
Universidad Politécnica de Madrid · ETSI Informáticos
Experimental Software Engineering
We bring the scientific method to software development: controlled experiments, replications and observational studies that turn practice into reliable evidence.
About us
Evidence for software engineering
GRISE investigates how to generate and aggregate reliable evidence about software development methods, techniques and tools. Our work ranges from the methodology of software engineering experimentation (design, analysis, replication and synthesis) to its application in testing, test-driven development, requirements engineering, usability and continuous software engineering.
We collaborate with universities and companies across Europe and Latin America, and we have trained a large community of researchers through our PhD programme.
Work with us
Interested in a Bachelor's or Master's thesis (TFG/TFM), a PhD or a research stay on empirical software engineering? Write to any member of the group.
Get in touchResearch
Research lines
- 01.
Experimentation methodology, replication and evidence synthesis
We develop and evaluate methods to design, analyse, report and replicate software engineering experiments, including families of experiments, crossover designs and meta-analysis. We also study what makes evidence unreliable, such as statistical errors, publication bias and researcher bias.
- 02.
Causal inference and observational studies in mining software repositories
We adapt study designs from epidemiology, such as retrospective and prospective cohort studies and matching, to draw causal conclusions from software repository data. Recent studies examine how microservice adoption, code smells and static-analysis configurations affect development outcomes.
- 03.
Test-driven development and software testing
We run families of experiments in universities and companies to measure how test-driven development and testing techniques affect quality and productivity. We also examine how developers perceive testing effectiveness and how measurement instruments influence experimental results.
- 04.
Requirements engineering and human factors
We study empirically how the experience, domain knowledge and cognitive abilities of analysts shape requirements elicitation and inspection. We also investigate cognitive biases in software engineering.
- 05.
Usability and human-computer interaction
We investigate how to build usability into software development, from usability mechanisms and patterns to the evaluation of chatbots and collaborative modelling tools. In digital health, we design and evaluate mobile applications for older adults.
- 06.
Agile, rapid and continuous software engineering
We study how companies measure and manage quality, especially quality requirements, in agile and rapid software development. This work combines case studies and design science with industrial partners.
- 07.
Artificial intelligence and software engineering
We assess the methodological rigour of experiments that apply deep learning to software engineering tasks and build evidence-based AI tools. We also explore combining large language models with domain-specific languages for web engineering and teaching.
Latest
News
New national project on the empirical evaluation of LLMs in software engineering
The State Research Agency's proposed definitive resolution of the 2025 Knowledge Generation call includes project PID2025-173082NB-I00, Evaluación empírica de LLM en ingeniería del software: de experimentos con humanos a buenas prácticas metodológicas, led by Sira Vegas and Óscar Dieste (2026–2029, with a predoctoral contract).
Sira Vegas gives a keynote at MODELS 2026
Sira Vegas closed the keynote programme of MODELS 2026 in Málaga with the talk 'The Fragility of Software Engineering Experiments: What We Fail to See'.
Prospective cohort study of SonarQube Cloud at ASE 2026
Sabato Nocera, Sira Vegas and Giuseppe Scanniello present at ASE 2026 a prospective cohort study on the effects of customised configurations of static code analysis tools.
Francy D. Rodríguez becomes permanent faculty
Francy Diomar Rodríguez Tibocha holds a permanent lecturer position (Profesora Permanente Laboral) in DLSIIS from 1 September 2026, according to the UPM research portal.
Recent work
Publications
Building and validating deep learning models for forecasting the quality of cloud services
Automated Software Engineering, 33(3)
Enhancing Interoperability of Ubiquitous Devices through WS-CDL-Based Coordination
International Journal of Web and Grid Services, 22(3), pp. 350–378
Investigating Causality in Software Repositories Using Cohort-Based Methods
IEEE Transactions on Software Engineering, pp. 1–21
Motivation for international academic mobility: Comparing incoming and outgoing engineering students
Journal of International Students, 16(19)
Visibility of Domain Elements in the Elicitation Process Interviews: A Family of Empirical Studies
IEEE Transactions on Software Engineering, 52(4), pp. 1332–1351
People
The team





