Professor, Demographer, Social Explorer Developer.
New York City
Joined April 2009
- Tweets235
- Following336
- Followers315
- Likes84
Steven Ruggles was just name a MacArthur Fellow for the Class of 2022. Below is a Brief Description and a Link to More Details.
macfound.org/fellows/class-o…
3
Andrew A Beveridge retweeted
A few years ago, the US Census Bureau adopted a controversial method for maintaining the privacy of personal data, Differential Privacy. Now it turns out that the rationale for doing so was unfounded, amazing story, tinyurl.com/mr4b7uas
He just put up a full blown analysis:
Kansas Result Suggests 4 Out of 5 States Would Back Abortion Rights in Similar Vote
The referendum in a reliably Republican state shows which party has the most energy on the issue.
Full article with a Map here :nytimes.com/2022/08/04/upsho…
Of course, there are other differences between theses states, but abortion's overperformance in Kansas survives controls for race, education, Evangelicals, etc.
The Census Has Gone Out of Business Until 2023--
The data on race and Hispanic status by sex, and age along with much other data will be delayed until 2023 or in some cases later, and what is released will be drastically reduced. nytimes.com/2022/04/29/us/20… 1/2
2/2 The main excuse, the new privacy system based upon so-called DP. But it was unjustified and does not protect privacy. See link.springer.com/article/10… and science.org/doi/10.1126/scia… and slate.com/technology/2022/03… and arxiv.org/pdf/2202.04872.pdf
They haven't backed down. Instead, no data.
Andrew A Beveridge retweeted
So thrilled yesterday to have an Authors-Meets-Critics session on my new book, w/ great comments from @ProfCandis, Rachel Ann May, & Gloria Cox @SPSAnews! Co-authored w/ Amy Lutz & Lakshmi Jayaram. tcpress.com/parenting-in-pri…. Free ship. w/ code TCP2022 to check it out! @TCPress
ALT Pamela R Bennett holding her book titled Parenting in Privilege or Peril: How Social Inequality Enables or Derails the American Dream.
Looks like Yale’s Grand strategy is to do anything to get 6 billion dollars. Yale BA 67; MPhil 71; PhD 73.
NEW: Yale president Peter Salovey issues letter citing Yale's "unwavering commitment to academic freedom."
Says he "should have tried harder to improve the situation" w donors that led Beverly Gage, head of the Grand Strategy seminar, to resign... (1/3)
president.yale.edu/president…
1
It is not just Carrolton MS. Virginia’s new census data is distorted at local levels, analysts say: ‘It can’t be accepted as fact’ dailypress.com/government/vi…
Census Data less noisy, but still some issues, privacy not so much: From Krish Muralidhar
to figure out the approximate disclosure risk is as follows:
But a quick way of trying to figure out the approximate disclosure risk is as follows:
Epsilon = 17.14
Exp(Epsilon) = 27784809
Frank McSherry, one of the inventors of differential privacy "Apple has put some kind of handcuffs on in how they interact with your data," he says. "It just turns out those handcuffs are made out of tissue paper."
wired.com/story/apple-differ…
And given that any increase in Epsilon is exponential, the Census' privacy level is 23 times worse than even Apple's epsilon level.
Best I can do for now. Hope it helps.
Krish Muralidhar @histdem
More: In summary, if the article by Garfinkel et al.proves anything, it proves that even for a very small data set, even when a lot of information is released, if simple disclosure prevention techniques are properly applied, reconstructing the data set is practically impossible.
But I certainly do mean to imply that differential privacy is not the only solution.
The Abowd paper says (above the title on the first page): “These attacks on statistical databases are no longer a theoretical danger.” Freudian slip?
1
2
New analysis by Krish Muralidhar, shows that Abowd Garfinkel paper upon the new DAS is based has fundamental errors. dropbox.com/s/7plwq3mxe6gt5o…
