Revisiting fixed- and random-effects models: some considerations for policy-relevant education research

Clarke, Paul; Crawford, Claire; Steele, Fiona; Vignoles, Anna

doi:10.1080/09645292.2013.855705

Cited by 79 publications

(76 citation statements)

References 20 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…54 Alternatively, the random-effect estimate itself can be regarded as a precision-weighted estimate. 55 Based on estimates of γ j , we can now proceed to describe hospital performance. For non-linear models, we do this in two different ways.…”

Section: Methodology and Statistical Approachmentioning

confidence: 99%

Variations in outcome and costs among NHS providers for common surgical procedures: econometric analyses of routinely collected data

Street

Gutacker

Bojke

et al. 2014

Health Services and Delivery Research

View full text Add to dashboard Cite

BackgroundIt is important that NHS resources are used to their full extent, but efforts to reduce costs may have an adverse effect on patient outcomes. Our research is designed to provide a better understanding of the inter-relationship between costs and health outcomes among NHS providers (hospitals) for common surgical procedures.ObjectivesIn England, patient-reported outcomes measures (PROMs) are collected from patients undergoing one of four elective procedures: unilateral hip replacement, unilateral knee replacement, groin hernia repair and varicose vein surgery. We identify variation in patient-reported outcomes (PROs) across hospitals, assess the relationship between the cost and outcomes among NHS hospitals for these procedures, and determine the extent to which variations in outcomes and costs are due to differences in hospital performance.Data sourcesWe link Hospital Episode Statistics (HES) data with reference cost data and PROM data for patients having the four treatments between April 2009 and March 2010.MethodsThe first part of the empirical analysis focuses on variation in different dimensions of self-reported health status. We argue that each of the EuroQol-5D questionnaire (EQ-5D; European Quality of Life-5 Dimensions) dimensions should be assessed separately. Our graphical summary of the differential impact that hospitals have on PROs indicates the probability of reporting a given health outcome and shows how these probabilities vary across EQ–5D dimensions and hospitals. The second part of the empirical analysis focuses on the performance of hospitals and the inter-relationship between PROs and resource use.ResultsWe find that poorer post-treatment health status is associated with lower initial health status, higher weighted Charlson score, more diagnoses and lower socioeconomic status. We find significantly unexplained variation among hospitals in outcomes for patients undergoing hip replacement, knee replacement or varicose vein surgery, but not for hernia patients. For all four treatments we find significant unexplained variation in resource use among hospitals, whether measured by cost of treatment or length of stay. This suggests that hospitals can improve their utilisation of resources.LimitationsOur analyses are based on the HES. If data are missing from the medical record, or extracted and coded inaccurately, HES will contain errors. Hospitals should minimise these errors. Our study suffers from a large number of missing data, mainly because some hospitals were better than others at administering the baseline survey.ConclusionsThere is no general evidence that hospitals with lower resource use have worse health outcomes. There is a significant positive correlation for varicose veins, but this is sensitive to the choice of resource use and PRO measures. For hip and knee replacement the correlation is either insignificant or negative (depending on the resource use and PRO measures), implying that promoting health outcomes and controlling costs are not contradictory objectives. Indeed, we are able to identify hospitals with better than expected outcomes where resource use is below average. Future research should address how to handle missing data, evaluate hospital performance within the broader health economy, communicate PROMs to prospective patients, evaluate the impact of PROMs on patient choice and provider behaviour and evaluate PROMs for people with chronic conditions.FundingThe National Institute for Health Research Health Service and Delivery Research programme.

show abstract

Section: Methodology and Statistical Approachmentioning

confidence: 99%

Variations in outcome and costs among NHS providers for common surgical procedures: econometric analyses of routinely collected data

Street

Gutacker

Bojke

et al. 2014

Health Services and Delivery Research

View full text Add to dashboard Cite

show abstract

“…Therefore, the equation for the fixed effect (FE) model is as follows: (3) (4) where T and C are time and country specific effects, and μ is the error term. The FE model is useful to investigate the relationship between the predictor and outcome variables within an entity (e.g., country and person), and checks for individual influences on the predictor and outcome variables (Clarke et al 2010).…”

Section: Empirical Methodsmentioning

confidence: 99%

A study on the relationship between international trade and food security: Evidence from less developed countries (LDCs)

Kang¹

2015

AGR ECON-CZECH

View full text Add to dashboard Cite

Hunger and poverty are main critical issues in less developed countries (LDCs) and can be attributed to the fact that almost 70% of the LDC populations are employed in the agricultural sector. Although the international trade can reduce food insecurity in LDCs, for example, facilitating the access to larger markets, the volatility in food prices since the late 2000s has negatively aff ected the food availability. Th is study examines the eff ects of trade on the food security and traces a U-shaped relationship between two variables, using 2000-2010 panel data for the LDCs. Th e results are robust across different methods and show that the food security decreases in the initial stages of the trade expansion but increases beyond a given threshold. Th e key fi nding of this study is the U-shaped relationship between trade and food security, which suggests that the policy makers should pay more attention to expanding investments in the agricultural productivity, particularly from the viewpoints of the self-suffi ciency and self-reliance.

show abstract

“…Equally, at the higher level, the mean term is no longer constrained by Level 1 effects, so it is free to account for all the higher-level variance associated with that variable. As such, the estimate of b 1 in 12 Instead of using the higher-level unit mean (an aggregate variable), Clarke et al (2010) suggest using global (Diez-Roux 1998) unit characteristics that are correlated with that mean. These global variables express the causal mechanism underlying the association expressed by b 4 , which may not be linear, as is assumed by Models 10 and 11.…”

mentioning

confidence: 99%

“…These global variables express the causal mechanism underlying the association expressed by b 4 , which may not be linear, as is assumed by Models 10 and 11. Including x j would be over-controlling in this case, and such a model has a different interpretation of the higher-level residual, but it is harder to reliably control out all (or even most) of the between effect from the within effect without using x j (Clarke et al 2010) in Equation 11. However, this is not a problem when using the formulation in Equation 12, as the within variable is already group mean centered, so the inclusion of x j is optional depending on the research question at hand, as in the 'within' model in Table 1. 13 Because of this, the number of higher-level units in the sample must be considered, and as such caution should be taken regarding how many higher-level variables (including x j s) the model can estimate reliably.…”

mentioning

confidence: 99%

Explaining Fixed Effects: Random Effects Modeling of Time-Series Cross-Sectional and Panel Data

Bell¹,

Jones²

2014

PSRM

1,091

933

View full text Add to dashboard Cite

This article challenges Fixed Effects (FE) modeling as the ‘default’ for time-series-cross-sectional and panel data. Understanding different within and between effects is crucial when choosing modeling strategies. The downside of Random Effects (RE) modeling—correlated lower-level covariates and higher-level residuals—is omitted-variable bias, solvable with Mundlak's (1978a) formulation. Consequently, RE can provide everything that FE promises and more, as confirmed by Monte-Carlo simulations, which additionally show problems with Plümper and Troeger's FE Vector Decomposition method when data are unbalanced. As well as incorporating time-invariant variables, RE models are readily extendable, with random coefficients, cross-level interactions and complex variance functions. We argue not simply for technical solutions to endogeneity, but for the substantive importance of context/heterogeneity, modeled using RE. The implications extend beyond political science to all multilevel datasets. However, omitted variables could still bias estimated higher-level variable effects; as with any model, care is required in interpretation.

show abstract

Revisiting fixed- and random-effects models: some considerations for policy-relevant education research

Cited by 79 publications

References 20 publications

Variations in outcome and costs among NHS providers for common surgical procedures: econometric analyses of routinely collected data

Variations in outcome and costs among NHS providers for common surgical procedures: econometric analyses of routinely collected data

A study on the relationship between international trade and food security: Evidence from less developed countries (LDCs)

Explaining Fixed Effects: Random Effects Modeling of Time-Series Cross-Sectional and Panel Data

Contact Info

Product

Resources

About