* Colonel, United States Marine Corps (Retired). Ph.D (Doctor of Philosophy w/Judicial Studies major), University of Nevada, Reno (2017); M.J.S. (Master of Judicial Studies w/Trial Judges major), University of Nevada, Reno (1998); LL.M. (Military Law) 1989, The Judge Advocate General's School, U.S. Army; LL.M. (Environmental Law) 1986, George Washington University Law School; J.D. 1978, George Washington University Law School; B.A. 1975, Amherst College. Colonel Anderson served as a judge on the U.S. Navy-Marine Corps Court of Criminal Appeals from July, 1998 until his retirement in April, 2002. He currently serves as the Director of the Central Legal Staff and Chief Deputy Clerk at the U.S. Court of Appeals for the Armed Forces. Author of Let Jurors Talk: Authorizing Pre-Deliberation Discussion of the Evidence During Trial, 174 Mil. L. Rev. 92 (2002); Summary Contempt Power in the Military: Amend or Repeal Article 48, UCMJ, 160 Mil. L. Rev. 158 (1999); and Spying in Violation of Article 106, UCMJ: The Offense and the Constitutionality of Its Mandatory Death Penalty, 127 Mil. L. Rev. 1 (1990). This article is an edited version of a paper submitted in partial fulfillment of the requirements for his Ph.D. The views expressed are the personal views of the author. The author wishes to thank Dr. James T. Richardson, Foundation Emeritus Professor of Sociology and Judicial Studies, University of Nevada, Reno; Emily Wood, Doctoral Student and Research Assistant, Nevada Center for Surveys, Evaluation, and Statistics, Grant Sawyer Center for Justice Studies, University of Nevada, Reno; and Sherri Barker, Program Officer, Judicial Studies Program, University of Nevada, Reno, for their assistance.
1 Justice Oliver Wendell Holmes, The Path of the Law, 10 Harv. L. Rev. 457, 469 (1897).
 10 U.S.C. §§ 801-946 (2012).
 10 U.S.C. § 942(a)-(b) (2012).
 U.S. Const. art. I, § 8, cl. 14 (The Congress shall have power “[t]o make Rules for the Government and Regulation of the land and naval Forces.”).
 Earl Warren, The Bill of Rights and the Military, 37 N.Y.U. L. Rev. 181, 188 (1962).
 The CAAF has jurisdiction to hear (1) all cases in which a sentence to death has been approved, (2) all cases reviewed by a Court of Criminal Appeals which the Judge Advocate General orders sent to it; and (3) all cases reviewed by a Court of Criminal Appeals in which, upon petition of the accused and on good cause shown, the CAAF has granted review. 10 U.S.C. § 867(a) (2012). The Judge Advocate General shall refer to a Court of Criminal Appeals the record in each case of trial by court-martial in which the sentence, as approved, extends to death, dismissal of a commissioned officer, cadet or midshipman, dishonorable or bad-conduct discharge, or confinement for one year or more. 10 U.S.C. § 866(b) (2012). A Court of Criminal Appeals is the first line appellate court for court-martial convictions, and each panel of that court, established by the Judge Advocate General of each service, is composed of not less than 3 appellate military judges. 10 U.S.C. § 866(a) (2012).
 See 10 U.S.C. § 867a(a) (2012) (“Decisions of the United States Court of Appeals for the Armed Forces are subject to review by the Supreme Court by writ of certiorari as provided in section 1259 of title 28. The Supreme Court may not review by a writ of certiorari under this section any action of the Court of Appeals for the Armed Forces in refusing to grant a petition for review.”).
 Jonathan Lurie, Military Justice 50 Years After Nuremberg: Some Reflections on Appearance v. Reality, 149 Mil. L. Rev. 189, 191 (Summer, 1995).
 10 U.S.C. § 942(b)(3) (2012).
 Art. 67(a)(1), Uniform Code of Military Justice, Pub. L. 506 (81st Cong.), ch. 169 (2d Sess.), 64 Stat. 107, 129 (Act of May 5, 1950, Pub. L. 506, 1950 U.S.C.C.A.N., Vol. 1 at 130).
 §541(c), National Defense Authorization Act for Fiscal Year 2017, Pub. L. 114-328 (114th Cong. 2d Sess.), 130 Stat. 2000 (Dec. 23, 2016).
 Email from Judge Scott Stucky entitled Political Party, to the author (Mar. 22, 2017) (on file with author).
 References to the Court will include both the CAAF and COMA.
 The United States Court of International Trade, an Article III court, is comprised of nine judges appointed by the President with the advice and consent of the Senate, and “[n]ot more than five of such judges shall be from the same political party.” 28 U.S.C. § 251 (2012). The United States Court of Appeals for Veterans Claims, an Article I court, is comprised of at least three and not more than seven judges, and “[n]ot more than the number equal to the next whole number greater than one-half of the number of judges of the Court may be members of the same political party.” 38 U.S.C. § 7253 (2012).
 Matthew A. Samberg, Note, ‘Established by Law’: Saving Statutory Limitations on Presidential Appointments From Unconstitutionality, 85 N.Y.U. L. Rev. 1735, 1750-51 (2010) (noting that independent federal regulatory agencies “play critical roles in promoting the national welfare, and Congress has decided that the important decisions they make require bipartisan input”).
 2 U.S.C. § 437c.(a)(1) (2012).
 47 U.S.C. § 154(b)(5) (2012).
 15 U.S.C. § 41 (2012).
 46 U.S.C. § 301(b)(1) (2012).
 46 U.S.C. § 1111(b) (2012).
 42 U.S.C. § 5841(b)(2) (2012).
 15 U.S.C. § 78d.(a) (2012).
 Samberg, supra note 15, at 1737.
 Common Legislative Encroachments on Executive Branch Authority, 13 Op. O.L.C. 248, 250 (1989) (superseded by The Constitutional Separation of Powers Between the President and Congress, 20 Op. O.L.C. 124, 124 n.* (1996)).
 U.S. Const. art. II, § 2, cl. 2.
 Common Legislative Encroachments, supra note 24, at 250.
 Samberg, supra note 15, at 1752.
 Samberg, supra note 15, at 1735-36. See Hanah M. Volokh, The Two Appointments Clauses: Statutory Qualifications for Federal Officers, 10 U. Pa. J. Const. L. 754, 747 & n.12 (2008); Note, Congressional Restrictions on the President’s Appointment Power and the Role of Longstanding Practice in Constitutional Interpretation, 120 Harv. L. Rev. 1914, 1926 (2007); Donald J. Kochan, The Unconstitutionality of Class-Based Statutory Limitations on Presidential Nominations: Can a Man Head the Women’s Bureau at the Department of Labor?, 37 Loy. U. Chi. L.J. 43, 46 (2005); Adam J. Rappaport, Note, The Court of International Trade’s Political Party Diversity Requirement: Unconstitutional Under Any Separation of Powers Theory, 68 U. Chi. L. Rev. 1429 (2001); Michael J. Gerhardt, Toward a Comprehensive Understanding of the Federal Appointments Process, 21 Harv. J.L. & Pub. Pol’y 467, 534-35 (1998); Richard P. Wulwick & Frank J. Macchiarola, Congressional Interference with the President’s Power to Appoint, 24 Stetson L. Rev. 625, 643-45 (1995).
 Samberg, supra note 15, at 1754.
 U.S. Const. art. II, § 2, cl. 18.
 Free Enterprise Fund v. Public Co. Accounting Oversight Bd., 561 U.S. 477, 130 S.Ct. 3138, 3155 (2010) (“No one doubts Congress’s power to create a vast and varied federal bureaucracy.”) and see id. at 3165 (Breyer, J. dissenting) (“[T]he Necessary and Proper Clause affords Congress broad authority to ‘create’ governmental ‘offices' and to structure those offices ‘as it chooses.’”) (quoting Buckley v. Valeo, 424 U.S. 1, 138 (1976) (per curiam)).
 Samberg, supra note 15, at 1753.
 Id. at 1737, 1740-42, 1747.
 See id. at 1756 (“[W]hat if the Commission’s political balance was thrown off because an existing member changed his party affiliation from Republican to Democratic?”). However, the de facto officer doctrine would appear to validate the decision of a court with a defective member. This doctrine “confers validity upon acts performed by a person acting under the color of official title even though it is later discovered that the legality of [his] appointment to office is deficient.” Ryder v. United States, 515 U.S. 177, 180 (1995).
 Myers v. United States, 272 U.S. 52 (1926). The Myers court invalidated the statute because it held that under Article II of the Constitution, the President had sole power to remove as an incidence of his power to appoint.
 Myers, 272 U.S. at 128-29.
 Id. at 265-71 (footnotes omitted).
 U.S. Const. art. II, § 2, cl. 14 and 18.
 Gallagher v. Quinn, 363 F.2d 301, 303-04 (D.C. Cir.), cert. denied, 385 U.S. 881 (1966) (internal citations omitted).
 Samberg, supra note 15, at 1748.
 Jonathan Lurie, Arming Military Justice (Volume 1 – the Origins of the United States Court of Military Appeals, 1775-1950)128-35 (1992).
 Id. at 154-61.
 United States v. Merritt, 1 C.M.A. 56, 61, 1 C.M.R. 56, 61 (1951).
 Lurie, supra note 42, at 193-203.
 Uniform Code of Military Justice (No. 37): Hearings on H.R. 2498 Before a Subcomm. of the House Comm. on Armed Services, 81st Cong., 1st Sess. 565-1307 (March 7, 8, 9, 10, 14, 16, 17, 18, 21, 22, 23, 24, 25, 26, 30, 31, April 1,2, and 4, 1949) at 582.
 Id. at 604.
 95 Cong. Rec. 5719-20 (1949).
 Uniform Code of Military Justice, supra note 46, at 1272-73.
 Id. at 1277-80.
 H.R. 4080, 81st Cong., 1st Sess. (April 7, 1949).
 Id. at 54-55.
 Full Committee Hearings on H.R. 3341 and H.R. 4080 (No. 44): House of Representatives, Committee on Armed Services (April 27, 1949).
 Id. at 1335-36, 1340, 1350.
 Id. at 1340, 1349-50.
 Lurie, supra note 42 at 229.
 Id. at 228-29 (footnote omitted).
 H.R. Rep. No. 491 (81st Cong., 1st Sess. 1949).
 Id. at 9.
 95 Cong. Rec. 5744 (1949).
 H.R. 4080, 81st Cong. 1st Sess., H.R. Rep. No. 491, Union Calendar No. 190 (April 28, 1949) at 56.
 Uniform Code of Military Justice: Hearings on S. 857 and H.R. 4080 Before a Subcomm. of the Senate Comm. on Armed Services, 81st Cong., 1st Sess. 1-334 (April 27, and May 4, 9, and 27, 1949). See H.R. 4080 in the Senate of the United States, 81st Cong., 1st Sess. (May 6, 1949).
 Compare id. at 18 with Uniform Code of Military Justice, supra note 46, at 582.
 Uniform Code of Military Justice, supra note 64, at 43, 311-15.
 H.R. 4080 in the Senate of the United States, S. Rep. No. 486, Calendar No. 481, 81st Cong., 1st Sess. (June 10, 1949), Art. 67(a)(1) at 159.
 96 Cong. Rec. 1292-1310; 1353-70; 1412-17; 1430-47 (1950).
 See Conf. Rep. No. 1946, Uniform Code of Military Justice Conference Report to accompany H.R. 4080 (April 24, 1950) at 4.
 Uniform Code of Military Justice, Pub. L. 506 (81st Cong.), ch. 169 (2d Sess.), 64 Stat. 107 (Act of May 5, 1950, Pub. L. 506, 1950 U.S.C.C.A.N., Vol. 1 at 110). The original 140 articles of the UCMJ were codified at 50 U.S.C. (Chap. 22) §§ 551-736 and enacted into positive law at 10 U.S.C. §§ 801-810 in 1956 (Act of August 10, 1956, Pub.L. 1028 (84th Cong.), ch. 1041 (2d Sess.), 70A Stat. 36, 1956 U.S.C.C.A.N., Vol. 1 at 1336, 1379-1431. See Lurie, supra note 42, at 255. See also Art. 140, Sec. 5, UCMJ, Act of May 5, 1950, Pub. L. 506, 1950 U.S.C.C.A.N., Vol. 1 at 145 (providing that the UCMJ will become effective on the last day of the twelfth month after approval of the Act).
 Art. 67(a)(1), UCMJ, Act of May 5, 1950, Pub. L. 506, 1950 U.S.C.C.A.N., Vol. 1 at 130.
 Jonathan Lurie, Pursuing Military Justice (Volume 2 – The History of the United States Court of Appeals for the Armed Forces, 1951-1980) 125-27, 138-39 (1998).
 Draft, Reform of the Court of Military Appeals, Office of the General Counsel, Department of Defense, May 7, 1979 (located in the Law Library, United States Court of Appeals for the Armed Forces, 450 E St, NW, Washington, D.C.).
 Report of Department of Defense Study Group on the United States Court of Military Appeals, Office of the General Counsel Department of Defense, July 25, 1988 at 11 (located in the Law Library, United States Court of Appeals for the Armed Forces, 450 E St, NW, Washington, D.C.).
 Revision of the Laws Governing the U.S. Court of Military Appeals and the Appeals Process, Hearings on H.R. 6406 and H.R. 6298 before the Military Personnel Subcommittee of the Committee on Armed Services, House of Representatives, 96th Cong., 2d Sess., Feb, 7, Mar. 6, and Sept. 23, 1980 (H.A.S.C. No. 96-55, G.P.O. Washington, DC 1980) at 77, 80 (testimony and written statement of Hon. A. B. Fletcher, Jr., Chief Judge of the U.S. Court of Military Appeals).
 Id. at 99-100.
 Id. at 100.
 Report of Department of Defense Study Group, supra note 75, at 12.
 United States Court of Military Appeals Committee Report, January 27, 1989 (located in the Law Library, United States Court of Appeals for the Armed Forces, 450 E St, NW, Washington, D.C.).
 Id. at 12.
 Id. at 26.
 Eugene R. Fidell, Going on Fifty: Evolution and Devolution in Military Justice, 32 Wake Forest L. Rev. 1213, 1218 (1997) (also reprinted in Eugene R. Fidell & Dwight H. Sullivan, eds., Evolving Military Justice (2002) at 18).
 Eugene R. Fidell, The Next Judge, 5 J. Nat’l Security L. & Pol’y 303, 308 (2011).
 Francis A. Gilligan & Frederic I. Lederer, Court-Martial Procedure (3d ed. 2006), Vol. 2 at § 25-61.00 at 25-33 (footnotes omitted).
 Id. at § 25-61.00 at 25-33 n.214.
 William T. Generous, Jr., Swords and Scales: The Development of the Uniform Code of Military Justice 49 (1973). See also Uniform Code of Military Justice (No. 37): Hearings on H.R. 2498 Before a Subcomm. of the House Comm. on Armed Services, supra n.46, at 1275-76; Full Committee Hearings on H.R. 3341 and H.R. 4080 (No. 44): House of Representatives, Committee on Armed Services, supra note 54 at 1339-40; Lurie, supra note 42, at 227.
 See National Defense Authorization Act for Fiscal Year 1991, Pub. L. No. 101-501, 104 Stat. 1485 (Nov. 5, 1990) at § 541(f) (providing that “[f]or purposes of appointment of judges to the court, a person retired from the armed forces after 20 or more years of active service (whether or not such person is on the retired list) shall not be considered to be in civilian life”). This change did not alter the President’s ability to appoint retired reserve officers to the Court, as long as they had not served on active duty for 20 or more years.
 See National Defense Authorization Act for Fiscal Year 2014, Pub. L. No. 113-66, 127 Stat. 672 (Dec. 26, 2013) at § 531(a) (providing that “a person may not be appointed as a judge of the court within seven years after retirement from active duty as a commissioned officer of a regular component of an armed force”). This cooling off period is similar to that used for the appointment of the Secretary of Defense. See 10 U.S.C. § 113 (2012) (“A person may not be appointed as Secretary of Defense within seven years after relief from active duty as a commissioned officer of a regular component of an armed force.”).
 Jonathan Lurie, Presidential Preferences and Aspiring Appointees: Selections to the U.S. Court of Military Appeals 1951-1968, 29 Wake Forest L. Rev. 521, 555-56 (1994).
 Email from Court Executive to the author entitled Tentative Legislative Package dated 5 January 2017 (on file with author).
 Copy of proposed amendment rationale provided by the Chief Judge Erdmann to the author on 2 March 2017 (on file with author).
 Emails from Chief Judge Erdmann to the author entitled CAAF Legislative Package, Legislative Sitrep, and Legislative Proposals dated 2 March 2017 (on file with author); email from Dwight Sullivan, DoD GC’s office, entitled A Little Help dated 15 February 2017 (on file with author).
 Emails from Chief Judge Erdmann to the author entitled CAAF Legislative Package, Legislative Sitrep, and Legislative Proposals dated 2 March 2017 (on file with author); email from Judge Stucky to author entitled Political Party dated 22 March 2017 (on file with author).
 §541(c), National Defense Authorization Act for Fiscal Year 2017, Pub. L. 114-328 (114th Cong. 2d Sess.), 130 Stat. 2000 (Dec. 23, 2016).
 John A. Sautter & J. Derek Randall, A Jury of One’s Peers: An Empirical Analysis of the Choice of Members in Contested Military Courts-Martial, 217 Mil. L. Rev. 91, 100 (2013).
 Id. at 100 n.56.
 See CAAFLog, http://www.caaflog.com/category/end-o-term-stats (last visited Nov. 20, 2017).
 See http://supremecourtdatabase.org/ or http://scdb.wustl.edu/ or http://artsandsciences.sc.edu/poli/juri/sct.htm [United States Supreme Court Database]; http://artsandsciences.sc.edu/poli/juri/appct.htm [U.S. Courts of Appeals Database]; http://artsandsciences.sc.edu/poli/juri/attributes.htm [Judicial Attributes Database of U.S. Courts of Appeals].
 The Supreme Court Database, Wash. U. L., http://supremecourtdatabase.org (last visited Nov. 20, 2017).
 U.S. Appeals Courts Database, U. S. C., http://artsandsciences.sc.edu/poli/juri/appct.htm (last visited Nov. 20, 2017); Attributes of U.S. Federal Judges Database, U. S. C., http://artsandsciences.sc.edu/poli/juri/ attributes.htm (last visited Nov. 20, 2017).
 U.S. Appeals Courts Database, U. S. C., http://artsandsciences.sc.edu/poli/juri/ appct.htm (last visited Nov. 20, 2017).
 U.S. Federal Judges Database, U. S. C., http://artsandsciences.sc.edu/poli/juri/ attributes.htm (last visited Nov. 20, 2017).
 Glendon A. Schubert, Quantitative Analysis of Judicial Behavior 25 (1959).
 Id. at 77-129.
 Id. at 129-142.
 Id. at 77-142.
 Id. at 16.
 Id. at 129.
 Id. at 142.
 Jeffrey A. Segal & Harold J. Spaeth, The Supreme Court and the Attitudinal Model (1993).
 Jeffrey A. Segal & Harold J. Spaeth, The Supreme Court and the Attitudinal Model Revisited (2002).
 Cass R. Sunstein, David Schkade, Lisa M. Ellman, & Andres Sawicki, Are Judges Political?: An Empirical Analysis of the Federal Judiciary vii (2006).
 Id. at 6-9.
 Id. at 8-18.
 Id. at vii.
 Frank B. Cross, Decision Making in the U.S. Courts of Appeals (2007).
 Id. at 11.
 Id. at 7.
 Id. at 22.
 Id. at 24.
 Id. at 6.
 Id. at 5. He also commented:
An empirical researcher does not need a perfect measure of variables to reach conclusions. Imperfections in measurement tend to obscure results rather than produce spurious positive results. If research with imperfect measurements nevertheless produces a statistically and substantively significant finding, that research probably understates the true result.
Id. at 20-21.
 Id. at 6.
 Id. at 7-9.
 Id. at 9.
 Id. at 7-8.
 Id. at 8.
 Lee Epstein, William M. Landes, & Richard Posner, The Behavior of Federal Judges: A Theoretical and Empirical Study of Rational Choice (2013).
 Id. at 5.
 Id. at 8-9.
 Id. at 385.
 Id. at 385-86.
 Daniel P. Pinello, Linking Party to Judicial Ideology in American Courts: A Meta-Analysis, 20 Just. Sys. J. 219, 220 (1999).
 Id. at 221.
 Id. at 221-24. Meta-analysis is defined as “quantitative statistical analysis that is applied to separate but similar experiments or studies of different and usually independent researchers and that involves pooling the data and using the pooled data to test for statistical significance.” Meta-analysis, Merriam-Webster Unabridged, http://unabridged.merriam-webster.com/browse/meta-analysis (last visited Nov. 20, 2017).
 Pinello, supra note145, at 240-43.
 Joshua B. Fischman & David S. Law, What is Judicial Ideology, and How Should We Measure It?, 29 Wash. U. J.L. & Pol’y 133, 136 (2009) (“Empirical scholarship on the subject of judicial ideology is vulnerable to two sets of difficulties, which tend to blend into one another. The first set is theoretical; the second set is methodological. … [T]he theoretical problem [is] that scholars use the term ‘judicial ideology’ in the absence of any widespread agreement or clear understanding as to what the term means in the first place. It is difficult for scholars to devise appropriate and broadly acceptable measures of judicial ideology when they and their readers have different concepts—or perhaps no coherent concept at all—of ‘judicial ideology’ in mind. As a result, bona fide intellectual disagreement over the nature of judicial behavior is too easily compounded by outright misunderstanding. … [As to the methodological difficulty, there are] three … significant and common practical obstacles to the measurement of judicial ideology. First, ideology is not a tangible phenomenon that can be directly observed. Second, judicial behavior is often open to multiple interpretations. Third, judicial ideology may be a multidimensional phenomenon, such that a judge who is liberal in one context may be moderate or conservative in another, or the labels ‘liberal,’ ‘moderate,’ and ‘conservative’ may not seem applicable at all.”).
 Id. at 137.
 Id. at 204-05.
 Id. at 167-68.
 Id. at 149-50.
 Corey Rayburn Yung, Judged by the Company You Keep: An Empirical Study of Ideologies of Judges on the United States Court of Appeals, 51 B.C. L. Rev. 1113, 1144-53 (2010).
 Id. at 1148.
 Pinello, supra note 145, at 220.
 Yung, supra note 156, at 1138 (“By identifying voting blocs, assessments can be made about the ideologies of the judges that form those blocs.”). See also id. at 1143 (“This study compares judges to determine which ones are more conservative or liberal relative to their colleagues based upon whom they most often vote with and against.”).
 Id. at 1191.
 Id. at 1153.
 Id. at 1159-60.
 Id. at 1201.
 Harry T. Edwards & Michael A. Livermore, Pitfalls of Empirical Studies That Attempt to Understand the Factors Affecting Appellate Decisionmaking, 58 Duke L.J. 1895 (2009).
 Id. at 1899.
 Mark Klock, Cooperation and Division: An Empirical Analysis of Voting Similarities and Differences During the Stable Rehnquist Court Era – 1994 to 2005, 22 Cornell J.L. & Pub. Pol’y 537, 540-41 (2013).
 Id. at 542-43.
 Id. at 554.
 Judge Sparks did not participate in any of the counted dissenting opinions in this empirical study.
 Other issues identified in the database included general categories of due process/legal procedure, legal sufficiency, admissibility of evidence, substantive offenses and defenses, providence of guilty pleas, instructions, lesser-included offenses, prosecutorial misconduct, rights to counsel and confrontation, Article 31 rights and the right against self-incrimination, multiplicity, unreasonable multiplication of charges, mental responsibility, and pretrial and post-trial processing. With regard to many of the cases, the votes fell with the category of due process/legal procedure with a subtopic of “material prejudice” or “waiver.” In such cases, the real issues in the case were avoided with the following rationales: “Even if error, there was no material prejudice”; and “Even if error, there was no objection and the matter was waived.”
 The top ten list was based on the latest ranking in US News and World Report. See https://www.usnews.com/best-graduate-schools/top-law-schools/law-rankings?int=a1d108. Although this type of list did not exist throughout the history of the Court, I used it as a rough measure of what law schools are considered elite.
 I coded the variables as follows: Political affiliation: 0 = Republican; 1= Democrat ; Prior military experience: 0 = No prior military experience; 1 = Prior military experience; Attendance at elite law school: 0 = Did not attend elite law school; 1 = Did attend elite law school. Votes were coded as follows: 0 = Government; 1= Appellant. And case subtypes were coded as follows: 5 = Speedy trial; 6 = Challenges to court members; 7 = Command influence; 8 = Ineffective assistance of counsel; 9 = Jurisdiction.
 Sautter & Randall, supra note 104, at 106. Of course, one drawback to this method is that “because case coding relies upon a wholly binary construction of concept,” it may lack “significant nuance in particular cases.” Yung, supra note 156, at 1146-47.
 Sautter & Randall, supra note 104, at 106.
 This coding is consistent with how criminal cases are coded in The Supreme Court Database Codebook. See http://scdb.wustl.edu/_brickFiles/2013_01/SCDB_2013_01 _codebook.pdf.
 In a Gallup poll, military veterans of all ages tend to be more Republican than are those of comparable ages who are not veterans. http://www.gallup.com/poll/118684/military-veterans-ages-tend-republican.aspx (last visited on Apr. 11, 2017).
 Of course, there are drawbacks to this method. One such drawback is that it cannot make use of unanimous opinions, because “it is impossible to draw any inference about the relative positions of the judges from the voting alignment in a unanimous decision.” Another drawback is that “ideology in this context only measures how often particular judges vote with each other, and not how often they support particular types of outcomes.” And a third drawback is that this measure is a one-dimensional approach, when judicial “ideology is never perfectly one-dimensional.” Fischman & Law, supra note 150, at 165-66.
 The chi-square (χ2) test tests if the overall model is significant. That is, it tests if there is an effect of the independent variables taken together on the dependent variable. In this case, it is significant, which indicates that the independent variables, political affiliation, military experience, and elite law school, when taken together have an effect on the dependent variable. If the chi square test was not significant (as in some of the logistic regressions looking at each case type individually), this means that it is not a good model and the predictor variables (independent variables) are not affecting the dependent variable.
 Conventionally, a p value .05 or less is considered significant, from .05 to .1 is considered marginally significant, and anything larger than .1 is not significant.
 The Nagelkerke R2 value provides an indication of how large an effect the independent variables have on the dependent variable. In this case, political affiliation, military experience, and elite law school are only explaining about 8% of the variance in vote. This is relatively low and means that there is a large degree of unexplained variance in vote (i.e., political affiliation, military experience, and elite law school do not explain all of the variation in vote). There are likely other factors such as case facts.
 With an extremely large sample size, sometimes differences (e.g., differences between votes cast by those who attended an elite law school and those who did not) will be significant (p < .05) even when the difference is really small. This is because as the sample size increases, the power to detect even tiny differences between groups increases. Therefore, just because something is statistically significant (p < .05) with a large sample size is not always meaningful or practically relevant. Typically, to tell if a result is “practically meaningful,” effect sizes are examined. This is a little less clear in logistic regression, but one way to do this is the Nagelkerke R2, which is a pseudo-R2 measure (a measure designed to evaluate goodness-of-fit logistic models). As already discussed, the R2 value (about 9%) is quite small.
 B is the regression coefficient. The Wald Z statistic tests the statistical significance (indicated by the associated p value). The Exp(B) value is the odds ratio (e.g., “The odds of voting for the appellant decrease by a factor of .424 for votes cast by judges with prior military experience.”).
 As a hypothetical, suppose two judges are identical with respect to all other variables except that one did not attend an elite school and one did. Because the elite school variable is coded as 0 for did NOT attend elite school and 1 for DID attend elite school, “changing” from did NOT attend to DID attend is a one-unit change in the elite school variable. If the odds ratio value for this variable is 2.15, this means that the odds that the judge who DID attend an elite law school votes for the appellant (liberal) are about 2.15 times the odds that the “equivalent” judge who did NOT attend elite school would vote for the appellant. If the Odds Ratio = 1, then elite school attendance does not affect the odds of outcome (voting for appellant). If the Odds Ratio is > 1, elite school attendance increases the odds of outcome (voting for appellant). If the Odds Ratio is < 1, then elite school attendance decreases the odds of outcome (voting for appellant). See: http://stats.idre.ucla.edu/spss/output/logistic-regression.
 Interestingly, by running the same logistic regression on just the three-judge cases, the political affiliation would have been significant (i.e., the odds of voting for the appellant would increase for votes cast by Democrat judges); however, by considering just the five-judge cases, the political affiliation would not have been significant. And with all the cases considered together, the result is that political affiliation was not significantly related to vote.
 Note that 27 three-judge cases in the database in which there was a dissent were not counted because all of the judges on the panel were Republican.
 As noted in footnote 185 above, if just the three-judge cases would have been considered, the political affiliation would have been significant; this result differs for the five-judge cases and with the cases consolidated. Why a three-judge court would be more apt to be influenced by the political party balance requirement than a five-judge court is a matter left for future study.
 Epstein, Landes, & Posner, supra note 140, at 385.