Model Specification Research Articles

Remote assessment of the association between early executive function and mathematics skills

AbstractExecutive function (EF) skills are consistently associated with global mathematics assessments. However, less is known about which specific mathematics skills invoke EF in early childhood. We adapted batteries of EF, numerical, and patterning tasks to be conducted via synchronous video conferencing with typically developing 4‐year‐old children (N = 115, 56.6% female, 85.2% White, non‐Hispanic, 7.8% Multiracial, 0.9% Hispanic, 0.9% Black, 1.7% American Indian/Alaska Native, 2.6% Asian and 0.9% missing) from primarily upper middle class households (maternal education range: some college to graduate or professional degree, mode = graduate or professional degree) in the United States. We found feasibility of remote adaptation varied greatly by task. Just over half of enrolled children (57.02%) completed all tasks and there were higher rates of missingness compared to an in‐person sample with a similar age range collected in schools. Nonetheless, working memory predicted various mathematics skills, including patterning, verbal counting, and number comparison, controlling for age and maternal education. Relations between cognitive flexibility and patterning varied across model specifications. Overall, these findings provide insight into the costs and benefits of remotely administered direct assessments of EF and mathematics in early childhood and add to the growing body of literature suggesting working memory plays an important role in several foundational early mathematics skills. Suggestions are offered for remote assessment with preschoolers.Highlights Executive function and mathematics tasks were adapted for remote use and associations among skill sets were examined. Remotely administered data collection sessions had higher rates of missingness than comparable in‐person studies. Findings replicate and extend prior work on the associations between working memory and patterning, verbal counting, and number comparison in a remote setting.

Class Enumeration in Mixture Modeling with Nested Data: A Brief Report

Given the increased focus of educational research on what works for whom and under what circumstances over the last decade, educational researchers are increasingly turning toward mixture models to identify heterogeneous subgroups among students. Such data are inherently nested, as students are nested within classrooms and schools. Yet there has been limited guidance on which specifications are most appropriate for enumerating latent classes when data are nested. This study utilized longitudinal, state-collected student data to demonstrate the impact of different specifications (i.e., ignoring nested data, using a post-hoc adjustment, and a parametric and non-parametric approach) of a latent class model when analyzing nested data. The overarching goal of this study was to provide the implications of four different model specifications commonly used to adjust for clustering in the context of mixture modeling. We highlight factors that may influence researchers’ decisions to employ one approach over another when conducting multilevel mixture modeling. We conclude with a set of recommendations that may be particularly helpful for the use of these methods in educational settings, where nested data is common.

Unlocking Dividends: The impact of managerial social capital on international corporate payouts

We present robust international evidence that managerial social capital is a significant determinant of dividend policy worldwide. Our analysis reveals that social capital mitigates information asymmetry and financial constraints, thereby resulting in increased dividend payouts. The effect is particularly pronounced for firms anticipating high cash retention costs. These findings are consistent with the pecking-order perspective of corporate payouts. Moreover, we identify a significant moderating role of shareholder legal protection and national cultural characteristics. Our results remain robust across alternative model specifications and tests for endogeneity.

Firm-level political risk and stock price crashes

In this study, we examine the relationship between firm-level political risk and stock price crash risk. Using a broad dataset of 4230 U.S. firms, 38,097 firm-year observations from 2002 to 2019, we reveal a positive association between political risk and stock price crash risk. These findings are robust to several model specifications and endogeneity checks. By using the Brexit referendum as a quasi-natural experiment, we provide evidence of a causal relationship between political risk and crash risk. Through channel analysis, we identify that this relationship is mediated via higher idiosyncratic volatility, lower price informativeness, and higher distress risk. We also find that our results are more pronounced in intangible-intensive firms. Interestingly, we show that managers of these firms respond to political risk by engaging in bad news hoarding. Finally, strong (external or internal) corporate governance mechanisms can moderate the positive relationship between political risk and stock price crash risk.

Hypokalemia, hypomagnesemia, and hyponatremia are associated with acute kidney injury in patients treated with cisplatin.

Cisplatin-associated acute kidney injury (C-AKI) is common. Predictive factors include age >60 years, hypertension, cisplatin dose, diabetes, and serum albumin < 3.5 g/L. The association between C-AKI and hypokalemia, hypomagnesemia or hyponatremia has not been well characterized. Data from a previous retrospective observational study was obtained. Patients were separated into three groups with similar cisplatin doses and schedules. Group A received cisplatin 60-100 mg/m2 every three weeks with laboratory assessments before treatment, group B received cisplatin 60-75 mg/m2 every three weeks with laboratory assessments before days 1 and 8 and group C had weekly cisplatin 40 mg/m2 with weekly laboratories assessments. The association between hypomagnesemia, hypokalemia, hyponatremia, and risk of AKI was determined using a counting process specification of Cox's regression models. A total of 1301 patients were separated into groups A (n = 713), B (n = 204), and C (n = 384). The proportion of patients with at least one event of hypokalemia, hypomagnesemia, or hyponatremia was lower in group A (29.2%, 57.6%, 36.2%) compared to groups B (43.6%, 67.2%, 59.8%) and C (49.0%, 78.7%, 51.0%). The incidence of all grade C-AKI was 35.6% (group A), 46.6% (group B), and 18.2% (group C). In group A, the risk of AKI doubled with hyponatremia or hypomagnesemia and tripled with hypokalemia. This association was not seen with other groups. Among patients with the highest doses of cisplatin, the presence of one electrolyte disorder was associated with an increased risk of C-AKI. Other studies are needed to characterize the presence of an electrolyte disorder as a predictive risk factor of C-AKI in this subpopulation.

Relational hyperevent models for the coevolution of coauthoring and citation networks

Abstract The development of appropriate statistical models has lagged behind the ambitions of empirical studies analysing large scientific networks—systems of publications connected by citations and authorship. Extant research typically focuses on either paper citation networks or author collaboration networks. However, these networks involve both direct relationships, as well as broader dependencies between references linked by multiple citation paths. In this work, we extend recently developed relational hyperevent models to analyse networks characterized by complex dependencies across multiple network modes. We introduce new covariates to represent theoretically relevant and empirically plausible mixed-mode network configurations. This model specification allows testing hypotheses that recognize the polyadic nature of publication data, while accounting for multiple dependencies linking authors and references of current and prior papers. We implement the model using open-source software to analyse publicly available data on a large scientific network. Our findings reveal a tendency for subsets of papers to be cocited, indicating that the impact of these papers may be partly due to endogenous network processes. More broadly, the analysis shows that models accounting for both the hyperedge structure of publication events and the interconnections between authors and references significantly enhance our understanding of the mechanisms driving scientific production and impact.

Does financial market liberalization promote corporate radical innovation? Evidence from China

This paper investigates the impact of financial market liberalization on corporate radical innovation using Chinese non-financial firm-level data from 2000 to 2022. The results show that increased financial market liberalization promotes corporate radical innovation by alleviating financing constraints, enhancing capital allocation efficiency, and improving credit allocation quality. These findings remain robust even after replacing variable measurements, changing model specifications, and addressing endogeneity concerns. The positive impact of financial market liberalization on corporate radical innovation is more pronounced for firms facing higher financing constraints, governance efficiency, information asymmetry, marketization, and capital intensity. Our study underscores the critical role of thoughtful financial market reforms in driving high-quality corporate development.

The nexus of conventional, religious and ethical indexes during crisis

This study examines the interconnectedness between conventional and ethical indexes. Using a Bayesian graphical vector autoregressive model, we derive the contemporaneous and temporal interdependencies among these stock index returns before and during the Covid-19 pandemic. Our model specification strategy combines vector autoregressive models with networks. The findings provide empirical evidence of increased interconnectedness during the Covid-19 period across all networks. Notably, the religious and FTSE Islamic networks exhibited greater resilience during the pandemic. This could be attributed to the rigorous screening processes for religious portfolios, which focus on lower-leveraged equity stocks, contributing to their stability. Additionally, our results show that the Covid-19 crisis affected network density and the roles of key player shock transmitter entities, as indicated by changes in hub and authority scores, with new key players emerging during the crisis.

What is it good for? On the inflationary effects of military conflicts

Military conflicts and wars affect the development of a country in various dimensions. Rising inflation rates are a potentially important economic effect associated with conflict. High inflation can undermine investment, weigh on private consumption, and threaten macroeconomic stability. Furthermore, these effects are not necessarily restricted to the locality of the conflict but can also spill over to other countries. Therefore, to understand how conflict affects the economy and to make a more comprehensive assessment of the costs of armed conflict, it is important to take inflationary effects into account. To disentangle the conflict-inflation nexus and to quantify this relationship, we conduct a panel analysis for 151 countries over the period 1950–2019. To capture indirect inflationary effects, we construct a distance-based spillover index. In general, the results of our analysis confirm a statistically significant positive direct association between conflicts and inflation rates. Furthermore, we document a statistically significant positive indirect association between conflicts and inflation rates in uninvolved countries. These findings are robust across various model specifications. Moreover, our results indicate that conflict-induced inflation is not driven solely by the increase in the money supply.

A stochastic production frontier model for evaluating the performance efficiency of artificial intelligence investment worldwide

As artificial intelligence (AI) begins to take center stage in technological innovations, it is essential to understand the business value of AI innovation efforts and investments. While some early work at the firm level exists, there is a shortage of literature that takes a larger country-level perspective. This study investigated the effect of AI innovation efforts on production efficiency across countries using stochastic production frontier approaches. In addition, our model also included the traditional economic inputs of capital and labor. We used both the Cobb–Douglas function and Constant Elastic Substitution model specifications. The significant findings of this study are as follows: Innovation efforts in AI measured by the number of AI-related patents and capital investment in AI have a substantial effect on economic output. The significance of AI investments indicates the need for a robust digital infrastructure as a prerequisite for harnessing AI capabilities. The complementary relationship between labor and AI-related patents implies that high-skilled labor is often necessary to incorporate AI inputs into production. However, as AI capabilities develop, they weaken the effect on labor input. The study also distinguishes between AI innovation (research and development activities indicated by AI patents) and the production efficiency of AI investments (return on every dollar invested), highlighting that more AI innovation does not always translate into better production efficiency. The findings indicate that while the United States leads innovation in AI, the UK has the best production efficiency. China ranked fourth in AI innovation and has the lowest production efficiency among the countries included in the study.

Inference with cross-lagged effects-Problems in time.

The interpretation of cross-effects from vector autoregressive models to infer structure and causality among constructs is widespread and sometimes problematic. I describe problems in the interpretation of cross-effects when processes that are thought to fluctuate continuously in time are, as is typically done, modeled as changing only in discrete steps (as in e.g., structural equation modeling)-zeroes in a discrete-time temporal matrix do not necessarily correspond to zero effects in the underlying continuous processes, and vice versa. This has implications for the common case when the presence or absence of cross-effects is used for inference about underlying causal processes. I demonstrate these problems via simulation, and also show that when an underlying set of processes are continuous in time, even relatively few direct causal links can result in much denser temporal effect matrices in discrete-time. I demonstrate one solution to these issues, namely parameterizing the system as a stochastic differential equation and focusing inference on the continuous-time temporal effects. I follow this with some discussion of issues regarding the switch to continuous-time, specifically regularization, appropriate measurement time lag, and model order. An empirical example using intensive longitudinal data highlights some of the complexities of applying such approaches to real data, particularly with respect to model specification, examining misspecification, and parameter interpretation. (PsycInfo Database Record (c) 2024 APA, all rights reserved).

Comparing Functional Trend and Learning among Groups in Intensive Binary Longitudinal Eye-Tracking Data using By-Variable Smooth Functions of GAMM.

This paper presents a model specification for group comparisons regarding a functional trend over time within a trial and learning across a series of trials in intensive binary longitudinal eye-tracking data. The functional trend and learning effects are modeled using by-variable smooth functions. This model specification is formulated as a generalized additive mixed model, which allowed for the use of the freely available mgcv package (Wood in Package 'mgcv.' https://cran.r-project.org/web/packages/mgcv/mgcv.pdf , 2023) in R. The model specification was applied to intensive binary longitudinal eye-tracking data, where the questions of interest concern differences between individuals with and without brain injury in their real-time language comprehension and how this affects their learning over time. The results of the simulation study show that the model parameters are recovered well and the by-variable smooth functions are adequately predicted in the same condition as those found in the application.

The role of loan loss provisions in income inequality: Evidence from a sample of banking institutions

This study explores the empirical link between income inequality and banks’ Loan Loss Provisions (LLPs) through a sample of banking institutions from 132 countries, applying a panel regression methodology. The evidence reveals that higher LLPs have a positive impact on income inequality, and the findings remain valid across various model specifications and income inequality measures. The results also hold against various robustness tests, such as different bank sizes, developed vs. emerging countries, the impact of the 2008 global financial crisis, and the controlling for risk. The implications are relevant for stakeholders, including regulators that endeavor to protect banking systems against expected and unexpected losses via LLPs. Specifically, since credit decisions have substantial effects on income inequality, regulators should mitigate the accumulation of LLPs, allowing more funds to be available for other banking system activities and functions.

Effect of Chinese outward FDI on youth unemployment in sub-Saharan Africa.

This paper investigates the effect of Chinese outward foreign direct investment (FDI) on youth unemployment in sub-Saharan Africa (SSA) by using a modified labour demand model to identify the investment sources that are helpful for reducing youth unemployment in the region. The model is applied to a panel of 42 countries for the period 2003-2021 using random-effect, and generalized method of moment (GMM) estimators for robustness check. Our results suggest that Chinese FDI has direct short-term reducing effect on youth unemployment in SSA. The direction of China's capital investment to infrastructure development and other labour-intensive activities leads to immediate reduction in youth unemployment. However, overtime, due to absence of linkages with domestic firms, and thus lack of demand effects, Chinese FDI becomes insignificant for employment creation. Our results also indicate that Other FDI does not lead to significant reduction in youth unemployment both currently and overtime. Our analysis gives partial support to the argument that the impact of Chinese FDI may differ from those of developed countries. Finally, we could not find evidence that the effect of FDI on employment depends on host country human capital and institutional quality. Several specifications of the empirical model are tested, and explanations are provided for the results. Policy implications are highlighted, especially the need to attract more job absorbing FDI and improve domestic absorptive capacity.

License to drive: State driver’s licensing laws and the travel of unauthorized immigrants

Since 1993, eighteen states and the District of Columbia have adopted laws allowing unauthorized immigrants to acquire driver’s licenses. California is one of these states, passing the Safe and Responsible Drive Act (AB 60) in 2013. In this paper, we examine the relationship between the adoption of state driver’s licensing laws and the commute mode of unauthorized immigrants. We focus specifically on the proportion of unauthorized Latin American immigrants in California who commute by car, drive alone to work, and commute by public transit, testing a number of model specifications and proxies for identifying undocumented immigrant residents. We find that the implementation of AB 60 was associated with a 2.0 percentage-point increase in car commuting and a 4.2 percentage-point increase in driving to work alone by unauthorized immigrants. Results also show a 1.4 percentage-point decline in the rate of commuting by public transit after AB 60′s implementation. This drop in transit use, however, was less statistically robust than findings regarding car commuting.

A network model of stress contagion: evidence from the vocational classroom

PurposeOur study focuses on stress contagion in vocational school classes, examining how students’ stress experiences affect their spatial classmates. For this purpose, we apply a novel formal network model that allows us to differentiate between endogenous and exogenous peer effects in the stress contagion process. Using the network model, we investigate effects on students’ stress levels, considering the stress and coping experiences of spatial peers as well as didactic-methodological context factors.MethodsWe test our statistical model using secondary data collected in a study involving 53 students from two classes at a public German vocational training school. The students’ experiences of stress (time pressure, pressure to succeed) and coping strategies (understanding of the subject matter, self-confidence) were captured using the experience sampling method, while classroom characteristics (e.g., teacher instruction, cooperative work) were recorded through video-based analysis of lessons. Utilizing the panel data, we employ maximum likelihood estimation to assess the spatial peer effects model for both classrooms.ResultsAmong other findings, all model specifications revealed significant peer effects for both stress measures, indicating that the higher the stress experience of immediate peers in the classroom, the higher the individual stress experienced by the students. Concerning the considered context factors, we found, for example, that increased cooperative work leads to higher levels of stress experience.ConclusionFrom a substantive perspective, our results underscore the role of peer-to-peer contagion in the vocational classroom and thus suggest a nuanced examination of cooperative practices. From a research methodology perspective, our approach illustrates how various methods (such as experience sampling, video-based classroom observation, and spatial network information) complement and enrich each other, highlighting the added value of our network analytical approach

Exploring for the Change in Change of Employment Status Over Time

The study seeks to explore for presence of dissimilar developmental paths in employment status of American individuals. The empirical longitudinal exercise is conducted with hybrid latent class growth model and latent class growth mixtures. For testing the hypothesis of heterogeneous subgroups within employment status data three mainstream modelling variants of growth modelling are tested and elaborated in terms of model performance. Progressive employment status based typology is found robust to alternative model specifications voting for presence of clustered growth patterns in experiencing employment status over time. The resulted employment status based typology describes three clusters of active, inactive and mediocre active participants in terms of moving between various stages of employment status for the age span of 18years to 35 years i.e., from adult to post adult prime working life years. The presented approach to understand and explore employment status is novel perspective compared to classic survey based employment status data.

Field observation and verbal exchange as different peer effects in farmers’ technology adoption decisions

AbstractFarmers’ adoption of novel technologies is influenced by other farmers’ behavior, a phenomenon known as peer effects. Although such effects have been studied intensively, the literature does not clearly distinguish between those that result (1) from verbal exchanges with other farmers and (2) from field observations, including the application of technology, its outcomes, and field conditions. We extend existing theoretical concepts and hypothesize that verbal information exchanges and field observations are two types of peer effects. Using data from an online survey of German sugar beet farmers’ application of mechanical weeding from early 2022, we find that the likelihood of adopting mechanical weeding increases across all model specifications by around 26%–28% if at least one adopter is known and by approximately 30%–32% if at least one field is observed. The two types of peer effects complement and reinforce each other in explaining adoption decisions. The effects increase with the number of adopters known and fields observed but decrease with larger distances to the observed fields. The findings can support designing extension services and future peer effects research that should consider the distinction between peer effects arising from verbal exchanges and field observations.

Why does uncovered interest parity fail empirically?

The Uncovered Interest Parity theory, predicated on perfect capital mobility and a floating exchange rate regime, faces challenges in real-world contexts marked by capital restrictions and diverse exchange rate regimes. This research investigates the validity of the UIP hypothesis in selected OECD countries, considering the role of capital control and de facto exchange rate regimes. By analyzing annual and monthly data, including immediate rates, short-term borrowing rates, and 10-year government bond rate differentials, several panel estimation techniques and alternative empirical models are employed. The results are robust to sub-sample analyses, alternative specifications of the empirical model, and different estimation methods. The study provides evidence supporting the UIP, particularly for immediate and short-term interest rates. The empirical examination underscores the necessity of accounting for the de facto floating exchange rate regime and capital flow restrictions when testing the UIP model. Consequently, the UIP hypothesis moves beyond being a theoretical fallacy and emerges as empirically verifiable, holding significant implications for policymakers and advancing the understanding of international finance.

The revolution next door.

This paper explores the cascading influence of revolutionary moments on democracy and inequality, not at home, but across borders. We use data on revolutions and other social upheavals over the past 120years and examine their cross-national impact on a range of variables in neighboring countries. Engaging with debates on whether substantial democracy and equality increases require extraordinary circumstances, our research investigates whether revolutionary activities induce consequential spillovers, such as policy concessions from elites in neighboring contexts. In exploring spillover effects, the paper examines how significant events in one nation influence social life in adjacent ones. It encompasses an analysis of 171 countries over two centuries, connecting data on revolution with democracy and equality metrics, and hypothesizing that elite fear of revolutionary contagion may necessitate democracy and equality concessions to mitigate potential uprisings. Findings suggest neighboring revolutions positively impact domestic democracy and equality levels. We observe significant increases in an index of democracy and two indices of economic egalitarianism, although one of the egalitarianism measures is robust to all model specifications. Additionally, we find that isolated "protest-led ousters" can moderately increase suffrage and one of our indices of egalitarianism, while coups do not seem to impact democracy or inequality variables. By examining various upheaval types and outcomes across time and space, the study illuminates the causal relationship between global mobilizations and local changes, providing insights into how global events inform domestic outcomes.