Inference in differences-in-differences with few treated groups and heteroskedasticity

Differences-in-Differences (DID) is one of the most widely used identification strategies in applied economics. However, how to draw inferences in DID models when there are few treated groups remains an open question. We show that the usual inference methods used in DID models might not perform well when there are few treated groups and errors are heteroskedastic. In particular, we show that when there is variation in the number of observations per group, inference methods designed to work when there are few treated groups tend to (under-) over-reject the null hypothesis when the treated groups are (large) small relative to the control groups. This happens because larger groups tend to have lower variance, generating heteroskedasticity in the group x time aggregate DID model. We provide evidence from Monte Carlo simulations and from placebo DID regressions with the American Community Survey (ACS) and the Current Population Survey (CPS) datasets to show that this problem is relevant even in datasets with large numbers of observations per group. We then derive an alternative inference method that provides accurate hypothesis testing in situations where there are few treated groups (or even just one) and many control groups in the presence of heteroskedasticity. Our method assumes that we can model the heteroskedasticity of a linear combination of the errors. We show that this assumption can be satisfied without imposing strong assumptions on the errors in common DID applications. With many pre-treatment periods, we show that this assumption can be relaxed. Instead, we provide an alternative inference method that relies on strict stationarity and ergodicity of the time series. Finally, we consider two recent alternatives to DID when there are many pre-treatment periods. We extend our inference methods to linear factor models when there are few treated groups. We also derive conditions under which a permutation test for the synthetic control estimator proposed by Abadie et al. (2010) is robust to heteroskedasticity and propose a modification on the test statistic that provided a better heteroskedasticity correction in our simulations.

Inference in differences-in-differences with few treated groups and heteroskedasticity

Differences-in-Differences (DID) is one of the most widely used identification strategies in applied economics. However, how to draw inferences in DID models when there are few treated groups remains an open question. We show that the usual inference methods used in DID models might not perform well when there are few treated groups and errors are heteroskedastic. In particular, we show that when there is variation in the number of observations per group, inference methods designed to work when there are few treated groups tend to (under-) over-reject the null hypothesis when the treated groups are (large) small relative to the control groups. This happens because larger groups tend to have lower variance, generating heteroskedasticity in the group x time aggregate DID model. We provide evidence from Monte Carlo simulations and from placebo DID regressions with the American Community Survey (ACS) and the Current Population Survey (CPS) datasets to show that this problem is relevant even in datasets with large numbers of observations per group. We then derive an alternative inference method that provides accurate hypothesis testing in situations where there are few treated groups (or even just one) and many control groups in the presence of heteroskedasticity. Our method assumes that we know how the heteroskedasticity is generated, which is the case when it is generated by variation in the number of observations per group. With many pre-treatment periods, we show that this assumption can be relaxed. Instead, we provide an alternative application of our method that relies on assumptions about stationarity and convergence of the moments of the time series. Finally, we consider two recent alternatives to DID when there are many pre-treatment groups. We extend our inference method to linear factor models when there are few treated groups. We also propose a permutation test for the synthetic control estimator that provided a better heteroskedasticity correction in our simulations than the test suggested by Abadie et al. (2010).

Labor rights, formality and spillovers: evidence from Brazil

Nós analisamos o efeito da emenda constitucional 72/13 no Brasil, que igualou direitos trabalhistas de empregadas domésticas a aqueles de outros empregados. Mostramos que, após a legislação, uma considerável cobertura midiática e um interesse público intensificado aumentou o conhecimento geral de direitos trabalhistas de empregadas domésticas. Como consequência, o não-seguimento de legislações trabalhistas no setor de serviços domésticos ficou mais difícil. Ao mesmo tempo, a necessidade de regulamentar adicionalmente a emenda fez com que custos trabalhistas ficassem praticamente inalterados. Usando uma abordagem de diferença-em-diferenças que compara ocupações selecionadas ao longo do tempo, mostramos que a emenda -- e a discussão que ela causou -- levou a um aumento na formalização e nos salários de empregados domésticos. Então, usando a heterogeneidade do impacto da emenda em grupos demográficos, nossos resultados mostram que emprego doméstico foi reduzido e que mulheres pouco qualificadas saíram força de trabalho e foram para empregos de menor qualidade. Testes de placebo e análises de robustez indicam que nossos resultados não são explicados por diversas interpretações alternativas.

Cherry picking with synthetic controls

The synthetic control (SC) method has been recently proposed as an alternative method to estimate treatment e ects in comparative case studies. Abadie et al. [2010] and Abadie et al. [2015] argue that one of the advantages of the SC method is that it imposes a data-driven process to select the comparison units, providing more transparency and less discretionary power to the researcher. However, an important limitation of the SC method is that it does not provide clear guidance on the choice of predictor variables used to estimate the SC weights. We show that such lack of speci c guidances provides signi cant opportunities for the researcher to search for speci cations with statistically signi cant results, undermining one of the main advantages of the method. Considering six alternative speci cations commonly used in SC applications, we calculate in Monte Carlo simulations the probability of nding a statistically signi cant result at 5% in at least one speci cation. We nd that this probability can be as high as 13% (23% for a 10% signi cance test) when there are 12 pre-intervention periods and decay slowly with the number of pre-intervention periods. With 230 pre-intervention periods, this probability is still around 10% (18% for a 10% signi cance test). We show that the speci cation that uses the average pre-treatment outcome values to estimate the weights performed particularly bad in our simulations. However, the speci cation-searching problem remains relevant even when we do not consider this speci cation. We also show that this speci cation-searching problem is relevant in simulations with real datasets looking at placebo interventions in the Current Population Survey (CPS). In order to mitigate this problem, we propose a criterion to select among SC di erent speci cations based on the prediction error of each speci cations in placebo estimations

Lesões pulmonares provocadas pela nicotina por via subcutânea em ratos.

RESUMO: Objetivo: Estudo com o objetivo de observar a influência da nicotina, aplicada pela via subcutânea, em pulmões de ratos. Métodos: Foram utilizados 20 ratos Wistar pesando 235±35g, separados aleatoriamente em 2 grupos iguais. O grupo I (n=10) recebeu nicotina na dose de 2 mg/ Kg/dia pela via subcutânea durante 20 dias e o grupo II (n=10) recebeu placebo pela mesma via de administração. Resultados: Os resultados mostraram que no grupo I ocorreu broncopneumonia em 3 (30%) ratos, leucocitose alveolar em 10 (100%) e leucocitose septal em 7 (70%). Atelectasia foi encontrada em 2 (20%). Transformados em escores, os dados totalizaram 52 pontos. Os escores das alterações observadas nos pulmões do grupo II atingiram 11 pontos (p<0,05). Conclusão: Os dados permitiram concluir que o uso da nicotina por via subcutânea contribuiu para o aparecimento de lesões pulmonares em ratos, em número e intensidade significativamente maiores do que nos animais considerados controles.

Lesões pulmonares provocadas pela nicotina por via subcutânea em ratos

Objetivo: Estudo com o objetivo de observar a influência da nicotina, aplicada pela via subcutânea, em pulmões de ratos. Métodos: Foram utilizados 20 ratos Wistar pesando 235±35g, separados aleatoriamente em 2 grupos iguais. O grupo I (n=10) recebeu nicotina na dose de 2 mg/Kg/dia pela via subcutânea durante 20 dias e o grupo II (n=10) recebeu placebo pela mesma via de administração. Resultados: Os resultados mostraram que no grupo I ocorreu broncopneumonia em 3 (30%) ratos, leucocitose alveolar em 10 (100%) e leucocitose septal em 7 (70%). Atelectasia foi encontrada em 2 (20%). Transformados em escores, os dados totalizaram 52 pontos. Os escores das alterações observadas nos pulmões do grupo II atingiram 11 pontos (p<0,05). Conclusão: Os dados permitiram concluir que o uso da nicotina por via subcutânea contribuiu para o aparecimento de lesões pulmonares em ratos, em número e intensidade significativamente maiores do que nos animais considerados controles

Efetividade do clareamento dentário com peróxido de carbamida a 10% e avaliação dos efeitos adversos sobre o esmalte dentário

The aim of the present study was to assess the effectiveness and adverse effects on dental enamel caused by nightguard vital bleaching with 10% carbamide peroxide. This was accomplished through the interaction of researchers from different areas such as dentistry, materials engineering and physics. Fifty volunteers took part in the doubleblind randomized controlled clinical trial. They were allocated to an experimental group that used Opalescence PF 10% (OPA) and a control group that used a placebo gel (PLA). Fragments of human dental enamel from the vestibular surface of healthy premolars, extracted for orthodontic reasons, were fixed to the vestibular surface of the first upper molars of the volunteers for in situ observation. Bleaching was performed at night for 21 days. The observation periods included Baseline (BL), T0 (21 days), T30 (30 days after treatment) and T180 (180 days after treatment, only for the OPA group). Tooth color was assessed by comparing it with the Vita® scale and by the degree of satisfaction expressed by the volunteer. We also assessed adverse clinical effects, dental sensitivity and gingival bleeding. The study of adverse effects on enamel was conducted in vivo and in situ, using the DIAGNOdent® laser fluorescence device to detect mineral loss. Scanning electron microscopy (SEM) was used to check for superficial morphological alterations, energy dispersive spectrophotometry (EDS) to semiquantitatively assess chemical composition using the Ca/P ratio, and the x-ray diffraction (XRD) technique to observe alterations in enamel microstructure. The results showed that nightguard vital bleaching with 10% carbamide peroxide was effective in 96% of the cases, versus 8% for the PLA group. Dental sensitivity was present in 36% (9/25) of the cases. There was no significant association between gingival bleeding and the type of gel used (p = 1.00). In vivo laser fluorescence analysis showed no difference in values for the control group, whereas in the OPA group there was a statistically significant difference between baseline values in relation to the subsequent periods (p<0.01), with lower mean values for post-bleaching times. There was a significant difference between the groups for times T0 and T30. Micrographic analysis showed no enamel surface alterations related to the treatment performed. No significant alteration in Ca/P ratio was observed in the OPA group (p = 0.624) or in the PLA group (p = 0.462) for each of the observation periods, nor between the groups studied (p=0.102). The XRD pattern for both groups showed the presence of three-phase Hydroxyapatite according to JCPDS files (9-0432[Ca5(PO4)3(OH)], 18-0303[Ca3(PO4)2.xH2O] and 25-0166[Ca5(PO4)3(OH, Cl, F)]). No other peak associated to other phases was found, independent of the group analyzed, which reveals there was no disappearance, nucleation or phase transformation. Neither was there any alteration in peak pattern location. With the methodology and protocol used in this study, nightguard vital bleaching with 10% carbamide peroxide proved to be an effective and safe procedure for dental enamel

Suplementação oral de zinco pode diminuir o perfil sérico de ferro em crianças saudáveis

Conselho Nacional de Desenvolvimento Científico e Tecnológico

Ação da cafeína sobre o rendimento esportivo de ciclistas em condições de calor e umidade

Objetivo: Verificar a ação da cafeína no tempo de rendimento, a taxa de esforço percebido (RPE), os níveis plasmáticos de glicose, sódio e potássio, a temperatura timpânica (Tt), o peso corporal (PC), freqüência cardíaca (FC) e concentração urinária da cafeína com a ingestão de doses de 5 e 9 mg/kg de cafeína e placebo, em provas ciclísticas sob condições de alto risco térmico. Métodos: Foram estudados 8 ciclistas treinados e aclimatizados em 3 provas de 45 km utilizando o modelo experimental e duplo-cego com randomização intra-sujeitos. Resultados: Não foram observadas diferenças significativas entre as variáveis avaliadas, entretanto o tempo de rendimento e a RPE foram menores com as doses de 5 e 9 mg/kg de cafeína que com a dose placebo. Conclusões: Estes dados indicam que as condições de calor e umidade podem ser suficientes para mascarar o benefício ergogênico da cafeína, entretanto deve-se considerar que a cafeína pode exercer influencia sobre a percepção subjetiva de esforço podendo levar à redução dos sinais de fadiga durante o exercício e conseqüente melhora do desempenho esportivo

Influência do tratamento com ácido lipoico sobre o metabolismo ósseo e perfil antioxidante de mulheres na perimenopausa

Bone is a dynamic tissue that is in constant process of remodeling in response to mechanical stress and hormonal changes. This study aimed to understand the relationship between the biochemical changes, which women in the menopausal transition are subject to, and how the use of an alternative therapy with lipoic acid (LA) could influence these changes. The study of double-blind, was carried out in perimenopausal women that underwent a three month treatment with 600 mg of AL compared with another group that received placebo during the same period. This study showed that women had a waist circunference and body mass index above the values recommended by WHO (WC ≥ 80 cm; BMI > 25kg/m2). Associated with this, these women had increased concentrations of total cholesterol and triglycerides, and borderline LDL (Total Cholesterol > 200mg/dL; Triglycerides > 150mg/dL; LDL >130mg/dL). These changes were not affected by treatment with AL. There were no shifts in liver profile (ALT, AST and GGT), kidney profile (urea, creatinine, total protein and albumin), mineral profile (Total Calcium, Ionized Calcium, Phosphorus and Magnesium) as well in bone markers (osteocalcin, Total Alkaline Phosphatase and Tartrate Resistant Acid Phosphatase) after treatment with LA. The results of the oxidative profile showed that treatment with LA decreased GPx activity (p < 0,01), while for the TBARS, GSH and SOD activity there were no differences. With regard to SOD, this enzyme will submit to be high in the placebo group after 3 months of study (p<0,05). The expression of RANKL mRNA was reduced (p < 0,05) and of RANK increased (p <0.001), after treatment with LA, while the expression of IL-6 and TNF-ɑ genes were no changed. We conclude that women already in the perimenopause stage have changes in lipid profile and body composition that could induce shifts in oxidative and bone metabolism. However, LA treatment has provided an effective effect in the oxidative and bone profile since the earliest markers such as GPx activity and mRNA expression of RANKL, respectively, were reduced associated with no change in SOD activity. These results suggest a beneficial and protective effect of LA, indicating it potential as an alternative treatment to help the to prevent the complications associated with estrogen deficiency

Effects of levomepromazine and different desflurane concentrations upon electrocardiographic variables in dogs

Objectives To investigate the effects of levomepromazine and different desflurane concentrations upon electrocardiographic variables.Animals Twenty adult mongrel dogs of both sexes weighing 6-28 kg.Methods Dogs were divided into two groups of 10 animals. Group I received 1 mg kg(-1) lV of levomepromazine and 15 minutes later anesthesia was induced with propofol (3 mg kg(-1) IV). Desflurane end-tidal concentration was set at 1.6 MAC. After 30 minutes at this concentration, measurements were taken and the end-tidal concentration was reduced to 1.4 MAC. Thereafter, it was reduced to 1.2 and then 1.0 MAC at 1.5-minute intervals. The same procedure was followed for group 2, except that levomepromazine was replaced with 0.2 mL kg(-1) of 0.9% saline solution and more propofol was needed for induction (7 mg kg(-1)). The animals' body temperature was maintained between 38.3 and 39 degreesC using a heating pad. The electrocardiographic tracing was obtained from lead II throughout the experimental period. The measurements were taken immediately before the administration of levomepromazine or placebo (T-1), 15 minutes after pre-medication (T-2) and 30 minutes after the establishment of 1.6 MAC (T-3)The other measurements were made at the concentrations of 1.4, 1.2, and 1.0 MAC, respectively (T4-6). The numerical data were submitted to analysis of variance plus F-test (p < 0.05).Results the dogs that received levomepromazine had a decrease in heart rate. However, in both groups it increased with desflurane administration. Levomepromazine, in association with desflurane, did not induce significant electrocardiographic changes, and all mean values (except P-wave duration) were within the reference range for this species.Conclusions and clinical relevance This study documented that levomepromazine, in association with desflurane, does not induce significant changes in electrocardiographic variables, suggesting that this drug combination has minimal effect on myocardial conduction.

Efeitos cardiorrespiratórios da buprenorfina em cães anestesiados pelo desfluorano

Fundação de Amparo à Pesquisa do Estado de São Paulo (FAPESP)

Efeito do ácido alfa-lipóico sobre biomarcadores de estresse oxidativo, inflamação e fatores de risco cardiovascular em hipertensos

Alpha-lipoic acid (ALA) is a potent antioxidant with favourable anti-inflammatory, metabolic and endothelial effects, and has been widely investigated due to its potential against cardiovascular risk factors. This study aimed to evaluate the effect of oral ALA supplementation on oxidative stress biomarkers, inflammation and cardiovascular risk factors in patients with hypertension. This is a double-blind placebo-controlled randomized clinical trial, where the intervention was evaluated prospectively comparing results in both groups. The sample consisted of 64 hypertensive patients who were randomly distributed into ALA group (n = 32), receiving 600 mg / day ALA for twelve weeks and control group (n = 32), receiving placebo for the same period. The following parameters were evaluated before and after intervention: lipid peroxidation, content of reduced glutathione (GSH), enzymatic activities of glutathione peroxidase (GPx) and superoxide dismustase, ultrasensitive C-reactive protein (hs-CRP), triglycerides, total cholesterol and fractions, fasting glucose and anthropometric indicators. There was a statistically significant reduction (p <0.05) in serum concentrations of total cholesterol, very low density lipoprotein (VLDL), high density lipoprotein (HDL), triglycerides and blood glucose. There was a reduction in body weight and waist, abdominal and hip circumferences in the group that received ALA. In addition, there was a statistically significant increase (p <0.05) in the contents of reduced glutathione (GSH) and glutathione peroxidase (GPx) in the group receiving ALA. Oral administration of ALA appears to be a valuable adjuvant therapy, which may contribute to decrease the damage caused by oxidative stress and other risk factors associated with the atherosclerotic process

Efeito da dexmedetomidina sobre a arritmia cardíaca induzida pela adrenalina em cães anestesiados pelo sevofluorano

Avaliou-se o efeito da dexmedetomidina sobre o ritmo cardíaco em 20 cães, sem raça definida, de ambos os sexos e considerados sadios, anestesiados pelo sevofluorano e submetidos a doses crescentes de adrenalina. Os animais foram, aleatoriamente, distribuídos em dois grupos (placebo e dexmedetomidina). No grupo placebo, os animais receberam, por via intravenosa, solução de NaCl a 0,9%, na dose de 0,3ml/kg. Foram considerados dois momentos, M0 e M1, imediatamente antes e após a aplicação, respectivamente. Após 10 minutos, realizou-se a indução anestésica com sevofluorano, por meio de máscara facial vedada, até a perda do reflexo laringotraqueal. em seguida, procedeu-se à intubação orotraqueal e a manutenção da anestesia foi realizada com a administração de sevofluorano na concentração de 1,5CAM, em circuito anestésico com reinalação parcial de gases. Decorridos 20 minutos da indução anestésica, iniciou-se a administração intravenosa contínua de solução de adrenalina a 2% em doses crescentes de 1, 2, 3, 4 e 5mg/kg/min, por meio de bomba de infusão, com aumento da dose em intervalos de 10 minutos. Imediatamente antes desse acréscimo eram feitas as mensurações (M2 a M6). No grupo dexmedetomidina empregou-se a mesma metodologia substituindo-se a solução de NaCl a 0,9% por hidrocloridrato de dexmedetomidina, na dose de 1µg/kg. Foram registradas as pressões arteriais, em M0 e em M2 a M6, e o traçado eletrocardiográfico, na derivação DII (M2 a M6), considerando-se para efeito estatístico o número total de bloqueios atrioventriculares (BAV) de primeiro e segundo graus e de complexos ventriculares prematuros (ESV), coincidentes com cada dose de adrenalina. Os dados foram submetidos à análise de variância seguida pelo teste de Tukey (P<0,05). Verificou-se que a dexmedetomidina interfere significativamente na condução atrioventricular levando a maior ocorrência de BAV e reduz o número de ESV nas doses infundidas de 2 e 3mg/kg/min de adrenalina. Logo após a aplicação de dexmedetomidina, observaram-se redução da freqüência cardíaca e da pressão arterial, cuja diminuição persistiu por até uma hora.

Efeitos cardiorrespiratórios da associação de tiletamina/zolazepam em cães (Canis familiaris) pré-tratados ou não pela acepromazina

Avaliou-se o uso da acepromazina como pré-tratamento à associação de tiletamina/zolazepam. Para tanto, utilizaram-se 20 animais da espécie canina, machos e fêmeas, adultos, hígidos, divididos em 2 grupos de igual número. O grupo 1 (controle) foi pré-tratado com 0,1 ml/kg de solução salina a 0,9 % e o grupo 2 com 0,2 mg/kg de acepromazina, ambos por via intravenosa. Decorridos 20 minutos, todos os animais receberam, pela mesma via, 10 mg/kg da associação tiletamina/zolazepam. Imediatamente antes da medicação pré-anestésica (M1), antes da aplicação da associação (M2) e aos 15, 30, 45 e 60 minutos após a administração da tiletamina/zolazepam, realizou-se mensuração de: freqüência cardíaca (FC); pressão arterial sistólica (PAS), diastólica (PAD) e média (PAM); débito (DC) e índice cardíaco (IC); volume sistólico (VS); eletrocardiograma (ECG); freqüência respiratória (FR); CO2 ao final da expiração (ETCO2); saturação da oxiemoglobina (SpO2); e temperatura retal (T0). Observou-se estabilidade cardiovascular, miorrelaxamento e aumento do período hábil anestésico com o uso da acepromazina na medicação pré-anestésica. O tratamento estatístco dos valores numéricos pela análise de perfil mostrou que a acepromazina diminuiu a FR; entretanto, a SpO2 e ETCO2 não sofreram alterações estatisticamente significativas, permitindo concluir que o emprego da fenotiazina apresenta vantagens sobre o uso isolado da associação tiletamina/zolazepam, em cães.