Bridging the gap between multilevel modeling and economic methods

Oshchepkov, Aleksey; Shirokanova, Anna

doi:10.1016/j.ssresearch.2021.102689

Cited by 15 publications

(5 citation statements)

References 77 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…However, there are tradeoffs – weaker assumptions also mean less efficiency. Thus, we recommend looking at comparisons of approaches (e.g., Oshchepkov & Shirokanova 2022) or conducting sensitivity tests to check the robustness of inferences to choices of approach for modeling standard errors. A full discussion of clustered and robust standard errors is beyond the scope of this paper, and we refer applied researchers to the documentation for the ‘ sandwich’ package in R and other comprehensive reviews (e.g., Cameron & Miller 2015).…”

Section: Statistical Model Designs To Coping With Omitted Variablesmentioning

confidence: 99%

“…As such, we highlight the need for a site level random effect with either of these two designs or a clustered standard error. For estimating standard errors, in general, we urge researchers to incorporate random effects or clustered robust standard errors as needed to accommodate clustering in the error, per the study design, recognizing the tradeoffs of using both and appropriate context ( reviewed in Oshchepkov & Shirokanova 2022).…”

Section: Comparison Of Approachesmentioning

confidence: 99%

See 1 more Smart Citation

Causal inference with observational data and unobserved confounding variables

Byrnes,

Dee

2024

Preprint

View full text Add to dashboard Cite

As ecology tackles progressively larger problems, we have begun to move beyond the scale at which we can conduct experiments to derive causal inferences. Randomized controlled experiments have long been seen as the gold standard for quantifying causal effects in ecological systems. In contrast, observational data, though available at larger scales, has primarily been used to either explore ideas derived from experiments or to generate patterns to inspire experiments - not for causal inference. This avoidance of using observational data for causal inference arises from the valid fear of confounding variables - variables that influence both the causal variable of interest and the studied effect that can lead to spurious correlations. Unmeasured confounders can lead to incorrect conclusions - a problem known as Omitted Variable Bias - that leads to the common saying, "Correlation is not causation." However, many other scientific disciplines that cannot do experiments for reasons of ethics or feasibility have developed rigorous approaches for causal inference from observational data. Here we show how Ecologists can harness these approaches, starting by using causal diagrams to identify potential known and unknown sources of confounding. We use a motivating example of assessing the effects of warming on intertidal snails to discuss how ecologists currently handle observational survey data and inference - often incorrectly with mixed models that produce biased coefficient estimates. We present alternative sampling designs and the statistical model designs that make use of them, discuss how they work using the language of causal path diagrams, demonstrate how easily they can be applied to common ecological datasets, and finally how well they are able to overcome problems of unmeasured confounding variables. We show how all of these techniques out-perform common approaches via simulation with respect to both bias and power. Our goal is to enable researchers to advance the field of Ecology at scale using observational data both on its own and as an important complement to experiments.

show abstract

Section: Statistical Model Designs To Coping With Omitted Variablesmentioning

confidence: 99%

Section: Comparison Of Approachesmentioning

confidence: 99%

Causal inference with observational data and unobserved confounding variables

Byrnes,

Dee

2024

Preprint

View full text Add to dashboard Cite

show abstract

“…Applying single-level modeling will ignore such clustering and nesting effects, which will assume individuals within a country are independent observations. MLME has demonstrated greater efficacy in accurately estimating standard errors compared to traditional economic modeling approaches (Cheah, 2009;Moulton, 1990;Oshchepkov & Shirokanova, 2022). The integration of aggregate-level data with microlevel data often results in ordinary least squares (OLS) estimates on the aggregate data, even with clustered standard errors, yielding standard errors that are excessively small.…”

Section: Empirical Strategiesmentioning

confidence: 99%

“…Despite these advantages, endogeneity problems persist for MLME, similar to their occurrence in widely‐used OLS methods in economic analysis. Regarding group‐level omitted variables, MLME surpasses popular economic solutions like the fixed effect model by providing equivalent estimates while retaining the flexibility of the random effect model (Oshchepkov & Shirokanova, 2022). It is crucial to note that while MLME reduces omitted variable biases by accounting for regional and country group characteristics with estimation flexibility, it cannot address reverse relations resulting from the impact of tax morale on corruption and any potential measurement errors.…”

Section: Empirical Strategiesmentioning

confidence: 99%

Corruption spills over on citizens: The case of self‐reported motivation to cheat on tax

Siddique

2024

Bulletin of Econ Res

View full text Add to dashboard Cite

Corruption poses a significant threat to societies’ social and economic fabric, eroding trust and contractual relationships between governments and citizens. This study investigates the repercussions of corruption spillover on individuals and explores how they adjust their motivation to engage in tax cheating by exploiting two distinct types of corruption heterogeneities at both individual and national levels. Leveraging newly collected individual corruption experiences locally from the World Value Survey alongside national‐level corruption data from Transparency International, we employ multilevel hierarchical modeling to estimate the effects of corruption on individual tax behavior. Our findings reveal that corruption negatively impacts people's willingness to comply with tax laws, with individuals who have personally experienced corruption exhibiting a more pronounced erosion of tax morale compared to those who have not. By shedding light on the intricate dynamics of corruption's influence on tax compliance, this paper contributes to our understanding of the broader implications of corruption for people and society. This paper can also inform policymakers and stakeholders in designing targeted interventions to mitigate the adverse consequences of corruption and restore citizens’ trust in their government.

show abstract

“…Such a task would be indeed unfeasible in these pages, not only because of the length constraints, but also, and more importantly, because of data problems. Apart from the need of collecting harmonized information on these dimensions, the number of countries available for our analysis (which is the relevant level of variation of the potential explanatory variables) is well below the minimum required for sound statistical inference, usually established around 50 (Angrist & Pischke, 2008;Bryan & Jenkins, 2016;Oshchepkov & Shirokanova, 2022).…”

Section: Introductionmentioning

confidence: 99%

Gender Gaps in Working Conditions

et al. 2023

View full text Add to dashboard Cite

The evolution of the gender pay gap has received a great deal of attention from academia and public opinion alike, yet our understanding of the differences in non-monetary working conditions is much scarcer. Through the use of the European Working Conditions Survey and six composite indicators of job quality, this research aims to expand our knowledge on the gender gaps in job attributes across the European Union over the period 2005–2015 in three ways: firstly, we explore gender differences in working conditions, revealing the distinct patterns identified by the dimensions of job quality and country; secondly, we find that, on average, women’s relative position deteriorates in terms of the physical environment and working time quality; a result that again reflects considerable heterogeneity across the countries in the sample. Lastly, we find clear evidence of a catch-up process in job quality in all the non-monetary dimensions of working conditions across the European Union.

show abstract

Bridging the gap between multilevel modeling and economic methods

Cited by 15 publications

References 77 publications

Causal inference with observational data and unobserved confounding variables

Causal inference with observational data and unobserved confounding variables

Corruption spills over on citizens: The case of self‐reported motivation to cheat on tax

Gender Gaps in Working Conditions

Contact Info

Product

Resources

About