Testing for time-invariant unobserved heterogeneity in generalized linear models for panel data

Bartolucci, Francesco; Belotti, Federico; Peracchi, Franco

doi:10.1016/j.jeconom.2014.09.002

Cited by 25 publications

(8 citation statements)

References 30 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…We adjusted the panel-data regression for time to account for underlying trends and the total TCS score of each EU MS. GDP per capita was not included in the model as it did not improve the fit of the model. The fixed-effects specification accounts for time-invariant unobserved factors within each country [32].…”

Section: Discussionmentioning

confidence: 99%

Are smokers “hardening” or rather “softening”? An ecological and multilevel analysis across 28 European Union countries

et al. 2019

View full text Add to dashboard Cite

BackgroundTobacco control policies can reduce smoking prevalence. These measures may be less effective where smoking prevalence has significantly declined, as the remaining smokers have “hardened”. Our aim was to empirically evaluate the “hardening hypothesis” at the population level in the European Union (EU) and explore factors associated with hardcore smoking.MethodsWe conducted two separate analyses in the EU using data on smoking from the Eurobarometer surveys (2009–2017, n=112 745). 1) A panel-data fixed-effects linear regression to investigate changes over time in the percentage of hardcore smokers in relation to standardised smoking prevalence at the country level. 2) A multilevel logistic regression analysis with hardcore (daily smokers, ≥15 cigarettes per day who have not attempted to quit in the last 12 months) or light (<5 cigarettes per day) smoking as the dependent variable and time as the main independent variable, controlling for individual and ecological variables.ResultsWe studied 29 010 current smokers (43.8% hardcore smokers and 14.7% light smokers). The prevalence of hardcore smoking among adult smokers increased by 0.55 (95% CI 0.14–0.96) percentage points per each additional percentage point in the overall smoking prevalence. The odds of being a hardcore smoker increased over time and were higher in middle-aged males and people with financial difficulties, while the odds of being a light smoker significantly declined among females.ConclusionThis study does not support the “hardening hypothesis” in the EU between 2009 and 2017, but suggests a softening of the smoking population. Existing tobacco control policies are likely to be suitable to further decrease smoking prevalence in Europe.

show abstract

Section: Discussionmentioning

confidence: 99%

Are smokers “hardening” or rather “softening”? An ecological and multilevel analysis across 28 European Union countries

et al. 2019

View full text Add to dashboard Cite

show abstract

“…Next, we estimate the FE and RE model on our panel data samples, and Hausman-test was used to test the orthogonality between regressors and unobserved time-invariant effects 15…”

Section: Discussionmentioning

confidence: 99%

<p>Does Absenteeism Affect Academic Performance Among Undergraduate Medical Students? Evidence From “Rashid Latif Medical College (RLMC).”</p>

Khan¹,

Lodhi²,

Bhatti³

et al. 2019

AMEP

View full text Add to dashboard Cite

BackgroundDespite the strictness in attendance policies, absenteeism is an important current issue among medical and health sciences that affects the performance of undergraduate students worldwide. Attending lectures is the key to reaping the rewards of academic achievements in undergraduate medical students. As attendance pattern and performance go hand in hand– the physical engagement of students in the classroom and practical teaching approach have a synergistic impact on the output of lower-performing students to do their best.MethodsA total of 404 full-time undergraduate MBBS male and female students of Rashid Latif Medical College (RLMC) of age 18 or above were included in this study. The principal outcome variable, such as class grades or performance, is the dependent variable, while absenteeism is a unique independent variable.ResultsOur results capture the positive and significant impact of attendance coefficient in all proxy variables of regression models but exhibit significant decline from OLS3 to OLS4 when proxy variables for ability, effort, and motivation were included. However, we found the effect disappears, when we incorporate panel data fixed effect estimators to minimize the time-invariant student-specific unobservable traits on performance.ConclusionAttendance of the student is only the mandatory variable that must be monitored and regulated by corrective actions to achieve the better academic performance of the students.

show abstract

“…However, this may be unrealistic where time varying unobserved heterogeneity could arise as a result of unobserved time varying omitted variables or macro level shocks which influence each observational unit (N) in a different way. As a consequence, parameter estimates may be biased when the individual effects are assumed to be time invariant when in fact they are really time varying, see Bartolucci et al (2013). This is also likely to be more problematic in long panels.…”

Section: Random Time Invariant Unobserved Heterogeneitymentioning

confidence: 99%

The existence and persistence of household financial hardship: A Bayesian multivariate dynamic logit framework

Brown

Ghosh

Taylor

2014

Journal of Banking & Finance

View full text Add to dashboard Cite

Abstract:We investigate the existence and persistence of financial hardship at the household level using data from the British Household Panel Survey. Our modelling strategy makes three important contributions to the existing literature on household finances. Firstly, we model nine different types of household financial problems within a joint framework, allowing for correlation in the random effects across the nine equations. Secondly, we develop a dynamic framework in order to model the persistence of financial problems over time by extending our multi-equation framework to allow the presence or otherwise of different types of financial problems in the previous time period to influence the probability that the household currently experiences such problems. Our third contribution relates to the possibility that experiencing financial problems may be correlated with sample attrition. Hence, we explicitly model missing observations in the panel in order to allow for such attrition. Our modelling framework allows us to identify any persistence in financial problems over time as well as any interdependence that may exist between different types of financial problems. The results reveal interesting variations in the determinants of experiencing different types of financial problems including demographic and regional differences. Our findings also highlight persistence in experiencing financial problems over time as well as the role that saving on a regular basis can play in mitigating financial problems.

show abstract

Testing for time-invariant unobserved heterogeneity in generalized linear models for panel data

Cited by 25 publications

References 30 publications

Are smokers “hardening” or rather “softening”? An ecological and multilevel analysis across 28 European Union countries

Are smokers “hardening” or rather “softening”? An ecological and multilevel analysis across 28 European Union countries

<p>Does Absenteeism Affect Academic Performance Among Undergraduate Medical Students? Evidence From “Rashid Latif Medical College (RLMC).”</p>

The existence and persistence of household financial hardship: A Bayesian multivariate dynamic logit framework

Contact Info

Product

Resources

About