Designing research instruments

Laterite works in close collaboration with clients to design effective research instruments. When developing research instruments our teams follow a number of principles:

(i) never re-invent the wheel, study the literature;

(ii) always think about the local context;

(iii) test the research instruments, don’t assume they make sense; and

(iv) keep it short, ideally under 1h for both ethical and budget reasons. 

Research instruments are back-translated into the local language, piloted in live conditions and tested through semi-structured interviews and focus groups. This ensures that the research instruments are consistent with the research objectives and that questions are well understood and interpreted by the target population. 

 

To ensure data accuracy and quality, Laterite checks a number of quality criteria including the flow of questions, the wording of questions, the precision of the translation and the degree to which the research instruments capture the local context.

Designing weighting strategies

Laterite’s research team helps clients select the most appropriate weighting strategy for their sample. Weighting considerations that we can help you think through:

(i) help clients estimate population totals (using weights reflecting the inverse probability of selection);

(ii) adjust samples so that they better match the characteristics of the population (using post-stratification weights);

(iii) correct for attrition bias (if there is a known source of attrition); or

(iv) balance the treatment and control group (weighting on the inverse of the propensity score).

Econometric analysis

Laterite’s research team helps clients make sense of data using econometric analysis techniques and statistical software tools.

Our teams work with a number of principles:

(i) structure the analysis following a clear set of priors/hypothesis;

(ii) work collaboratively, as that is the best way to generate additional ideas and hypotheses;

(iii) start simple, focusing on uni-variate and bi-variate analysis to understand variables and patterns in the data;

(iv) automate analysis wherever possible to accelerate delivery to the client and not focus too much energy on trivial calculations;

(v) estimate coefficients using several techniques/models to test the robustness of results;

(vi) think twice about how standard errors and test-statistics are estimated, and

(vii) convey findings graphically.

 

We have experience working with all sorts of datasets, including household datasets, child datasets, agriculture-focused datasets, export datasets, national tax databases, etc.

Network analysis and complexity techniques

Laterite is constantly looking for partners interested in applying network analysis or complexity methods to development issues. The type of techniques we routinely adapt to our project work, include:

(i) studying the relationship between variables not only in the framework of regression analysis but also through the lense of a network of variables, which has proven to lead to very interesting insights;

(ii) applying measures of complexity to calculate socio-economic indexes, such as wealth;

(iii) generating similarity networks, to identify and visualise clusters of individuals or clusters; and

(iv) using network visualisations to convey messages/findings.  

 

One example of how Laterite links network analysis and development, is what we call “Proximity Controls”, a new impact evaluation method proposed by Laterite and published in the Journal of Globalization and Development  (click here to read paper). We are always working to expand our capabilities and project portfolio in this space.

Designing & implementing sampling strategies

Laterite helps clients design the most optimal sampling strategies, given research objectives and budget constrains. Laterite also works with clients to operationalize these strategies, while minimizing the risks of compromising the validity of the sample and staying within budget parameters.

 

Sampling effectively is key to the success of research projects; it determines:

(i) which research questions can be addressed;

(ii) the validity of the results;

(iii) the analytic options available;

(iv) the precision of results; and

(v) the required budget, as sample size is one of the biggest cost-drivers in research projects.

 

Sampling always comes with uncertainty, risks and limitations. Laterite helps clients understand some of the analytic and financial trade-offs involved when opting for different sampling strategies as well as the sensitivity of statistical power to various known/unknown parameters using established formulas or Monte Carlo Simulations.

Program evaluations from design to analysis

Our research teams specialise in a variety program evaluation techniques, including randomised controlled trials (RCT), cluster-randomized controlled trials (CRTs) with balancing or pair-wise matching, regression discontinuity design (RDD), synthetic and proximity control methods, matching techniques and instrumental variable approaches.

 

Laterite has conducted or contributed to the design and analysis of impact evaluations in Rwanda, Kenya, Tanzania, South Sudan, Burundi and Ethiopia, covering gender, education, teacher training, gender violence, cash transfers, poverty reduction and agriculture support programs. 

 

Through the Laterite Lab, Laterite has also contributed to the program evaluation literature, proposing a new technique called Proximity Controls (click here to read paper), that enables flexible evaluations at the cluster level (published in the Journal of Globalization and Development).

Generating testable research questions

Laterite supports clients in the design of relevant and testable research questions using its proprietary design-thinking approach. 

 

This iterative and collaborative method, allows us to design research projects in a structured way, starting from policy assumptions or what we refer to as policy levers, through to identifying testable priors/hypotheses, symptoms that we would expect to observe if these priors were true, and associated metrics. The process can involve semi-structured interviews and focus groups.

 

The aim of the design-thinking-approach is to:

(i) think strategically and in a structured way about how a research project can inform policy and program design/implementation;

(ii) to generate ideas on the best hypothesis to test;

(iii) to engage with the client through a structured thinking process; and

(iv) to ensure research projects are as relevant as possible.

GIS mapping and data visualizations

Laterite has the in-house capability to conduct simple GIS mapping for clients, including choropleth maps, graduated symbol maps, and point maps, that account for the vast majority of our clients’ needs.

 

Laterite also puts a large emphasis on data visualisation and conveying statistics in a graphical and simple way. Data visualisation plays a double role:

(i) it makes data easier to understand and read; but

(ii) also helps our research teams generate new insights about patterns and correlations.

 

Laterite is continuously expanding its capabilities in the GIS and Data Visualization space and experimenting with new ways to convey information, in particular when it comes to network visualisations.



Competent Research Team

Laterite’s core team brings strong individual and collective technical expertise in delivering research projects, from start to finish. Our research team, trained in some of the world’s best universities, specialises in econometric analysis, program evaluation, and sampling techniques.



Structured Research Approach

Laterite takes a very structured approach to research using a proprietary method. This iterative and collaborative method, allows our teams to design research projects in a structured way, starting from policy assumptions, through to identifying testable priors/hypotheses, symptoms that we would expect to observe if these priors were true, and associated metrics.


 

Full Research Package

Laterite offers a “one-stop-shop” for organisations interested in conducting research, from design, through to electronic data collection and analysis. With excellent logistics systems in place, a large team of enumerators, electronic equipment, a strong research team, and an excellent understanding of the local context, we have the capacity to mobilize quickly.


 

Analytic Rigour & Transparency

Laterite strives to produce analysis of the highest quality. We achieve this by taking a structured approach, testing hypotheses both quantitatively and qualitatively and systematically stress-testing our results and methods. This also means being forthcoming about the limitations of the research undertaken and the assumptions under which results hold.


 

Local Knowledge and Presence

We have worked on many different topics, with clients in both public and private sectors, Laterite has an excellent standing of the local context. As a company, we invest in always improving our local knowledge and integrating the specifics of the local context into our project design and project execution.


 

Testing Innovative Methods

At Laterite, we try to innovate with every new opportunity, integrating the latest available research methods, testing approaches borrowed from other sciences (in particular, complex networks) and applying recently acquired ideas from previous projects. Through our work and the Laterite Lab, we aim to make a significant contribution to development economics.

Explore our other services:

Data Research Advisory