United American Insurance Co. v. Smith, 371 S.W.3d 685, 2010 Ark. 468 (2010)

Dec. 2, 2010 · Arkansas Supreme Court · No. 10-9
371 S.W.3d 685, 2010 Ark. 468

2010 Ark. 468

UNITED AMERICAN INSURANCE COMPANY; Heartland Alliance of America Association; and Farm & Ranch Healthcare, Inc., Appellants, v. Jean SMITH and Loria Ivie, Individually and on Behalf of All Others Similarly Situated, Appellees.

No. 10-9.

Supreme Court of Arkansas.

Dec. 2, 2010.

Rehearing Denied Jan. 13, 2011.

*688Friday, Eldredge & Clark, LLP, by: Robert S. Shafer and William A. Waddell, Jr., Little Rock, and Fellers, Snider, Blankenship, Bailey & Tippens, by: Eric S. Eissenstat and Brooks A. Richardson, Oklahoma City, OK, for appellant United American Insurance Company.

Barber, McCaskill, Jones & Hale, P.A., by: Micheál L. Alexander and Rick Behring, Jr., Little Rock, for appellant Heartland Alliance of America Association.

Bourland & Kirkman, L.L.P., by: William L. Kirkman and Susanna Johnson, Fort Worth, and Huckabay, Munson, Row-lett and Moore, P.A., by: Mike Huckabay and Beverly A. Rowlett, Little Rock, for appellant Farm & Ranch Healthcare, Inc.

Zuckerman Spaeder LLP, by: Michael R. Smith, Carl S. Kravitz, Ellen D. Marcus, and Jason M. Knott, Washington, DC, John Doyle Nalley, Benton, and Brian G. Brooks, Greenbrier, for appellees.

JIM GUNTER, Justice.

|[This is an interlocutory appeal from an order of the Saline County Circuit Court certifying a class action entered on September 9, 2009. Appellants assert on appeal that the circuit court erred in certifying the class pursuant to the requirements provided in Arkansas Rule of Civil Procedure 28. Specifically, appellants contend that the circuit erred in finding that common issues of fact and law predominate over individual issues; that the circuit court erred in finding that the numerosity requirement was met where many of the class members’ claims were barred by the doctrine of res judicata; that the circuit court made improper findings on the merits; and that the class defined by the circuit court is overly broad. We affirm.

bln 2004, appellees Jean Smith and Lo-ria Ivie filed suit, individually and on behalf of all others similarly situated, against appellants United American Insurance Co. (“UA”), Heartland Alliance of America Association (“Heartland”), and Farm & Ranch Healthcare, Inc. (“F&R”). UA is an insurance company that issues limited health benefits policies, which provide capped or limited coverage for certain hospital, surgical, or other medical costs. Heartland issues membership providing discounts, sometimes called “repricing,” for medical and related products and services. F&R is an insurance agency whose agents sell UA policies, Heartland memberships, life insurance, and other insurance-related products.

In their second amended complaint, filed on July 22, 2008, appellees alleged that appellants sold UA health insurance policies to appellees and other class members by systematically misrepresenting the scope and nature of the policies. Specifically, appellees maintained that F&R agents advertised UA’s limited benefits policy as better than or equivalent to a major medical policy when coupled with a Heartland membership; misrepresented that a membership in Heartland and a life insurance policy were included free with a UA policy; and explained that a six-dollar charge was a one-time application fee when, in fact, it was charged monthly. Appellees asserted four causes of action in their complaint, including breach of contract, violation of the Arkansas Deceptive Trade Practices Act, unjust enrichment, and common law fraud.

On February 17, 2009, appellees filed a motion for class certification, requesting that the circuit court define the class as all *689persons who, between January 1998 to the present, |swere residents of Arkansas, Louisiana, Texas, Georgia, or California and purchased certain UA limited benefits policies coupled with Heartland memberships from F&R agents. Appellees maintained that for nearly a decade, F&R Area Marketing Director John Mills was responsible for recruiting, training, and supervising sales agents in the five indicated states and that he designed and implemented a sales presentation that was highly effective at perpetrating the alleged fraud. Appellees alleged that Mills distributed an instructive manual called “the Collage” that agents in his region were required to study and utilize and that F&R agents were repeatedly encouraged to implement Mills’s techniques in their sales presentations in order to be successful. The Collage included various sales presentation illustrations an agent could use to make a sale that appellees contend misrepresented the worth and benefits of the limited benefits policies, including “working the T,” “milking the cow,” the “calculator method,” and an “explanation of benefits” illustration to show the effect of a Heartland membership on a medical bill.

Appellees attached several exhibits to their motion for class certification, including the deposition testimony of F&R agents Glenn Bearden, Chris Bailey, and Ronnie Walthall. Also included were emails that Bearden and Bailey sent in 2004 to the F&R president, their regional manager, and their area marketing director. In his email, Bearden recalled that he had attended the required training school for new agents in Fort Worth, Texas, and that after being subjected to four days of “strenuous” training, he began to have doubts about the coverage of the UA limited benefits policy. He was concerned that new agents were not |4allowed to see an actual policy but were taught to sell it using the “milking the cow” and “caleulator method” illustrations. He explained that after selling for a couple of months, he finally saw a policy and realized that “the way we were taught how the policy pays out and what it actually does is a blatant misrepresentation.” He stated that when he realized this, he immediately ceased selling the policies the way he had been instructed and, consequently, was unable to make a sale.

In his email, Chris Bailey explained that he felt uncomfortable with the new-agent training he had received in Fort Worth, including the use of several illustrations that were designed to “put the customers’ mind at ease about the policy and explain to them why these hospital surgical policies were ‘better’ than major medical.” Bailey stated that when he returned from new-agent training, he had one “ride along” with his district manager, Barry Collinsworth. He felt Collinsworth misrepresented the policy to the client they met with that day. Bailey shared the same concern as Bearden regarding not being shown an actual policy, so he purchased one for himself. Once he read the policy, he realized that “all of the questions that we had were not being answered correctly and that the way we were taught was not a true representation of what the policy actually does.”

In his deposition testimony, Ronnie Walthall stated that during his new-agent training, he was never shown a UA policy and no agent he knew had seen an actual policy. He testified that the Collage was edited, updated, and distributed to agents regularly and that they were encouraged repeatedly to utilize it. Walthall stated that he was taught to sell UA limited | .^benefits policies by comparing them to major medical policies with the purpose of making “the [UA] policies look more attractive than major medical.” He explained that several of the illustrations *690contained in the Collage were intended to show how a customer could replace existing major medical coverage with a UA limited benefits policy, including the “working the T” illustration that he used regularly to make sales. He stated that he was told to instruct customers purchasing UA policies that they would receive a telephone call following their purchase explaining that it was not the same as major medical coverage. He testified that he was taught to tell those customers not to “let it freak you out.” He also testified that he was encouraged to bundle UA policies with a Heartland membership to receive his commission and bonus and that he was instructed not to let customers see the separate rates being figured for the items in the bundle. He testified that he sold UA limited benefits policies for four years using the training he was provided by appellants.

Smith and Ivie both live in Arkansas and purchased UA health insurance policies coupled with Heartland memberships and life insurance policies from F&R agents who had been trained by Mills. Smith and Ivie claim that the agents who sold those policies represented that the UA policies and Heartland memberships would provide more coverage than the major medical policies Smith and Ivie already owned. Based on those representations, Smith and Ivie cancelled their existing major medical health insurance, later incurred substantial medical bills, and were informed, that their UA policies would not pay their expenses. Appellees allege that at least 25,000 other persons living in the five-state area suffered similar 1 (¡damage.

In its fifty-two-page order granting certification of the class, the circuit court found that the requirements of Arkansas Rule of Civil Procedure 23 were satisfied. The court found that the class of persons affected could be upwards of 25,000 persons, which would be so numerous that joinder of all members was impractical. With regard to numerosity, the court specifically noted that it had previously denied Heartland’s motion for summary judgment based on res judicata. In that motion, Heartland contended that all class members other than Smith and Ivie were foreclosed from suit because of the settlement of a lawsuit in Texas. The circuit court stated that the claims in the present case were different from those litigated in the Texas case and that even if res judicata applied, it was an affirmative defense on the merits, which were not to be considered for class certification.

As to commonality, the circuit court found several questions of law and fact common to the class, including the following: whether F&R deliberately recruited new agents without insurance knowledge and experience; whether agents were trained to bundle UA policies with Heartland membership and life insurance policies, represent that package as equal to or better than major medical coverage, and disguise the price of the bundle so that customers thought the Heartland membership and life insurance policy were included free; whether F&R deliberately failed to provide its agents with copies of UA’s policies and to inform them of the actual terms of those policies; whether Mills trained agents in the five-state area to use the sales presentation provided in the Collage and required adherence to that presentation; 1 ./whether the sales presentation provided in the Collage was deceptive; whether F&R’s compensation plan for agents was intentionally designed to encourage the bundling of UA policies with Heartland memberships and life insurance policies; whether agents misled customers to believe UA limited benefits policies coupled with Heartland memberships provided the same or better coverage than major medical policies; whether agents misled *691customers to believe that the price of a UA policy included a Heartland membership and a life insurance policy; whether the defendants allowed their agents to sell UA policies as replacements for pre-existing major medical policies; whether the defendants are liable for fraud for misrepresenting the coverage and price of a UA policy; whether the defendants violated the Arkansas Deceptive Trade Practices Act; whether the defendants were unjustly enriched by their actions; and whether the defendants breached oral contracts with class members when UA policies turned out to be less beneficial than major medical policies.

The court also found that the claims and defenses were typical for all members of the class and that the representative parties and counsel would fairly and adequately protect the interests of the entire class. Furthermore, the court found that common questions of law and fact predominated over questions affecting only individual members of the class, specifically whether appellants engaged in a uniform scheme to convince customers that a UA-Heartland bundle was superior to major medical coverage and whether appellants misrepresented to customers that a Heartland membership and life insurance policy were included free with the purchase of a UA policy. Finally, the circuit court found that a class-action lawsuit was the |Rsuperior method for adjudicating the claims.

In its order certifying the class action, the circuit court defined the class as

[a]ll persons who, between January 1998 and the present, were residents of Arkansas, California, Georgia, Louisiana or Texas and during the time that they were residents of those states purchased through Farm & Ranch, (i) a health insurance policy issued by United American known as Flexguard Plan, CS-1 Common Sense Plan, GSP Good Sense Plan, SHXC Surgical & Hospital Expense Policy, HSXC 7500 Hospital/Surgical Plan, MMXC Hospital/Surgical Plan, SMXC Surgical/Medical Expense Plan and/or SSXC Surgical Safeguard Expense Plan, and (ii) a membership in Heartland.

We have held that circuit courts are given broad discretion in matters regarding class certification and that we will not reverse a circuit court’s decision to grant or deny class certification absent an abuse of discretion. Union Pac. R.R. v. Vickers, 2009 Ark. 259, 308 S.W.3d 573. When reviewing a circuit court’s class-certification order, we review the evidence contained in the record to determine whether it supports the circuit court’s decision. Id. We have stated:

[T]he issue of whether to certify a class is not determined by whether the plaintiff or plaintiffs have stated a cause of action or will prevail on the merits, but rather whether the requirements of Rule 23 are met. We have also observed that it is totally immaterial whether the petition will succeed on the merits or even if it states a cause of action. An order denying or granting class certification is separate from the merits of the case. Although we do not delve into the merits of the underlying claims in a potential class-action case, we will review the.trial court’s order to determine whether the requirements of Rule 23 are satisfied.

Am. Abstract & Title Co. v. Rice, 358 Ark. 1, 9, 186 S.W.3d 705, 710 (2004) (internal citations omitted).

Rule 23 of the Arkansas Rules of Civil Procedure governs class actions and class certification. The rule provides, in pertinent part, as follows:

l/a) Prerequisites to Class Action. One or more members of a class may sue or be sued as representative parties on be*692half of all only if (1) the class is so numerous that joinder of all members is impracticable, (2) there are questions of law or fact common to the class, (3) the claims or defenses of the representative parties are typical of the claims or defenses of the class, and (4) the representative parties and their counsel will fairly and adequately protect the interests of the class.
(b) Class Actions Maintainable. An action may be maintained as a class action if the prerequisites of subdivision (a) are satisfied, and the court finds that the questions of law or fact common to the members of the class predominate over any questions affecting only individual members, and that a class action is superior to other available methods for the fair and efficient adjudication of the controversy.

This court has reviewed the provisions of Rule 28 on numerous occasions and has held that, in order for a class-action suit to be certified, six factors must be met. Specifically, the party seeking certification must establish (1) numerosity; (2) commonality; (3) predominance; (4) typicality; (5) superiority; and (6) adequacy. Williamson v. Sanofi Winthrop Pharm., Inc., 347 Ark. 89, 60 S.W.3d 428 (2001).

Each of the three separate appellants in this case filed a separate brief asserting various points on appeal. All three appellants argue that the circuit court erred in finding common issues of law and fact predominated over individual issues. Heartland maintains that all plaintiffs except Smith and Ivie are barred by the doctrine of res judicata from relitigating the issues and that after removing those other plaintiffs, the numerosity requirement is not met. F&R argues that the class defined in the order was overly broad. Finally, UA argues that the circuit court made merits findings in its order.

I. Predominance

Appellants contend that the findings that the circuit court made to support | ^predominance are not substantiated by the evidence. Appellants claim that liability will hinge on the specific sales presentations given during face-to-face meetings with each member of the class by different F&R agents and that a uniform, class-wide establishment of proof on that issue is impossible.

Rule 23(b) requires that “the questions of law or fact common to the members of the class predominate over any questions affecting only individual members.” Predominance is a more stringent requirement than commonality. Baker v. Wyeth-Ayerst Labs. Div., 338 Ark. 242, 992 S.W.2d 797 (1999).1 In Georgia-Pacific Corp. v. Carter, 371 Ark. 295, 265 S.W.3d 107 (2007), this court noted that the starting point in examining the predominance issue is whether a common wrong has been alleged against the defendant. This court has repeatedly recognized that conducting a trial on the common issue in a representative fashion can achieve judi*693cial efficiency. Ark. Blue Cross & Blue Shield v. Hicks, 349 Ark. 269, 78 S.W.3d 58 (2002). If a case involves preliminary, common issues of liability and wrongdoing that affect all class members, the predominance requirement of Rule 23 is satisfied even if the circuit court must subsequently determine individual issues in bifurcated proceedings. ChartOne, Inc. v. Raglon, 373 Ark. 275, 283 S.W.3d 576 (2008). We have recognized that a bifurcated process of certifying a class to resolve preliminary, common issues and then decertifying the class to resolve individual issues is consistent with Rule 23. Id. In addition, we have said that

[t]he predominance element can be satisfied if the preliminary, common issues may be resolved before any individual issues. In making this determination, we do not merely compare the number of individual versus common claims. Instead, we must decide if the issues common to all plaintiffs “predominate over” the individual issues, which can be resolved during the decertified stage of bifurcated proceedings.

Carter, 371 Ark. at 301, 265 S.W.3d at 111. If, however, the preliminary issues are individualized, then the predominance requirement is not satisfied. Raglon, 373 Ark. at 286, 283 S.W.3d at 585.

Appellants claim that this court’s decision in Union Pacific Railroad v. Vickers, 2009 Ark. 259, 308 S.W.3d 573, is disposi-tive of the predominance issue in this case. In Vickers, the class plaintiffs alleged that Union Pacific employees practiced law without authorization through its practice of reaching settlements and obtaining releases from individuals who had claims against the company. We held that although there was testimony that Union Pacific had a general plan or goal in handling claims, there was no common pattern of facts to be applied to every case. Because Union Pacific claims representatives would have to consider various factors and issues for each claimant in reaching a settlement, we held that individual issues predominated over common ones.

However, in Lemarco, Inc. v. Wood, 305 Ark. 1, 804 S.W.2d 724 (1991), we addressed a fact scenario more similar to our present case. In Lemarco, a class estimated at 800 members | ^was certified that consisted of persons who entered into retail installment membership contracts with Lemarco, a private buyers club. Appel-lees asserted that Lemarco defrauded them of membership fees by promising free gifts for attending sales presentations and misrepresenting the value of membership. This court affirmed the trial court’s decision on the basis that predominance was satisfied because there were common issues of fact arising out of Lemarco’s sales training, solicitation mailing, sales presentations, installment contracts, and its intentions to defraud. We concluded by discussing the predominance and superiority questions in tandem. We held that common issues should be resolved first to attain real efficiency. We then stated that

[e]ven if the trial court eventually determines that the cases have to splinter with respect to some individual claims, efficiency would still have been achieved by resolving those common questions which predominate over individual questions. Finally, pursuing this case as a class action is fair to both sides. Le-marco can offer evidence concerning its sales training, solicitation mailings, sales presentations, installment contracts, and its intentions regarding all of them. It can also present individual defenses to the claims of individual class members, if necessary, once the common questions have been determined. A class action approach is also fair to the class members in that they probably would not sue *694if they could not do so as a class since it would not be economically feasible to do so.

305 Ark. at 4, 804 S.W.2d at 726.

Additionally, in SEECO, Inc. v. Hales, this court affirmed a lower court’s predominance finding where “[t]he issue of a fraudulent scheme is central ... and a common starting point for all class members.” 330 Ark. 402, 412, 954 S.W.2d 234, 240 (1997). There, the class plaintiffs sought recovery based on a fraud theory that required proof of reliance by each plaintiff. Id,., 954 S.W.2d at 240. We noted that fraud cannot be presumed and must be | ^established by clear and convincing evidence. Id. However, after citing to our own precedent and persuasive authority from other states, we concluded that the individualized reliance element did not defeat the class certification because the overarching issue was the existence of the alleged scheme and that the defendants would be allowed to present their defense asserting lack of reliance at the merits stage of the litigation. Id. at 413-14, 954 S.W.2d at 240-41; see also Arthur v. Zearley, 320 Ark. 273, 895 S.W.2d 928 (1995) (reversing a trial court’s certification of a class where individual issues regarding informed consent essential to the claims of each class member predominated over questions common to the class); Summons v. Mo. Pac. R.R., 306 Ark. 116, 813 S.W.2d 240 (1991) (concluding that trial court erred in not granting class certification where repeated litigation on the liability question outweighed the fact that each claimant would have different damage evidence); Int’l Union of Elec., Radio & Mach. Workers v. Hudson, 295 Ark. 107, 747 S.W.2d 81 (1988) (affirming class certification where defendant could present defense to the claims of individual class members after those members had presented evidence of general question of liability); Ark. La. Gas Co. v. Morris, 294 Ark. 496, 744 S.W.2d 709 (1988) (holding that even where evidence established individual questions relating to reliance or mistake, the trial court could defer those questions until after it had disposed of the questions common to the class).

In addressing predominance in the present case, the circuit court found that the case “at its center involves a claim of class-wide oral fraud pursuant to uniform training, uniform corporate policies[,] and uniform written scripting.” The circuit court found that appellees had | upresented sufficient evidence to show that appellants were potentially liable for a scheme to uniformly market that the UA-Heartland “bundle” was equal to or better than coverage provided by major medical insurance and that a Heartland membership and life insurance policy were free with the purchase of a UA health policy. The court found that those two issues were the overarching ones in the litigation. The court pointed to the evidence that it believed was common to all class members to prove those allegations: the training given to F&R agents in Fort Worth; the failure to provide agents with actual copies of UA policies; the five-state-specific training class provided by Mills after F&R’s other required training; the policy of F&R to never allow checks to be made out to Heartland; the incentives to agents to sell UA-Heartland bundles; and the written training materials known as the Collage. The circuit court dismissed appellants’ argument that predominance was destroyed because each potential customer had to make a purchase decision based on an individualized, oral, in-home presentation by a sales agent. The court found that there was sufficient evidence that appellants had trained, supervised, and encouraged their agents to create “the two core impressions concerning superiority to major medical coverage and the inclusion of *695free products with the purchase of a UA [policy].” The circuit court recognized that appellants might defend the lawsuit on the basis that not every agent executed the same sales script, but the circuit court found that such a defense did not destroy predominance. The court acknowledged our holding in Vickers, but noted that this case involved more than a general company goal or plan and instead involved demonstrative evidence of a policy and scheme |1sto convince potential customers of the two core impressions.

We hold that the circuit court did not abuse its discretion in finding that common issues of law and fact predominate over individual issues in this case. Although individual issues may be raised by appellants to defend against appellees’ claims of fraud and misrepresentation, those challenges. do not override the crucial issue common to all members of the class— whether appellants engaged in a general scheme to defraud and misrepresent to potential customers that a UA-Heartland bundle was equal to or better than major medical coverage and whether appellants falsely implied that a Heartland membership and life insurance policy were included free with the purchase of a UA healthcare policy.

II. Res Judicata/Numerosity

Heartland asserts that the circuit court erred in “failing to apply res judicata in this case” and abused its discretion in certifying a class action because if it had applied the doctrine of res judicata, the class membership would not be so numerous as to require certification. Appellees respond by asserting that an interlocutory appeal from class certification is not appropriate for the evaluation of affirmative defenses, that Heartland is attempting to argue res judicata at this stage as an attempt to appeal from the denial of its summary judgment motion, and that res judicata does not bar the present action.

In January 2005, Eric B. Rodriguez filed a class action and derivative suit in Texas against appellants. That lawsuit, Rodriguez v. Burdine, Case No. DC 2005-08, Starr County District Court, sought compensatory damages and injunctive relief based on allegations that 11fiappellants had fraudulently concealed Heartland membership prices in UA policy premiums. It resulted in a national class certification and settlement. The class was defined as all persons who were Heartland members between January 1, 2001, and June 5, 2006, except for any individual (1) who has a lawsuit pending against any defendant as of June 5, 2006, and is not represented by the plaintiffs counsel; (2) who, while represented by counsel, previously signed a document that releases defendants, or any of them, from any of the released claims; or (3) as to all defendants except Heartland, who timely submit a proper request for exclusion. On October 6, 2006, the judge in Rodriguez signed a final order and judgment finding that, as to Heartland, the class should be designated as mandatory and that no class member could be excluded from the settlement as to Heartland.

In the present case, the circuit court noted in its certification order that it had previously denied Heartland’s motion for summary judgment based on res judicata wherein it found that the claims in Rodriguez were not the same as the ones asserted by appellees and that the Rodriguez settlement reserved and did not release the claims in this case. The circuit court found that regardless, the issue of res judicata was an affirmative defense that would be taken up at the merits stage, not the certification stage.

This court has been clear it does not delve into the merits of the un*696derlying claims when addressing certification. FirstPlus Home Loan Owner 1997-1 v. Bryant, 372 Ark. 466, 277 S.W.3d 576 (2008). A trial court may not consider whether the plaintiffs will ultimately prevail, or even whether they have a cause of action. Id. Moreover, in Fraley v. Williams Ford Tractor & Equipment Co., 339 Ark. 322, 5 S.W.3d 423 (1999), we held that evaluating the merits of an affirmative defense was not appropriate at the certification stage. There, the trial court found that the numerosity requirement had not been satisfied because many putative class members had released Williams Ford from liability and because the releases were properly obtained by Williams Ford. The Fraleys argued that the trial court’s findings regarding the affirmative defenses of release and consent were an improper intrusion into the merits of the case at the certification stage. In deciding in the Fraleys’ favor, we noted that certification challenges based on the statute of limitations, fraudulent concealment, releases, causation, or reliance do not bar satisfaction of Rule 23(b)’s predominance requirement. Id.; see also SEECO, Inc. v. Hales, 330 Ark. 402, 954 S.W.2d 234 (1997).

Furthermore, this court in Direct General Insurance Co. v. Lane, 328 Ark. 476, 944 S.W.2d 528 (1997), held that a party could not acquire impermissible interlocutory review of the denial of a summary judgment motion by arguing the merits of an affirmative defense in an appeal from a class certification. The denial of a motion for summary judgment is not a final order and is not appropriate for interlocutory appeal. Id. In Lane, Direct General attempted to reargue on appeal from certification its summary-judgment assertion that the plaintiffs claim should be rejected on the merits because he failed to prove damages. Id. We refused to consider the merits of the defense and explained that an interlocutory appeal from class certification was not an appropriate vehicle for relitigating issues denied on summary judgment. Id.

|,8We affirm the circuit court’s decision on this point. First, determining whether the doctrine of res judicata applies to the claims asserted against Heartland would require this court to evaluate the merits of that affirmative defense, and this court has been clear it will not delve into the merits on appeal from certification. Second, the issue as to whether claims asserted by appellees against Heartland are barred by the doctrine of res judicata was ruled upon by the circuit court when it denied Heartland’s motion for summary judgment. Our law does not permit Heartland to file an interlocutory appeal from that denial. Rather, it must be taken up on direct appeal after a final order has been entered. Therefore, it is not proper to reliti-gate that issue at this stage.

III. Class Definition

F&R asserts in its separate brief on appeal that the circuit court’s definition of the class is overly broad because there is no evidence to support that any alleged fraudulent training or sales took place after 2004; that John Mills ceased doing training at the Fort Worth training school in March 2004; that he ceased supervising agents in Arkansas, Texas, and Louisiana in September 2006; and that he ceased supervising agents in California and Georgia in April 2004. Because appellees’ claims are based significantly on the alleged fraudulent training and supervision of Mills,.F&R maintains that the class should have been limited to the time period and geographic location corresponding to his direct supervision. Appellees respond and claim that F&R’s point is not a basis for reversing the certification order because *697the circuit court can trim the class definition to conform to the evidence presented if it later becomes 119necessary.

In order to maintain a class action, there must be a defined class that will make it administratively feasible for a court to determine membership in the class. Ark. Blue Cross & Blue Shield v. Hicks, 349 Ark. 269, 281, 78 S.W.3d 58, 65 (2002). In other words, class identity must be feasible, and the class cannot be excessively broad or amorphous. Id. Clearly defining the class insures that those people who are actually harmed by the defendant’s wrongful conduct will participate in the relief ultimately awarded. Id.

The circuit court did not abuse its discretion in this regard. Although the training and supervision provided by Mills was an integral part of the allegations of fraud and misrepresentation, his participation alone was not the only factor. Furthermore, many of the training materials he utilized were distributed to agents he did not directly come into contact with but were filtered to them by their managers and supervisors. The main determination for class size is that it not be excessively broad. However, the class should include all those that might have been harmed. F&R argues that the class defined by the circuit court in this case exceeds the proof presented. However, the circuit court clearly evaluated the claims made and attempted to limit the class based on the evidence it had before it. While the class may later be limited once the merits of the claim are evaluated, the circuit court clearly used its discretion to try to contain the class but to make sure its limitations did not exclude any harmed party.

IV. Merit Findings

hnUA maintains in its separate brief on appeal that the circuit court’s certification order improperly includes findings on the merits of the appellees’ claims. We have held that a circuit court is not to delve into the merits of the underlying claim when deciding whether the procedural strictures of Rule 23 have been met. Advance Am., Cash Advance Ctrs. of Ark., Inc. v. Garrett, 344 Ark. 75, 40 S.W.3d 239 (2001). Here, however, the circuit court specifically stated in its order granting certification that it was “not ruling on the merits at this stage or foreclosing any merit issues” and that its “findings of fact and conclusions of law” were “with respect to class certification.” From our review of the certification order, we are convinced that the circuit court avoided making a decision on the merits of appel-lees’ claims and instead evaluated the evidence put forth by appellees to examine only the question of class certification. In so much as any findings of the circuit court are alleged to be determinations of merit, the circuit court expressly stated those findings were not to be considered merit findings.

Affirmed.

Special Justice JERRY LARKOWSKI joins in this opinion.

CORBIN, DANIELSON, and WILLS, JJ., dissent.

SHEFFIELD, J., not participating.

ELANA CUNNINGHAM WILLS, Justice,

dissenting.

Because I believe that Rule 23’s predominance requirement is lacking in this case, I respectfully dissent. The quantum of | ¾1 evidence in this case falls somewhere between the evidence presented in Union Pacific Railroad v. Vickers, 2009 Ark. 259, 308 S.W.3d 573, on the one hand, and Lemarco, Inc. v. Wood, 305 Ark. 1, 804 S.W.2d 724 (1991), on the other. It is the *698quality, however, rather than the quantity of the evidence that is critical.

In Vickers, swpra, this court stated that the common “goal” or plan of the railroad for settling claims in that case was not sufficient, standing alone, to establish a uniform set of operative facts, such that common issues would predominate over individualized ones. Vickers, 2009 Ark. 259, at 18, 308 S.W.3d at 582. The actual encounters between the railroad representatives and the potential plaintiffs in Vick-ers contained individualized facts and thus rendered the case improper for class treatment.

In Lemarco, supra, the defendant’s president testified that the salesmen who were selling memberships in a private buyers club were trained by videotape and were instructed on what to say and how to make the sales presentations. 305 Ark. at 4, 804 S.W.2d at 726. Notably, however, he also testified that the sales presentations were “virtually the same each time,” and that the mail solicitations sent to prospective members and the retail sales contracts were “virtually the same.” Id. This last-mentioned testimony is not discussed by the majority and clearly distinguishes this case from Lemarco.

In this case, there is ample evidence of a uniform plan or scheme for selling the insurance products — that is, this case is like Lemarco in the sense that there is evidence that the salesmen were trained to say the same thing and make the same presentation to potential |22buyers. But in my view, that is not the determinative factor for purposes of the predominance requirement. As the circuit court found, this case “at its center involves a claim of class-wide oral fraud.” It does not matter how ample or how specific the evidence is as to the salesmen’s training. It matters whether that training was uniformly implemented in the actual oral encounters with potential buyers. Only then is there one set of operative facts, which would lead to common issues predominating over individualized ones.

There was evidence in Lemarco, by way of the company president’s admission, that the sales training and the plan for selling the private buyers club memberships was uniformly implemented. That element was missing in Vickers, and it is missing in this case. Certainly, we did not hold in Vickers that a case alleging oral misrepresentations or fraud can never be certified as a class action. Nor do our decisions, including Lemarco v. Wood, require defendants to admit that sales presentations or other relevant transactions were virtually the same in each case. But there has to be evidence that the plan or scheme was uniformly implemented, such that there is a common set of operative facts.1 That evidence is simply missing in this case.

The circuit court found, in paragraph 28 of its “Findings of Fact” that “[a]s found more fully below, Mills succeeded in personally training all of his agents in a uniform, scripted sales [^presentation that they presented in a substantially uniform way in all sales calls.... ” This finding is unsupported by any citation to the evidence. At the hearing below, the trial judge inquired as to what evidence showed *699that the class members heard the same sales presentation. Appellees’ counsel responded, “We contend that the Farm and Ranch agents would not have been able to sell these policies by selling them honestly,” and counsel agreed with the trial court’s restatement of their argument that “had they not heard the same thing, there is no way they would have bought.” This is not evidence.

On appeal, the appellees continue this contention, repeatedly asking this court to infer a similar pattern of operative facts: “Because the central message drummed into agents was fraudulent, it is reasonable to conclude that agents conveyed that message to customers, regardless of the precise words used with any particular customer”; “That customers were led to act against their own interests is evidence that agents misled them”; “Evaluation of the effect of Appellant’s fraud training has to take notice of the complete absence of training in anything other then fraud, providing further support for the finding that agents used the sales tools they have been given”; and “A customer would surrender major medical coverage only filed to a contrary understanding — that the UA-Heartland package is as good or better than that coverage.” (Emphasis added.)

In my view, it is error to infer this crucial link in determining whether class certification is warranted. I would therefore hold that the circuit judge abused his discretion and would reverse the certification order, rendering consideration of the other issues on appeal | ^unnecessary. See Vickers, 2009 Ark. 259, at 20 n. 4, 308 S.W.3d at 583 n. 4.

CORBIN and DANIELSON, JJ., join in this dissent.