- Email Address(es)
- email@example.com, firstname.lastname@example.org
- Office Phone
- 416 535 8501 x6674
- Office Address
- 33 Russell St. Room 1032 Toronto, ON M5S 2S1
- Riding Numbers (Stats), Looping Infinito (About everything - In Portuguese)
- Biostatistics Division
Office of Global Public Health Education & Training
- Adjunct Lecturer
- SGS Status
- Associate Member
- Sampling Methodology
We often find ourseves not thinking about the sampling design used for collecting the data we analyze. But the fact is that most of the inference we use is design based inference, which assumes probability sampling design to be generalized to a target population. From clinical trials to big data we are flooded with data collected through non-probability sampling and we need to get better on understanding and modeling this sort of data.
An excellent resource on missing data and sampling is the website of Prof. Roderick J.A. Little.
- Respondent Driven Sampling
Respondent Driven Sampling (RDS) is a methodology for sampling hard-to-reach populations. RDS has enjoyed great popularity as it allows us to reach some small marginalized or at risk populations that that are so important in our society. But it uses a non-probability samplign design and still faces many technical challenges.
Prof. Douglas Heckathorn was who defined the methodology in 1997. Since then there has been many theoretical development, including a software to analyze RDS. This page contains the software and some resources.
- Reproducible Research
The key to good science goes many times through good data analysis. Te concept of reproducible research, based on which not only you but anybody else has the tools to reproduce the analyses and results you published, is so improtant to this proccess of making good science. Reproducible research is a set of innitiatives that have come about very recently. More information can be found here (the Journal Biostatistics calls for reproducibility from its authors) and here. The R-project has an entire task view dedicated to reproducible research. If you want to get a feel for it, a very nice R package is knitr.
- Causal Inference
A difficult yet important question is “How to draw causal conclusions from observational data?”. Experiments are not always feasible and the amount of observational type of data available has increased exponentially with modern technology. It is important to understand techniques and assumptions that make causal claims possible in such data. Judea Pearl has been very prolific on writing about causal inference, specially using Direct Acyclic Graphs. I find the Wikipedia entry on the Potential Outcome framework for causal inference to be a good reference for this most used framework, which includes matching, propensity scores and conterfactuals.
- Latent Variable Models
- Data Visualization
Education & Training History
Degree in Statistics – University of Campinas (2000)
MSc in Statistics – University of Sao Paulo (2005) – Thesis: Formative Indicator in Structural Equaation Models
Primary Teaching Responsibilities
- Practicum Supervisor
Professional Summary & Appointments
2013 – Adjunct Lecturer at DLSPH
2012 to now – Biostatistician at Centre for Addiction and Mental Health (CAMH).
2007 to 2012 – Senior Statistician at Ipsos-Reid Canada.
2004 to 2007 – Senior Statistician at Ipsos Brasil.
2000 to 2004 – Statistician at The Gallup Organization.
- I always like biostatistics and tried to keep myself close to it:
2011 to 2012 – Volunteer work as Biostatistican at CAMH.
2005 to 2012 – Volunteer work as Biostatistician at Federal University of Sao Paulo (UNIFESP)
SANCHES, M and CALLAGHAN, R. (2013) – Impact of Minimum Legal Drinking Age Legislation on Alcohol Related Inpatient Morbidity in Canada, 1997-2007: A Regression-Discontinuity Approach. Addiction, 108.9 (2013); 1590-1600.
SANCHES, M. and WEINER, J. (2012). Being Creative in the Choice Design – Performance of Hierarchical Bayes with Sparse Information Matrix. Paper presentend at the Sawtooth Conference in March, 2012.
SANCHES, M., LARANJEIRA, R. et al (2010) . Alcohol use patterns among Brazilian adults. Rev. Bras. Psiquiatr, 32.3 (2010); 231-241.
SANCHES, M., PINSKY, I. et al (2010) . Exposure of adolescents and young adults to alcohol advertising in Brazil. Journal of Public Affairs 10, 50-58.
SANCHES, M., MOREIRA-ALMEIDA, A. et al, (2010) . Religious Involvement and Socio-Demographic Factors – A Brazilian National Survey. Rev. Psiquiatria Clinica, 37(1):12-5.
SANCHES, M., TAVARES, H. et al, (2010). Gambling in Brazil: lifetime prevalence and socio-demographics correlates. Psychiatry Research, ahead of print.
SANCHES, M., GOMES, L. et al, (2007). Effect of the substrate for pupation in the postfeeding larval dispersal of Chrysomya albiceps. Iheringea. Série Zoologia, 97(3);239-342.
SANCHES, M., GOMES, L. et al, (2007). Behavior of the combined radial post-feeding larval dispersal of the Blowflies Chrysomia megacephala and Chrysomya albiceps. Brazilian Archives of Biology and Technology, 50(2);279-288.
SANCHES, M., GOMES, L. et al, (2007). Visual and olfactory factors interaction in resource location by the blowfly, Chrysomya megacephala in natural condition. Neotropical Entomology, 36(5);633-639.
SANCHES, M., GOMES, L. et al, (2006). Influence of photoperiod on body weight and depth of burrowing in larvae of Chrysomya megacephala and implications for forensic entomology. Revista Brasileira de Entomologia, 50(1);76-79.
SANCHES, M., CALLIGARIS, I. et al, (2005). Morphometric Analysis of a population of Diplopods of the Genus Rhinocricus Karsch, 1881. Folia Biologica (Praha), 51;40-46.