Won NIJ competition on surveys

The submission Gio and myself put together, Using Every Door Direct Mail Web Push Surveys and Multi-level modelling with Post Stratification to estimate Perceptions of Police at Small Geographies, has won the NIJ Innovations in Measuring Community Attitudes Survey challenge.

Specifically we took 1st in the non-probability section of the competition. The paper has the details, but using every door direct mail + post-stratifying the estimates is the approach we advocate. If you are a city or research group interested in implementing this and need help, feel free to get in touch.

Of course if you want to do this yourself go for it (part of the reason it won was because the method should be doable for many agencies in house), but letting me and Gio know we were the inspiration is appreciated!

Second, for recruiting for criminology PhDs, CRIME De-Coder has teamed up with the open access CrimRXiv consortium

This example shows professor adverts, but I think the best value add for this is for more advanced local govt positions. Anymore many of those civil service gigs are very competitive with lagging professor salaries.

For people hiring advanced roles, there are two opportunities. One is advertising – so for about the same amount as advertising on LinkedIn, you can publish a job advert. This is much more targeted than LinkedIn, so if you want PhD talent this is a good deal to get your job posting on the right eyeballs.

The second service is recruiting for a position. This is commission based – if I place a candidate for the role then you pay the recruiter (me and CrimRXiv) a commission. For that I personally reach out to my network of people with PhDs interested in positions, and do the first round of vetting for your role.

Third, over on Crime De-Coder I have another round of the newsletter up, advice this round is that many smaller cities have good up and coming tech markets, plus advice about making fonts larger in python/R plots. (Note in response to that post, Greg Ridgeway says it is better to save as vector graphics as oppossed to high res PNG. Vector is slightly more work to check everything is kosher in the final produced plot, but that is good advice from Greg. I am lazy with the PNG advice.)

No more newsletters this year, but let me know if you want to sign up and I will add you to the list.

Last little tidbit, in the past I have used the pdftk tool to combine multiple PDFs together. This is useful when using other tools to create documents, so you have multiple outputs in the end (like a cover page or tech appendix), and want to combine those all together into a single PDF to share. But one thing I noticed recently, if your PDF has internal table of content (TOC) links (as is the case for LaTeX, or in my case a document built using Quarto), using pdftk will make the TOC links go away. You can however use ghostscript instead, and the links still work as normal.

On my windows machine, it looks something like:

gswin64 -q -sDEVICE=pdfwrite -o MergedDoc.pdf CoverPage.pdf Main.pdf Appendix.pdf

So a few differences that if you just google. Installing the 64 bit version on my windows machine, the executable is gswin64, not gs from the command line. Second, I needed to manually add C:\Program Files\gs\gs10.02.1\bin to my PATH for this to work at the command prompt the way you would expect, installing did not do that directly.

Quarto is awesome by the way – definitely suggest people go check that out.