Harvard University The Politics of Personal Data Gov 1430

Home | Schedule | Syllabus

Back to Lab

Lab 3 Supplement: Learning Personal Information From Fragments of Data

Goal.

In this activity you will get some experience with personal online databases and linking and working with disparate pieces of data.

Suggested Readings and References

• theDataMap.org
• Re-identification of Washington State Health Data
• 50-state Survey of Statewide Data
• Uniqueness of Personal Demographics
• Uniqueness of Personal Demographics at Harvard

Experiment.

Professor Alice was reading a news article about four children who had cancer. She remembered hearing about our class and the public availability of health data that has medical details, but does not have the names or the patients. She also knows that medical terminology is not transparent. So, Professor Alice thinks that no further medical information about Brandon Steele, one of the children in the new story, can be learned than that which appears in the news story. Here is the article Professor Alice read.

News article

In the article, Professor Alice learned that "Brandon Steele of Taylorville, Illinois was diagnosed with neuroblastoma in August 1991 and later died." Professor Alice has heard about our class and knows there are other kinds of data available on individuals, but she does not believe any further information can be found about Brandon.

Perform an experiment that proves or disproves Professor Alice's belief. Your experiment can prove Professor Alice's belief by showing how none of the data can reliably relate to Brandon Steele. Or, your experiment can disprove her belief by conducting an experiment that compiles relevant data fragments using the data sources below, and reports comprehensive details specifically about Brandon Steele's medical history.

In the sample data below, all samples include all entries relevant to the reported demographics. No other records exist that have the same demographics.

• Sample medical profiles: by DOB, ZIP, Sex and by by ZIP, DOB, Sex. These medical profiles are a sample drawn from semi-publicly available hospital discharge data.

• Death index. Here is a sample drawn from our copy of the Social Security Death index: sample.

• ZIP codes

• Sample of ICD-9 diagnosis codes. There are more than 15,000 International Classification of Diseases version 9 (ICD-9) codes. These are used to record the diagnosis for a medical visit.

• Sample of hospital identity codes.

Write a summary of your experiment as a 3-5 page scientific paper. The basic parts of the scientific paper you will write for this assignment will have five parts: an Abstract, Introduction, Methods, Results and Discussion. The Abstract section should be a one paragraph summary stating the goal of predicting Social Security numbers. The Introduction describes why your experiment is important. You should not assume the person reading the paper knows anything about this assignment. Include references and use authoritative sources to make your points. You should not make any sweeping or unsubstantiated statements in your writing. The Methods section is where you describe the actual experiment and your findings appear in the Results section. Use the Discussion section to explain the significance of this experiment, its shortcomings and broader implications. Include at least one statement about the limits of your approach. See the write-up on the Gov1430 class-wide experiment on the course wiki for an example of writing in this format.

Submit

Upload your work to the Course Wiki.

Presentation

Present your experiment and findings in class.

Copyright © 2013-2015 President and Fellows Harvard University | Data Privacy Lab