Anonymity and Sensitivity

41
Anonymity and Sensitivity Fritz Scheuren, Speaker NORC University of Chicago February 2013

description

Anonymity and Sensitivity. Fritz Scheuren, Speaker NORC University of Chicago February 2013. Scope of USA Migration Studies. Emigration from USA to World Immigration from World to USA Direct Indirect. Outline of Talk. Problem Science of Approach Engineering of Its Implementation - PowerPoint PPT Presentation

Transcript of Anonymity and Sensitivity

Page 1: Anonymity and Sensitivity

Anonymity and Sensitivity

Fritz Scheuren, SpeakerNORC University of Chicago

February 2013

Page 2: Anonymity and Sensitivity
Page 3: Anonymity and Sensitivity

Scope of USA Migration Studies

• Emigration from USA to World• Immigration from World to USA

Direct Indirect

Page 4: Anonymity and Sensitivity

Outline of Talk

• Problem• Science of Approach• Engineering of Its Implementation• Results So Far!• Next Steps?

Page 5: Anonymity and Sensitivity

Immigrants Sans Paper

• Physical Border Protection Worldwide?•Mostly “Thought” to be South to North?• Only An Indirect Estimate Available

Page 6: Anonymity and Sensitivity

Legal Immigrants First?

• Over a Million Come Annually, still Largest Legal Flow Of Any Country• About 150 Million Made citizens since 1940

Page 7: Anonymity and Sensitivity

Undocumented May Be Much Smaller?

• Flow is believed to average Less than Legal • Stock Estimated, perhaps, at 12 million• Flow is Business Cycle Driven, but usually less

Page 8: Anonymity and Sensitivity

Physical Border Protection

• Extremely long Borders with Canada and Mexico• Mexico Border Not Well Guarded • Drug Trafficking a Major Problem for USA/Mexico

Page 9: Anonymity and Sensitivity

Mainly South to North?

• Mexican Border Crossers but This Model is too Simple• Student and Visitor Visa Overstays • Still Often Tied to US Jobs

Page 10: Anonymity and Sensitivity

Who Comes Legally Now?

• Traditional Americas and Western Europeans but Less legally than earlier• Eastern Europeans Still in Large Numbers• Koreans and Chinese up, Africa coming up

Page 11: Anonymity and Sensitivity

Science Approach

• Two Card Method (Handout)• Two Independent Replicates• Difference Estimate of Number of Undocumented

Page 12: Anonymity and Sensitivity

Two Card Example(Green Card)

Page 13: Anonymity and Sensitivity

First Card

• Two options Given --• First option: US Citizenship or Undocumented• Second Option: Green Card or Other

Page 14: Anonymity and Sensitivity

Two Card Example(Other than Green)

Page 15: Anonymity and Sensitivity
Page 16: Anonymity and Sensitivity

Second Card

• Two Options Given –• First Option: US Citizenship• Second Option: Green Card or Other• Both Cards: Safe Answer

Page 17: Anonymity and Sensitivity
Page 18: Anonymity and Sensitivity

Difference Undocumented

• Estimate Unbiased, if Question Understood and Answered Honestly• But Variance Twice as Large as Direct Question• However Bias Eliminated

Page 19: Anonymity and Sensitivity

Engineering Solution

• Human Factors Design Study• To test Factor Importance• Protecting Anonymity to Gain Unbiased Answers

Page 20: Anonymity and Sensitivity

A Credible Balancing Bar?

Page 21: Anonymity and Sensitivity

Balance of Trust?• Key is Creating High Anonymity to Overcome PERCEIVED level of risk.• Given the Sensitivity of the Subject• Need Data so we tested.

Page 22: Anonymity and Sensitivity

Sensitivity Loss

• The Two Card method has a Variance Cost and is a challenge in moving Univariate analysis to a Multivariate form• The Newness remains for Anonymity form

Page 23: Anonymity and Sensitivity

Anonymity Loss

• Credibility of Promise is a real concern?• So Move to a Promise that CONSTRUCTS Established Trust? • Key Improvement!

Page 24: Anonymity and Sensitivity

Results So Far• About 350 tests of the Two-Card method and its variations have been undertaken.• The tests show that the results are robust to minor wording changes

Page 25: Anonymity and Sensitivity

More Results So Far

• The Results are not robust to being wordy. • Potential respondent are not paid, probably a mistake. • If version was lengthy, more break-offs likely.

Page 26: Anonymity and Sensitivity

Still More Highlights

• Location changes, also are not robust. • Intercepts near the Mexican Border (near San Diego) were seen as more risky, so stopped.

Page 27: Anonymity and Sensitivity

Other Changes?

• If scaled up, more characteristics and more intercept locations are definitely needed.• Respondent based area probabilities needed to adjust for self selection.

Page 28: Anonymity and Sensitivity

Interviewer Training

• The extensive paradata calculated needs to be turned into a model for training.• Intercept Interviews are changing NORC Team. • How Measure/Control?

Page 29: Anonymity and Sensitivity

Intercept Scalability

• How replicable are the Interceptors we chose?• How replicable and scalable is this process?•What controls are there and how scientific are we?

Page 30: Anonymity and Sensitivity
Page 31: Anonymity and Sensitivity

Selected Respondent Comments

• [Interview] was easy, because none of the questions are hard and explained it very well.• [Glad you] understand the immigrants and their education level.

Page 32: Anonymity and Sensitivity

More Respondent Comments

• He was hesitant at first and I asked him why. • He said because people [might identify him].I asked him how I was going to find out. I told him [many] are similar].

Page 33: Anonymity and Sensitivity

Next Steps Coming?

•Write up Results of this small NSF grant?• Expand to another Undocumented group: Asians, Russian Jews?• Take Sting off Hispanics

Page 34: Anonymity and Sensitivity

Portable to Europe?Your Question!

• In a register based society is this needed?• Some evidence says Yes but little leakage?• More references needed? Short list follows?

Page 35: Anonymity and Sensitivity

Many, Many Thanks!

• Contact Details Fritz Scheuren +1-202-320-3446 [email protected]

Page 36: Anonymity and Sensitivity

References• Cochran, William G. Sampling Techniques.

New York: Wiley, 1977. Print.• Dyson, Freeman J. "Is Science Mostly

Driven by Ideas or by Tools?" Science 338.6113 (2012): 1426-427. Print. • Galison, Peter. Image and Logic: A Material

Culture of Microphysics. Chicago: University of Chicago, 1997.

Page 37: Anonymity and Sensitivity

References

• GAO. A "Grouped Answers" Approach to Surveying Foreign-Born Respondents. Rep. no. GAO-06-775. N.p.: GAO, 2006. Print.• Kuhn, Thomas S. The Structure of

Scientific Revolution. Chicago: University of Chicago, 1996. Print.

Page 38: Anonymity and Sensitivity

References• Lehoux, Daryn, and Jay Foster. "History

and Philosophy of Science: A Revolution of Its Own." Science 338.6109 (2012): 885-86. Print.

• Pasic, Peter.” What the Romans Really Knew,” Science 338:273-274.

• Reynolds, Gregg. Immigrant Experience Survey Interim Report. , 2013.

Page 39: Anonymity and Sensitivity
Page 40: Anonymity and Sensitivity
Page 41: Anonymity and Sensitivity