SAM D. JOHNSON, Circuit Judge:
This is an appeal in the second trademark infringement action brought by Exxon Corporation against Texas Motor Exchange of Houston. The central issue in this appeal is whether there is a likelihood of confusion between the trademark EXXON and the words Texon or Tex-On. The district court found that there was no likelihood of confusion and refused to enjoin defendant from using either Texon or Tex-On. We affirm in part and reverse in part.
[502]*502I. Facts
In late 1972 and early 1973 the owner of Texas Motor Exchange of Houston began searching for a new name for his corporation. He felt that Texas Motor Exchange was too common a name in Texas. After some thought and consultation with his attorney, the owner settled on the name Texxon. In May 1973 he established a new corporation-Texxon Motor Centers, Inc. At this point the owner began using all forms of advertising to promote the new corporation. He substituted Texxon for Texas in two large neon signs located in front of the former place of business so that both of the signs, which previously read Texas Motor Exchange, read Texxon Motor Exchange. The owner also began using Texxon as the most prominent name in the large painted sign appearing on one side of his place of business. Finally, Texxon was prominently displayed in newspaper and yellow page advertisements, and Texxon Motor Exchange was used as the firm’s trade name in radio advertising. Unfortunately for the owner, his effort to find an innovative name that would attract more business served as the basis for the instant cause of action.
In November 1972 Standard Oil Company changed its name to Exxon Corporation. Exxon Corporation owns several federal registrations of EXXON as a trademark for petroleum and other automotive products that it markets.1 Additionally, after adopting EXXON as its national trademark and changing its name in November 1972, Exxon Corporation went to great expense to advertise and promote its EXXON mark and name throughout the United States. It is safe to say that this effort has made EXXON virtually a household phrase.
In February 1975 Exxon Corporation filed an action for trademark infringement and unfair competition against Texxon Motor Center. Exxon Corporation sought damages and, most importantly, a judgment restraining defendant from using Texxon or Texon in connection with its business. The district court held a hearing in January 1977. The court found that defendant’s use of Texxon infringed plaintiff’s rights in its federally registered EXXON mark and name and that defendant had engaged in unfair competition. The trial judge then issued the following order:
1. Defendant Texxon Motor Centers Incorporated be, and hereby is, affirmatively directed to:
(a) Change its corporate name by eliminating “Texxon” therefrom;
(b) Remove TEXXON from any and all signs painted on or otherwise affixed to the building in which said defendant’s place of business is located; and
(c) Deliver up for destruction all other signs, labels, prints, packages, wrappers, receptacles, and advertisements in its possession on which TEXXON appears.
2. Defendants Texas Motor Exchange of Houston, Inc. and Texxon Motor Centers Incorporated and their officers, agents, representatives, employees, attorneys and any and all persons in active [503]*503concert or participation with them be and hereby are, permanently enjoined and restrained from:
(a) Using TEXXON as part of any corporate or trade name or mark or in the advertising, offering for sale or sale of any services or products; or
(b) Using any signs, advertising or promotional materials of any kind or nature whatsoever on which the trade name or mark is displayed in red letters on a white background above a blue bar or rectangle or in any other manner imitative of plaintiff’s distinctive EXXON logotype.
In accordance with the district court’s order, the defendant ceased using the name Texxon on any of its signs or in any of its advertisements. In its place, defendant began using the marks Texon and Tex-On. Defendant, however, did not formally change its corporate name by eliminating Texxon. Instead, defendant conducted business under the names Texon and Tex-On and decided to wait to formally change its name until Exxon Corporation filed a second lawsuit and the district court ruled on the propriety of the use of either Texon or Tex-On. As the defendant suspected, in September 1978 Exxon Corporation moved for a modification and enforcement of the permanent injunction. Specifically, plaintiff urged the court to modify the injunction and prohibit defendant from using Texon or Tex-On. In November 1978 the district court held a second hearing, this time on plaintiff’s motion to modify the permanent injunction. The court ruled that the names Texon and Tex-On do not infringe upon plaintiff’s EXXON mark and do not constitute unfair competition. The district court denied plaintiff’s motion to modify the permanent injunction. Plaintiff then instituted this appeal.
II. Standard for Modification of the Injunction
Exxon Corporation seeks to modify the original injunction and impose additional limitations upon the defendant.2 The standard that Exxon Corporation must meet in order to justify the imposition of additional restrictions is, under present law, uncertain. The starting point for this Court’s analysis must be the Supreme Court’s decision in United States v. United Shoe Machinery Corp., 391 U.S. 244, 88 S.Ct. 1496, 20 L.Ed.2d 562 (1968).
United Shoe Machinery involved the Government’s effort to take stronger steps to remedy the corporation’s monopolization of the shoe machinery market. The district court, which issued the initial injunction enjoining the corporation from monopolizing the shoe machinery business, refused to modify the injunction at the Government’s request. The district court held that its power to modify was limited to cases involving “(1) a clear showing of (2) grievous wrong (3) evoked by new and unforeseen conditions.” United States v. United Shoe Machinery Corp., 266 F.Supp. 328, 330 (1967), citing United States v. Swift & Co., 286 U.S. 106, 52 S.Ct. 460, 76 L.Ed. 999 (1932). On appeal the Supreme Court rejected the use of the Swift test for deciding whether to impose additional restrictions on a defendant. The Supreme Court instead instructed the district court to determine whether the relief originally ordered had produced the intended results. “If it has not, the District Court should modify the decree so as to achieve the required result with all appropriate expedition.” 391 U.S. at 252, 88 S.Ct. at 1501.
The holding in United Shoe Machinery indicates that an injunction may be modified to impose more stringent requirements on the defendant when “the original purposes of the injunction are not being fulfilled in any material respect.” 11 C. Wright and A. Miller, Federal Practice and Procedure § 2961 (1973). In the case at bar, the district court designed the original injunction to prevent defendant from inf ring[504]*504ing on plaintiff’s rights and engaging in unfair competition. Thus, if plaintiff can establish that the new marks adopted by the defendant, Texon and Tex-On, also infringe on plaintiff’s rights in its federally registered EXXON mark, then appropriate relief should be granted.
III. Likelihood of Confusion
15 U.S.C. § 1114(1) governs lawsuits for the infringement of a registered mark. That section provides that a defendant shall be liable for the use of an imitation of a registered mark if the use “is likely to cause confusion, or to cause mistake, or to deceive.” Id. The standard established by this provision is known as the “likelihood of confusion” test. In determining whether there is a likelihood of confusion, the finder of fact evaluates a variety of factors: the type of trademark, the similarity of design, the similarity of the product, the identity of retail outlets and purchasers, the similarity of advertising media used, defendant’s intent, and actual confusion. Roto-Rooter Corp. v. O’Neal, 513 F.2d 44, 45 (5th Cir. 1975). A finding on likelihood of confusion is a fact finding and is reviewed by this Court under the clearly erroneous test. T.G.I. Fridays v. International Restaurant Group, Inc., 569 F.2d 895 (5th Cir. 1978); Amstar Corp. v. Domino’s Pizza, Inc., 615 F.2d 252 (5th Cir. 1980). This Court’s review of the record and evaluation of the factors indicates that the trial court was clearly erroneous in concluding that there was not a likelihood of confusion between EXXON and Texon, but that the trial court did not err when it held that there was no likelihood of confusion between EXXON and Tex-On.
A. Type of Trademark
The first factor to be considered is whether EXXON is a “strong” or a “weak” trademark. A strong trademark is one that is rarely used by parties other than the owner of the trademark, while a weak trademark is one that is often used by other parties. In short, the more distinctive a trademark, the greater its “strength.” The strength of a trademark is important in determining the scope of protection that is granted. “The greater the number of identical or more or less similar trade-marks already in use on different kinds of goods, the less is the likelihood of confusion ... . ” Restatement of Torts § 729, Comment g (1938).
Our review of the record establishes that EXXON is a strong trademark. There is no evidence in the record indicating that any other party besides Exxon Corporation uses EXXON. Defendant did offer evidence in an attempt to prove that Texon was a name commonly used in Texas. In order to prove this defendant introduced the corporate charters of five companies using the name Texon, one company using the name Texxon, and one town in Texas named Texon. Plaintiff successfully rebutted this contention. Plaintiff established that Texon Oil & Land Company went out of business in the 1950’s; Texon Royalty Company never actually engaged in business in Texas or anywhere else under that name; Texon Oil Company ceased selling gasoline under that name as part of a settlement with Exxon; Texon Energy Corporation and Texon Petroleum Corporation were engaged in trademark infringement litigation with Exxon Corporation; and Texxon Corporation forfeited its charter in March 1975. Additionally, plaintiff pointed out that defendant failed to establish any connection between the towij of Texon and any of the corporations that used the name Texon. In summary, the evidence indicates that EXXON is a strong trademark deserving wide protection. Compare Amstar Corp. v. Domino’s Pizza, Inc., 615 F.2d 252 (5th Cir. 1980) (72 third-party registrations of the mark “Domino”); Armstrong Cork Co. v. World Carpets, Inc., 597 F.2d 496 (5th Cir. 1979) (multiple uses of the trademark “World”).
B. Similarity of Design
This factor has been described as “really nothing more than a subjective eyeball test.” 2 J. McCarthy, Trademarks and Unfair Competition § 237:7 (1973). The [505]*505similarity of design is determined by considering the overall impression created by the mark as a whole rather than simply comparing individual features of the marks, Armstrong Cork Co. v. World Carpets, Inc., 597 F.2d 496, 502 (5th Cir. 1979); Restatement of Torts § 729, Comment b (1938). Obviously, the greater the similarity in the design of the trademarks, the greater the likelihood of confusion.
An examination of defendant’s two marks establishes that Texon is used in a manner quite similar to plaintiff’s use of EXXON. Texon is printed in red with all block letters on a white background. The street address is printed in blue. EXXON is printed in red with all block letters on a white background. It is underlined with a blue bar. Defendant’s mark Tex-On, however, is quite different from plaintiff’s use of the trademark EXXON.
C. Similarity of Products
The third factor to be evaluated when determining whether there is a likelihood of confusion is the similarity between the products and services provided by the defendant and plaintiff. The greater the similarity between the products and services, the greater the likelihood of confusion. Plaintiff Exxon Corporation is most noted for its wide range of petroleum products. These products are predominantly dispensed to the buying public by Exxon service stations. In addition to selling Exxon products, many of these service stations also provide other car care services, including mechanical work. Defendant is engaged primarily in providing automotive repair services on transmissions and engines. Defendant also appears to sell some car parts. Both plaintiff and defendant are heavily involved in car care. There is a strong similarity between their wares and services. Compare Amstar Corp. v. Domino’s Pizza, Inc., 615 F.2d 252 (5th Cir. 1980) (holding that there is minimal similarity between a company that sells pizza and a company that manufactures sugar, salt, mustard, ketchup and other condiments).
D. Identity of Retail Outlets and Purchasers
“Dissimilarities between the retail outlets for and the predominant consumers of plaintiff’s and defendant’s goods lessen the possibility of confusion, mistake, or deception.” Id. at 262. In the case at bar, there is a strong similarity between the predominant consumers of the parties’ goods and services, but the retail outlets are quite dissimilar.
The evidence indicates that both plaintiff and defendant appeal to a wide variety of customers. These customers, however, all have one trait in common-they are members of the car driving public. They turn to the plaintiff and defendant in order to keep their cars functional. Thus, there is an identity of purchasers.
There is a substantial difference in the nature of the parties’ retail outlets. Plaintiff typically offers it goods and services [506]*506through a large number of outlets that have come to be known as service stations. Defendant, on the other hand, now operates out of a large warehouse-type building constructed out of aluminum siding. It would be difficult to mistake defendant’s ear care center for an Exxon service station.
E. Identity of Advertising Media
The fifth factor used in evaluating the likelihood of confusion is the similarity between the parties’ advertising campaigns. The greater the similarity in the campaigns, the greater the likelihood of confusion. In the case at bar, both plaintiff and defendant use virtually identical advertising media. Both parties used radio, television, newspaper ads, yellow page ads, and large signs identifying their places of business. Although plaintiff’s advertising is national in nature, both parties aim at the Houston market. For the purposes of this lawsuit, there is a substantial identity of advertising media.
F. Defendant’s Intent
A defendant’s intent to deceive buyers is merely one factor to be considered in determining whether there is a likelihood of confusion. 2 J. McCarthy, Trademarks and Unfair Competition § 23:31 (1973). If, however, a plaintiff can show that a defendant adopted a mark with the intent of deriving benefit from the reputation of the plaintiff, that fact alone “may be sufficient to justify the inference that there is confusing similarity.” Restatement of Torts § 729, Comment f (1938); Amstar Corp. v. Domino’s Pizza, Inc., 615 F.2d 252 (5th Cir. 1980). In the case at bar, the evidence on defendant’s intent is mixed. There is some evidence in the record indicating that the defendant intentionally adopted the marks Texon and Tex-On in order to appropriate some of the good will associated with the EXXON mark. There is other evidence in the record indicating that defendant’s actions were not taken with the intent of benefiting from the EXXON mark. The district court impliedly found that the defendant did not adopt the marks Texon and Tex-On with the intent of confusing, misleading, or deceiving the public. We do not find this finding of fact- to be clearly erroneous. Thus, in evaluating the various factors, this Court will assume that defendant did not adopt its marks with the intent of deriving benefit from Exxon Corporation’s reputation.
G. Actual Confusion
The best evidence of likelihood of confusion is provided by evidence of actual confusion. Roto-Rooter Corp. v. O’Neal, 513 F. 2d 44 (5th Cir. 1975). In the case at bar, plaintiff did not introduce any evidence to establish actual confusion. Such evidence, however, is not essential to a finding of likelihood of confusion. Id. Parties often introduce survey evidence in an effort to demonstrate that there is a likelihood of confusion. At trial, Exxon Corporation introduced a survey designed to show confusion among the car driving public. That survey indicated that there is a high possible confusion level between Texon and EXXON. The survey did not attempt to determine the confusion engendered by the mark Tex-On.
Prior to examining the findings of the survey, it is important to determine exactly what weight is to be given to the evidence. There are two important factors which affect the weight to be granted to these surveys: the format of the questions and the manner of conducting the survey. Holiday Inns, Inc. v. Holiday Out in America, 481 F.2d 445, 447 (5th Cir. 1973). In this case, an examination of these two variables indicates that the survey is to be given substantial weight.
The first factor to be considered in evaluating the validity of a survey is the format of the questioning. Surveys that involve nothing more than showing an individual a trademark and asking if it brings anything else to mind are given little weight in this Circuit. Id.; Amstar Corp. v. Domino’s Pizza, Inc. Id. at 448. These surveys are seen as little more than word-association tests. An example of this type survey is found in Holiday Inns. There the plaintiff conduct[507]*507ed a survey by showing people a placard bearing the logo “Holiday Out” and asking the individuals what their state of mind was upon seeing the placard. This Circuit granted the survey little weight because it failed to account for the number responses attributable to the word “Holiday” as distinguished from the service mark “Holiday Out.” Since the word “Holiday” appeared so frequently in service marks, this Court found the survey to be almost meaningless.
In the case at bar, the plaintiff’ survey did more than just ask what was brought to mind when the viewer saw a photograph of one of defendant’s Texon signs. The survey also probed what there was about the sign that elicited the response. The format of the questions in Exxon Corporation’s survey was designed to ensure a valid opinion poll. (See Appendix for a complete copy of the survey.)
The second factor to evaluated in assessing the validity of a survey is the adequacy of the universe. “The appropriate universe should include a fair sampling of those purchasers most likely to partake of the alleged infringers goods or services.” Amstar Corp. v. Domino’s Pizza, Inc., 615 F.2d at 264. An example of an insufficient universe is found in Amstar Corp. In that case, the issue was whether “Domino’s Pizza” infringed on the trademark “Domino” for sugar sold in grocery stores. Domino Sugar conducted a survey in ten cities. Eight of those cities had no “Domino’s Pizza” outlets and the outlets in the other two cities had been opened for less than three months when the survey was taken. Additionally, the only persons questioned were women found at home who identified themselves as the member of the household primarily responsible for grocery buying. The survey completely neglected the primary customers of “Domino’s Pizza”-young, single, male, college students. Id. Since the survey failed to examine the proper universe, this Court discounted the results of that poll. Id.
The survey conducted by Exxon Corporation focused on the appropriate universe. The survey was conducted between 10:00 a. m. and 6:00 p. m. in two high-traffic shopping centers on four different days. Plaintiff interviewed 515 individuals. The individuals interviewed were approximately two-thirds male, evenly divided among age groups, and employed in a wide range of occupations. All were licensed drivers. While this universe is not perfect, it is close enough so that, when combined with the format of the questions, it is clear that the survey is entitled to great weight.
The survey results themselves indicate a high possibility of confusion between Texon and EXXON. Approximately 15 percent of the individuals surveyed associated the Tex-on sign with EXXON. Another 23 percent associated the sign with gasoline, a gas station, or an oil company. Only seven percent associated the sign with defendant’s corporation. This survey constitutes strong evidence indicating a likelihood of confusion between Texon and EXXON. Once again, it should be noted that no survey evidence was introduced indicating a likelihood of confusion between Tex-On and EXXON.
H. Evaluation
An examination of the various factors indicates that the district court erred in concluding that there was no likelihood of confusion between Texon and EXXON.3 EXXON is a strong mark; the Texon design is quite similar to the EXXON design; the products, purchasers, and advertising media utilized are identical; and, most importantly, the survey evidence indicates a high confusion level between Texon and EXXON. On the other hand, this Court concludes that the district court did not err [508]*508in holding that there is no likelihood of confusion between Tex-On and EXXON. The record indicates that the Tex-On design is not similar to the EXXON design and there is no evidence of confusion between the marks Tex-On and EXXON. After reviewing the evidence, we are not “left with the definite and firm conviction that a mistake has been committed” by the district court in holding that there is no likelihood of confusion between Tex-On and EXXON. United States v. United States Gypsum Co., 333 U.S. 364, 395, 68 S.Ct. 525, 541, 92 L.Ed. 746 (1948).
IV. Conclusion
This Court’s review of the record indicates that the district court was clearly erroneous when it concluded that there is no likelihood of confusion between Texon and EXXON. Accordingly, the district court is instructed to issue an injunction, similar to the one issued in January 1977, prohibiting defendant from using the mark Texon in connection with its business. Our review of the record also indicates that the district court did not err in holding that there was no likelihood of confusion between Tex-On and EXXON. Accordingly, the judgment of the district court is
AFFIRMED IN PART, REVERSED IN PART, AND REMANDED.
APPENDIX
COMPANY RECOGNITION STUDY Hello, I’m - with Houston Interviewing a national marketing research and public opinion organization. We’re conducting a research study of people’s reactions to company logo’s and signs. Are you a licensed driver and do you drive a car or truck? (IF NO, TERMINATE AND TALLY. IF YES, ASK.....)
No _
(Tally)
la. Are you . . .
(READ) Under 25 ( )
25-34 ( )
35-44 ( )
45-54 ( )
55 and over ( )
lb. (SIGHT CHECK) Sex . . .
Male ( )
Female ( )
lc. And, what is your occupation?
Industry_
Job Title_
(ASK RESPONDENT TO PARTICIPATE IN THE STUDY AND TO COMPLETE THE INTERVIEW)
Refused -
2a. (SHOW RESPONDENT PHOTOGRAPH AND ASK . . . .)
First, would you please look at this photograph. Take as much time as you need and I’ll ask you some questions.
What is the first thing that comes to mind when looking at this sign? (RECORD VERBATIM)
2b. What was there about the sign that made you say that? (PROBE FULLY)
3a. (IF A COMPANY NAME IS NOT MENTIONED IN QUESTION 2, ASK QUESTION 3a. IF A COMPANY NAME IS MENTIONED, SKIP TO QUESTION 3b.)
What is the first company that comes to mind when you look at this sign? (RECORD NAME EXACTLY AS RESPONDENT STATES IT)
3b. What was there about the sign that made you mention (COMPANY) ? (CLARIFY AND PROBE FULLY)