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).

Phylogenetic structure of Guinea-Bissau ethnic groups for mitochondrial DNA and Y chromosome genetic systems

The maternal and paternal genetic profile of Guineans is markedly sub-Saharan West African, with the majority of lineages belonging to L0-L3 mtDNA sub-clusters and E3a-M2 and E1-M33 Y chromosome haplogroups. Despite the sociocultural differences among Guinea-Bissau ethnic groups,marked by the supposedly strict admixture barriers, their genetic pool remains largely common. Their extant variation coalesces at distinct timeframes, from the initial occupation of the area to later inputs of people. Signs of recent expansion in mtDNA haplogroups L2a-L2c and NRY E3a-M2 suggest population growth in the equatorial western fringe, possibly supported by an early local agricultural centre, and to which the Mandenka and the Balanta people may relate. Non-West African signatures are traceable in less frequent extant haplogroups, fitting well with the linguistic and historical evidence regarding particular ethnic groups: the Papel and Felupe-Djola people retain traces of their putative East African relatives; U6 and M1b among Guinea-Bissau Bak-speakers indicate partial diffusion to Sahel of North African lineages; U5b1b lineages in Fulbe and Papel represent a link to North African Berbers, emphasizing the great importance of post-glacial expansions; exact matches of R1b-P25 and E3b1-M78 with Europeans likely trace back to the times of the slave trade.

On invariant Rings of Sylow subgroups of finite classical groups

In this thesis we study the invariant rings for the Sylow p-subgroups of the nite classical groups. We have successfully constructed presentations for the invariant rings for the Sylow p-subgroups of the unitary groups GU(3; Fq2) and GU(4; Fq2 ), the symplectic group Sp(4; Fq) and the orthogonal group O+(4; Fq) with q odd. In all cases, we obtained a minimal generating set which is also a SAGBI basis. Moreover, we computed the relations among the generators and showed that the invariant ring for these groups are a complete intersection. This shows that, even though the invariant rings of the Sylow p-subgroups of the general linear group are polynomial, the same is not true for Sylow p-subgroups of general classical groups. We also constructed the generators for the invariant elds for the Sylow p-subgroups of GU(n; Fq2 ), Sp(2n; Fq), O+(2n; Fq), O-(2n + 2; Fq) and O(2n + 1; Fq), for every n and q. This is an important step in order to obtain the generators and relations for the invariant rings of all these groups.

Fecal androgen levels in common marmoset (Callithrix jacchus) males living in captive family groups

In captive common marmoset groups, the reproductive inhibition observed in subordinate female seems to be a result of olfactory, visual and behavioral cues from the dominant female. However, few studies have examined the relationship among adult males living in the same social group. These studies have shown that reproductive failure among peer males seems to be based on hormonal and behavioral mechanisms. New insights on sexual strategies in primates have been shown using fecal steroids, but so far no information is available for common marmoset males. In the present study, we evaluated the influence of light-dark cycle, age and reproductive condition on the profile of fecal androgens in males living in the same family group. Feces were collected from six fathers and six sons for androgen determination during the light phase of the 24-h cycle for eight days randomly distributed over a 4-week period. Androgen levels were determined by enzyme immunoassay technique. Adult sons showed higher androgen levels (166.97 ± 22.95 ng/g) than fathers (80.69 ± 44.38 ng/g) and juveniles (49.06 ± 23.15 ng/g; P < 0.05). No diurnal variation (P > 0.05) in fecal androgen profile was observed in adults or juveniles. No indication of androgen-mediated social competition between fathers and adult sons was demonstrable. These results provide basic information on fecal androgen profile useful to investigate the socioendocrinology of free-ranging common marmoset males and verify that, in contrast to daughters, the reproductive suppression of sons is not based on physiological inhibition of their gonads

Effect of energy intake on performance and carcass composition of broiler chickens from two different genetic groups

In order to evaluate the effect of energy intake and broiler genotype on performance, carcass yield, and fat deposition, 600 one-day-old male chicks from two different genetic groups (AgRoss 308 - commercial line and PCLC - Empresa Brasileira de Pesquisa Agropecuária (EMBRAPA) non-improved line) were fed diets with different metabolizable energy level (2950, 3200 and 3450 kcal/kg). A completely randomized experimental design in a 2X3 factorial arrangement with four replications of 25 birds per treatment was applied. In order to ensure different energy intake among treatments within each strain, feed intake was daily adjusted by pair-feeding schemes. AgRoss 308 broilers had better performance and carcass yield, and presented lower abdominal fat deposition rate. In both genetic groups, the highest dietary energy level increased weight gain, heart relative weight, and fat deposition. However, it reduced the difference between AgRoss 308 and PCLC for feed conversion ratio and carcass protein deposition. These findings allow concluding that genetic improvement had a significant effect on broiler energy metabolism, and that the highest performance differences between genetic groups are found when low-energy intake is imposed.

Reproductive and behavioral aspects of red-winged tinamous (Rhynchotus rufescens) in groups with different sex ratios

The aim of this research study was to evaluate the reproductive performance of tinamous submitted to five different male:female ratios. The study was carried out with 72 birds in a randomized experimental design with 4 replications. Tinamous were housed in cages, using the ratios of one (1:1), two (2:1), three (3:1) and four (4:1) females per male, and also one male was housed with three females individually (3R:1), in a rotational system. Reproductive records of the breeding season from September 2004 to March 2005 were used. The reproductive traits studied were: number of eggs laid, fertility, and percentage of eggs damaged and cracked by pecking. Nonparametric analyses of these traits were performed using Kruskal-Wallis test. Two replications of treatments 1:1 and 4:1, and one of treatment 2:1 were video-taped for three days, 12 hours/day. The videotapes were sampled according to the scan method to fit an ethogram. Birds were also watched for one hour per day to study dominance and agonistic behavior. None of the reproductive traits was affected by mating sex ratio (p<0.05). Female dominance could be related to displacement behavior (r=1.00), and male sitting in immobility plus sitting in activity behaviors were related to lower number of damaged eggs (r=-0.90). Social dominance was indirectly determined by displacement behavior in the study situation. A large number of damaged eggs occurred in all treatments, thereby not allowing a clear conclusion on the best male:female ratio.

Gastrointestinal nematode infection in beef cattle of different genetic groups in Brazil

Resistance to natural infection by gastrointestinal nematodes was compared in 67 female calves of the following genetic groups: Nelore (NX); 1/2 Senepol + 1/2 Nelore (SN); and 1/2 Aberdeen Angus + 1/2 Nelore (AN). The NX (n = 26), SN (n = 23) and AN (n = 18) animals were monitored for 14 months, during which they remained without treatment, allowed to graze in a tropical environment. Eggs per gram of feces (EPG), coprocultures and packed cell volume (PCV) were carried out monthly. No significant effects of the interaction between the genetic groups and month/year of collection and the genetic group on the EPG were found, but there was a significant influence of the month of collection (P < 0.01). The monthly PCV measurements did not differ for the animals of the three genetic groups and there was no association found between the EPG and PCV. The animals of the SN and NX groups showed similar numbers of EPG with results zero, while for the AN group these numbers were significantly lower (P < 0.05). Although the NX group had a large number of EPG with results zero, it also contained many animals with high counts, meaning this group had higher averages during the entire study period. The following nematode genera were found in the coprocultures: Haemonchus, Cooperia, Oesophagostomum and Trichostrongylus, the latter in smallest proportion. There was no significant difference between the genetic groups for averages of all parasites identified, except Cooperia, which were present in higher numbers in the animals of the NX group (P < 0.05). The results obtained in this experiment suggest that the use of Bos taurus x Bos indicus crossbreeds can be a good strategy to reduce the use of chemical control in Brazil. (C) 2009 Elsevier B.V. All rights reserved.

Resistance of cattle of various genetic groups to the tick Rhipicephalus microplus and the relationship with coat traits

Conselho Nacional de Desenvolvimento Científico e Tecnológico (CNPq)

Efeito do tamanho de tubetes na qualidade de mudas de Jatobá (Hymenaea courbaril L. var. stilbocarpa (Hayne) Lee et Lang.), Ipê-amarelo (Tabebuia chrysotricha (Mart. ex DC.) Sandl.) e Guarucaia (Parapiptadenia rigida (Benth.) Brenan)

Conselho Nacional de Desenvolvimento Científico e Tecnológico (CNPq)

What Role Should Government Regulation Play in Ecological Restoration? Ongoing Debate in São Paulo State, Brazil

Conselho Nacional de Desenvolvimento Científico e Tecnológico (CNPq)

Dominance of legume trees alters nutrient relations in mixed species forest restoration plantings within seven years

Failures in reforestation are often attributed to nutrient limitation for tree growth. We compared tree performance and nitrogen and phosphorus relations in adjacent mixed-species plantings of contrasting composition, established for forest restoration on Ultisol soil, originally covered by tropical semi-deciduous Atlantic Forest in Southeast Brazil. Nutrient relations of four tree species occurring in both planting mixtures were compared between a legume-dominated, species-poor direct seeding mixture of early-successional species ("legume mixture"), and a species-diverse, legume-poor mixture of all successional groups ("diverse mixture"). After 7 years, the legume mixture had 6-fold higher abundance of N(2)-fixing trees, 177% higher total tree basal area, 22% lower litter C/N, six-fold higher in situ soil resin-nitrate, and 40% lower in situ soil resin-P, compared to the diverse mixture. In the legume mixture, non-N(2)-fixing legume Schizolobium parahyba (Fabaceae-Caesalpinioideae) had significantly lower proportional N resorption, and both naturally regenerating non-legume trees had significantly higher leaf N concentrations, and higher proportional P resorption, than in the diverse mixture. This demonstrate forms of plastic adjustment in all three non-N(2)-fixing species to diverged nutrient relations between mixtures. By contrast, leaf nutrient relations in N(2)-fixing Enterolobium contortisiliquum (Fabaceae-Mimosoideae) did not respond to planting mixtures. Rapid N accumulation in the legume mixture caused excess soil nitrification over nitrate immobilization and tighter P recycling compared with the diverse mixture. The legume mixture succeeded in accelerating tree growth and canopy closure, but may imply periods of N losses and possibly P limitation. Incorporation of species with efficient nitrate uptake and P mobilization from resistant soil pools offers potential to optimize these tradeoffs.

Ecological distribution of stream macroalgae in different spatial scales using taxonomic and morphological groups

Conselho Nacional de Desenvolvimento Científico e Tecnológico (CNPq)

The reactivity of oytho-methoxy-substituted catechol radicals with sulfhydryl groups: Contribution for the comprehension of the mechanism of inhibition of NADPH oxidase by apocynin

Redox processes are involved in the mechanism of action of NADPH oxidase inhibitors such as diphenyleneiodonium and apocynin. Here, we studied the structure-activity relationship for apocynin and analogous ortho-methoxy-substituted catechols as inhibitors of the NADPH oxidase in neutrophils and their reactivity with peroxidase. Aiming to alter the reduction potential, the ortho-methoxy-catechol moiety was kept constant and the substituents at para position related to the hydroxyl group were varied. Two series of compounds were employed: methoxy-catechols bearing electron-withdrawing groups (MC-W) such as apocynin, vanillin, 4-nitroguaiacol, 4-cyanoguaiacol, and methoxy-catechol bearing electron-donating groups (MC-D) such as 4-methylguaiacol and 4-ethylguaiacol. We found that MC-D were weaker inhibitors compared to MD-W. Furthermore, the radicals generated by oxidation of MC-W via MPO/H(2)O(2), but not for MC-D, were able to oxidize glutathione (GSH) as verified by the formation of thiyl radicals, depletion of GSH, and recycling of the ortho-methoxy-catechols during their oxidations. The capacity of oxidizing sulfhydryl (SH) groups was also verified when ovalbumin was incubated with MC-W, but not for MC-D. Since the effect of apocynin has been correlated with inactivation of the cytosolic fractions of the NADPH oxidase complex and its oxidation during the inhibitory process develops a special role in this process, we suggest that the close relationship between the reactivity of the radicals of MC-W compounds with thiol groups and their efficacy as NADPH oxidase inhibitor could be the chemical pathway behind the mechanism of action of apocynin and should be taken into account in the design of new and specific NADPH oxidase inhibitors. (c) 2007 Elsevier B.V. All rights reserved.