News! We are happy to announce our two keynote speakers, Johan Ugander and Ingmar Weber!
The OSSM workshop will focus on observational studies (namely, causal analyses) of data traces created by people, whether directly or indirectly, through their interactions with social networks, messaging services, and other applications and devices.
Human generated content in general, including social media, has been shown to be a rich repository of data for observational studies across many areas: public health, with research on prevalence of disease and on the effects of media on the development of disease; medicine, showing the ability to detect mental disease in individual using social media; education, to optimize teaching and exams; and sociology, to prove theories previously tested on very small populations. These studies have been conducted from data including social media, search engine logs, location traces, and other forms of human generated content.
While many past studies showed a correlation between variables of interest, some studies were able to show causal relationships through natural experiments and careful analyses. Natural experiments are empirical studies in which people are assigned to control or experimental conditions according to factors not under the control of investigators, where these factors resemble a random assignment. With Internet data, natural experiments occur frequently through policy changes, outside influences, and A/B testing. Our workshop will focus on all aspects of causal inference from human generated content, with studies that developed novel methods of identifying and using natural experiments or other methods for inferring causality.
Studies based on human generated content require investigators to preserve the privacy of users and abide by proper ethical codes. We will encourage discussion of how these goals may be attained in secondary analysis of data such as that used in the aforementioned studies.
We are soliciting participation both from domain-experts interested in specific applications of methods, as well as methods-experts.
Topics of interest broadly include:
- Applications and domain-specific explorations
- Interpreting user-generated data, including text, structured data, and temporal data.
- Causal analyses in social media, for example, through conditioning analyses, instrumental variables or regression discontinuity analysis.
- Identifying natural experiments and using them to understand causal inferences
- Identifying and/or mitigating population, behavioral and other systematic biases in social media
- Novel methods for preserving privacy in such experimentation
- Ethical implications of such experiments
- Qualitative, experimental and other evaluation and validation methods for observational study results
General Format
The core of the workshop will consist of structured discussion and breakouts on topics preselected by participants. Broad discussion topics will be chosen based on initial submissions, and refined during a pre-workshop video chat open to all workshop participants several weeks prior to the workshop. Potential outcomes of these breakouts will include a position paper, or potential research directions/ideas that could lead to interdisciplinary collaborations, particularly between computer and social scientists. We will support breakout groups with a facilitator as well as a scribe.