discrete-time survival models with embedded neural networks as estimators of time to default. This provides flexibility to express nonlinearities and interactions between variables and hence allows for models with better overall model fit. Additionally, the neural networks are used to estimate age–period–cohort (APC) models so that default risk can be decomposed into time components for loan age (maturity), origination (vintage), and environment (e.g., economic, operational, and social effects). These can be built as general models or as local APC models for specific customer segments. The local APC models reveal special conditions for different customer groups. The corresponding APC identification problem is solved by a combination of regularization and fitting the decomposed environment time risk component to macroeconomic data since the environmental risk is expected to have a strong relationship with macroeconomic conditions. Our approach is shown to be effective when tested on a large publicly available US mortgage dataset. This novel framework can be adapted by practitioners in the financial industry to improve modeling, estimation, and assessment of credit risk.
... or Cox & Snell R square = .07 (Hemmert et al., 2018). However, the model was significantly related to the outcome of muscle building supplement use, χ 2 (5) = 32.34, ...
... Hosmer and Lemeshow tests found that the data were a good fit for the model, p = .271. The overall model did not indicate good fit to the data according to Hemmert et al. (2018); Cox & Snell R square = .05 and Nagelkerke R square = .09, ...
Introduction
The demand for appearance and performance enhancing substances, including muscle building supplements and anabolic androgenic steroids, is increasing in Australia. However, little is known about the associations between appearance and performance-based factors and appearance and performance enhancing substances (APES), particularly among adolescent boys. This study sought to examine (a) the prevalence of muscle building supplement use in a sample of adolescent boys and (b) how both performance and appearance factors relate to muscle building supplement use and favourable attitudes towards anabolic androgenic steroids in this sample.
Method
N = 488 adolescent boys aged 13–16 ( Mage = 14.59) from nine Australian schools completed measures of supplement use, favourable attitudes towards using steroids, muscle dissatisfaction, body fat dissatisfaction, mesomorphic ideal internalisation, weight training, and sports participation. Hierarchical logistic regressions were used to examine cross-sectional correlates of muscle building supplement use and favourable attitudes towards using anabolic androgenic steroids.
Results
In the past three months, 12.7% of the sample had used muscle building supplements. Both appearance and performance-related factors – mesomorphic ideal internalisation and weight training – were related to muscle building supplement use. Only one appearance-related factor – body dissatisfaction – was related to favourable attitudes towards anabolic androgenic steroids.
Discussion
The findings from this study are important as they may help to guide intervention strategies regarding appearance and performance enhancing substance use by Australian adolescent boys, with the ultimate goal of ensuring this population’s safety.
... Pseudo r² was calculated for the adjusted models to compare them. Unlike linear regressions, pseudo r² does not represent the proportion of variance explained by the model but rather the improvement of the model compared to a null model (Hemmert et al., 2018). The analyses were conducted using STATA version 13. ...
Various factors may affect the likelihood of individuals who commit offenses during adolescence continuing to offend into adulthood. This study aimed to: (1) Describe and compare recidivism rates among 350 adult men who had gone through the juvenile system in the countryside of São Paulo; (2) Evaluate the prediction of recidivism according to psychosocial profiles; (3) Assess the influence of race on recidivism. Official recidivism data from the sample were collected and analyzed using logistic regression analysis, revealing that a more markedly problematic psychosocial profile was associated with greater chances of recidivism, while being Black was linked to higher chances of criminal prosecution. This study highlights the importance of identifying which psychosocial profiles are associated with a higher likelihood of persistent offending to target more effective interventions. It also reveals the presence of racial bias in the Brazilian criminal justice system, indicating structural racism.
... This method (see Budescu 1993;Azen andBudescu 2000, 2001 for more details) was utilized to rank the predictors by their relative importance by calculating the incremental pseudo-R 2 of each predictor when added to a predictor model. Unlike its linear analogue, pseudo-R 2 is computed by a logistic regression, which maximises the likelihood of the model involving a combination of predictors through multiple sample sizes from the observations (Hemmert et al. 2018). It indicates the improvement of the model over the null model (the null model assumes no dependency on any predictor). ...
Terrestrial ecosystems are one of the major sinks of atmospheric CO2 and play a key role in climate change mitigation. Forest ecosystems offset nearly 25% of the global annual CO2 emissions, and a large part of this is stored in the aboveground woody biomass. Several studies have focused on understanding the carbon sequestration processes in forest ecosystems and their response to climate change using the eddy covariance (EC) technique and remotely sensed vegetation indices. However, very few of them address the linkage of tree-ring growth with the ecosystem-atmosphere carbon exchange, and nearly none have tested this linkage over a long-term (> 100 years) — limited by the short-term (< 50 years) availability of measured ecosystem carbon flux. Nevertheless, tree-ring indices can potentially act as proxies for ecosystem productivity. We utilise the Coupled Climate Carbon Cycle Model Intercomparison Project (C4MIP) model outputs for its 140-year-long simulated records of mean monthly gross primary productivity (GPP) and compare them with the tree-ring growth indices over the northwestern Himalayan region in India. In this study, we examine three coniferous tree species: Pinus roxburghii and Picea smithiana wall. Boiss and Cedrus deodara and find that the strength of the correlation between GPP and tree ring growth indices (RWI) varies among the species.
... This corresponds with recent range shifts of Bewick's Swans, as individuals will winter further east when temperatures are higher (Linssen et al. 2023), exhibiting the same individual plasticity in habitat selection at a larger scale. Although the low R 2 of 0.13 for the GAMM fitted to these data could arise from missing explanatory factors, models often produce low R 2 values with larger sample sizes (Hemmert et al. 2018). Therefore, swans will spend more time outside PAs during periods of flooding or freezing, making them more vulnerable to threats, such as disturbance, under these conditions. ...
Protected areas are one of the major tools used in the conservation of biodiversity, but animals are unlikely always to remain within these human-made boundaries. Understanding when and why species choose to leave protected areas can help us to improve the effectiveness of these management tools. Here, we investigate the use of protected and non-protected areas by two migratory species undergoing rapid wintering population changes in northwest Europe: Whooper Swans Cygnus cygnus and Bewick's Swans Cygnus columbianus bewickii. Global positioning system tags were fitted to 15 Whooper Swans in winter 2008/09 and to 18 Bewick's Swans from winter 2013/14 to 2014/15 at the Ouse Washes Special Protection Area (an internationally important roost for wintering waterbirds) and on adjacent fields in southeast England. Here, swans feed on farmland during the day but return to designated reserves to roost at night, where they receive protection from predators and disturbance within managed roost habitats. When swans roost elsewhere at alternative sites, they may face more adverse conditions, and so understanding the extent and causes of the use of alternative roosts is important for swan conservation efforts. The alternative roosting proportion, defined as the proportion of nights spent outside protected reserves, was 0.237 for Bewick's Swans and challenging to quantify accurately for Whooper Swans. A generalized additive mixed model to model repeated measurements on individuals showed that the proportion of time that Bewick's Swans spent at alternative roosts correlated positively with river level and negatively with temperature. Competition and foraging flight distances are thought to drive these relationships, as swans seek access both to roost space and to nearby feeding habitats. Our findings improve our understanding of the environmental conditions under which migratory waterbirds may choose to roost outside protected areas.
... Furthermore, the percent correct predictions (PCP) and pseudo-R 2 are used to interpret the predictive ability and fitting degree of the model respectively (Li et al., 2013). In the case where the pseudo-R 2 > 0.2, it shows a relatively good fit (Allison, 2013;Hemmert et al., 2018). To assess how well the LR model fits the data the likelihood function (−2log(likelihood), and pseudo-R 2 values (McFadden R 2 , Cox and Snell R 2 , Nagelkerke R 2 ) were used, indicating the proportion of variability in the response variable explained by the model. ...
Urban sprawl is at the centre of contemporary urban debates and is very often criticised for its negative social and environmental externalities. Building on geo-spatial methodology and field observations, this study sets out to develop a holistic approach to understanding urban sprawl through a particular focus on a rapidly growing metropolitan urban agglomeration of northern India, that is, Varanasi. Following an analysis of principal urban growth modes through landscape expansion index and changing patterns of urban landscape through landscape metrics, this research refers to a geo-spatially grounded logistic regression model in order to shed light on the dynamic impact of multiple growth factors on the trajectory of urban sprawl in the study area. Moreover, in order to address the evolving nature of urban sprawl in Varanasi and its immediate surroundings this study focuses on a temporal span of 20 years, encompassing two consecutive decades of the 21st century (2001–2011 and 2011–2021). The research concludes by paving a platform for an amalgamation of the findings from different geo-spatial metrics as well as by corroborating the principal findings with the propositions of a number of contemporary hypotheses concerning the spatial growth of urbanised territories.
... These measures included pseudo-R 2 measures, the relative importance of independent The pseudo-R 2 measures were calculated, indicating that the IEN characteristics explain approximately 11%-18% of the variability in successful outcomes (Table 3). This suggests that the model exhibits a moderate level of goodness-of-fit (Hemmert et al. 2018). ...
Aims
This article describes the sociodemographic characteristics of internationally educated nurses since the change in the registration examination in 2015. It aims to investigate the association between internationally educated nurses' sociodemographic characteristics and their successful integration into the nursing workforce in Canada.
Design
Cross‐sectional and secondary data survey questions.
Methods
This study adopts a cross‐sectional and secondary data analysis, utilising data from IENs who engaged with internationally educated nurse initiatives such as the Creating Access to Regulated Employment Centre for Internationally Educated Nurses (CARE) or initiated the registration process with the College of Nurses of Ontario (CNO) in 2015 and after.
Results
There were 259 participants, with 155 participants from primary data collection and 104 participants from secondary data sources. Quantitative analysis reveals that most participants are females, under 40 years old, educated in English and hold at least a bachelor's degree in nursing, with 47.3% of internationally educated nurses migrated from India and the Philippines. Significant associations were identified between internationally educated nurses having CARE membership and the currency of nursing practice and their successful outcomes.
Conclusion
Recognising and addressing the unique needs of IENs is essential for their successful integration into the Canadian healthcare workforce, thereby ensuring resilience and cultural competence in nursing for the future.
Implications for the Profession
This analysis highlights the impact of sociodemographic characteristics of internationally educated nurses on their successful outcomes and underscores the diversity and richness they bring to the healthcare landscape. Since internationally educated nurses continue to experience challenges while integrating into the Canadian nursing workforce, these findings have substantial implications for nursing policy, practice, professional development and research.
... Finally, the model can provide valuable information for predicting future outcomes, making it a valuable tool in practical applications. Generally, the logistic regression model is a versatile and powerful statistical tool for analyzing binary outcomes and is widely used in medicine, social sciences, marketing, and finance, among others [18]. ...
... Goodness of fit was described using McFadden pseudo-R 2 . 16 The null hypothesis was that neither tube load, reconstruction strength, nor slice thickness influenced the Table 2 Image quality criteria and grading scores. ...
Introduction: Low contrast resolution in abdominal computed tomography (CT) may be negatively affected by attempts to lower patient doses. Iterative reconstruction (IR) algorithms play a key role in mitigating this problem. The reconstructed slice thickness also influences image quality. The aim was to assess the interaction and influence of patient dose, slice thickness, and IR strength on image quality in abdominal CT. Method: With a simultaneous acquisition, images at 42 and 98 mAs were obtained in 25 patients. Multiplanar images with slice thicknesses of 1, 2, and 3 mm and advanced modeled iterative reconstruction (ADMIRE) strengths of 3 (AD3) and 5 (AD5) were reconstructed. Four radiologists evaluated the images in a pairwise manner based on five image criteria. Ordinal logistic regression with mixed effects was used to evaluate the effect of tube load, ADMIRE strength, and slice thickness using the visual grading regression technique. Results: For all assessed image criteria, the regression analysis showed significantly (p < 0.001) higher image quality for AD5, but lower for tube load 42 mAs, and slice thicknesses of 1 mm and 2 mm, compared to the reference categories of AD3, 98 mAs, and 3 mm, respectively. AD5 at 2 mm was superior to AD3 at 3 mm for all image criteria studied. AD5 1 mm produced inferior image quality for liver pa-renchyma and overall image quality compared to AD3 3 mm. Interobserver agreement (ICC) ranged from 0.874 to 0.920. Conclusion: ADMIRE 5 at 2 mm slice thickness may allow for further dose reductions due to its superiority when compared to ADMIRE 3 at 3 mm slice thickness. Implications for practice: Combination of thinner slices and higher ADMIRE strength facilitates imaging at low dose.
... Overall model performance was evaluated using several different metrics. These included evaluating the null and residual deviance, Akaike Information Criteria (AIC) value, Bayesian Information Criteria (BIC) value, and Tjur's and Nagelkerke's Pseudo-R 2 to assess for changes in model likelihood compared to a null model (Hemmert et al., 2018). Moreover, prediction or classification performance was also evaluated for each model using the area under the curve (AUC) value for a receiver operating characteristic (ROC) analysis of both the training and testing datasets. ...
Numerous differences exist between and within research projects related to assessment and operationalization of potentially traumatic events (PTEs) for youth, especially when measuring polyvictimization. However, few studies have systematically examined how polyvictimization measurement differences influence PTE’s relation to functioning. This study sought to address these knowledge gaps by conducting a secondary data multiverse replication (SDMR) to systematically (re)evaluate PTE polyvictimization measurement approaches. Participants included 3297 adolescents (Mage = 14.63; 50.59% female; 65.15% white) from the National Survey of Adolescents-Replication study who completed a structured interview on PTE exposure and emotional and behavioral health (i.e., posttraumatic stress and major depressive disorder, drug and alcohol use, and delinquency). Results indicated that PTE operationalizations using a count variable tended to demonstrate better model performance and prediction of youth at-risk of emotional and behavioral health challenges, compared to models using a binary (yes/no) PTE operationalization. Differences in model performance and prediction were less distinct between models examining multiple forms of a single type of PTE (e.g., maltreatment, community violence), compared to models examining multiple PTE types. These findings emphasize the importance of using multidimensional approaches to PTE operationalization and the need for more multiverse analyses to improve PTE evidence-based assessment.
... While we find highly significant coefficients for our main explanatory variables, overall, the models explain 2%-3% of the variation in the dependent variable, based on McFadden R 2 -squared values (Hemmert et al., 2018). In this sense, there might be rival explanations not accounted for in our modeling. ...
Citizens experience onerous encounters with the bureaucracy for various reasons, often political. Administrative burden reduction (ABR) has been pursued to improve citizen-state interactions, especially for vulnerable populations who are disproportionately impacted by burdens. This study seeks to explain the degree of ABR by bureaucrats when the burdens are deployed by their political superiors. We conceptualize it as a function of client vulnerability and bureaucrats’ sense of job security and organizational commitment. We examine these linkages in the context of a COVID-19 rental assistance program for two vulnerable groups—elderly and Blacks. The findings from the two single factorial experiments show that clients’ vulnerability increases the degree of ABR, but only for the elderly. Moreover, bureaucrats who make decisions based on their organizational commitment approach ABR more slowly and only in the context of age vulnerability.
... Table 5 shows the results of the ordinal logistic regression. (Hemmert et al., 2018), showing a 25.6% improvement in the prediction of the outcome relative to the intercept-only model. The variable Interest is the main predictor for the overall value rating in the model. ...
A relevant learning space for academics, especially junior researchers, is the academic conference. While conference participation has long been associated with personal attendance at the conference venue, virtual participation is becoming increasingly important. This study investigates the perceived value of a purely virtual academic conference for its participants by analyzing the evaluation feedback (N = 759) from three virtual and two face-to-face LAK conferences. For the purposes of this study, we derive a definition of conference value and identify factors contributing to the overall value rating of virtual academic conferences based on the existing literature. Results indicate a perceived value of virtual conferences comparable with that of face-to-face events, satisfaction with social interaction and topics of interest being the most important predictors. Our insights show that virtual conferences are valuable events for academic professional development and conference organizers can utilize these results to design a valuable event for their participants.
... In the second analysis, we used a logistic regression with a quasibinomial generalized linear model (GLM) to predict the frequency of encountering a caiman using gillnet frequency of encounter as a predictor in each grid cell and modeling parameters by the maximum likelihood method (Zuur et al., 2009). We chose a quasibinomial distribution because of overdispersion and confirmed model fits by log-likelihood (pseudo R 2 = 0.12; Hemmert et al., 2018) and identified a significant improvement in the model over the base model (likelihood ratio p < 0.01) (Archer et al., 2007). Areas not surveyed were excluded from our analysis. ...
1. Artisanal fishing is an important subsistence practice in freshwater habitats
worldwide, but overexploitation threatens the conservation of several nontarget
species including crocodylians. We investigated the effects of artisanal fishing on
the distribution of a population of broad-snouted caiman (Caiman latirostris)
inhabiting the Tapacurá reservoir, within the highly altered and threatened
Atlantic Forest biome.
2. We conducted spotlight surveys to detect caimans and gillnets deployed in the
reservoir from April 2015 to June 2022. We evaluated temporal differences in
gillnet encounter rates and the relationship between caimans and gillnet
distribution.
3. Gillnet encounter rates remained consistent year-round, while caiman encounter
rates were highest near gillnets, especially in the river channel and in forested
margins. Caimans are opportunistic predators attracted by tangled fish in gillnets
and likely prefer habitats with increased fish abundance.
4. Future research should continue monitoring the interaction between caimans and
fishing and include local communities in conservation efforts.
... In addition, pseudo-R 2 measures produce lower R 2 values compared to those associated with good fits in linear regression models [39]. When compared to another pseudo R 2 of the same type, on the same data, predicting the same outcome, a higher pseudo R 2 only indicates that the model better predicts the outcome [40]. Taking all of this into account, the confidence intervals and the Nagelkerke pseudo R 2 were reported as indices of the strength of association. ...
Objective:
To investigate if there are dose-response relationships between self-reported waking-state oral behaviours, including awake bruxism, and three indicators of psychological distress (depression, anxiety, stress).
Methods:
The study sample consisted of 1,886 patients with function-dependent TMD pain. Relationships between six non-functional and six functional waking-state oral behaviours, scored on a 5-point ordinal scale, and the psychological factors were investigated using ordinal logistic regression.
Results:
Mean age was 42.4 (±15.3) years, 78.7% being female. The odds of reporting the higher categories of non-functional oral behaviours depended on the severity of depression, anxiety, and stress. Most OR coefficients followed a quadratic dose-response distribution, the others increased linearly as the severity of the psychological scales increased. Almost no such associations were found with normal jaw function behaviours.
Conclusion:
Within the limitations of this study, it may be concluded that non-functional waking-state oral behaviours, including awake bruxism, and psychological distress have a dose-response relationship, with higher levels of distress being associated with higher reports of oral behaviours.
... The McFadden pseudo-R 2 value of 0.31 indicates good model fit, since the pseudo-R 2 metric doesn't equal to the conventional R 2 metric. 42 Compared to the factors found in the literature the second regression model outperformed model 1 regarding the p-value as well as the pseudo R 2 metric. ...
Background
Weight gain is a common side effect in psychopharmacology; however, targeted therapeutic interventions and prevention strategies are currently absent in day‐to‐day clinical practice. To promote the development of such strategies, the identification of factors indicative of patients at risk is essential.
Methods
In this study, we developed a transdiagnostic model using and comparing decision tree classifiers, logistic regression, XGboost, and a support vector machine to predict weight gain of ≥5% of body weight during the first 4 weeks of treatment with psychotropic drugs associated with weight gain in 103 psychiatric inpatients. We included established variables from the literature as well as an extended set with additional clinical variables and questionnaires.
Results
Baseline BMI, premorbid BMI, and age are known risk factors and were confirmed by our models. Additionally, waist circumference has emerged as a new and significant risk factor. Eating behavior next to blood glucose were found as additional potential predictor that may underlie therapeutic interventions and could be used for preventive strategies in a cohort at risk for psychotropics induced weight gain (PIWG).
Conclusion
Our models validate existing findings and further uncover previously unknown modifiable factors, such as eating behavior and blood glucose, which can be used as targets for preventive strategies. These findings underscore the imperative for continued research in this domain to establish effective preventive measures for individuals undergoing psychotropic drug treatments.
... The McFadden Pseudo R-squared corresponds to 0.54. Note, hereby, that values beyond 0.5 indicate an excellent fit [59]. The predictive accuracy of the model is 0.98. ...
Super apps allow users to access messaging, payments, e-commerce, deliveries, ridesharing, and many other services within the same app. While there are some very successful and dominant super apps in Asia such as WeChat, KakaoTalk, Alipay, or Grab, others, including Elon Musk with X (Twitter), are aiming to establish super apps in the U.S. and Europe. This explanatory study analyzes the super app phenomenon from a firm-level perspective. It provides preliminary insights on how digital platforms are reaching the super app status, and are evolving from single-purpose to multipurpose apps. Using data from 380 platforms in the mobility sector, a regression model is estimated to understand which platforms are capable of pursuing a super app strategy: young, agile, and risk-taking firms. I also discuss the case of Uber to illustrate the motivations and the various growth strategies that are incrementally paving the way to becoming a super app. Finally, testable propositions and a conceptual model are forwarded to stimulate future research on this timely topic.
... In the statistical analysis, all categories in the respective groups were tested against the reference categories tube load of 65 mAs and FBP reconstruction regardless of which comparisons were actually made by the observers (8) . The goodness of fit was reported using McFadden's pseudo R 2 (10) . ...
Non-linear properties of iterative reconstruction (IR) algorithms can alter image texture. We evaluated the effect of a model-based IR algorithm (advanced modelled iterative reconstruction; ADMIRE) and dose on computed tomography thorax image quality. Dual-source scanner data were acquired at 20, 45 and 65 reference mAs in 20 patients. Images reconstructed with filtered back projection (FBP) and ADMIRE Strengths 3–5 were assessed independently by six radiologists and analysed using an ordinal logistic regression model. For all image criteria studied, the effects of tube load 20 mAs and all ADMIRE strengths were significant (p < 0.001) when compared to reference categories 65 mAs and FBP. Increase in tube load from 45 to 65 mAs showed image quality improvement in three of six criteria. Replacing FBP with ADMIRE significantly improves perceived image quality for all criteria studied, potentially permitting a dose reduction of almost 70% without loss in image quality.
... Panel Generalized Quantile Regression estimation results of the research model are given in Table 3. [135]). ...
Through the examination of the ecological consequences of human actions, policymakers are able to distinguish certain areas in which resource use can be increased and the generation of waste diminished. This study examines the effects of foreign direct investment, gross domestic product, industrialization, renewable energy consumption, and urban population on the ecological footprints in 131 countries between 1997 and 2020. The objective of this study is to establish a thorough understanding of the relationship between these variables and ecological footprints while considering temporal changes from economic and environmental aspects. The analysis of a substantial dataset encompassing many countries aims to uncover recurring patterns and trends that can provide valuable information for the formulation of policies and strategies pertaining to sustainable development on a global level. The study fills a significant gap in the knowledge on the ecological impact of different variables, providing a nuanced understanding of the interdependencies among these factors, thus guiding sustainable development strategies, and promoting global sustainability. The study utilizes quantile regression analysis, a nonparametric estimator, to estimate consistent coefficients. The statistical analysis reveals that FDI, urbanization, and GDP have statistically significant and positive effects on ecological footprints. Industrialization and renewable energy consumption show significant and negative relationships with ecological footprints. The findings of this study contribute to the understanding of the relationships among these variables and provide insight to inform policy and decision-making efforts focused on reducing ecological consequences and advancing sustainable development goals.
... Furthermore, we applied Nagelkerke R square and Hosmer & Lemeshow tests to assess the goodness-of-fit of the models. Firstly, we conducted Nagelkerke R square, this helped us see how likely our model was to produce the observed data, and a lower log likelihood meant a better fit [63]. Additionally, we conducted the Hosmer-Lemeshow goodness-of-fit test. ...
Background
Over one-third of women worldwide suffer from anaemia. The prevalence of anaemia is particularly pronounced among women of reproductive age (WRA) in developing countries, such as India. No prior study has ever exclusively studied the prevalence of anaemia across the Aspirational Districts of India. Therefore, the purpose of this study was to examine the prevalence of anaemia across Aspirational Districts of India and to identify the determinants of anaemia among WRA in these districts.
Methods
From the National Family Health Survey (NFHS)-4 (2015-16) and NFHS-5 (2019-21), data on 114,444 and 108,782 women aged 15–49 from Aspirational Districts were analyzed in our study, respectively. Bivariate statistics and multivariable binary logistic regression were used to identify the determinants of anaemia.
Results
The national prevalence of anaemia among WRA has increased from 53% in NFHS-4 to 57% in NFHS-5 whereas anaemia among WRA in Aspirational Districts has increased from 58.7% in NFHS-4 to 61.1% in NFHS-5. Between 2015 and 2021, over 60% of Aspirational Districts experienced an increase in the prevalence of anaemia and one-fourth, specifically 29 out of 112, observed a rise by at least 10 percentage points (pp). Notably, there are significant variations in anaemia prevalence among districts, with Simdega and Udalgiri having the highest anaemia prevalence in NFHS-4 and NFHS-5 at 78.2% and 81.5%, respectively. During this period, Barpeta followed by Udalgiri of Assam have witnessed the maximum increase with 29.4% and 26.7% respectively. Moreover, pooled regression results show women with three to four children [AOR: 1.13, 95% CI: 1.08–1.17], women who breastfeed [AOR: 1.17, 95% CI: 1.13–1.20], Scheduled Tribe women [AOR: 1.39, 95% CI: 1.35–1.44], poorest women [AOR: 1.27, 95% CI: 1.22–1.33] and women those who consume fish occasionally [AOR: 1.14, 95% CI: 1.12–1.17] were more likely to be anaemic.
Conclusion
The significant increase in anaemia among WRA in Aspirational Districts of India is a matter of concern. Given the rise in anaemia among WRA, determinants-based and district-specific measures must be designed and implemented to reduce the prevalence of anaemia among Aspirational Districts of India.
... where ( ) and ( ) indicate the likelihood of the target prediction model and baseline model (null model), respectively. Similar to R-squared used in linear regression to calculate the proportion of explained variance, pseudo-R-squared measures the degree of improvement of the model likelihood over a null model, which is a simple baseline model containing no predictor variables (Hemmert et al., 2018). ...
Survival models have become popular for credit risk estimation. Most current credit risk survival models use an underlying linear model. This is beneficial in terms of interpretability but is restrictive for real-life applications since it cannot discover hidden nonlinearities and interactions within the data. This study uses discrete time survival models with embedded neural networks as estimators of time to default. This provides flexibility to express nonlinearities and interactions between variables, and hence allows for models with better overall model fit. Additionally, the neural networks are used to estimate Age-Period-Cohort (APC) models so that default risk can be decomposed into time components for loan age (maturity), origination (vintage) and environment (e.g., economic, operational and social effects). These can be built as general models, or as local APC models for specific customer segments. The local APC models reveal special conditions for different customer groups. The corresponding APC identification problem is solved by a combination of regularization and fitting the decomposed environment time risk component to macroeconomic data, since the environmental risk is expected to have a strong relationship with macroeconomic conditions. Our approach is shown to be effective when tested on a large publicly available US mortgage data set. This novel framework can be adapted by practitioners in the financial industry to improve modelling, estimation and assessment of credit risk.
... For binary logistic regression (BLR) models we can only establish a pseudo r 2 . This does not represent the proportion of variance explained and thus there are no clear guide values for an adequate model (Hemmert et al., 2018); however, some authors indicate that good models should present a value > 0.2 (McFadden, 1979). ...
Defining nutrient thresholds that protect and support the ecological integrity of aquatic ecosystems is a fundamental step in maintaining their natural biodiversity and preserving their resilience. With increasing catchment pressures and climate change, it is more important than ever to develop clear methods to establish thresholds for status classification and management of waters. This must often be achieved using complex data and should be robust to interference from additional pressures as well as ameliorating or confounding conditions. We use both artificial and real data to examine challenges in setting nutrient thresholds in unbalanced and skewed data. We found significant advantages to using binary logistic regression over other techniques. However, one of the key challenges is objectively selecting a probability from which to derive the nutrient threshold. For this purpose, the examination of the proportions of matching and mismatching status classifications of nutrients and a biological quality element using a confusion matrix is a key step that should be more widely adopted in threshold selection. We examined a large array of statistical measures of classification accuracy and their performance over combinations of skewness and imbalance in the data. The most appropriate threshold probability is a compromise between maximising overall classification accuracy and reducing mismatches expressed as commission (false positives) without excessive omission (false negatives). An application to a lake type indicated total phosphorus thresholds that would be around 50 μg l−1 lower than the threshold achieved by an ‘unguided’ approach, indicating that this approach is a very significant development meriting attention from national authorities responsible for water management.
... this only confirms that occupational injuries depend on multiple and random variables, so we consider it a reasonable fit for our explanatory model. Furthermore, pseudo-R 2 based on log-pseudolikelihood represent the improvement in model likelihood over a null model (Hemmert et al., 2018), indicating an increase in the likelihood of our final model by 23,81 % (regression I pseudo R2 = 0.0407 and regression III pseudo R2 = 0.0508). ...
This paper analyzes the effect of worker under-skilling on occupational safety. We estimate the impact of skill deficits on the probability of suffering an accident at work and, second, on the duration of sick leave. In addition, we test whether the company's measures to control the actions of these workers reduce this effect. We propose two moderation models in a sample of 42,871 workers obtained from the Sixth European Working Conditions Survey (EWCS6). The results show that under-skilled workers suffer more accidents and longer periods of sick leave. Furthermore, the results suggest that on-the-job training, safety information, and teamwork weaken the relationship between under-skilling and accidents. However, the duration of sick leave is only reduced by teamwork. Our analysis shows that certain organizational and regulatory practices need to be modified to address the health effects of a lack of skills. The article includes some proposals in this regard.
... The change in model improvement was assessed using the pseudo-R-squared. Unlike ordinary least square regression, logistic regression does not have a true R-squared (Hemmert et al., 2018). Hence, some level of caution should be exercised when interpreting the pseudo-R-squared. ...
The purpose of this cross-sectional study was to investigate risk (violence and victimization, symptoms of depression, substance use, and obesity) and protective factors (physical activity and academic performance) associated with suicidal ideation and suicide attempts among Black adolescents. Data were obtained from the 2017 Youth Risk Behavior Survey. The analytic sample consisted of 658 adolescents ages 14–18 years (51.8% female) who self-identified as Black. Multivariable binary logistic regression was conducted to examine risk and protective factors associated with suicidal ideation and suicide attempts. About 16% of the sample reported suicidal ideation, and 9.1% made a suicide attempt during the past 12 months. Factors associated with suicidal ideation and suicide attempts included weapon-carrying on school property and symptoms of depression. Physical activity ( AOR = 0.28, 95% CI = 0.11–0.68) and higher academic performance had protective effects on suicide attempts ( AOR = 0.35, 95% CI = 0.17–0.71). Clinicians and school counselors working with Black adolescents may want to inquire about suicide when these risk factors are experienced and strengthen the protective factors identified.
... where the sample size and m is number of normal patients, is the maximum value of the log-likelihood function the null model [6,8]. ...
The restricted intracranial volume in syndromic craniosynostosis is due to limited skull expansion caused by early fusion of multiple cranial sutures. This led to progressive increased in intracranial pressure which has long been established as the cause of optic nerve damage. Optic nerve damage secondary to the narrowing of optic canal in syndromic craniosynostosis has been reported but not comprehensively explored. The objective of this study is to predict using logistic regression the cause of optic nerve atrophy is caused by increased intracranial pressure or structural narrowing of optic canal. The study involved 11 measurements of features of the optic canals as predictor variables. A binary logistic regression and variable selection method were applied to the 11 measurements to choose the best combination of the predictors. The results show good models that could be considered a suitable representation of the data. The height at the optical cranial side is the most dominant feature in the top 20 models that specifies the syndromic patients, followed by the area and the perimeter for both optic canals, then the length of the medial wall and the diameter at the mid canal for right and left canal respectively. The paper's findings provide significant evidence for using this method as an alternative to determine if the cause of optic nerve atrophy is related to either increased intracranial pressure or narrowing optic canal structure among syndromic craniosynostosis patients, thus saving the patient from ineffective operations, additional risks, and waste of resources.
Background and Objectives
Basal cell carcinoma (BCC) is the most common form of skin cancer, originating from basal cells in the skin’s outer layer. It frequently arises from prolonged exposure to ultraviolet (UV) radiation from the sun or tanning beds. Although BCC rarely metastasizes, it can cause significant local tissue damage if left untreated. Early detection is essential to prevent extensive damage and potential disfigurement. The United States Preventive Services Task Force (USPSTF) currently remains uncertain about the benefits and potential harms of routine skin cancer screenings in asymptomatic individuals. This paper evaluates the accuracy of predicting BCC using patients’ medical histories to address this uncertainty and support early detection efforts.
Methods
We analyzed the medical histories of 405,608 patients, including 7733 with BCC. We categorized 25,154 diagnoses into 16 body systems based on the hierarchy in the Systematized Nomenclature of Medicine (SNOMED) ontology. For each body system, we identified the most severe condition present. Logistic Least Absolute Shrinkage and Selection Operator (LASSO) regression was then employed to predict BCC, using demographic information, body systems, and pairwise and triple combinations of body systems, as well as missing value indicators. The dataset was split into 90% for training and 10% for validation. Model performance was evaluated using McFadden’s R ² , Percentage Deviance Explained (PDE), and cross-validated with the area under the receiver operating characteristic curve (AUC).
Results
Diagnoses related to the Integument system showed an 8-fold higher likelihood of being associated with BCC compared to diagnoses related to other systems. Older (age from 60 to 69) white individuals were more likely to receive a BCC diagnosis. After training the model, it achieved a McFadden’s R ² of 0.286, an AUC of 0.912, and a PDE of 28.390%, reflecting a high level of explained variance and prediction accuracy.
Conclusions
This study underscores the potential of LASSO Regression models to enhance early identification of BCC. Extant medical history of patients, available in electronic health records, can accurately predict the risk of BCC. Integrating such predictive models into clinical practice could significantly improve early detection and intervention.
Purpose
This study explores how various officer and event-level factors influence Milwaukee Police Department officers’ decision to activate their body-worn cameras (BWCs) across both community member-initiated services and officer-initiated activities.
Design/methodology/approach
Across the 1,052 officers and 1,066,112 officer-events in the sample, we use descriptive and logistical regressions to assess differences in BWC activations across calls for service and officer-initiated activities.
Findings
We found similar activation rates between calls for service (41.5%) and officer-initiated activities (44.1%). However, our logistic regression analysis results suggest the explanatory power of the event and officer-level variables was substantially better in models examining officer-initiated activities. Among calls for service, officers were more likely to activate BWCs during calls involving crimes against persons compared to other crimes or non-criminal incidents. Activation was more frequent during traffic stops than other self-initiated activities. Activation increased when the event resulted in an advisement, citation, detention or arrest.
Originality/value
The success of police BWC programs hinges on whether officers activate their cameras when interacting with community members. Findings suggest that officers are more likely to activate their BWCs during activities that involve direct interactions with community members, especially in situations with a higher potential for volatility or serious criminal implications.
Despite global efforts on meeting sustainable development goals by 2030, persistent and widespread sanitation deficits in rural, underserved communities in high-income countries—including the United States (US)—challenge achieving this target. The recent US federal infrastructure funding, coupled with research efforts to explore innovative, alternative decentralized wastewater systems, are unprecedented opportunities for addressing basic sanitation gaps in these communities. Yet, understanding how to best manage these systems for sustainable operations and maintenance (O&M) is still a national need. Here, we develop an integrated management approach for achieving such sustainable systems, taking into account the utility structure, operational aspects, and possible barriers impeding effective management of decentralized wastewater infrastructure. We demonstrate this approach through a binomial logistic regression of survey responses from 114 public and private management entities (e.g., water and sewer utilities) operating in 27 states in the US, targeting the rural Alabama Black Belt wastewater issues. Our assessment introduces policy areas that support sustainable decentralized wastewater systems management and operations, including privatizing water-wastewater infrastructure systems, incentivizing/mandating the consolidation of utility management of these systems, federally funding the O&M, and developing and retaining water-wastewater workforce in rural, underserved communities. Our discussions give rise to a holistic empirical understanding of effective management of decentralized wastewater infrastructure for rural, underserved communities in the US, thereby contributing to global conversations on sustainable development.
Health care systems are increasingly partnering with community-based organizations to address social determinants of health (SDH). We established a program that educates and connects patients with SDH needs at a primary care clinic to community services and facilitated referrals.
To evaluate the effect of addressing SDH soon after discharge on hospital readmission in a clinic population.
Pre/post, quasi-experimental design with longitudinal data analysis for quality improvement.
Clinic patients (n = 754) having at least one hospital discharge between June 1, 2020, and October 31, 2021, were included. Of these, 145 patients received the intervention and 609 served as comparison.
A primary care liaison was employed to assess and educate recently discharged clinic patients for SDH needs and refer them for needed community services from June 1, 2020, to October 31, 2021.
Hospital readmissions within 30, 60, and 90 days of discharge were tracked at 6-month intervals. Covariates included patient age, sex, race/ethnicity, insurance status, income, Hierarchical Condition Category risk scores, and Clinical Classification Software diagnosis groups. Data for all hospital discharges during the intervention period were used for the main analysis and data for the year before the intervention were extracted for comparison.
Overall, patients in the intervention group were older, sicker, and more likely to have public insurance. The reductions in 30-, 60-, and 90-day readmissions during the intervention period were 14.39%, 13.28%, and 12.04% respectively in the intervention group, while no significant change was observed in the comparison group. The group difference in reduction over time was statistically significant for 30-day (Diff = 12.54%; p = 0.032), 60-day (Diff = 14.40%; p = 0.012), and 90-day readmissions (Diff = 14.71%; p = 0.036).
Our findings suggest that screening clinic patients for SDH, and educating and connecting them to community services during post-hospital care may be associated with reductions in hospital readmissions.
Purpose
This study aims to analyse the outcomes of Indian insolvency proceedings for their ex-post economic efficiency. Ideally, insolvent yet viable companies should witness resolution, whereas insolvent-unviable companies should be liquidated. This study aims to ascertain the key forces that ensure or prevent the application of the first part of this maxim in practice.
Design/methodology/approach
The study uses logistic regression on a sample of 320 corporate insolvencies (out of 942 insolvencies) reported under the Insolvency and Bankruptcy Code (IBC), 2016. Two-stage least squares regression is used to check endogeneity issues.
Findings
The results claim high levels of rationality from the financial creditors and acceptable levels of viability from the plan proposers for precluding liquidation of insolvent yet viable companies. The findings reveal that an excess of value from resolution over that from liquidation, controls the outcomes of insolvency proceedings. Further examinations indicate that financial creditors’ focus on upfront recovery prevents them from judging the plans on other viability-related factors. Based on the findings, this study recommends that IBC must focus on the importance of both long-term recovery rates and resolution.
Originality/value
To the best of the authors’ knowledge, this is one of the first studies to empirically analyse Type 2 efficiency-related errors prevalent in the Indian insolvency proceedings since the enactment of its new code. The empirical explorations offered in this research can prove to be unique for policy-making.
Sexual homicides (SHs) demand nuanced research for effective prevention, treatment, risk assessment and theoretical insights. Intimate‐partner sexual homicides (IPSHs), comprising approximately 20% of SHs, have received limited attention. This study compares IPSHs (n = 56) and non‐intimate partner sexual homicides (NIPSHs) (n = 236) in Australia and New Zealand by investigating offender, victim, and crime‐scene characteristics. While IPSH perpetrators were typically older, separated, and had prior domestic violence convictions, victims were more often non‐white with histories of domestic violence and substance use. Although crime‐scene locations and post‐offence behaviours differed, similar crime scene behaviours were displayed across offender groups, which seemed to be routed in different underlying motives. Whereas drivers of IPSH commonly were grievance and anger, associated with offences occurring after arguments, drivers for NIPSH were more often sexual deviance and sadism. Overall, IPSH encompasses aspects of domestic violence, homicide, and sexual violence, distinguishing it from SH.
Just-In-Time (JIT) defect prediction aims to identify defects early, at commit time. Hence, developers can take precautions to avoid defects when the code changes are still fresh in their minds. However, the utility of JIT defect prediction has not been investigated in relation to crashes of mobile apps. We therefore conducted a multi-case study employing both quantitative and qualitative analysis. In the quantitative analysis, we used machine learning techniques for prediction. We collected 113 reliability-related metrics for about 30,000 commits from 14 Android apps and selected 14 important metrics for prediction. We found that both standard JIT metrics and static analysis warnings are important for JIT prediction of mobile app crashes. We further optimized prediction performance, comparing seven state-of-the-art defect prediction techniques with hyperparameter optimization. Our results showed that Random Forest is the best performing model with an AUC-ROC of 0.83. In our qualitative analysis, we manually analysed a sample of 642 commits and identified different types of changes that are common in crash-inducing commits. We explored whether different aspects of changes can be used as metrics in JIT models to improve prediction performance. We found these metrics improve the prediction performance significantly. Hence, we suggest considering static analysis warnings and Android-specific metrics to adapt standard JIT defect prediction models for a mobile context to predict crashes. Finally, we provide recommendations to bridge the gap between research and practice and point to opportunities for future research.
Secondary education for all is one of the UN’s attainable goals. Many countries, including Pakistan, are struggling to achieve this target. Earlier research has attempted to analyze the determinants of secondary schooling by taking the total income of households. However, households of different income groups respond differently to varying socio-economic factors. This study attempts to identify the household-level socio-economic determinants of secondary schooling across different income groups in Pakistan. It utilizes national survey data from the Pakistan Social and Living Standards Measurement (PSLM) Survey 2019-20. We selected households from the dataset that had at least one member of secondary school age (13-20 years). Households that enrolled a secondary school-age member in school or whose member achieved secondary schooling were categorized as having demand for secondary schooling. Furthermore, instead of taking the total income of households, study takes six categories of income. The results of logit estimation show that demand for secondary schooling increases across successive income groups, indicating secondary schooling is a normal commodity. The proportion of male school-going age members and having a female head increase the likelihood of demand for secondary schooling. The study recommends that policies to increase enrollment at the secondary level may focus more on lower-income groups and on the education of females.
In this paper, we assess the impact of the COVID‐19 pandemic on the Australian visual arts sector. We base our analysis on the responses of over 1500 visual artists and arts workers to a survey conducted by the National Association for the Visual Arts ( NAVA ), the national peak body for the visual and media arts, craft and design sector in September 2021. NAVA employed this online survey to study the relationship between the pandemic and both the incomes and mental health of artists and arts workers. Using regression analysis, we find that there has been a significant impact for both artists and arts workers, with the severity of the impacts varying by gender, age and the availability of state‐based and Australian Government support programmes. Reduced hours and loss of contracted work and commissions due to the pandemic were both related to declines in income and mental health outcomes for artists and for arts workers. Housing stress was associated with a higher likelihood of a significant or extreme mental health impact for artists and arts workers. In addition, artists' incomes and mental health outcomes were impacted when faced with a reduced ability to sell, although some artists were able to increase their online profiles.
Guiding wind energy sector growth through suitability analysis is a growing priority. We present in this work a logistic regression model that predicts suitable sites for state-level and nationwide wind energy development in the United States. The model's aggregation of publicly available data to 20 different grid cell resolutions, along with four predictor configurations, allows end-users to investigate commercial wind farm site suitability for their region, project size, and predictors of interest. Model performance maximizes at higher grid cell resolutions and when using a complete and/or refined predictor set. Validation of the model's performance against existing commercial wind farm locations demonstrates its ability to consistently diagnose over 75% of grid cell states correctly. As such, high suitability grid cells that currently lack wind farms could represent candidate locations for wind farm construction. This model and its aggregated datasets can be applied in other suitability analysis contexts, particularly solar energy development.
One of the most commonly accepted models of relationships among three variables in applied industrial and organizational psychology is the simple moderator effect. However, many authors have expressed concern over the general lack of empirical support for interaction effects reported in the literature. We demonstrate in the current sample that use of a continuous, dependent-response scale instead of a discrete, Likert-type scale, causes moderated regression analysis effect sizes to increase an average of 93%. We suggest that use of relatively coarse Likert scales to measure fine dependent responses causes information loss that, although varying widely across subjects, greatly reduces the probability of detecting true interaction effects. Specific recommendations for alternate research strategies are made.
Errors can be introduced into scientific research when continuous concepts are measured on scales that rank the concepts into a few categories. This presents a potential problem because measures of association between two variables may differ depending on whether continuous or collapsed measures are used. We analyzed simulated data and examined differences in the correlation between two normally distributed continuous variables and the same two variables collapsed into a small number of categories. In general, the differences in correlation coefficients computed on continuous variables and the same variables collapsed into a few categories are small. The greatest differences in the correlations between the two types of variable occur when the continuous variables' correlation is high and only a few categories are used for the collapsed variables. When as few as five categories are used to approximate the continuous variables, the correlation coefficients and their standard deviations for the collapsed and continuous variables are very close. These findings suggest that under certain conditions it may be justifiable to analyze categorical data as if it were continuous.
Three research theories used to explain firm boundaries are transaction cost economics, an options perspective, and a resource-based view of the firm. Our integrated model addresses the degree to which each of these three perspectives explains firm boundaries for technology sourcing is contingent on managerial risk taking, which is partly determined by organizational context. Our results suggest that, in general, management stockholdings, firm risk orientation, and slack resource availability moderate the extent to which the perceived threat of opportunism, the threat of commercial failure, and opportunity for sustainable advantage all influence firm boundaries.
Early work in strategic management emphasized single case studies, followed by research on corporate diversification strategy, firm heterogeneity, strategic groups, and generic business strategies. Intermediate work added the foci of environmental determinants and strategic choice, often using secondary data from large, multi-industry firm samples. Recently, the most prominent new theoretical paradigm is the resource-based view of the firm, using smaller sample studies. Future research is likely to integrate and contrast multiple theories and to develop more fine-grained and complex models. Quantitative research will emphasize longitudinal data, dynamic analysis, and greater focus on specific strategic decisions/actions. Future research will use more specialized tools such as panel data analysis, dynamic models of partial adjustment, logistic and Poisson regression analyses, event history analysis, network analysis, and structural equation modeling. Nontraditional research designs will also gain popularity, such as combined qualitative/quantitative data approaches and comparison of outliers.
This study investigated 3 broad classes of individual-differences variables (job-search motives, competencies, and constraints) as predictors of job-search intensity among 292 unemployed job seekers. Also assessed was the relationship between job-search intensity and reemployment success in a longitudinal context. Results show significant relationships between the predictors employment commitment, financial hardship, job-search self-efficacy, and motivation control and the outcome job-search intensity. Support was not found for a relationship between perceived job-search constraints and job-search intensity. Motivation control was highlighted as the only lagged predictor of job-search intensity over time for those who were continuously unemployed. Job-search intensity predicted Time 2 reemployment status for the sample as a whole, but not reemployment quality for those who found jobs over the study's duration. (PsycINFO Database Record (c) 2012 APA, all rights reserved)
Multimarket (or multipoint) contact has been shown to deter aggressive actions by rivals toward each other, producing a situation of mutual forbearance among firms. To create this deterrent capability, however, firms must enter each others’ markets, which is just the kind of action that the deterrent is supposed to limit. This study explores the questions: Under what conditions are firms likely to behave aggressively toward their multimarket rivals by entering their markets and when will they engage in mutual forbearance? We describe how the effect of multimarket contact on the market-entry moves of a firm changes as the level of contact a firm has with its rivals increases. We draw on competitive intelligence and decision-making theory to argue that the competitive advantages associated with multimarket contact are supplemented
The authors examine the growth and survival of international joint ventures (IJVs) from a legiti-macy perspective. In a sample of 291 Sino-Japanese joint ventures in China, they found that Chi-nese parent age, Chinese parent size, and IJV industry relatedness to either parent had a positive effect on IJV growth and/or survival. However, IJV industry relatedness to both parents led to lower rates of IJV growth and survival. The findings highlight the importance for IJVs to obtain both external and internal legitimacy, as well as the difficulties IJVs face in acquiring internal legitimacy from both parents simultaneously.
Despite recent setbacks for Internet start-upactivity in the United States, global e-commerce sales are predicted to grow inthe next few years. The basis of market entry by internationally expanding U.S.Internet firms is the research focus. The entrepreneurial context of U.S.Internet firms is analyzed, and a model predicting market entry is proposed, inwhichcountry risk, cultural distance, the level of uncertainty avoidance,and power distance (related to authority, inequality and trust) in theinternational target market are predicted to have a negative direct effect onthe possibility of international market entry. This contingency model of international market entry is tested on a sampleof 7,000 country entry decisions by 179 U.S. Internet companies. Internationalmarket entry was considered the dependent variable; country risk, nationalcultural values, and cultural distance were the independent variables; andmarket size, technology infrastructure, firm-level factors, andfirm-and-industry level fixed effects were the control variables. The findings suggest that different cultural dimensions have an independenteffect on international market entry by Internet firms. The level ofuncertainty avoidance was found to have a negative effect on the probability ofmarket entry by Internet firms, while the society's degrees of individualismand masculinity had a positive effect. The findings also show that market size has a positive moderating effect onthe relations between country risk, cultural distance, and national culturalvalues. The findings suggest that market size moderates the relations betweencountry risk, cultural distance, national cultural values, and market entry.(CBS)
Dieses Buch behandelt Modelle zur Analyse kategorialer Daten. Kategoriale Daten sind Variablen, die eine begrenzte Anzahl von Ausprägungen (Kategorien) haben. Bei vielen der in Umfrageforschung und amtlicher Statistik erhobenen Merkmale handelt es sich um kategoriale Daten. Beispiele wären etwa das Geschlecht einer Befragungsperson, ihre Parteipräferenz, die Anzahl der Mitbewohner im Haushalt dieser Person, ihre Schichtzugehörigkeit und ähnliches mehr. In diesem Lehrbuch geht es um eine anwendungsorientierte Einführung in die multivariate Analyse kategorialer Daten. Konkret werden vier Ansätze vorgestellt: die gewichtete Regression nach Grizzle, Starmer und Koch (GSK-Ansatz), die Klasse der log-linearen Modelle, die logistische Regression und die Analyse latenter Klassen.
When compared to research on the association between immigration and crime, far less attention has been given to the relationship between immigration, citizenship, and criminal punishment. As such, several fundamental questions about how noncitizens are sanctioned and whether citizenship is a marker of stratification in U. S. courts remain unanswered. Are citizens treated differently than noncitizens-both legal and undocumented-in U. S. federal criminal courts? Is the well-documented Hispanic-white sentencing disparity confounded by citizenship status? Has the association between citizenship and sentencing remained stable over time? And are punishment disparities contingent on the demographic context of the court? Analysis of several years of data from U. S. federal courts indicates that citizenship status is a salient predictor of sentencing outcomes-more powerful than race or ethnicity. Other notable findings include the following: accounting for citizenship substantially attenuates disparities between whites and Hispanics; the citizenship effect on sentencing has grown stronger over time; and the effect is most pronounced in districts with growing noncitizen populations. These findings suggest that as international migration increases, citizenship may be an emerging and powerful axis of sociolegal inequality.
One of the accepted beliefs among strategic management researchers is that strategic decision consensus among decision-makers facilitates decision implementation speed and implementation success. Despite this belief, there has been little empirical research directly examining these relationships. Using hierarchical regression, this study examines the effects of strategic decision consensus and commitment on decision implementation speed and success. Results from a sample of 68 intact strategic decision-making teams in Southeastern United States hospitals support the normatively accepted ideal that decision consensus helps build decision commitment, which in turn positively affects implementation success. Contrary to our expectations, however, decision commitment serves to slow implementation speed. The implications of these results for consensus and implementation research are discussed.
This research examined the relationships between board composition, stock ownership, and the existence of director liability protection in a sample of industrial manufacturing firms. Results indicate that the impact of board leadership on liability protection depends on the proportion of independent directors on the board. Findings suggest that high levels of independent directors compromise the governing effectiveness of independent board leadership. Results demonstrate that stock ownership patterns affect liability protection as well. High levels of institutional ownership are associated with liability protection. While the theoretical paradigm of organizational economics predicts that the decision to provide protection is based on economic advantage, this study shows that liability protection for directors involves social dimensions as well as economic ones.
Despite the burgeoning research on lifestyles, we have surprisingly little evidence to answer one of the literature’s founding questions: Is the association between social class and lifestyles disappearing? I explore this inquiry with data from the past four decades. In analyzing the class-lifestyle association, I examine changes in the variability of lifestyles within and between social classes. Using data from the General Social Survey on informal social ties and formal membership ties to voluntary associations, I derive proxies for lifestyles and examine their relation to social class with latent class models. Results show that social classes’ contemporary sociability patterns are substantively similar to traditional descriptions from empirical studies on analogous data from as early as the mid-twentieth century. The association between social classes and sociability patterns shows no sign of having weakened over the past four decades. In fact, recent trends of civic disengagement and social isolation in contemporary U.S. society, which these data corroborate, reinforce class differences in sociability.
Boards experience elevated levels of turnover among outside directors following financial fraud. Scholars have proposed two mechanisms that may drive this turnover. The first views turnover as part of a board's efforts to repair organizational legitimacy and avert resource withdrawal. The second argues that turnover is a byproduct of individual directors' efforts to safeguard their own reputations and mitigate professional devaluation. We use data on director departures following 63 fraud events to explore the relative importance of these two mechanisms. The results clarify our understanding of responses to governance failures and the challenges of reconstituting board membership following financial improprieties. Copyright © 2013 John Wiley & Sons, Ltd.
It is argued that P-values and the tests based upon them give unsatisfactory results, especially in large samples. It is shown that, in regression, when there are many candidate independent variables, standard variable selection procedures can give very misleading results. Also, by selecting a single model, they ignore model uncertainty and so underestimate the uncertainty about quantities of interest. The Bayesian approach to hypothesis testing, model selection, and accounting for model uncertainty is presented. Implementing this is straightforward through the use of the simple and accurate BIC approximation, and it can be done using the output from standard software. Specific results are presented for most of the types of model commonly used in sociology. It is shown that this approach overcomes the difficulties with P-values and standard model selection procedures based on them. It also allows easy comparison of nonnested models, and permits the quantification of the evidence for a null hypothesis of interest, such as a convergence theory or a hypothesis about societal norms.
The effect of sample size on goodness of fit indices that are widely used in structural equation models is discussed and empirically illustrated. Many of the indices discussed in the. article are reported in such applications as LISREL and EQS. GR, AGR, Normed fit index (NFI), Non-normed fit index (NNFI), Incremental fit index (LFI) and Comparative Fit Index (CFI) are examined for iwo types of sample size effects. The means of the sampling distributions of GR, IR and CR were relatively stable across different sample sizes and these three indices were not as downward biased as NFI and NNFI.
In this article, the authors explore the social bases of cultural con- sumption by examining the association between newspaper read- ership and social status. They report a strong and systematic as- sociation between status and newspaper readership which is consistent with the expected link between status in the classical Weberian sense, on the one hand, and cultural level and lifestyle, on the other. This association persists in a multinomial logistic re- gression model in which the authors take into account, among other things, educational attainment which serves as a proxy for the re- spondent's information-processing capacity. The social status of re- spondent's father and best friend also have significant and sub- stantial effects on newspaper preference. Finally, the authors report results that indicate that the effects of status on newspaper read- ership are generally more important than those of class.
Scale coarseness is a pervasive yet ignored methodological artifact that attenuates observed correlation coefficients in relation to population coefficients. The authors describe how to disattenuate correlations that are biased by scale coarseness in primary-level as well as meta-analytic studies and derive the sampling error variance for the corrected correlation. Results of two Monte Carlo simulations reveal that the correction procedure is accurate and show the extent to which coarseness biases the correlation coefficient under various conditions (i.e., value of the population correlation, number of item scale points, and number of scale items). The authors also offer a Web-based computer program that disattenuates correlations at the primary-study level and computes the sampling error variance as well as confidence intervals for the corrected correlation. Using this program, which implements the correction in primary-level studies, and incorporating the suggested correction in meta-analytic reviews will lead to more accurate estimates of construct-level correlation coefficients.
At the present time very little can be claimed to be known as to the effects of weather upon farm crops. The obscurity of the subject, in spite of its immense importance to a great national industry, may be ascribed partly to the inherent complexity of the problem which it presents, and more especially to the lack of quantitative data obtained either under experimental or under industrial conditions, by the study of which accurate knowledge alone can be acquired. Of the industrial applications of such knowledge it is unnecessary to speak here in detail. It is sufficient to indicate that the present system of Life Insurance, which safeguards the economic stability of many thousands of families, and occupies the activities of many of the greatest financial corporations, was made possible by the studies of statistics of human mortality by the mathematicians of the eighteenth and nineteenth centuries, and that on the basis of adequate knowledge similar economic stability with its attendant security of capital should be within the reach of the industrial farmer. The inherent complexity of the relationships which it is sought to elucidate, between the yields of farm crops, and the previous weather which largely controls those yields, arises primarily from the complexity of the problem of specifying the weather itself. Meteorologists have, however, gradually devised a number of instrumental observations, which although far from specifying the total environment of the growing plant, as understood by the plant physiologist, do nevertheless give a sufficiently detailed account of the general environmental conditions of the growing crop, in so far as these vary from season to season. It is probable indeed that almost all the weather influences to which crop variations are due could be expressed in terms of the instrumental observations of a modern meteorological station. The actual difficulty of calculating the crop variations from given instrumental records is, however, immense; only an attack of the most preliminary kind upon the general problem can be attempted in this paper. The complete aim of agricultural meteorology should, however, be emphasised, for it is only by its substantial achievement that other causes of crop variation can be freed from much obscurity.
In this paper the author tries to identify various influences on the coefficient of determination (R2) which originate in the research designs of empirical studies, rather than in the research subjects within the framework of an exploratory meta-analysis. The following results are obtained: the larger the sample size and the smaller the number of regressors in a study, the smaller is R2; time-series studies achieve higher values for R2 than cross-sectional studies; studies with secondary data achieve higher values for R2 than studies with primary data; publications in the Journal of Marketing Research show on average lower values for R2 than publications in the International Journal of Research in Marketing and in Marketing Letters.
R-2 is widely relied on in linear regression to index a model's discriminatory power Many counterparts have been proposed for use in logistic regression, but no single measure is consistently used. Two potential criterion values are relevant: the explained variance in the latent scale underlying the binary indicator of event occurrence and the explained risk of the event itself, In this study, Monte Carlo methods were used to examine the performance, with respect to fixed theoretical levels of explained variance and explained risk, of eight R-2 analogues. The McKelvey-Zavoina measure appears to be best at estimating explained variance and either the sample-estimated explained risk or the ordinary least squares R-2 to be best at indexing explained risk. Other measures appear to be poor choices, primarily because asymptotic trends suggest they may be inconsistent estimators of the relevant criterion.
This paper develops a model, with assumptions similar to those of the linear model, for use when the observed dependent variable is ordinal. This model is an extension of the dichotomous probit model, and assumes that the ordinal nature of the observed dependent variable is due to methodological limitations in collecting the data, which force the researcher to lump together and identify various portions of an (otherwise) interval level variable. The model assumes a linear eflect of each independent variable as well as a series of break points between categories for the dependent variable. Maximum likelihood estimators are found for these parameters, along with their asymptotic sampling distributions, and an analogue of R (the coefficient of determination in regression analysis) is defined to measure goodness of fit. The use of the model is illustrated with an analysis of Congressional voting on the 1965 Medicare Bill.
We theorize that if law matters in Supreme Court decision making, it matters not as a mechanistic force that dictates decisions, but as an institutional construct created by justices who possess political attitudes. Jurisprudential regimes identify relevant case factors and/or set the level of scrutiny or balancing the justices will use. These jurisprudential regimes have the potential to make a significant difference in the decisions of the justices. We identify a candidate jurisprudential regime, content-neutrality, which appears to govern the general area of free expression law. The Court applies the strictest standard of review to regulations of expression that target the content or viewpoint of expression. Relying on a series of statistical tests using logistic regression, we find that the justices take seriously this jurisprudential regime.
This article evaluates and compares “president” and “presidency” centered explanations of presidential activity with respect to one important presidential power, the veto. Using individual bill data for nine congresses characterized by divided party government, I estimate a logistic regression model of presidential vetoes. This contrasts with previous research, which has used annual aggregate data. Using individual bill data allows controls for objectionable legislation passed by Congress and enables me to measure and compare the “propensity to veto” of different presidents. I conclude that presidential vetoes are in substantial measure caused by Congress passing objectionable bills, but that even controlling for congressional behavior, presidents exhibit strikingly different veto behavior. Ford was most prone to using the veto, and Reagan, the least. Although external factors exert great influence on the president's veto decision, the individual choices and strategies of presidents also have an important influence.
Data analysis is a key element of the research process. Accordingly, appropriate doctoral training in data analysis is vital to the strategic management field's future. We used a two-study design to evaluate quantitative data analysis trends and doctoral training. An analysis of Strategic Management Journal articles from 1980 to 2001 reveals that, contrary to some predictions, the use of general linear model techniques such as regression has increased over time. However, the use of more specialized techniques, including those suitable for examining longitudinal data, discrete events, and causal structure, has also grown substantially. A survey of recent doctoral graduates shows that, although skilled with general linear models, many are ill prepared to use specialized techniques. Based on our findings, we offer suggestions aimed at bridging gaps between what doctoral students (and other researchers) know and what they need to know about data analysis. Copyright © 2003 John Wiley & Sons, Ltd.
Using data on U.S. investment banking firms’ syndication in underwriting corporate stock offerings during the 1980s, this study explores the factors that drive alliance formation between two specific firms. We compare resource complementarity, status similarity, and social capital as a basis of alliance formation. The findings indicate that the likelihood of investment banks’ alliance formation is positively related to the complementarity of their capabilities, as well as their status similarity. Social capital arising from banks’ direct and indirect collaborative experiences also plays a very important role in alliance formation. The number of deals given by a lead bank to a potential partner over the past three years has an inverted U-shaped relationship to the probability that the lead bank will invite the potential partner to form an alliance. Our findings indicate that status similarity and social capital have a stronger effect on alliance formation in initial public offering deals than in secondary offering deals, as the former are more uncertain than the latter. Using these findings, we discuss the role of complementarity, status similarity, and social capital in alliance formation. Copyright © 2000 John Wiley & Sons, Ltd.
The logit and probit models have become critical parts of the management researcher's analytical arsenal, growing rapidly from almost no use in the 1980s to appearing in 15% of all articles published in Strategic Management Journal in 2005. However, a review of three top strategy journals revealed numerous areas in their use and interpretation where current practice fell short of ideal. Failure to understand how these models differ from ordinary least squares can lead researchers to misunderstand their statistical results and draw incorrect conclusions regarding the theory they are testing. Based on a review of the methodological literature and recent empirical papers in three leading strategy journals, this paper identifies four critical issues in their use: interpreting coefficients, modeling interactions between variables, comparing coefficients between groups (e.g., foreign and domestic firms), and measures of model fit. For each issue, the paper provides a background, a review of current practice, and recommendations for best practice. A concluding section presents overall implications for the conduct of research with logit and probit models, which should assist both authors and readers of strategic management research.
Strategy researchers are increasingly turning their attention from examining the impact of strategic choices on firm performance to examining the factors that determine strategic choices at the firm level. This shift of research orientation has meant that researchers are increasingly faced with a limited dependent variable (LDV) that takes a limited number of usually discrete values, for which LDV methods such as Logit or Probit are required. Despite their growing popularity, there appears to be widespread problems in the use of LDV methods. This paper complements recent papers that offer general guidelines by presenting and illustrating the practical steps needed to implement the methods essential for analyzing and interpreting the results from LDV models.
This article examines the adequacy of the “rules of thumb” conventional cutoff criteria and several new alternatives for various fit indexes used to evaluate model fit in practice. Using a 2‐index presentation strategy, which includes using the maximum likelihood (ML)‐based standardized root mean squared residual (SRMR) and supplementing it with either Tucker‐Lewis Index (TLI), Bollen's (1989) Fit Index (BL89), Relative Noncentrality Index (RNI), Comparative Fit Index (CFI), Gamma Hat, McDonald's Centrality Index (Mc), or root mean squared error of approximation (RMSEA), various combinations of cutoff values from selected ranges of cutoff criteria for the ML‐based SRMR and a given supplemental fit index were used to calculate rejection rates for various types of true‐population and misspecified models; that is, models with misspecified factor covariance(s) and models with misspecified factor loading(s). The results suggest that, for the ML method, a cutoff value close to .95 for TLI, BL89, CFI, RNI, and Gamma Hat; a cutoff value close to .90 for Mc; a cutoff value close to .08 for SRMR; and a cutoff value close to .06 for RMSEA are needed before we can conclude that there is a relatively good fit between the hypothesized model and the observed data. Furthermore, the 2‐index presentation strategy is required to reject reasonable proportions of various types of true‐population and misspecified models. Finally, using the proposed cutoff criteria, the ML‐based TLI, Mc, and RMSEA tend to overreject true‐population models at small sample size and thus are less preferable when sample size is small.
Firms simultaneously face the need to cooperate with and control an alliance partner. To complement the transaction cost perspective's emphasis on the need to control and limit opportunistic behavior, we examine the sources and impact of the cooperation costs incurred in order to work with a partner. We propose that these costs increase with greater joint task complexity and interpartner diversity, and perceptions of equitable behavior affect the perceptions of these costs. Hypotheses derived from the framework are tested in a sample of 231 contractual alliances between architects and general contractors in the Hong Kong construction industry. We find that both cooperation costs and transaction costs affect the level of time and effort a manager expends on an alliance, supporting our fundamental proposition that the costs of cooperation and control are conceptually and empirically distinct. We argue that cooperation costs should be incorporated into studies that compare the choice of alternative partners and alliance structures, as well as among the broader categories of market, hierarchy, and hybrid governance forms. Copyright © 2005 John Wiley & Sons, Ltd.
We develop and test a novel theory about strategic noise with regard to CEO appointments. Strategic noise is an anticipatory and preemptive form of impression management. At the time it announces a new CEO, a board of directors seeks to manage stakeholder impressions by simultaneously releasing confounding information about other significant events. Several CEO and firm characteristics affect the likelihood that this will happen. Strategic noise is most likely when long-term CEOs have a wide pay gap between other top managers at high stock price performance firms, and when a new CEO does not have previous CEO experience or comes from a less well-regarded firm. Results showing that CEO succession announcements are noisier than they would be by chance have some interesting implications for impression management theory, traditional event study methodology, and managerial and public policy. Interviews with public firm directors on CEO succession provide additional validity for the strategic noise construct and help us to articulate key elements of the theory. Copyright © 2011 John Wiley & Sons, Ltd.