OPINION OF THE THREE-JUDGE COURT
NIEMEYER, Circuit Judge:After the 2010 decennial census, Maryland enacted a new redistricting plan in October 2011 for its eight congressional districts. The plaintiffs, nine African-American residents of Maryland, commenced this action against election officials of Maryland (“Maryland” or “the State”), contending that the redistricting plan violates their rights under Article I, § 2, of the U.S. Constitution; the Fourteenth and Fifteenth Amendments of the U.S. Constitution; and § 2 of the Voting Rights Act of 1965 because the plan dilutes African-American voting strength within the State and intentionally discriminates against African-Americans. For the same reasons, plaintiffs also challenge Maryland’s “No Representation Without Population Act” (“the Act”), which purports to correct census data for the distortional effects of the Census Bureau’s practice of counting prison inmates as residents of their place of incarceration.
This three-judge court, convened pursuant to 28 U.S.C. § 2284(a), heard arguments on December 20, 2011, on the plaintiffs’ motion for preliminary injunction and Maryland’s motion to dismiss or for summary judgment, as well as on the merits of the case. By agreement of the parties, the court received the testimony of all witnesses by affidavit.
For the reasons given herein, we deny Maryland’s motion to dismiss based upon an inappropriate convening of the three-judge court, deny the plaintiffs’ motion for an injunction — preliminary or permanent — and grant Maryland’s motion for summary judgment, obviating its motion to dismiss under Federal Rule of Civil Procedure 12(b)(6).
I. Factual Context
The 2010 census determined Maryland’s population to be 5,773,552. This number entitled the State to eight congressional seats, the same number it had after the 2000 census.
*891On July 4, 2011, Governor Martin O’Malley appointed the Governor’s Redistricting Advisory Committee (the “GRAC”),1 and that committee held twelve public meetings across the state between July 28 and September 12, 2011. Over the course of these meetings, the GRAC received more than 350 comments from members of the public. Among these comments were several proposed redistricting plans from third-party groups, including one from the Fannie Lou Hamer Political Action Committee. The Fannie Lou Hamer plan differed from all the other third-party submissions in that it proposed the creation of three, rather than two, majority African-American districts.
The GRAC presented its proposed plan to the Governor on October 4, 2011. (Compl. Attach. B.) After posting the plan online and receiving additional comments from the public, the Governor announced that he would submit to the legislature a plan that was “substantially similar” to the GRAC proposal.
The Governor’s proposed redistricting map (Compl. Attach. C) was introduced as House Bill (H.B.) 1 and Senate Bill (S.B.) 1 in an emergency legislative session beginning October 17, 2011. H.B. 1 was assigned to the House Rules Committee but was never reported out of committee. The Senate Committee on Reapportionment and Redistricting, however, held a joint hearing on S.B. 1 with the House Rules Committee on the same day the bill was introduced. After the hearing, the Senate Committee approved the bill and sent it to the floor of the Senate. While the bill was being debated, State Senator E.J. Pipkin moved to amend the bill. Like the Fannie Lou Hamer plan, Senator Pipkin’s proposed amendment created three majority African-American districts. (Compl. Attach. C.) Specifically, the Pipkin map proposed the creation of a new Fifth District that would stretch from the southern portion of Charles County, through Prince George’s County, and into the western Baltimore suburbs. The Senate rejected Pipkin’s amendment and, after adopting minor technical amendments, passed the bill. The bill was then sent to the House of Delegates on October 18, 2011.
During the House debate, several substantive amendments to the bill were proposed and rejected. On October 19, after making some technical amendments, the House passed the bill. It then returned the bill to the Senate, which concurred in the House’s technical amendments and enacted the bill on October 20, 2011. The Governor signed S.B. 1 into law later that day.
Like the redistricting plan passed after the 2000 census, the enacted State Plan creates two majority African-American congressional districts. The Seventh District, which includes large portions of Baltimore City and its surrounding suburbs, has an African-American voting age population (“VAP”) of 53.75%, and a non-Hispanic white VAP of 35.75%. The Fourth District, which is centered in Prince George’s County, has an African-American VAP of 53.72% and a non-Hispanic white VAP of 28.65%.
The plaintiffs in this case, who are African-American residents of Maryland, commenced this action on November 10, 2011, naming Linda H. Lamone in her official capacity as Maryland’s Administrator of Elections and Robert L. Walker in his official capacity as Chairman of the State *892Board of Elections. In their complaint, the plaintiffs allege that the State Plan was deficient in several respects, claiming specifically (1) that the State Plan’s creation of two, rather than three, majority African-American districts intentionally discriminates against minorities, in violation of the Fourteenth and Fifteenth Amendments (Counts 1 and 2), and unlawfully dilutes African-American voting strength in violation of the Voting Rights Act of 1965, 42 U.S.C. § 1973 (Count 5); (2) that the State Plan is unconstitutionally malapportioned (Counts 3, 4, and 6); and (3) that the State Plan is a partisan gerrymander, in violation of the Fourteenth Amendment (Count 7). At the time the plaintiffs filed their complaint, they also filed a motion under 28 U.S.C. § 2284 to convene a three-judge court to adjudicate their claims, as well as a Motion for a Preliminary Injunction on Counts 3 and 6.
The State opposed convening a three-judge court. And after it was convened, the State filed a Motion to Dismiss for Failure to State a Claim or, in the Alternative, for Summary Judgment, and a Request for Review of the Order Convening a Three-Judge Panel.
The parties thereafter filed responsive and reply briefs and, upon agreeing to present the testimony of witnesses by affidavit, the affidavits of numerous witnesses.
II. Three-Judge Court
At the outset, Maryland requests that we review the single-district judge’s ruling that the plaintiffs’ complaint is sufficiently substantial to justify convening a three-judge court under 28 U.S.C. § 2284. It argues that the ruling failed to take into account the Fourth’s Circuit’s precedent in Duckworth v. State Administration Board of Election Laws, 332 F.3d 769 (4th Cir.2003). In that case, the Fourth Circuit held that when a complaint fails to state a claim upon which relief can be granted, as required by Federal Rule of Civil Procedure 12(b)(6), “by definition [it is] insubstantial and so properly [is] subject to dismissal by the district court without convening a three-judge court.” Id. at 772-73. The State’s argument rests on an assumed distinction between a complaint that “does not state a substantial claim for ... relief’ and the Rule 12(b)(6) standard. For purposes of construing § 2284, we find no material distinction, and deny Maryland’s motion.
The Rule 12(b)(6) standard is a threshold condition for proceeding with an action, testing the sufficiency of a complaint to state a claim for relief. See Brzonkala v. Va. Polytechnic & State Univ., 169 F.3d 820, 829 (4th Cir.1999) (en banc). In applying the standard, we have required more than formulaic, conclusory allegations. See Duckworth, 332 F.3d at 774-75. Indeed, the Supreme Court has required that a complaint must have sufficient “heft” in alleging facts so as to state a “plausible” claim for relief. Bell Atlantic Corp. v. Twombly, 550 U.S. 544, 555, 127 S.Ct. 1955, 167 L.Ed.2d 929 (2007).
Here, the single-judge court reviewed the complaint and concluded that the claims presented were sufficiently substantial to proceed with convening a three-judge court. Under the standard for convening a three-judge court, which is informed by the standard for granting Rule 12(b)(6) motions, we agree. See Ashcroft v. Iqbal, 556 U.S. 662, 129 S.Ct. 1937, 173 L.Ed.2d 868 (2009); Twombly, 550 U.S. at 555-56, 127 S.Ct. 1955; Duckworth, 332 F.3d at 773-75; Simkins v. Gressette, 631 F.2d 287, 295 (4th Cir.1980) (noting that convening a three-judge court is not required to address insubstantial claims). Accordingly, we deny the State’s motion to dismiss the three-judge court.
*893III. “No Representation Without Population Act” (Counts 3, 4, and 6)
The plaintiffs challenge as unconstitutional Maryland’s “No Representation Without Population Act” (“the Act”), 2010 Md. Laws, ch. 67, codified at Md.Code Ann., Art. 24 § 1-111, Election Law (“EL”) § 8-701. They contend that the adjustments made under the Act result in malapportionment, in violation of Article I, § 2, of the Constitution, and racial discrimination, in violation of the Fourteenth Amendment.
Maryland enacted the “No Representation Without Population Act” in 2010. According to the State, the Act is intended to “correct for the distortional effects of the Census Bureau’s practice of counting prisoners as residents of their place of incarceration.” (Defs.’ Mot. Summ. J. 1.) These distortional effects stem from the fact that while the majority of the state’s prisoners come from African-American areas, the state’s prisons are located primarily in the majority white First and Sixth Districts. As a result, residents of districts with prisons are systematically “overrepresented” compared to other districts. In other words, residents of districts with prisons are able to elect the same number of representatives despite in reality having comparatively fewer voting-eligible members of their community.2
To rectify this perceived imbalance, the Act requires that for purposes of drawing local, state, and federal legislative districts, inmates of state or federal prisons located in Maryland must be counted as residents of their last known residence before incarceration. Prisoners who were not Maryland residents prior to incarceration are excluded from the population count, and prisoners whose last known address cannot be determined are counted as residents of the district where their facility is located.
The Maryland Department of Planning (the “MDP”) accomplished the necessary population count adjustments by performing a multistep analysis of the records for prisoners housed in the Maryland Division of Correction.3 As a result of the MDP’s analysis, 1,321 inmates who reported a pre-incarceration address outside of Maryland were excluded from the redistricting database. Other prisoners were reassigned to their prior residences. The largest changes from these reassignments occurred in the Sixth District, which contains the majority of the prisons and lost 6,754 individuals, and the Seventh District, which includes Baltimore City and gained 4,832 individuals. In no case did the adjustments made by the MDP exceed 1% of a district’s population. (Defs.’ Mot. Summ. J. 10.)
After the 1,321 out-of-state prisoners were excluded, Maryland’s adjusted population base for redistricting fell to 5,772,-*894231. This figure resulted in a new ideal Congressional district size of 721,528.875 individuals. Because the new ideal district size was not a whole number, the State set the size of the Eighth District to 721,528 individuals and the size of the remaining districts to 721,529 individuals.
The plaintiffs first contend that Maryland’s adjustments to the census data result in malapportionment, in contravention of the “One Person, One Vote” standard established in Reynolds v. Sims, 377 U.S. 533, 84 S.Ct. 1362, 12 L.Ed.2d 506 (1964), under Article I, § 2, of the Constitution.
Article I, § 2 provides that the members of the House of Representatives are to be chosen “by the People of the several States.” U.S. Const, art. I, § 2. As interpreted by the Supreme Court, this provision mandates that “as nearly as is practicable one man’s vote in a congressional election is to be worth as much as another’s.” Wesberry v. Sanders, 376 U.S. 1, 7-8, 84 S.Ct. 526, 11 L.Ed.2d 481 (1964). “[T]he ‘as nearly as practicable’ standard requires that the State make a good-faith effort to achieve precise mathematical equality.” Kirkpatrick v. Preisler, 394 U.S. 526, 530-31, 89 S.Ct. 1225, 22 L.Ed.2d 519 (1969). “Unless population variances among congressional districts are shown to have resulted despite such effort, the State must justify each variance, no matter how small.” Id. States do not have unlimited discretion in performing the calculations required to meet the “One Person, One Vote” standard. In Kirkpatrick and again in Karcher v. Daggett, 462 U.S. 725, 103 S.Ct. 2653, 77 L.Ed.2d 133 (1983), the Supreme Court concluded that because the census count represents the “ ‘best population data available,’ it is the only basis for good-faith attempts to achieve population equality.” Karcher, 462 U.S. at 738, 103 S.Ct. 2653 (internal citation omitted) (quoting Kirkpatrick, 394 U.S. at 528, 89 S.Ct. 1225).
Relying on these statements in Karcher and Kirkpatrick, the plaintiffs contend that “for determining congressional districts the only [population] number that can be used is the number generated by the U.S. census.” (Pis.’ Mot. Supp. Mot. Prelim. Inj. 7.) Accordingly, they argue that Maryland’s decision to adjust the census number is unconstitutional.
We believe that the plaintiffs fail to read the Karcher and Kirkpatrick statements in their fuller context. Although Karcher and Kirkpatrick do require states to use census data as a starting point, they do not hold, as the plaintiffs maintain, that states may not modify this data to correct perceived flaws. A more complete reading of the opinion in Karcher makes this point clear. The Court there recognized that “the census may systematically undercount population, and the rate of undercounting may vary from place to place.” 462 U.S. at 738, 103 S.Ct. 2653. It cautioned, however, that “[i]f a State does attempt to use a measure other than total population or to ‘correct’ the census figures, it may not do so in a haphazard, inconsistent, or conjectural manner.” Id. at 732 n. 4, 103 S.Ct. 2653 (citing Kirkpatrick, 394 U.S. at 534-35, 89 S.Ct. 1225). Thus, the New Jersey redistricting plan at issue in Karcher was rejected not because the state used adjusted census data, but because the state failed to perform its adjustments systematically. See id. at 738, 103 S.Ct. 2653 (“Attempts to explain population deviations on the basis of flaws in census data must be supported with a precision not achieved here ” (emphasis added)). Taken together, these Karcher statements suggest that a State may choose to adjust the census data, so long as those adjustments are thoroughly documented and applied in *895a nonarbitrary fashion and they otherwise do not violate the Constitution.
Although the case law on this issue is sparse, the majority of the courts to consider the issue have similarly concluded that Karcher and Kirkpatrick do not bar the use of adjusted census data. For example, in City of Detroit v. Franklin, 4 F.3d 1367 (6th Cir.1993), the Sixth Circuit considered a challenge to the Census Bureau’s alleged undercounting of the primarily African-American residents of the City of Detroit. The court held that the plaintiffs lacked standing to pursue their claim against the Census Bureau because the allegedly harmful act — the decision to use unadjusted census data in the redistricting process — had been made by the Michigan legislature. In reaching this conclusion, the Sixth Circuit rejected the plaintiffs argument that the Michigan legislature was constitutionally compelled to use unadjusted census data:
The Court in Karcher did not hold that the states must use census figures to reapportion congressional representation. The Supreme Court merely reiterated a well-established rule of constitutional law: states are required to use the “best census data available” or “the best population data available” in their attempts to effect proportionate political representation. Nothing in the constitution or Karcher compels the states or Congress to use only the unadjusted census figures.
Id. at 1374 (quoting City of Detroit v. Franklin, 800 F.Supp. 539, 543 (E.D.Mich.1992)); see also Senate of State of Cal. v. Mosbacher, 968 F.2d 974, 979 (9th Cir.1992) (stating in dicta that “[i]f the State knows that the census data is unrepresentative, it can, and should, utilize noncensus data in addition to the official count in its redistricting process”); Perez v. Texas, No. 11-CA-360-OLG-JES-XR, slip op. at 24 (W.D.Tex. Sept. 2, 2011) (holding that “the State could enact a constitutional amendment or statute that modifies the count of prisoners as residents of whatever county they lived in prior to incarceration ... [but] there is no federal requirement to do so”).
The plaintiffs’ contrary argument rests primarily on the decision in Travis v. King, 552 F.Supp. 554 (D.Haw.1982), which, they argue, is the only case directly on point. In Travis, the Hawaii legislature had decided to exclude from its population measure the entire military population, without attempting individual assignment, but allowed “the presence of this large military population ... [to] aid[] in achieving its two congressional seats.” Travis, 552 F.Supp. at 571. A three-judge court subsequently held that Hawaii’s actions violated the “One Person, One Vote” principle. But Travis was decided before the Supreme Court’s decision in Karcher, and the district court in Travis therefore did not have the benefit of Karcher’s elaboration on the requirements of Article I, § 2. Further, after the categorical exclusion of all military personnel in Travis, the congressional districts still varied by over 300, id. at 569, whereas the Maryland legislature in this case drew districts as equally as possible after adjusting the census figures.
The conclusion that States may adjust census data during the redistricting process is also consistent with the practices of the Census Bureau itself. According to the Census Bureau, prisoners are counted where they are incarcerated for pragmatic and administrative reasons, not legal ones. The Bureau has explained that counting prisoners at their home addresses would require “collecting information from each prisoner individually” and necessitate “an extensive coordination procedure” with correctional facilities. U.S. Census Bu*896reau, Tabulating Prisoners at Their Permanent Home of Record” Address, at 10 (2006). Such an effort would likely cost up to $250 million. Id, And although the Census Bureau was not itself willing to undertake the steps required to count prisoners at their home addresses, it has supported efforts by States to do so. For the 2010 census, the Bureau released its population data for prisoners and other inhabitants of “group quarters” early to enable States to “leave the prisoners counted where the prisons are, delete them from redistricting formulas, or assign them to some other locale.” So, How Do You Handle Prisons?, Director’s Blog, U.S. Census Bureau (March 10, 2010), http:// blogs.census.gov/directorsblog/2010/03/sohow-do-you-handle-prisons. html.
The question remains whether Maryland’s adjustments to census data were made in the systematic manner demanded by Karcher. It seems clear to us that they were. As required by the regulations implementing the Act, see Md. Code. Regs. 34.05.01 (2011), the MDP undertook and documented a multistep process by which it attempted to identify the last known address of all individuals in Maryland’s prisons. The MDP and its redistricting contractor, Caliper Corporation, then used this information to make the relevant adjustments to the data it had received from the Census Bureau. (Defs.’ Mot. Summ. J. Exs. 2, 3, 4.) This process is a far cry from the “haphazard, inconsistent, or conjectural” alterations the Supreme Court rejected in Karcher.
The plaintiffs do not dispute that the MDP followed the prescribed process, but they raise two objections to the result. First, they argue that if Maryland wishes to correct for prisoner-related population distortions, it must also make similar adjustments to account for the distortionary effects of college students and members of the military. (Pis.’ Resp. Defs.’ Mot. Summ. J. 24.) Second, they contend that contrary to the MDP’s assumption, most prisoners do not return to their last known residence after release. (Pis.’ Resp. 26.)
Neither of these objections, however, is probative of whether the adjustments made were proper. To be sure, Maryland might come closer to its goal of producing accurate data if it assigned college students or active duty military personnel to their permanent home addresses for purposes of redistricting. But as with prisoners, Maryland is not constitutionally obligated to make such adjustments. Moreover, the State’s failure to improve its redistricting data even more by determining students’ and soldiers’ home addresses has little bearing on the merits of the plaintiffs’ Article 1, § 2 claim made with respect to prisoners.
We also observe that the plaintiffs’ argument on this point implies that college students, soldiers, and prisoners are all similarly situated groups. This assumption, however, is questionable at best. College students and members of the military are eligible to vote, while incarcerated persons are not. In addition, college students and military personnel have the liberty to interact with members of the surrounding community and to engage fully in civic life. In this sense, both groups have a much more substantial connection to, and effect on, the communities where they reside than do prisoners.
As to the plaintiffs’ second argument that the adjustments are improper because most prisoners do not return to their last known addresses after release, it would certainly be true that at least some prisoners will return to their old communities even if the plaintiffs are correct. See Nancy G. La Vigne et al., Urban Institute, A Portrait of Prisoner Reentry in Maryland 39 (reporting that 59% of Maryland prison*897ers returned to Baltimore City after their release from prison). Because some correction is better than no correction, the State’s adjusted data will likewise be more accurate than the information contained in the initial census reports, which does not take prisoners’ community ties into account at all.
In sum, we conclude that the State did not violate Article I, § 2 by adjusting the raw census data as it did.
The plaintiffs also contend that the Act’s exclusion of incarcerated non-Maryland residents from the population base constitutes intentional discrimination on the basis of race, in violation of the Fourteenth Amendment, because, of the 1,321 prisoners who were excluded, 71.08% are African American. (Compl. ¶¶ 49-53, 70.)
We find no support in the record for this contention. It is well-established that allegations of disparate impact alone are insufficient to state a claim under the Fourteenth Amendment. Washington v. Davis, 426 U.S. 229, 239, 96 S.Ct. 2040, 48 L.Ed.2d 597 (1976). Instead, plaintiffs are required to prove purposeful discrimination. Id. Our review of the record reveals no evidence that intentional racial classifications were the moving force behind the passage of the Act. In fact, the evidence before us points to precisely the opposite conclusion. As the amicus brief of the Howard University School of Law Civil Rights Clinic and other civil rights organizations makes clear, the Act was the product of years of work by groups dedicated to advancing the interests of minorities.
IV. Section 2 of the Voting Rights Act (Count 5)
The plaintiffs contend, as alleged in Count 5 of their complaint, that the State Plan violates the Voting Rights Act of 1965 (the ‘VRA”), 42 U.S.C. § 1973, because it fails to create a third majority African-American congressional district. To support this contention, the plaintiffs recite the census data that show that 30.9% of the Maryland population is African-American, representing 28% of the VAP, and they offer four maps that they contend demonstrate how a third African-American district could be created while still respecting traditional districting principles. These maps are the Fannie Lou Hamer Plan (Compl. Attach. A), Senator Pipkin’s Plan (Compl. Attach. D), and two plans by Antonio Campbell (Pis.’ Resp. Ex. 14). In each of these maps, a third majority African-American congressional district is formed by connecting residents of the Washington, D.C. suburbs with residents of the Baltimore suburbs.
Section 2(a) of the VRA prohibits the imposition of any electoral practice or procedure that “results in a denial or abridgement of the right of any citizen ... to vote on account of race or color.” 42 U.S.C. § 1973(a). “A violation of [§ 2(a) ] is established if, based on the totality of the circumstances, it is shown that the political processes are not equally open to participation by members of a class of citizens ... in that its members have less opportunity than other members of the electorate to participate in the political process and to elect representatives of their choice.” Id. § 1973(b).
In 1982, Congress amended § 2 to make clear that plaintiffs need not prove intentional discrimination. See Thornburg v. Gingles, 478 U.S. 30, 35, 106 S.Ct. 2752, 92 L.Ed.2d 25 (1986). Instead, a violation of § 2 may be demonstrated through discriminatory effect alone. Id. District line-drawing is therefore impermissible where its result, “interacting] with social and historical conditions, impairs the ability of a protected class to elect its candidate of choice on an equal basis with other vot*898ers.” Voinovich v. Quilter, 507 U.S. 146, 153, 113 S.Ct. 1149, 122 L.Ed.2d 500 (1993) (internal quotation marks omitted).
In the redistricting of a single member constituency, the most common means of manipulating the voting strength of a politically cohesive minority group are “cracking” and “packing.” “Cracking” occurs when redistricting lines are drawn in order to “divid[e] the minority group among various districts so that it is a majority in none.” Voinovich, 507 U.S. at 153, 113 S.Ct. 1149. “Packing” occurs when a redistricting plan results in an excessive concentration of minorities within a given district, thereby depriving the group of influence in surrounding districts. Id. at 153-54, 113 S.Ct. 1149.
In Thornburg v. Gingles, the Supreme Court established three preconditions (the “Gingles preconditions”) that a plaintiff must satisfy in order prove a violation of § 2 of the VRA:
First, the minority group must be able to demonstrate that it is sufficiently large and geographically compact to constitute a majority in a single-member district.
Second, the minority group must be able to show that it is politically cohesive. Third, the minority must be able to demonstrate that the white majority votes sufficiently as a bloc to enable it usually to defeat the minority’s preferred candidate.
Gingles, 478 U.S. at 50-51, 106 S.Ct. 2752. Only after all of these three preconditions are met will a court evaluate the totality of the circumstances.
Although the failure to create a majority-minority voting district may be the basis of a § 2 violation, § 2 does not obligate States to create the maximum possible number of majority-minority districts. See Johnson v. De Grandy, 512 U.S. 997, 1016-17, 114 S.Ct. 2647, 129 L.Ed.2d 775 (1994); Abrams v. Johnson, 521 U.S. 74, 90, 117 S.Ct. 1925, 138 L.Ed.2d 285 (1997). Rather, to succeed on their claim, plaintiffs seeking the creation of an additional majority-minority district must first independently establish the existence of each of the Gingles preconditions for their proposed district. Our analysis will address the first and third.
The first Gingles precondition requires plaintiffs to demonstrate that a minority group is “sufficiently large and geographically compact to constitute a majority in a single-member district.” Gingles, 478 U.S. at 50, 106 S.Ct. 2752. The Supreme Court has explained that “[w]hile no precise rule has emerged governing § 2 compactness, the ‘inquiry should take into account traditional districting principles such as maintaining communities of interest and traditional boundaries.’ ” League of United Latin Am. Citizens (“LULAC”) v. Perry, 548 U.S. 399, 433, 126 S.Ct. 2594, 165 L.Ed.2d 609 (2006) (quoting Abrams, 521 U.S. at 92, 117 S.Ct. 1925) (internal quotation marks omitted).
As their primary proof of compactness, the plaintiffs offer two affidavits from Dr. Ronald Keith Gaddie, a political science professor from Oklahoma. (Pis.’ Resp. Ex. 16; Pis.’ Mot. Prelim. Inj. Surreply Ex. 4.) Dr. Gaddie evaluates the majority-minority districts drawn in the Fannie Lou Hamer and Pipkin plans and concludes that they are sufficiently compact to warrant the creation of a third majority African-American district. Dr. Gaddie’s analysis also indicates that the additional majority African-American districts in both plans are more compact than several of the districts in the enacted State Plan.
*899Dr. Gaddie’s statistical analysis, however, blurs the meanings of the technical term “compactness.” Within the Supreme Court’s voting rights jurisprudence, the word “compactness” refers to two distinct concepts. “In the equal protection context, compactness focuses on the contours of district lines to determine whether race was the predominant factor in drawing those lines.” LULAC, 548 U.S. at 433, 126 S.Ct. 2594 (2006). By contrast, the Gingles “compactness” inquiry focuses on the compactness of the minority population itself, not the shape of the proposed minority district. Id. Thus, the evidence of statistical compactness offered by Dr. Gaddie is of only limited relevance to our § 2 analysis.
Even focusing our review to the statistics, however, that evidence does not prove the plaintiffs’ compactness conclusion. On page six of his report on demographic trends in Maryland, the plaintiffs’ expert, Dr. Peter Morrison, provides a graph entitled “Concentration of Population by Race and Hispanic Origin in Maryland’s Jurisdiction.” (Pis.’ Resp. Ex. 18.) The graph clearly shows that Maryland has two — and only two — distinct concentrations of African-Americans, one in the D.C. suburbs and another in the Baltimore area.
The plaintiffs attempt to shore up their § 2 claim by arguing that the areas combined by their proposed third majority-minority district constitute a single “community of interest.” Among other things, they argue that the Baltimore/Washington area forms an integrated transportation corridor and that residents of Howard County — which links Baltimore and Washington in their proposed plans — are equally likely to work in Baltimore City and the Washington D.C. area.
This argument is deficient in two respects. Although the distances at issue here are not as dramatic as in some cases courts have considered — the two Latino communities connected in LULAC were 300 miles apart, for instance — the differences between the two areas are real. While Baltimore’s economy has traditionally been based on industry, medical services, and its port, Washington’s economic strength derives primarily from the federal government. The two cities may share an airport, but they have separate cultural institutions and root for rival sports teams. And most importantly for election issues, both areas are in different media markets and have different newspapers. In light of these differences, we believe the plaintiffs have not shown sufficiently that residents of their proposed additional majority-minority district form a single community of interest.
We also emphasize that for purposes of the § 2 analysis, a plaintiff must demonstrate that the minority population at issue is sufficiently compact. Bush v. Vera, 517 U.S. 952, 997, 116 S.Ct. 1941, 135 L.Ed.2d 248 (1996) (Kennedy, J., concurring). The crucial weakness in the plaintiffs’ evidence is that it concerns residents of their proposed congressional district in general, and not minority residents specifically. In the absence of this kind of specific evidence, we may not accept bare assertions that the area’s African-American residents share the same characteristics, needs, and interests. See LULAC, 548 U.S. at 433, 126 S.Ct. 2594 (“[A] State may not assum[e] from a group of voters’ race that they think alike, share the same political interests, and will prefer the same candidates at the polls” (internal quotation marks omitted)) (Op. of Kennedy, J.).
Although the failure of plaintiffs to satisfy the first Gingles precondition is sufficient to dispose of the § 2 issue, we are also skeptical that the plaintiffs have satisfied the third Gingles precondition. Commonly referred to as “racially polar*900ized voting,” this precondition requires plaintiffs to prove that “the white majority votes sufficiently as a bloc to enable it ... usually to defeat the minority’s preferred candidate.” Gingles, 478 U.S. at 51, 106 S.Ct. 2752. Courts often conduct their review of this precondition on the basis of expert testimony.
The experts for both sides have analyzed a number of federal, state, and local elections in Maryland over the past decade to evaluate trends in racial voting. The State’s expert, Dr. Bruce E. Cain, a resident of Maryland and a professor at the University of California, Berkeley, reports that African-American candidates can in fact win contested elections in predominately white areas. (Defs.’ Mot. Summ. J. Ex. 1 at 7-11.) Dr. Cain’s analysis found high levels of white support for minority candidates in several races, including the 2004 Democratic primary race for the Fourth Congressional District. Dr. Cain further notes that three statewide Democratic primary races in the past decade have paired white candidates against African-American candidates: the 2006 race for Attorney General, the 2006 race for U.S. Senate, and the 2008 race for President. In each of these cases, the African-American candidate’s share of the white vote differed significantly, suggesting that factors other than race influenced electoral decisions.
The results of the 2006 Democratic primary elections in Montgomery County are of particular interest, as they provide, what Dr. Cain terms, a natural social science experiment. In that election, African-American Attorney General candidate Stuart Simms received 25% of the white vote in Montgomery County against a white opponent. In the same election, African-American candidate for County Executive, Ike Leggett, who was also running against a well-funded white opponent, received 65% of the white vote in Montgomery County. Thus, the evidence at the local level also demonstrates a pattern of varying support for African-American candidates among the white electorate.
Dr. Gaddie offers a competing take on the electoral evidence. He contends that Barack Obama’s primary victory over Hillary Clinton was atypical and should be attributed to resource and organization. More instructive, Dr. Gaddie insists, is the low share of the white vote received by then-Congressman Kweisi Mfume, an African-American, during his 2006 U.S. Senate campaign. Dr. Gaddie also points out that most African-Americans who hold state legislative offices have been elected from majority African-American districts.
To be sure, the evidence suggests that some instances of racial voting occur in Maryland. Even Dr. Cain concludes that Maryland experiences “moderate” racial polarization. But proof of occasional racial block voting is insufficient to fulfill the third Gingles precondition, which requires the showing that the white majority must be able “usually to defeat the minority’s preferred candidate.” Gingles, 478 U.S. at 51, 106 S.Ct. 2752; see also Rollins v. Fort Bend Indep. Sch. Dist., 89 F.3d 1205, 1213 (5th Cir.1996) (concluding that although special circumstances “may explain a single minority candidate’s victory ... [e]very victory cannot be explained away as a fortuitous event”).
On the record before us, we cannot conclude that the white electorate in Maryland is sufficiently racially polarized to satisfy the third Gingles precondition for a § 2 claim.
V. Equal Protection Violations (Counts 1 and 2)
The plaintiffs also contend that Maryland’s redistricting plan violates the Four*901teenth and Fifteenth Amendment.4 They assert that the state has intentionally discriminated against minorities by splitting minority communities when drawing district lines. According to the plaintiffs, the “bizarre” shapes of some of Maryland’s congressional districts and the State’s decision to create two minority influence districts in the Second and Fifth District, are only explainable by discriminatory intent. The plaintiffs also maintain that the Pipkin map better serves many of the State’s alleged redistricting goals. Specifically, they argue that Pipkin map splits fewer county lines and “census designated places,” and that the Pipkin map concentrates Maryland’s military installations in fewer districts.
As interpreted by the Supreme Court, the Equal Protection Clause prohibits states from using race as the sole or predominant factor in constructing district lines, unless doing so satisfies strict scrutiny. Easley v. Cromartie, 532 U.S. 234, 241, 121 S.Ct. 1452, 149 L.Ed.2d 430 (2001); Bush v. Vera, 517 U.S. 952, 958-59, 116 S.Ct. 1941, 135 L.Ed.2d 248 (1996) (plurality op. of O’Connor, J.). At the same time, however, the Court has held that the Equal Protection Clause does not preclude any consideration of race in the redistricting process. Indeed, the Court has acknowledged that “[r]edistrieting legislatures will ... almost always be aware of racial demographics.” Miller v. Johnson, 515 U.S. 900, 916, 115 S.Ct. 2475, 132 L.Ed.2d 762 (1995). The question presented here is whether the State has subordinated traditional, legitimate districting principles to racial considerations. Vera, 517 U.S. at 959, 116 S.Ct. 1941.
The plaintiffs have submitted several affidavits and reports discussing lamentable incidents of racism in Maryland. They also highlight various areas on the map where they claim the line-drawing has moved African-Americans in and out of districts in order to strengthen the Democratic Party’s advantage. For example, Professor Todd Eberly, a political scientist working in Maryland, suggests that African-American voters were used to balance out the effect of Republican-leaning white voters in the Baltimore suburbs. (Pis.’ Resp. Ex. 16.)
But this evidence does not suggest, much less prove, that the political process in general or the redistricting process in particular is so infected with racial considerations that a desire to dilute African-American voting strength was the predominate factor in the creation of the State Plan.
As an initial matter, we note that the plaintiffs have not shown that the State moved African-American voters from one district to another because they were African-American and not simply because they were Democrats. Moving Democrats for partisan purposes does not establish a violation of the Fourteenth Amendment under Shaw v. Reno, 509 U.S. 630, 113 S.Ct. 2816, 125 L.Ed.2d 511 (1993), “even if it so happens that the most loyal Democrats happen to be African-American Democrats and even if the State were conscious of that fact.” Hunt v. Cromartie, 526 U.S. 541, 551, 119 S.Ct. 1545, 143 L.Ed.2d 731 (1999). Distinguishing racial from political motivations is all the more important in a State like Maryland, where the vast major*902ity of African-American voters are registered Democrats.
Moreover, the plaintiffs offer little evidence suggesting that African-Americans are especially disadvantaged by the State Plan. The State Plan makes two out of the eight congressional districts majority African-American districts. This ratio of minority/majority seats — 25%—is thus in proportion to African-Americans’ share of the total voting-age population — 28%>. The State Plan also creates two districts, the Second and the Fifth, with significant and growing minority populations. Assuming population trends remain consistent, both of these districts could conceivably elect minority candidates on the basis of majority/minority coalition voting. Thus, while the State Plan may not be maximizing African-American political power in Maryland, it does give the African-American community a strong electoral position, which will continue to strengthen according to current trends.
This result is unsurprising given that the redistricting map drew the support of many members of Maryland’s African-American community. Indeed, even a brief review of the process leading to the State Plan’s enactment reveals that Maryland’s African-American political leadership played an integral part in its creation. Both GRAC Chair Jeanne Hitchcock and GRAC member Richard Stewart are African-American. While the GRAC was developing its initial proposal, Maryland’s Legislative Black Caucus submitted two proposed redistricting plans, both of which contained only two majority African-American congressional districts. The Legislative Black Caucus also submitted a document outlining its goals for the redistricting process, which included:
Maintaining] the existing opportunity for Black voters to elect candidates of their choice in current District 1.... Maintaining] the existing opportunity for Black voters to elect candidates of their choice in current District 4.... Unifying] Prince George’s County into Districts 4 and 5.
(Defs’. Mot. Summ. J. Ex. 14.) Each of these requests was fulfilled in the enacted State Plan. Furthermore, during the legislative hearings on S.B. 1, the Redistricting Committee heard testimony from several prominent African-Americans who spoke in favor of the bill — Prince George’s County Executive Rushern Baker, Montgomery County Executive Ike Leggett, and Baltimore Mayor Stephanie Rawlings-Blake.
Of course, not every member of the African-American community supports the State Plan. The plaintiffs in this ease clearly do not. But as counsel for the State suggested at oral argument, accepting the plaintiffs’ argument that discriminatory motivations predominated in the redistricting process would require us to conclude that “the entire African-American leadership in the State of Maryland was hoodwinked.” We cannot reach such a conclusion on this record.
Our finding that the plaintiffs have failed to carry their burden of proof should not be read as a complete endorsement of the State Plan. Admittedly, the shapes of several of the districts in the State Plan are unusually odd.5 Many obvious commu*903nities of interest are divided. For instance, Baltimore City, which could be placed in one congressional district given its current population, is instead split between three. Other districts combine citizens with widely divergent interests. That a farmer in Oakland should share a Representative with a federal contractor living in Potomac is, we think, a suspect proposition.
Yet the Supreme Court has made clear that “[t]he Constitution does not mandate regularity of district shape”; rather, for strict scrutiny to apply, “traditional districting criteria must be subordinated to race.” Vera, 517 U.S. at 962, 116 S.Ct. 1941. In its briefs and during oral argument, the State offered several plausible, nonracial reasons why the districts ended up looking as they do. For example, the decision to split majority African-American Prince George’s County between two districts was necessary because the county is more populous than the ideal district. Additionally, the basic shape of some districts has not changed substantially since the last redistricting, suggesting that incumbent protection and a desire to maintain constituent relationships might be the main reasons they take their present forms.
And even if the plaintiffs are correct that the Pipkin map is more effective than the State Plan in vindicating some legitimate redistricting interests, that fact alone does not render the State Plan illegitimate. As the Supreme Court has explained, States may consider a wide variety of factors during the redistricting process. Easley, 532 U.S. at 242, 121 S.Ct. 1452 (“[T]he legislature ‘must have discretion to exercise the political judgment necessary to balance competing interests,’ and courts must ‘exercise extraordinary caution in adjudicating claims that a State has drawn district lines on the basis of race’ ” (quoting Miller, 515 U.S. at 915, 115 S.Ct. 2475)). The State’s constitutionally permissible decision to prioritize certain interests over others, without more, does not establish that racial motivations predominated in the state’s decisionmaking.
Accordingly, we conclude that the plaintiffs failed to carry their burden.
VI. Political or Partisan Gerrymandering (Count 7)
In the final count of their complaint, the plaintiffs allege that Maryland’s redistricting plan is an impermissible partisan gerrymander. Specifically, they argue that the redistricting map was drawn in order to reduce the number of Republican-held congressional seats from two to one by adding Democratic voters to the Sixth District, which covers Western Maryland and portions of the Washington, D.C. suburbs.
Although this claim is perhaps the easiest to accept factually — Maryland’s Republican Party regularly receives 40% of the statewide vote but might well retain only 12.5% of the congressional seats — it is also the plaintiffs’ weakest claim legally, if they have standing to assert it at all. Since it first recognized the issue’s justiciability in Davis v. Bandemer, 478 U.S. 109, 106 S.Ct. 2797, 92 L.Ed.2d 85 (1986), the Supreme Court has struggled to define the parameters of a successful partisan gerrymandering claim. Recent cases have reaffirmed the conceptual viability of such claims, but have acknowledged that there appear to be “no judicially discernible and *904manageable standards for adjudicating political gerrymandering claims.” Vieth v. Jubelirer, 541 U.S. 267, 281, 124 S.Ct. 1769, 158 L.Ed.2d 546 (2004) (plurality opinion of Scalia, J.); see id. at 307-08, 124 S.Ct. 1769 (Kennedy, J., concurring); see also LULAC, 548 U.S. at 447, 126 S.Ct. 2594 (op. of the Court by Kennedy, J.) (finding no “reliable measure of impermissible partisan effect”). Faced with “an unbroken line of cases declining to strike down a redistricting plan as an illegal partisan gerrymander,” Henderson v. Perry, 399 F.Supp.2d 756, 761 (E.D.Tex.2005), all of the lower courts to apply the Supreme Court’s Vieth and LULAC decisions have rejected such claims. See, e.g., Perez, et al. v. Texas, No. 11-360, slip. op. at 21-22 (W.D.Tex. Sept. 2, 2011); see also Radogno v. Ill. Bd. of Elections, No. 11-4884, slip op. at 5-7, 2011 WL 5868225 (N.D.Ill. Nov. 22, 2011) (reviewing seven proposed standards the Supreme Court has rejected).
The plaintiffs here likewise offer no reliable standard by which to adjudicate their gerrymandering claim. At best, they appear to argue for a sort of hybrid equal protection/political gerrymandering cause of action, which would be judged under the standards applicable to discrimination challenges. The Supreme Court, however, has previously dismissed similar claims, emphasizing that although “[r]ace is an impermissible classification ... [politics is quite a different matter.” Vieth, 541 U.S. at 307, 124 S.Ct. 1769 (op. of Kennedy, J.). Absent a clear standard to apply, we must reject the plaintiffs’ arguments on this count.
VII. Conclusion
Based on the foregoing findings and conclusions, we conclude that the plaintiffs have not satisfied their burden of proof with respect to any of the Counts alleged in their complaint. Accordingly, we will deny the State’s motion to dismiss the three-judge court; deny the plaintiffs’ motion for a preliminary injunction; and grant the State’s motion for summary judgment. A separate paper entering judgment for the State is filed herewith.
Judge NIEMEYER wrote the opinion, in which Judges WILLIAMS and TITUS joined. Judges WILLIAMS and TITUS wrote concurring opinions.. The GRACs five members were: Chairperson Jeanne Hitchcock, Maryland’s Secretary of Appointments, State Senate President Thomas V. Mike Miller, State House Speaker Michael Busch, Prince George’s County businessman Richard Stewart, and James King, a small business owner and former member of the House of Delegates.
. The manner in which counting prisoners where they are incarcerated results in over-representation can be seen most clearly at the local level. For example, District 1 of the Somerset County Council was created as a majority-minority district in order to settle a Voting Rights Act lawsuit brought in the 1980s. However, because the largely minority population of Eastern Correctional Institute was counted in the district’s population for redistricting purposes, only a small number of African Americans who "reside” in the district were actually eligible to vote. As a result, an African-American was not elected to fill the seat until 2010.
. Because the Federal Bureau of Prisons rejected Maryland’s Freedom of Information Act request for information on the home addresses of individuals incarcerated in federal prisons within Maryland, the MDP counted approximately 1,500 individuals in federal custody as residents of the Federal Correctional Institute Cumberland and Camp Cumberland facilities.
. It is unclear whether the Fifteenth Amendment applies to vote dilution claims like the one being brought by the plaintiffs here. The Supreme Court has raised the issue, but has not yet issued a definitive holding. See Voinovich v. Quitter, 507 U.S. 146, 159, 113 S.Ct. 1149, 122 L.Ed.2d 500 (1993). But we need not resolve this question in the present circumstances.
. Maryland's Third Congressional District merits special discussion. The District begins in Pikesville, a northwest suburb of Baltimore City; leaks eastward to capture the northeast suburbs of Baltimore City; then drops down into Baltimore City, taking a slice of the City on its way to Montgomery County, a northwest suburb of Washington, D.C.; then veers eastward in a serpentine manner to include Annapolis, a city on the Chesapeake Bay. In form, the original Massachusetts Gerrymander looks tame by comparison, as this is more *903reminiscent of a broken-winged pterodactyl, lying prostrate across the center of the State. The Third District is rated at or near the bottom of all congressional districts in multipie measures of statistical compactness. See Redistricting the Nation, Top 10, http://www. redistrictingthenation.com/top 10 .aspx.