Our next Seminars
Discover the Datasios seminars.
Don’t cry over spilt soup, mop it up: methodological approaches to correct for sampling errors after data collection
Almost every cook tastes their soup before it has finished cooking to know whether to add more seasoning. Of course, the natural assumption during this ritual is that the small scoop of soup tasted after stirring faithfully represents whatever is in the larger pot! A similar process happens in survey research: when we take a sample, we are (typically) interested in getting a flavor of the larger population from that relatively small sample. Unfortunately, sometimes, things do not work well, and we may end up with a sample that is non-representative of the larger target population. This session will provide hands-on approach to correcting for such issues (to the extent possible).
Learning Objectives:
- Learn the difference between sampling and non-sampling sources of error in survey design.
- Appreciate the difference between coverage and non-response bias and how to address these issues before the survey
- Get an introduction to computing non-response weights after data collection.
Session length: 1.5 hours
Cost: Free ($0)
Date: TBD
An introduction to analyses of weighted survey data
Large national surveys like the National Health and Nutrition Examination Survey (https://www.cdc.gov/nchs/nhanes/index.htm) and the Behavioral Risk Factor Surveillance System (https://www.cdc.gov/brfss/index.html) have historically provided reliable sources for tracking national or state objectives and exploring topics of public health interest. Several new surveillance systems have sprung up since the beginning of COVID-19 and provide opportunities for exciting new research on disparities and the social aspects of the pandemic. This workshop provides an applied introduction to analyses of complex survey data.
Learning Objectives:
- Learn the different sources of freely available, large surveys available in the US and globally, including those specific to COVID-19.
- Get a clear understanding of why weighting is necessary and the implications of conducting unweighted analyses with complex survey data
- Learn what the “weight”, “psu”, and “strata” variables available in most complex surveys do
- Learn how to implement weighted analyses in R and Stata (along with a step-by-step manual).
Session length: 1.5 hours
Cost: Free ($0)
Date: TBD
Probability and non-probability sampling approaches in survey design
So, you have decided you need to conduct a survey to shed more light on those key issues you have in mind. You have a clear idea of who your target population is, and have even crafted your survey questions. What next? Well, now you have to decide how to select the participants in your survey, also known as sampling. This workshop will provide a hands-on approach to this important process.
Learning Objectives:
- To sample or not to sample: a logical framework.
- Appreciate the difference between probability vs non-probability sampling and which situations call for which sampling approach.
- Learn how to implement simple random, stratified, and cluster sampling techniques.
Session length: 2 hours
Cost: Free ($0)
Date: TBD
Sample size calculations for simple and complex surveys
The rapidly evolving public health landscape has created new data needs that may necessitate collection of primary data. A deep knowledge of sample size calculation is an invaluable skill in the toolkit of any public health practitioner. This workshop provides hands-on training in sample size calculation for cross-sectional surveys which remain the staple data source in public health. Participants will be provided access to free tools for sample size calculations.
Learning Objectives:
- Distinguish between validity and precision and appreciate how the sampling approach influences validity while sample size influences precision.
- Learn how to compute sample sizes for simple random and complex surveys.
- Get introduced to various open access (free) tools for sample size calculation
Session length: 1 hour
Cost: Free ($0)
Date: TBD
Referencing without tears: How to manage references in the New PubMed with Mendeley (A free referencing tool)
“Manual” referencing could perhaps be easy when there are only a handful but becomes highly inefficient with a much larger number of references (e.g., in a full-length manuscript). This is especially worsened when the reference style must be changed completely (e.g., reformatting for another manuscript after rejection). Yet, some researchers still manage references manually, perhaps because of not being familiar with using a reference manager or not being able to afford one. In this session, you will learn how to use an open source (free) reference manager alongside the new PubMed.
Learning Objectives:
- Learn how to install Mendeley as a desktop application and as a Microsoft plugin.
- Learn how to export references from PubMed in the appropriate format.
- See a demonstration of how to use Mendeley for referencing and how to change from one referencing style to another automatically.
Session length: 1 hour
Cost: Free ($0)
Date: TBD
Developing high-resolution figures and charts for scientific publication: Best practices and open source resources
A picture is worth a thousand words, but only if it is legible. Figures are an amazingly effective means of communication and biomedical journals are increasingly providing specific requirements on format and resolution. Researchers must know best practices for creation/presentation of figures and be able to access freely available tools to achieve this.
Learning Objectives:
- Learn best practices for presenting figures, charts, and images in scientific documents.
- Get an introduction to key open access tools for creating, transforming, and enhancing figures, charts, and images to journal specifications.
Session length: 1 hour
Cost: Free ($0)
Date: TBD