A faculty search is one of the most important processes overseen by a department chair. I've been involved with faculty searches, either as search committee member or chair, or as division head or department head, for 20 years. Over these years in my department the attention to affirmative action has grown. This absolutely does not mean less-qualified candidates are interviewed or hired. It means that we work to assemble the largest and most diverse pool of qualified applicants that we can, and we strive to identify the most promising candidates regardless of race, gender, or other qualities unrelated to the job description. We do so with explicit awareness of factors that discriminate against underrepresented groups -- in particular, our own implicit biases. Here are some of the things we have done, and my assessment of their utility.
For several years, search committee chairs have met with an Assistant Dean to review university affirmative action procedures and to be alerted to implicit bias and best practices for faculty recruitment based on the Michigan STRIDE materials. Last year I required all search committee members to attend a similar session which I led. Some faculty bristled but I let them know it was required. I will repeat this for new committee members as I am optimistic about its educational value over the long term.
For several years, I have required search committees to assemble lists of promising women and underrepresented minority postdocs to be invited for visits (which the department pays for), and who should be encouraged to apply to faculty searches. This process has led to some success, although the numbers to date are few. I strongly believe this proactive search process is important to building a strong faculty. In the words of Shirley Malcom of the AAAS, it converts a "sort committee" into a "search committee".
Before short lists are finalized, they are reviewed first by me as department head, then by a committee assembled by the Assistant Dean. Each search committee is required to justify their choices, especially why candidates are not on the short list who would increase the diversity of the pool. Occasionally I call for changes in a short list. The committees know that I have an eye on inclusion and they respect that spirit. I feel that this oversight is important in fields with serious underrepresentation problems (and not just for women). I would recommend this practice to others; it sends a clear signal to the faculty and that, in turn, helps change the climate.
This past year I ran a separate "open search" with no preference for subfield, and with explicit encouragement of applications in any area of physics or astronomy. I had hoped it might encourage talented individuals who feared that other searches in the department were so narrowly focused as to exclude them. In this it succeeded, although it did not increase the gender diversity of our total applicant pool. Most of those who applied were from areas of physics that were not explicitly mentioned in the other searches (e.g., string theory and optics). In astrophysics, our search was broad as it has been for many years. All our searches are open in the sense that we never have determined in advance whom we want to hire. In the future, it may be more efficient not to run a separate wide-open search but instead for the overall departmental ad to mention that candidates in fields not explicitly listed are invited to apply to the most closely related search and to contact search chairs if ever in doubt.