Power calculation for causal inference in social science: sample size and minimum detectable effect determination

Djimeu, Eric W.; Houndolo, Deo-Gracias

doi:10.1080/19439342.2016.1244555

Cited by 28 publications

(20 citation statements)

References 6 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…To determine sample size, we set an a priori power level and a minimum detectable effect that the intervention was expected to have on temporal discounting (Djimeu and Houndolo, 2016). We used the formula recommended by Bloom (2008):…”

Section: Participantsmentioning

confidence: 99%

Correcting misperceptions of relative income: impact on temporal discounting and social trust.

Guillou¹,

Grandin²,

Chevallier³

2020

Preprint

View full text Add to dashboard Cite

Social trust and income are associated both within and across countries, such that higher income typically correlates with increased trust. While this correlation is well-documented, the psychological mechanisms sustaining this relationship remain poorly understood. One plausible candidate is people's temporal discounting: on the one hand, trust has a strong time component - it exposes the individual to immediate costs in exchange of uncertain and delayed benefits; on the other hand, temporal discounting is robustly influenced by income. The goal of our studies was to test whether temporal discounting mediates the relationship between income and trust and whether experimentally manipulating perceived income has a downstream impact on temporal discounting and trust. To do so, participants underestimating their relative income position were randomized to receiving information about their true position in the income distribution or no information. In line with our hypothesis, our results indicate that temporal discounting mediates the effect of income on social trust in a pre-registered online study on British participants (N = 855). However, receiving a positive information shock on one's income position had no impact on either temporal discounting or social trust. In a second pre-registered study, we replicated the finding that temporal discounting mediates the effect of income on social trust in a representative sample of the British population (N = 1130).

show abstract

Section: Participantsmentioning

confidence: 99%

Correcting misperceptions of relative income: impact on temporal discounting and social trust.

Guillou¹,

Grandin²,

Chevallier³

2020

Preprint

View full text Add to dashboard Cite

show abstract

“…The power calculations of our evaluation largely followed Djimeu and Houndolo (2016), a manual prepared by 3ie Senior Evaluation Specialists. We adopted a two-stage sampling process.…”

Section: Sample Size Calculation and Determinationmentioning

confidence: 99%

Effectiveness of Evaluation Evidence to Influence Change in Primary Education Sub-sector in Uganda

Wokadala¹

2020

ASSRJ

View full text Add to dashboard Cite

A well-designed impact evaluation study in the education sector is relevant in informing decision and policymaking based on high-quality evidence generated from the research. In this paper, we provide a chronology on how we developed and implemented an impact evaluation on school facilities grant (SFG) programme designed to support Government aided primary schools in Uganda -to construct classrooms, sanitation facilities as well as teacher houses. We describe the appropriate methodology, randomization designs implemented and the target groups. Results revealed that selecting a clear intervention with identified observable and unobservable characteristics is important for the robust results. Measuring attribution on outcomes may also require a balance on internal and external validity checks and analysis. The internal validity, in addition to statistical measures also requires stakeholders' engagement in the impact study process as a way of promoting social audits of public programmes. Finally, a well-planned and designed evaluation often leads to increased public trust in the findings and evidence uptake hence informing policy decision in management of programmes and projects. 353 populated and neediest schools acquire physical infrastructure, in response to the national and international (EFAs and MDGs) commitments of universal primary education. The SFG aims at: a) promoting equitable access to primary education; b) building capacity within districts and the local communities as well as supporting supervision on sites; and c) alleviating poverty. The evaluation sought to answer the following question: Do schools that receive SFG register better learning outcomes than those that do not receive SFG? According to Khandker et al. (2010), to attribute an outcome to a particular intervention, impact evaluation should determine the instruments through which beneficiaries are responding to the intervention. In his study, education programmes usually appear promising at their outset, but with time they do not generate the expected outcomes. Therefore, in such cases conducting an impact evaluation would enable policy makers understand what works, when, where and how. This requires use of appropriate research methods.It is a universal fact that more education is associated with higher earnings with the global average rate of returns to schooling estimated at 10%. However, the positive benefits of education differs from country to another depending on factors related to gender, socio-economic status, geographical location, ethno-linguistic background among others (Montenegro and Patrinos, 2014; World Bank, 2012). The rigorous evaluation studies often isolate the factors and are able to measure attribution in the context. Well-conducted impact evaluations of programmes often lend empirical support to the various hypotheses on the education interventions needed by countries to adopt education reforms. According to Gertler et al. (2007); first, an appropriate model of behavior; second, detailed quantitative and qualitativ...

show abstract

“…To determine the sample size, we followed the general principles laid out by Singleton and Straits (2017) who suggest several interrelated principles -(1) the heterogeneity of the population; (2) the desired precision of generalisation; (3) the choice of sampling technique or method; (4) time/cost factors; and (5) the planned stratification of the data. On the desired precision of generalisation, we calculated the required sample size based on the power calculation (Djimeu & Houndolo, 2016). 7 For the purposes of this study, the need to over-sample (Salkind, 1997) to allow for non-response is not relevant, as the questionnaires were completed during face-to-face meetings with the producers in the field (rather than being carried out via mail/email or telephone)., In this way errors in data entries or interpretational mistakes were minimised to ensure the quality of data.…”

Section: Sampling and Data Collectionmentioning

confidence: 99%

Fair Trade and Wellbeing Improvements: Evidence from Sri Lanka

Holmes

Imai

2019

SSRN Journal

View full text Add to dashboard Cite

This paper investigates whether and how Fair Trade certification improves the wellbeing of small-scale producers by drawing upon a field study carried out by the authors in 2009 in the Central Province of Sri Lanka. A point of departure from earlier studies is to use a mixed methods approach, combining qualitative and quantitative data to assess the impact of Fair Trade on a broader set of development indicators to capture both the monetary and non-monetary progress of farmers. Methodologically, to overcome the limitation of small sample sizes of non-experimental survey data, we propose the use of propensity-score weighted linear and non-linear regression models with and without instrumenting the farmers' participation in Fair Trade. Here we have made treatment and control groups observationally comparable by applying propensity score matching (PSM) to match and weight the data, following Hirano and Imbens. We have found that Fair Trade certification increased farmers' actual income from tea production significantly, with fewer hours of work per day and accelerated perceived improvement in overall household income, as well as empowering women in decision making. Our mixed methods approach led us to conclude that Fair Trade certification benefits Fair Trade tea farmers through increased tea income and risk reduction.

show abstract

Power calculation for causal inference in social science: sample size and minimum detectable effect determination

Cited by 28 publications

References 6 publications

Correcting misperceptions of relative income: impact on temporal discounting and social trust.

Correcting misperceptions of relative income: impact on temporal discounting and social trust.

Effectiveness of Evaluation Evidence to Influence Change in Primary Education Sub-sector in Uganda

Fair Trade and Wellbeing Improvements: Evidence from Sri Lanka

Contact Info

Product

Resources

About