We've heard over and over that the Obama Administration has been the most diverse administration ever including for LGBT people. We're told that they have appointed more than 260 openly LGBT people to positions within the administration; however, the continuing failure of the President to issue an executive order barring the federal government and contractors with more than $10,000 in federal contracts from discriminating on the basis of sexual orientation and gender identity, an executive order that then Senator Obama promised during his Presidential campaign to the Houston LGBT Caucus in 2008, despite loud and persistent calls from LGBT leaders to do so and multiple letters from Democratic members of Congress now signed on by almost 200 Representatives and Senators, opens the President's own record of LGBT diversity in hiring to examination under the microscope.
When completing his second term cabinet appointments last year, a spokesperson with a major LGBT rights organization, the Human Rights Campaign (HRC), said it was "quite disappointing" that Obama had not named an openly LGBT person to the Cabinet:
While the Cabinet is full of staunch allies, there is no reason why qualified LGBT Americans willing to serve their country should be overlooked, especially in a day and age when LGBT people are an integral part of the fabric of our nation as everyone from doctors to teachers to professional basketball stars,” Cole-Schwartz said. “The president has said it’s our job to remind him when he’s fallen short and while there’s much for which to applaud him, on this issue this president has fallen far short.
Just months earlier, after the election, LGBT leaders, including openly gay Congressman-Elect Sean Patrick Maloney of New York, were lobbying for an openly LGBT cabinet appointee, particularly for Fred Hochberg of the Export-Import Bank to be named Commerce Secretary. This of course did not come to pass.
The administration can cite 260 openly LGBT appointees, but the question is how statistically significant is that number and is this claim of diversity diffuse throughout the administration or is it bottom heavy. I mean would we take seriously a law firm with 200 employees claiming to be diverse if all their "diversity" was in the mailroom, interns and clerical staff and none in the partners and associates?
Enter the statistical analysis. This might be a little math heavy for some. I've tried to remove the nitty-gritty computational stuff. To start with, I'll declare my assumptions. Several recent analyses of the U.S. population have given a figure of about 3.5% of the adult population of the United States being openly LGBT. To give President Obama the benefit of the doubt, I'll use Gallup's lower figure of 3.4% based on the thousands of interviews they conducted for their tracking polls throughout the 2012 elections. Now the pool of people that President Obama would be appointing from, however, is far more likely to be LGBT since LGBT people are far more likely to be Democrats as Gallup's analysis showed (13% R, 43% Ind and 44% D. When leaners were pushed, the numbers were 21% R, 13% I and 65% D). Assuming that Obama appoints approximately 5% Republicans, 25% Independents and 70% Democrats to the highest offices in his administration, this would yield a proportion range of 4.25% to 5.1% LGBT people for the pool of people that Obama is likely to pull appointments from (about 25 to 50% higher than the general population). For the analysis, I will test the hypothesis for all three values, 3.4%, 4.25% and 5.1%. These will be our proportions, p, in the statistical analysis.
Now that we have the proportion of the population that is LGBT "defined," let's define the types of appointments we're going to look at. First, and obviously, there are his Cabinet appointments. Next there are those positions that are designated Cabinet-level appointments, which for the Obama administration are the White House Chief of Staff, the Administrator of the EPA, the Director of the OMB, the U.S. Trade Representative, the Ambassador to the United Nations, the Chair of the Council of Economic Advisors and the Administrator of the Small Business Administration. For these purposes, I'm leaving out the Vice President since the President doesn't in the truest sense "appoint" someone Vice-President unless he's replacing the Veep under the 25th Amendment. Next are the highest level advisors and staff members to the President in the Executive Office of the President, that is those that carry the designation of Assistant to the President, Deputy Assistant to the President and Special Assistant to the President. Promotions of people from one title to another title or the significant addition or change of duties that result in a change in title are considered separate appointments. (Example: Emmett Beliveau has served the administration in three capacities, all as a Deputy Assistant to the President: Director of Advance and Operations, Director of the White House Military Office (or WHAMO "as we have apparently taken - over my strong objection - to calling it", see The West Wing, episode 2-07: "The Portland Trip") and Director of the Office of the Chief of Staff. Another is Jason Furman who has had four positions in the administration being a Deputy Assistant to the President for Economic Policy, Assistant to the President and Principal Deputy Director of the National Economic Council, Assistant to the President and Director of the National Economic Council, and finally Chair of the Council of Economic Advisors.) While some of the Vice President's staff carry one of the three high level staff designations (Assistant, Deputy Assistant or Special Assistant to the President), they are not being included. Appointments that span the two terms of President Obama are only being counted as one appointment rather than two.
Excluding the nomination of the new HHS Secretary, who has yet to be confirmed, excluding acting secretaries (since they aren't confirmed cabinet secretaries) and excluding appointments spanning the two terms (e.g. Eric Holder only counts once even though he's technically serving a second term like the President, i.e. the President is not being penalized for retaining cabinet secretaries just as he's not being penalized for retaining high ranking staff members) the President has had 26 confirmed cabinet appointees.
Excluding the Vice President and acting/interim officials, his other Cabinet-level officials total 19. Counting the top level EOP personnel not previously counted that carry the title of Assistant to the President (the Chief of Staff is an Assistant to the President, but being counted with the Cabinet-level officials) there have been not less than 67 people. The next level down, Deputy Assistant to the President, there have been not less than 75. And in the third tier, Special Assistant to the President, there have been not less than 211. I say not less than for those three because I have not found a comprehensive list of administration officials by title. The annual salary listings of White House staff doesn't include all persons with such titles since they are not always accredited to the White House. I have attempted to include as many as I could find. For transparency, I have uploaded an excel spreadsheet to Scribd of the Cabinet, Cabinet-level and EOP officials I counted for use in this analysis. I'm willing to re-run the numbers if anyone finds anyone else on the list that is openly LGBT or has more officials that meet the qualifications. (And before someone tries to "correct me" on leaving Brian Bond and Shin Inouye off the list since they have served as a public LGBT face for the administration, please note that neither is/was (Bond has left the administration) so much as a Special Assistant to the President. Both are lower ranking as evidenced by salary. Bond made about $92,000 a year and Inouye $78,000 a year whereas the lowest Special Assistant salary is $100,000 and the median is $130,000.) The Scribd spreadsheet
Obama Administration Officials
Next, we need to define our threshold...in statistical parlance our significance level or α. Depending on the application, these usually range from 0.0001 to 0.1 or 99.99% to 90% confidence. The most commonly used value is, however, 0.05 or 95% confidence level. Political and other polling generally uses this standard, so I see no reason not to accept its use here.
Thus, our hypotheses for this analysis will be
the null hypothesis, H0: p ≥ 0.034, p ≥ 0.0425 & p ≥ 0.051
the alternate hypothesis, Ha: p < 0.034, p < 0.0425 & p < 0.051
I will be modeling the analysis using the binomial distribution.
Some abbreviations used in the forthcoming tables:
p |
the proportion/probability of success. We are testing for p values of 0.034, 0.0425 and 0.051 |
n |
the number of "independent trials" in an experiment. For this analysis, this is the number of appointments |
x |
the number of "successes" in an experiment. For this analysis, this is the number of LGBT appointments |
μ |
the mean and is the value of n times p |
E[x] |
the "expected value" of x. This will be the rounded whole number found by multiplying n times p since you cannot have a fractional person |
F(x) |
the cumulative distribution function value of the binomial distribution for the value for x. For example, if x=3, then F(3) represents the probability that x=0 plus x=1 plus x=2 plus x=3, i.e. the probability it equals all whole number values up to and including the value of x. This will be the important number to watch. If it is above 0.05, we "fail to reject" the null hypothesis, i.e. there is insufficient evidence to conclude that the number of LGBT appointments is statistically smaller than their proportion in the population. If it falls below 0.05, we reject the null hypothesis in favor of the alternate hypothesis, i.e. the number of LGBT appointments is statistically smaller than their proportion in the population to a confidence level of 95%. For more information on the cumulative distribution function of the binomial distribution and how my numbers were calculated using the other numbers in my table, please see the wikipedia article |
First let's look at each group of appointments separately.
For p=0.034 |
Category |
n |
x |
E[x] |
μ |
F(x) |
Cabinet |
26 |
0 |
1 |
0.884 |
0.407 |
Cabinet-level |
19 |
0 |
1 |
0.646 |
0.518 |
Assistant to the President |
67 |
1* |
2 |
2.278 |
0.331 |
Deputy Assistant to the President |
75 |
0 |
3 |
2.550 |
0.075 |
Special Assistant to the President |
211 |
7^ |
7 |
7.174 |
0.572 |
* Bradley J. Kiley, Assistant to the President for Management and Administration
^ Jeremy M. Bernard, Special Assistant to the President and White House Social Secretary; Michael J. Gottlieb, Special Assistant to the President and Associate Counsel to the President; Kathleen R. Harnett, Special Assistant to the President and Associate Counsel to the President; Alison J. Nathan, Special Assistant to the President and Associate Counsel to the President; David L. Noble, Special Assistant to the President for Presidential Personnel; Ellie S. Schafer, Special Assistant to the President and Director of Visitors Office; Kamela M. Vasagam, Special Assistant to the President for Presidential Personnel
In each individual category, none falls below the 0.05 mark. While one would have by now expected at least once cabinet secretary and one cabinet-level appointment, the probabilities of having selected 0 are still quite high at 40.7% and 51.8% respectively. How does looking at the categories together look?
For p=0.034 |
Category |
n |
x |
E[x] |
μ |
F(x) |
Cabinet + Cab-level |
45 |
0 |
2 |
1.530 |
0.323 |
Cabinet + Cab-level + Asst |
112 |
1 |
4 |
3.808 |
0.103 |
Cabinet + Cab-level + Asst + Dep Asst |
187 |
1 |
6 |
6.358 |
0.012 |
Cabinet + Cab-level + Asst + Dep Asst + Sp Asst |
398 |
8 |
14 |
13.532 |
0.074 |
Taken as a group, we see a failure for the first time when considering the set of Cabinet, Cabinet-level, Assistants to the President and Deputy Assistants to the President taken as a group. Whereas one would expect 6 in a group of 187, there has been only one. There is a 1.2% chance in a group of 187 that only one or zero would be LGBT. The fact that there are 7 in the Deputy Assistants to the President category brings the probability back above 0.05 when all groups are combined. But this us using p=3.4%, which assumes that the proportion of gays in the pool of likely appointees is the same as their proportion of the general population. Gays are much more likely to be Democrats (or independents) and thus are more likely to be more than 3.4% of the pool of likely appointees. We estimated that the actual proportion in the population President Obama is likely to draw from to be between 4.25% and 5.1%. So here are the numbers for 4.25%:
For p=0.0425 |
Category |
n |
x |
E[x] |
μ |
F(x) |
Cabinet |
26 |
0 |
1 |
1.105 |
0.323 |
Cabinet-level |
19 |
0 |
1 |
0.808 |
0.438 |
Assistant to the President |
67 |
1 |
3 |
2.848 |
0.217 |
Deputy Assistant to the President |
75 |
0 |
3 |
3.188 |
0.038 |
Special Assistant to the President |
211 |
7 |
9 |
8.968 |
0.323 |
Cabinet + Cab-level |
45 |
0 |
2 |
1.913 |
0.142 |
Cabinet + Cab-level + Asst |
112 |
1 |
5 |
4.760 |
0.046 |
Cabinet + Cab-level + Asst + Dep Asst |
187 |
1 |
8 |
7.948 |
0.003 |
Cabinet + Cab-level + Asst + Dep Asst + Sp Asst |
398 |
8 |
17 |
16.915 |
0.012 |
Here we see the failures increase. At p=0.0425, we'd still expect one Cabinet secretary and one Cabinet-level official, but the probabilities of having zero have dropped to 32.3% and 43.8%, both still respectably high. With those two taken together, the probability of having zero is starting to be marginal at 14.2%. In other words, there is only a one in seven chance that there would be no openly LGBT persons in either of those categories. With 75 Deputy Assistants to the President, one would have expected 3 to be openly LGBT and the probability of having 0 is just 3.8%, less than our 5% significance level. When considering any groupings of the Cabinet and Cabinet-level officials with EOP staff, the President's hiring practices have fallen short with only a 4.6% chance of having 1 or 0 openly LGBT persons in the Cabinet, Cabinet-level and the Assistants to the President. One would expect to see 5. The probability of having 8 is higher than the probability of having 1 or 0 (5.5% vs 4.6%). Matters worsen significantly when adding the Deputy Assistants to the President into the mix with there being a 0.3% probability of having 1 or 0 open LGBT people out of those 187 appointments. Again, the numbers improve slightly due to a fair number of Special Assistants to the President, but still fall well below the 5% threshold with a probability of 1.2% of having 8 or fewer openly LGBT people out of 398 appointments. The probability of have 26 openly LGBT appointees is more than the probability of having only 8.
When p is taken to the higher end of the range, 0.051, there are no new failures. Because of this, I'll forgo lengthening this any further and not include that table, but I will note that the probability of having no openly LGBT Cabinet secretary or Cabinet-level official drops below 10% to 9.5%.
For a President that is quick to tout his record for diversity, these numbers, while not horrible, are by no means good. They show he is not hiring enough LGBT people in the highest levels of his administration and this is likely insulating him from having sufficient input and perspective from the LGBT community in the decisions regarding the LGBT community. He's also less likely to have LGBT people who feel comfortable speaking truth to power and having that message reach him. Consider a fast food restaurant chain. How likely is the CEO to take the advice of a part-time team member of store #1045 compared to a regional manager? And if the part-time team member of store #1045 does have the courage and opportunity to tell the CEO something, how likely is that person to still have their job if what they said is unfavorable compared to if the same observation is told to the CEO by a regional manager? Diversity is meaningless unless the diversity permeates through every level of the organization.
And it is not like the President has been incapable of doing this with respect to the LGBT community. While the judiciary does not serve at the pleasure of the President like the officials analyzed above, it is still instructive to see how the President has performed in his judicial appointments. He's been pretty much spot on with the numbers. At p=0.034, one would expect 6 U.S. District Court Judges, 1 Appeals Court Judge and 0 Supreme Court Justices. Out of his 189 District Court confirmations, 7 have been openly LGBT. Out of his 44 Appeals Court confirmations, 1 has been openly LGBT. Out of the total 235 judicial confirmations, one would expect 8 to be openly LGBT and exactly 8 have been openly LGBT. And those number will likely improve slightly as there are currently 2 openly LGBT judicial nominees pending in the Senate. However, with respect to openly LGBT people on the whole Federal bench, the numbers are WAYYYYYYY behind where they need to be. Since Deborah Batts took senior status and Vaughn Walker only came out only after taking senior status, the only non-senior openly LGBT Article III judges are the 8 appointed by President Obama and that's out of 607 District Court Judges, 162 Appeals Court Judges and 9 Supreme Court Justices (excluding seats currently vacant). Doing the math that's only 1.028% of the federal judiciary.
The bottom line is that if the President wants to have the moral authority to tell the LGBT community "no" on an executive order barring employment discrimination on the basis of sexual orientation and gender identity in the federal government and by federal contractors, perhaps he should first get his own office in order and improve his own hiring record for the people serving directly under him.
P.S. I have this diary posting during the daytime for readership reasons, but due to the security regulations at my new-ish job (at a federal contractor with more than $10,000 in federal contracts who, because President Obama has not signed the executive order, is at liberty to terminate my employment solely on account of me being gay), I will be unable to respond to comments until I get home from work. We have no Internet access at our desks and we can only have our cell phones on in the lunch room during our quick half hour lunch break.