Generalized Propensity Score Matching Research Articles

Effects of vegetable commercialization on household food security: a generalized propensity score matching approach

ABSTRACT The commercialization of smallholder farmers is seen as the most significant driver of improved food security and household welfare in most countries with low-agricultural economies. This study used a Generalized Propensity Score Matching (GPSM) approach to evaluate the average causal effects of vegetable commercialization levels on the food security status of households in Sabbata Hawas Woreda, Oromia Region, Ethiopia. Data were collected through surveys of 385 farm households, key informant interviews, focus group discussions, and observations. The results revealed that vegetable commercialization had both positive and negative effects on households’ food security. The MAHFP increased steadily as the levels of commercialization increased. However, the lower levels of commercialization are associated with negative outcomes for calorie intake, food consumption scores, and household dietary diversity scores. The caloric intake, HDDS, and FCS were improved after commercialization levels reached 50%, 40%, and 25% respectively. The findings indicate that an increase in commercialization may not always lead to consistent increases across all indicators of food security and it has heterogeneous effects on households with different levels of commercialization. The subsequent policy recommendations of the study results are that, even though smallholder commercialization strategies are crucial for ensuring increased food production, productivity, and incomes for food security, it still requires taking into account the heterogeneous food security effects of vegetable commercialization levels.

Propensity score matching for estimation of pairwise marginal hazard ratios

. There is a growing interest in using observational studies to estimate the effects of treatments on survival or time-to-event outcomes. However, few standard approaches can adequately accommodate multiple treatment levels, which are common in observational comparative effectiveness research. We study the asymptotic properties of the generalized propensity score matching estimators of the marginal hazard ratios between pairs of treatment levels. The estimates are obtained by fitting a marginal Cox proportional hazard model on the matched dataset. We evaluate our approach in a simulation study and a case study where we analyze the IQVIA electronic medical records data.

Reassessing the non-linear causal link between long-duration commuting and self-rated health: do behavioral preferences and built environment matter?

Long-duration commuting is widely recognized for its significant influence on health. However, while research has traditionally focused on direct impacts, there remains a critical need to explore the nonlinear dynamics of this relationship. This study aims to deepen our understanding of how behavioral preferences and built environments contribute to these complex interactions. This study was conducted in Jinan, China's most congested city, using data from the "Jinan Residents Commuting Survey" of 1,755 participants aged 19 to 59. We applied Generalized Propensity Score Matching (GPSM) to explore the nonlinear effects of commuting time on self-rated health, adjusting for participants' sociodemographic characteristics. Variables related to active commuting, physical activity, and perceived built environment were also examined for their potential moderating effects. Commuting for less than 21 minutes enhances health, but negative effects intensify and peak at 60 minutes. Heterogeneity analysis reveals that women and older adults, especially those with higher incomes, are more susceptible to long commutes, experiencing a delayed onset of adverse effects. While active commuting offers health benefits, it may exacerbate health issues if prolonged. Conversely, regular physical activity consistently improves health outcomes related to commuting. Additionally, factors like residential greenery and walkability help alleviate commuting-related stress, improving the overall commuting-health dynamic. This study clarifies the commuting-health relationship by identifying key time thresholds and the positive effects of active commuting and physical activity on mitigating longer commute impacts. The findings inform healthier commuting behaviors and offer practical guidelines for urban planning and policy-making to enhance commuter well-being.

Public interest or individual benefit: biased perception toward Chinese redistributive policies

PurposeRedistributive policies aim to reduce income disparities and improve social equity. This study investigates whether redistributive effects that successfully diminish objective income inequality also effectively alter people’s perceptions of inequality.Design/methodology/approachUtilizing data from the 2018 China Household Income Survey (CHIP), comprising 56,167 individuals, this study applies ordered probability regression (Oprobit) and ordinary least squares (OLS) for analysis. To address potential biases in estimates, we employed the generalized propensity score matching (GPSM) method to estimate the treatment effect of transfer income on perceptions of inequality.FindingsThe results indicate that while China’s redistribution policies effectively reduce income disparities, they do not improve perceptions of inequality. Individuals exhibit biased attitudes toward redistributive policies. Specifically, perceptions of inequality are insensitive to the overall redistributive effect; the relationship is negative among the poor but positive among the rich. This contradictory pattern may be attributed to perceived income losses among the rich and gains among the poor.Social implicationsThe findings have important implications for policy development. Redistribution policies should not only aim to mitigate income disparities but also address and improve people’s perceptions of inequality.Originality/valueExisting literature has largely overlooked the impact of redistribution on perceived income inequality. This study represents an early effort to explore whether redistributive policies that reduce income inequality also influence people’s perceptions of inequality.

Intergenerational care and rural childhood obesity in the digital era: Based on screen exposure perspective

BackgroundRural Chinese children are experiencing increasing obesity rates, yet studies often neglect the impact of IT and screen media growth on obesity risks in the context of intergenerational care, leading to incomplete strategies for the digital era. MethodsBy comprehensively utilizing the data on rural children aged 6–17 from the China Family Panel Studies (CFPS) and the China Health and Nutrition Survey (CHNS), this study aims to test the logical chain and specific mechanisms regarding “intergenerational care - screen exposure - rural childhood obesity”. We employ the Propensity Score Matching (PSM) and Generalized Propensity Score Matching (GPSM) methods to respectively address the self-selection biases associated with inter-generational care and children's screen exposure behaviors. Results1) Intergenerational care significantly increases screen exposure among rural children. 2) Gender bias increases the risk of screen exposure for rural boys under intergenerational care. 3) Children with higher screen exposure levels are more affected by intergenerational care, which further undermines parental supervision. 4) Children's screen exposure leads to increased sedentary time and higher probability of purchasing unhealthy foods, thereby exacerbating obesity. This process is facilitated by enhancing preferences for snacks, fast food, and beverages, and weakening preferences for physical activity. 5) GPSM analysis indicates that children's screen exposure has an inverted “U"-shaped impact on unhealthy dietary preferences and a “U"-shaped impact on activity preferences. It results in a nonlinear positive impact of screen exposure on obesity. This study reveals a positive association between screen exposure and obesity, offering new insights into how intergenerational care in the digital era may elevate obesity prevalence through excessive screen time for rural children.

Association between women’s empowerment and childhood obesity in China: a generalized propensity score matching analysis

ABSTRACT This study investigates the association between women’s empowerment and childhood obesity in China using Generalized Propensity Score Matching (GPSM) Analysis. Utilizing data from the China Family Panel Studies (CFPS), the research analyzes how varying degrees of women’s empowerment, defined by the wife’s income ratio to the couple’s total income, influence childhood obesity rates. Findings suggest an association between increased women’s empowerment and lower incidence of childhood obesity. The study also uncovers variations in the impact of women’s empowerment across different wife’s education levels and family structures. These results underscore the critical role of women’s empowerment in mitigating childhood obesity within China’s rapidly changing socio-economic landscape. This research offers valuable insights for policy interventions aimed at promoting women’s empowerment and addressing childhood obesity, highlighting a crucial link between gender dynamics and public health in contemporary China.

Do scholars' collaborative tendencies impact the quality of their publications? A generalized propensity score matching analysis

Recently, the research on the cooperation relationship between authors has received widespread attention. However, existing studies still have the following limitations: 1) They mainly study the impact of author collaboration patterns by correlation analysis without considering the existence of confounding factors. 2) Methods based on causal analysis primarily focus on exploring the impact of different cooperation models, while less considering the author's tendency to participate. 3) Previous studies fail to incorporate the structural attributes of the authors' cooperation network into covariates, which may lead to confounding bias. To overcome the above limitations, we further explore the causal effect of authors' participation levels on the quality of their publications by leveraging the Generalized Propensity Score Matching (GPSM) method. Moreover, to alleviate the influence of the structural features in the authors' cooperation network, we then take the typical structural features as covariates, preventing us from reaching incorrect conclusions caused by the variable bias. We conduct extensive experiments on a real-world dataset (collected from the Web of Science (WoS) core collection), and from the experimental results, we find that authors having different involvement tendencies usually have publications with different qualities. Specifically, we observe an “inverted U-shaped” curve on authors' participation tendencies. That is, the quality of papers first rises and then decreases with the increase of authors' participation tendencies, which means that researchers who excessively collaborate with others actually experience a decrease in average paper quality.

Is residential exposure to oil refineries a novel contextual risk factor for coronary heart disease?

Despite a multi-decade decrease in cardiovascular disease, geographic disparities have widened, with excess mortality concentrated within the United States (U.S.) South. Petroleum production and refining, a major contributor to climate change, is concentrated within the U.S. South and emits multiple classes of atherogenic pollutants. We investigated whether residential exposure to oil refineries could explain variation in self-reported coronary heart disease (CHD) prevalence among adults in southern states for the year 2018, where the majority of oil refinery activity occurs (Alabama, Mississippi, Louisiana, Arkansas, Texas, New Mexico, and Oklahoma). We examined census tract-level association between oil refineries and CHD prevalence. We used a double matching method to adjust for measured and unmeasured spatial confounders: one-to-n distance matching and one-to-one generalized propensity score matching. Exposure metrics were constructed based on proximity to refineries, activities of refineries, and wind speed/direction. For all census tracts within 10 km of refineries, self-reported CHD prevalence ranged from 1.2% to 17.6%. Compared to census tracts located at ≥5 km and <10 km, one standard deviation increase in the exposure within 5 km of refineries was associated with a 0.33 (95% confidence interval: 0.04, 0.63) percentage point increase in the prevalence. A total of 1119.0 (123.5, 2114.2) prevalent cases or 1.6% (0.2, 3.1) of CHD prevalence in areas within 5 km from refineries were potentially explained by exposure to oil refineries. At the census tract-level, the prevalence of CHD explained by exposure to oil refineries ranged from 0.02% (0.00, 0.05) to 47.4% (5.2, 89.5). Thus, although we cannot rule out potential confounding by other personal risk factors, CHD prevalence was found to be higher in populations living nearer to oil refineries, which may suggest that exposure to oil refineries can increase CHD risk, warranting further investigation.

The impact of adoption of milk safety practices on food and nutrition security: Evidence from smallholder dairy farmers in Ethiopia

Despite the progress observed over the past decades, food and nutritional insecurity remain the prominent challenges in Ethiopia. In this respect, improved farming practices and productivity-enhancing agricultural technologies have been identified as pathways out of these problems; these technologies are the priority agenda of the government policy in Ethiopia. Nonetheless, the impact of the adoption of milk safety practices on dairy farmers’ food and nutrition security has not yet been addressed, especially at the smallholder farmers’ level. The aim of this study was to evaluate the impact of the adoption of milk safety practices on dairy farmers’ food and nutrition security in Ethiopia, using primary data collected from 410 dairy farmers. Generalized propensity score matching was used to evaluate the heterogeneous impact of the adoption of milk safety practices on the household food and nutrition security indicators (per capita food consumption expenditure and food consumption score). The findings reveal that the adoption of milk safety practices had a moderate impact on food consumption expenditure but had no significant impact on food consumption score. An ordered probit endogenous switching regression was also used to account for unobserved bias and check the robustness of the generalized propensity score estimates. Accordingly, the results are similar to the generalized propensity score estimates, and unobserved bias was not a problem. In general, the impact of adoption of milk safety practices on food and nutrition security of smallholder dairy farmers is moderate in Ethiopia. This might be because of the high compliance costs with the adoption of milk safety practices and the lack of quality-based pricing systems. Hence, creating an enabling environment, which in turn, encourages the adoption of milk safety practices would help improve positive effects of the adoption of milk safety practices on food and nutritional security.

Addressing endogeneity between hazardous actions and motorcyclist injury severity by integrating generalized propensity score approach and instrumental variable model

Motorcyclist hazardous actions (e.g., particularly speed too fast or failing to stop in assured clear distance (ACD)) are commonly identified as risk factors that significantly impact the motorcyclist injury severity. However, endogenous effects resulting from motorcyclist hazardous actions have seldom been considered, which may cause the biased estimates. Specifically, two important sources of endogeneities (i.e., endogeneity arising from observed confounding factors and endogeneity caused by unobserved confounders) tend to yield a biased relationship between hazardous actions and motorcyclist injury severity. To jointly account for two sources of endogeneities and provide more robust estimates, the study tries to assess the effects of speed-too-fast and failing to stop in ACD on motorcyclist injury severity via a hybrid method by integrating the generalized propensity score approach with instrumental variable model. Specifically, we adopt a generalized propensity score matching method to reduce the endogeneity bias arising from observed confounders. Furthermore, the matched data are used to develop an instrumental variable model with random parameters to handle the endogeneity resulting from unobserved confounders and unobserved heterogeneity, which consists of random parameters binary logit models modelling the motorcyclist hazardous actions in the first stage and a random parameters logit model with heterogeneity in means modelling the motorcyclist injury severity in the second stage. The proposed approach is estimated based on Michigan motorcycle crash data from 2015 to 2018. Results suggest that alcohol use leads motorcyclists to engage in speed-too-fast, while alcohol use and signal control cause motorcyclists to be involved in failing to stop in ACD. Middle-aged and elderly motorcyclists, alcohol use, speed too fast, speed limit ≥50 mph, wet surface, and head-on/angle crashes significantly increase the injury severity of motorcyclists. Moreover, failing to stop in ACD produces a random parameter with heterogeneity in means, while intersection increases the mean effects of failing to stop in ACD on motorcyclist minor injury. These findings further provide insights for a better understanding of hazardous actions and motorcyclist injury severity via the impact analysis of various explanatory variables.

Social participation, psychological resilience and depression among widowed older adults in China.

Depression has become a challenging public health problem, and the loss of a spouse is one of the main causes of depression in older adults. Social participation and psychological resilience are protective factors that reduce depressive symptoms in adults. The purpose of this study was to explore the influences of social participation and psychological resilience on the depression of Chinese widowed older adults. We carried out a cross-sectional study on 790 community-dwelling widowed older adults in Shaanxi, China, in 2019. A structured face-to-face interview was used to collect data. We used the ordinary least squares model (OLS), the generalized propensity score matching (GPSM) and the mediation model to test the relationship between social participation, psychological resilience and depression. We find a U-shaped relationship between social participation and depression, with the increase in social participation scores, the average depression level decreased and then increased. The psychological resilience of the widowed older adults was negatively associated with depression. The higher their psychological resilience, the lower their depression level. Among the four factors of psychological resilience, factor 2 (optimism and positive acceptance of change) and factor 3 (secure relationships and tolerance of negative affect) of psychological resilience were protective factors for depression after adjusting for demographic variables and physical health conditions. Psychological resilience plays a mediating role, as evident in factor 2 and factor 3 mediating the relationship between social participation and depression. An appropriate level of social participation will relieve the depression of widowed older adults. Social participation can reduce the depression level of widowed older adults by improving their psychological resilience. Community and family could reduce depression by intervening in the social participation of widowed older adults. Active social participation is crucial because it has a protective and resilient impact, which can help people recover from the stress of losing a spouse.

The economic effects of unmanned aerial vehicles in pesticide application: evidence from Chinese grain farmers

Unmanned aerial vehicles (UAVs) are a recent innovation in precision agriculture technology. They are being used in a wide range of agricultural practices, whereby pesticide application is one of the most common uses of UAVs in China’s agriculture. However, the economic effects of UAVs in pesticide application have not been sufficiently investigated. To address the gap, this paper used propensity score matching to evaluate the economic effects of UAV adoption on outcome variables including revenue, pesticide costs, time spent on pesticide application, and pesticide application frequency based on a dataset covering over 2000 grain farmers across 11 provinces of China. Furthermore, generalized propensity score matching was used to evaluate the heterogeneity of outcome variables arising from differing UAV adoption intensities. The empirical results show that adoption of UAV increased revenue by approximately 434–488 dollars per hectare and reduced the time spent on pesticide application in the range of 14.4–15.8 h per hectare. Depending on the area with use of UAVs for pesticide spraying, UAV adoption has heterogeneous impacts on grain farmers’ revenue and the time spent on pesticide application. In terms of marginal revenue and marginal time spent on pesticide application, the optimal area with use of UAVs for pesticide spraying is estimated to be 20 hectares of arable land.

Promotion or inhibition of different incubation services? Evidence from government funding of China

Purpose The purpose of this paper is to investigate the relationship between the characteristics of the incubation industry, government funding, and the intensity of funding for different services. Because the incubation industry has particular characteristics, government funding varies for different services, and its intensity varies with service. Design/methodology/approach Government funding is classified as incubation subsidy and incubation incentive. Besides, incubation services include property management, business mentoring as well as investment and financing. Based on this, this study examines the influence mechanism of different subsidy and incentive on incubation services by using the generalized propensity score matching method. Findings The empirical results show that subsidy and incentive have an inverse-U shape effect on property management service, but a linear effect on business guidance service. Furthermore, subsidy does not affect investment and financing service, but incentive that can have a significant impact. Originality/value The theme of government funding and incubator services plays an important role in helping entrepreneurs expand their businesses. Incubation subsidy and incentive can provide important support to help enterprises obtain more preferential loans, technical services and technical support in the incubator. Applying it to incubator services can provide better technology and entrepreneurship guidance. These services can help new entrepreneurs understand products and markets, and how to develop more successfully in the early stage. In short, incubators supported by government funds can provide important support to entrepreneurs to help them successfully realize their business plans.

Research on the Impact of Rural Land Transfer on Non-Farm Employment of Farm Households: Evidence from Hubei Province, China

Agricultural scale operations and industrialization promote the transfer of the rural labor force to the industry sector, and the non-farm employment of farmers plays a great role in increasing their income and reducing poverty. It is of great significance to explore the non-farm employment of farmers for the governance of relative poverty and the achievement of common prosperity. The propensity score matching (PSM) and generalized propensity score matching (GPSM) were used to analyze the impact of rural land transfer on farm households' non-farm employment. According to the PSM estimation, compared to the farmers' land not transferred, the rural land transfer significantly increased the proportion of non-farm employment personnel in farm households and the months of per year non-farm employment per person. The total land transfer, paddy land transfer and dry land transfer could significantly increase the proportion of non-farm employment personnel in farm households by 0.074, 0.029 and 0.085 units, respectively, and could significantly increase the months of per year non-farm employment per person by 0.604, 0.394 and 0.617 units, respectively. According to the GPSM estimation, different types of rural land transfer areas have significant positive effects on the proportion of non-farm workers and the months of per year non-farm employment per person, and show an obvious increasing trend of returns to scale, that is, the proportion of non-farm workers and the months of per year non-farm employment per person of farmers are higher than the increase in rural land transfer area. Additionally, the return to scale effect of dry land transfer area is more obvious. In order to raise the income of farm households and narrow the gap between urban and rural areas, the land transfer system can be further improved, urbanization with the county town as an important carrier can be vigorously promoted, the participation of farm households in non-farm employment in the local area can be promoted and the support policy system for non-farm employment of rural labor force can be improved.

Too Much of a Good Thing? The Impact of Government Subsidies on Incubator Services: Empirical Evidence from China

In this paper, we investigate the impact of government subsidies on incubation services in incubators. Based on the use of the Generalized Propensity Score Matching (GPSM) method to effectively overcome the endogeneity problem, we find that there is an inverted U-shaped relationship between government subsidies and incubation services, and thus there is an optimal intensity of government subsidies. The inflection point of the inverted U shape for basic services is much smaller than that for value-added services and investment services. With a wider range of government subsidies, incubators will tend to provide better value-added and investment services rather than basic services. The inverted U-shaped relationship remains robust over time, and the range of appropriate subsidies to promote incubation services is tightening. The conclusions of this paper provide empirical evidence on the effectiveness of government subsidies in incubators, which is conducive to developing more effective industrial policies for future governments.

Effects of multiple cropping of farmland on the welfare level of farmers: Based on the perspective of poverty vulnerability

This paper aims to explore the impact of multiple cropping on farmers’ welfare level and provide the theoretical and empirical basis for solving relative poverty in rural areas in the future. The paper uses data from the field survey of 1,120 farmers in Hubei in 2018 and uses the Endogenous Transformation Regression Model (ESR) and generalized propensity score matching (GPSM) model to construct a counterfactual framework. The paper analyses the effect of multiple cropping on farmers’ relative poverty and examines its mechanism. The result shows that: Multiple cropping of farmland can reduce the relative poverty of farmers through the mechanism of yield improvement path and factor intensification path. Under the counterfactual hypothesis, the relative poverty of farmers would increase by 28.43% if the farmers who participated in the multiple cropping did not; and that the relative poverty of the farmers would decrease by 29.57% if the farmers who did not participate in the multiple cropping participated. From the perspective of multiple cropping, the poverty reduction effect of multiple cropping in paddy fields is higher than that of dry land. From the perspective of the degree of multiple cropping, the poverty reduction effects of paddy fields and dryland will experience an increasing return to scale as the multiple cropping index increases. When the household equivalent scale adjustment coefficient is not used to eliminate the impact of family population structure on the “family per capita comparable income,” the artificially high results estimated by the model cannot truly reflect the poverty-reducing effect of multi-cropping of farmland. This paper argues that the government can guide farmers to choose the suitable mode of multiple cropping to reduce farmers’ vulnerability to relative poverty.

Utilizing stratified generalized propensity score matching to approximate blocked randomized designs with multiple treatment levels

ABSTRACT Conducting causal inference in settings with more than one treatment level can be challenging. Classical methods, such as propensity score matching (PSM), are restricted to only a binary treatment. To extend propensity score methods beyond a binary treatment, generalized propensity score methods have been proposed, with generalized propensity score matching (GPSM) standing as the multi-level treatment analog to PSM. One drawback of GPSM is it is only capable of emulating a completely randomized trial (CRT) design and not the more efficient blocked randomized trial design. Motivated by the desire to emulate the more efficient design, we expand on GPSM estimating literature and develop a new estimator incorporating relevant stratifying variables into the GPSM framework. We examine the variance estimation methods available for GPSM and demonstrate how to extend the estimator to one where stratifying variables are included. While it would be straightforward to include relevant stratifying variables as covariates in the propensity score estimation, our method provides for researchers to conduct retrospective analyses more consistently with the prospective experiment they would have designed if permitted. Namely, our method permits researchers to approximate a stratified randomized trial as opposed to the CRT otherwise obtainable by GPSM. We apply our proposed method to an analysis of how the number of children in a household affects systolic blood pressure in adults. We conduct a simulation study assessing how the relationship between response, treatment, and strata affect the performance of our method and compare the results to non-stratified GPSM.

Effective Learning During COVID-19: Multilevel Covariates Matching and Propensity Score Matching.

In large-scale observational data with a hierarchical structure, both clusters and interventions often have more than two levels. Popular methods in the binary treatment literature do not naturally extend to the hierarchical multilevel treatment case. For example, most K-12 and universities have moved to an unprecedented hybrid learning module during the COVID-19 pandemic where learning modes include hybrid and fully remote learning, while students were clustered within a class and school region. It is challenging to evaluate the effectiveness of the learning outcomes of the multilevel treatments in a hierarchically data structured. In this paper, we study a covariates matching method and develop a generalized propensity score matching method to reduce the bias of estimation in the intervention effect. We also propose simple algorithms to assess the covariates balance for each approach. We examine the finite sample performance of the methods via simulation studies and apply the proposed methods to analyze the effectiveness of learning modes during the COVID-19 pandemic.

Duration of noninvasive respiratory support and risk for bronchopulmonary dysplasia or death.

Objective:To determine whether the duration of noninvasive respiratory support exposure is associated with bronchopulmonary dysplasia (BPD) or death in preterm infants.Methods:Multicenter, retrospective study of infants born at <29 weeks’ gestation. The association between days on noninvasive respiratory support and BPD or death was determined using instrumental variable techniques and generalized propensity score matching to account for potential confounding by illness severity.Results:Among 6,268 infants 36% developed BPD or died. The median duration of noninvasive respiratory support was 18 days. There was inconsistency in the association between noninvasive support and BPD or death when analyzed by instrumental variable techniques (Average Marginal Effect −0.37; 95% CI −1.23 to 0.50) and generalized propensity score matching (Average Marginal Effect 0.46; 95% CI 0.33 to 0.60).Conclusion:Findings on the association between duration of exposure to noninvasive respiratory support and the development of BPD or death were inconclusive.

Longer breastfeeding duration, better child development? Evidence from a large-scale survey in China

Based on a large-scale survey in China, we use the generalized propensity score matching method and instrumental variable method to explore the effect of breastfeeding duration on child development, and find that there is a relatively stable “U” relationship between breastfeeding duration and children’s early behavioral performance, with the optimal breastfeeding duration (6–10 months) promoting children’s early behavioral performance. However, more than the optimal breastfeeding duration will delay child development. We further find that the effect of breastfeeding duration on child development varies among different groups. There is evidence that the optimal breastfeeding duration has a greater positive effect on children from disadvantaged backgrounds. The conclusions of this paper have obvious practical value for the policy of child development strategies and parenting practices. Given the importance of early child development, there should be a reasonable breastfeeding duration, and the feeding interventions for disadvantaged children can bring higher returns.