HOWARD, Judge.

This is an appeal from a defense verdict in a products liability case involving an allegedly defectively designed three-wheeled Honda All Terrain Cycle (ATC).

On June 14, 1981, Mitchell Gosewisch was riding a Honda 185S ATC that he had purchased from his cousin, who had purchased the vehicle from a Tucson Honda dealer about three months before the accident involved here. Gosewisch was riding the vehicle in Pantano Wash when the vehicle flipped, throwing him to the ground. As a result of the accident Gosewisch is a quadriplegic.

It was Gosewisch’s theory that the vehicle had a design defect which caused it to flip unexpectedly. He contended that this ATC, which relies totally on three soft balloon tires for suspension, had design defects consisting of (1) very low-pressure, easily collapsible tires; (2) no mechanical suspension; (3) inherent instability of a three-wheel vehicle with a high center of gravity; (4) weak front forks and (5) a front brake vulnerable to being accidentally engaged.

The defendants contended that there were no design defects and that the accident occurred because Gosewisch was going too fast, 25 to 35 miles per hour, when he struck a large compacted pile of sand, causing him to sail into the air and crash the vehicle. Plaintiffs contend the trial court erred in (1) instructing the jury; (2) improperly hampering their discovery; (3) permitting defendants to change their theory of defense during the trial and (4) excluding evidence of other ATC accidents. We do not agree and we affirm.

I. FAILURE TO WARN

Plaintiffs contend that the trial court erred in refusing to give their jury instruction on the failure to warn. Singularly appropriate is our language in Embry v. General Motors Corporation, 115 Ariz. 433, 565 P.2d 1294 (App.1977):

“Appellant did not contend that the automobile was faultlessly constructed and designed. On the contrary, her theory and contention was that there was a design defect because if the motor mounts broke, under a certain mode of driving, the engine could bind the accelerator *392 linkage. Only with the design defect could one consider the automobile to be unreasonably dangerous. If there were no design defect, the failure to warn added nothing to appellant’s claim.

Under Restatement of Torts (Second), § 402A, products though faultessly made, may nevertheless be deemed “defective” if it is unreasonably dangerous to place the product in the hands of a user without a suitable warning. Gherna v. Ford Motor Company, 246 Cal.App.2d 639, 55 Cal.Rptr. 94 (1966). The failure to warn comes into play in a strict liability case when the product is perfectly manufactured and meets every requirement for its designed utility but nevertheless is unreasonably dangerous because of a failure to warn of is dangerous characteristics. Jackson v. Coast Paint and Lacquer Company, 499 F.2d 809 (9th Cir.1974); Tucson Industries, Incorporated v. Schwartz, 108 Ariz. 464, 501 P.2d 936 (1972). Here, however, we are not dealing with a product claimed to have been faultlessly manufactured and designed. Under such circumstances the court did not err in failing to give the foregoing instruction.” 115 Ariz. at 436, 565 P.2d at 1297.

Because plaintiffs here did not contend at trial that the ATC was faultlessly manufactured and designed—their sole contention being that the vehicle had design defects—there was no error in failing to give the instruction.

II. FAILURE TO GIVE PLAINTIFFS’ REQUESTED INSTRUCTION NO. 2

Plaintiffs requested the following instruction which was refused by the trial court:

“A product is unsafe for a reasonably foreseeable use if, at the time of its sale or manufacture, it is unsafe to an extent beyond that which would be contemplated by the average consumer, with the ordinary knowledge common to the community where the product is being sold.” (Emphasis added)

The trial court has no duty to give an instruction which misstates the law. Nichols v. Baker, 101 Ariz. 151, 416 P.2d 584 (1966). Nor should the trial court give instructions which are confusing or misleading. Evans v. Pickett, 102 Ariz. 393, 430 P.2d 413 (1967). It is not error to refuse an instruction which incorrectly states the law or which states it only partially, tending to mislead the jury. Valley National Bank v. Witter, 58 Ariz. 491, 121 P.2d 414 (1942).

Plaintiffs’ use of the word “unsafe” in their proffered instruction is erroneous and stems from their instruction No. 1, also refused by the court, which similarly used the word “unsafe”:

“A party who manufactures or sells a defective product is liable if the defect causes an injury and if the defect causes the product to be unsafe for a reasonably foreseeable use.” (Emphasis added)

Instruction No. 1 is identical to Recommended Arizona Jury Instructions, Products Liability 1 (1974) which was held to be erroneous in Byrns v. Riddell, Inc., 113 Ariz. 264, 550 P.2d 1065 (1976), stating that the requirement for liability was not that the product be “unsafe” but that it be “unreasonably dangerous.”

The trial court here instructed the jury that:

“A party who designs, manufactures and sells a product which is defective and unreasonably dangerous for a reasonably foreseeable use, is liable if the defect causes an injury.” (Emphasis added)

Plaintiffs’ instruction No. 2 was erroneous, misleading and confusing vis-a-vis the correct instruction given by the trial court. The trial court did not err in refusing plaintiffs’ instruction No. 2.

III. THE INSTRUCTION ON THE MISUSE OF THE ATC

Pursuant to plaintiffs’ request, the trial court instructed the jury as follows:

“A person who manufactures or sells a product which he has reason to foresee may cause an injury from a particular use, is required to give adequate di *393 rections for safe use of the product. If he fails to do so, he is liable for any injury resulting from the failure to give adequate directions.”

The trial court also gave the defendants’ requested instruction on misuse based on A.R.S. § 12-683(3):

“However, if the defendants prove that the particular use was contrary to any express and adequate instructions appearing on or attached to the product, or on its original container or wrapping, if the injured person knew or with exercise of reasonable and diligent care should have known of such instructions, the defendants are not liable.”

Plaintiffs contend that the defendants’ instruction on misuse should not have been given because there was no evidence to support the instruction. Specifically, they argue that the record lacks evidence of Gosewisch’s use of his ATC contrary to instructions “appearing on or attached to the product or on its original container or wrapper.” We do not agree. A sticker on the ATC admonishes the user to read the owner’s manual carefully. When the ATC was sold to Gosewisch’s cousin, it was accompanied by an owner’s manual. Mr. Gosewisch did not obtain the owner’s manual from his cousin nor did he read it. The manual contained instructions to riders to perform a pre-ride safety check, check the tire pressure before riding, know the terrain before riding, ride cautiously and with proper judgment, and drive slowly when unfamiliar with the terrain or the ATC. There is evidence in the record that Gosewisch did not follow any of these instructions. Where defendants present any evidence tending to support their theory, the instruction should be given. Cf. Correa v. Curbey, 124 Ariz. 480, 605 P.2d 458 (App.1979). The record provides ample justification for the instruction given to the jury.

IV. THE INSTRUCTION ON CONTRIBUTORY NEGLIGENCE

Plaintiffs contend that the trial court erred in not instructing the jury that contributory negligence is not a defense in a strict liability case. The record discloses that the trial court instructed the jury as follows:

“There may be several causes for an injury. If you find that the subject ATC is defective, causing it to be unreasonably dangerous for a reasonable foreseeable use, and that the plaintiff, Mitchell Gosewisch, was negligent in the operation of the ATC at the time of his accident, and that both the defective product and the plaintiff's negligence were causes of the accident, this would not defeat recovery by the plaintiffs.”

We find that the above instruction adequately covers the instruction proposed by the plaintiffs on the effect of contributory negligence in a products liability case.

V. THE STATE OF THE ART INSTRUCTION

A.R.S. § 12-683(1) provides that in any product liability action a defendant shall not become liable if the defendant proves that the following applies:

“The defect in the product is alleged to result from inadequate design or fabrication, and if the plans or designs for the product or the methods and techniques of manufacturing, inspecting, testing and labeling the product conformed with the state of the art at the time the product was first sold by the defendant.”

Pursuant to the statute the court instructed the jury:

“The defendants, Honda, shall not be liable if you find that the design and testing of the 1981 Honda ATC 185-S in this case conformed with the state of the art at the time the Honda product was sold by the defendants.”

The trial court further instructed the jury on the meaning of the phrase “state of the art”:

“State of the art means the technical, mechanical and scientific knowledge of manufacturing, designing, testing, or labeling the same or similar products which was in existence and reasonably feasible ... [for] use at the time of manufacture.” See A.R.S. § 12-681(6).

*394 The plaintiffs contend there was no evidence to justify the giving of any instructions on state of the art. We do not agree. Defendants point to evidence they introduced of custom and practice in the industry as probative of the state of the art. Although the customs of an industry may be relevant, because those customs may lag behind technological development, they are not identical with state of the art. O’Brien v. Muskin Corporation, 94 N.J. 169, 463 A.2d 298 (1983). Custom refers to what was being done in the industry; state of the art refers to what feasibly could have been done. Chown v. USM Corporation, 297 N.W.2d 218 (Iowa 1980). However, not every safety device for which technology exists is necessarily feasible. Chown v. USM Corporation, supra.

In order to show state of the art, the defendant can introduce testimony about the thoroughness of experimentation and research prior to manufacture. Olson v. Arctic Enterprises, Inc., 349 F.Supp. 761 (D.N.D.1972). Here the defendants introduced extensive evidence that the Honda 185S ATC was thoroughly tested and made safe prior to sale. The jury heard how dozens of test riders (including novices) rode the ATC in three nations and on some of the most diverse terrain in the world, from Mount Fuji, Japan to Melbourne, Australia and International Falls, Minnesota. The jury also heard about the extensive laboratory testing that preceded the production of every ATC model. The state of the art can also be established by testimony regarding the infeasibility of each of the testing and design alternatives suggested by the plaintiffs. See Starr v. J. Hacker Company, Inc., 688 F.2d 78 (8th Cir.1982). Here, plaintiffs claim the defendant should have used computer simulations to test the ATC. Defendants’ experts testified that computer simulations could not account for the variables involved in a rider-active vehicle. Plaintiffs claim that the ATC should have had mechanical suspension. Defendants’ experts tested ATCs with the mechanical suspension proposed by plaintiffs and found handling to be inferior. Original equipment front forks on the model 185S were shown to be substantially stronger than the replacement forks advocated by plaintiffs’ experts, even though the aftermarket forks were equipped with springs and mechanical shock absorbers. The jury also heard comparisons of the characteristics of the model 185S tire with those of the alternative tires manufactured by Goodyear and Dunlop and suggested by the plaintiffs’ experts. The trial court did not err in giving the instructions.

VI. PRETRIAL DISCOVERY

The trial court limited discovery in this case to claims and cases involving similar accidents on the 185S ATC. Pursuant to the trial court’s discovery order the defendants turned over information to the plaintiffs which was contained in 14 cases. The information which the plaintiffs received consisted, inter alia, of what the allegation of the plaintiff was in each case as to how the accident occurred, the name and address of counsel for the claimant, the date of the occurrence and its location and the plaintiff's name, address and telephone number. However, the trial court refused to allow the plaintiffs to see investigative reports made by the appellees’ insurance company. It also refused to allow the plaintiffs to see a “log of complaints” kept by one of appellees’ counsel. This “log of complaints” merely consisted of a list of claim numbers which were used by counsel to determine when a certain action had to take place in the claim or case. Plaintiffs claim that the trial court erred in not allowing them to discover the contents of the investigative reports and see the “log of complaints.” We do not agree.

The trial court’s broad discretion in matters of discovery will not be disturbed absent a showing of abuse. Brown v. Superior Court, 137 Ariz. 327, 670 P.2d 725 (1983). We do not believe the trial court abused its discretion by limiting plaintiffs’ discovery to cases involving similar accidents rather than all accidents involving the vehicle in question. Plaintiffs are entitled to discover only information in records which might contain facts that would lead to the discovery of admissible evidence. *395 They are not entitled to see the records and evidence on all complaints regarding the 185S ATC in order to determine whether the defendants gave them all the cases and claims involving similar accidents. Inquiry can be made into similar accidents caused by the same or similar models. See Kozlowski v. Sears, Roebuck and Company, 73 F.R.D. 73 (D.Mass.1976) (similar accidents relevant); see also 4 J. Moore, Moore’s Federal Practice, §§ 26.561, 26-109, 26-110 (2d ed. 1984).

Rule 26(b)(3), Rules of Civil Procedure, 16 A.R.S., provides that the discovery of relevant documents and tangible things which have been prepared in anticipation of litigation or for trial may be discovered only upon a showing that the party seeking discovery has substantial need of the materials in preparation of his case and that he is unable without undue hardship to obtain a substantial equivalent of the material by other means. This discovery extends to investigative reports by insurers. The rule also provides that in ordering discovery of such materials when the required showing has been made, the court shall protect against disclosure of the mental impressions, conclusions, opinions, or legal theories of an attorney or other representative of a party concerning the litigation.” Plaintiffs made no attempt in the trial court and do not on appeal show satisfaction of the condition placed upon the granting of discovery of the insurance investigation report—that is, they made no showing of the substantial need of the materials in the preparation of their case and that they were unable without undue hardship to obtain the substantial equivalent of the information by other means. In fact, it is clear from reading the depositions in this case that plaintiffs were not seeking facts and data contained in the investigative reports of the cases that were turned over to them, but rather were intent on discovering the defendants’ opinions and conclusions in the investigation files as to how the accidents had happened in those cases. Such information is not discoverable under the rule. 1

VII. FAILURE TO PRECLUDE TESTIMONY OF DEFENSE EXPERT WITNESS

Seven weeks before trial, defendant supplemented an interrogatory response to inform the plaintiffs that:

“With respect to the terrain, Dr. Packer is expected to express an opinion that the area ... contains a significant artifact (mound) not now visible in its June, 1981, form. This mound is associated with a race course utilized by motorcycle and ATC riders in the wash, extending from shore to shore. Dr. Packer will express the opinion that the terrain which plaintiff attempted to negotiate was not flat, but rather plaintiff struck this raised mound immediately prior to the accident. Dr. Packer will express an opinion that plaintiff MITCHELL GOSEWISCH travelled over this mound and went airborne somewhat analogous to a ski-jumper. He has concluded that the mound was not a barrier which halted the front wheel of the vehicle causing the rear wheels to pivot over the front wheel while it remained stationary____

******

Dr. Packer will also express an opinion on the effect of the terrain on the subject accident. He is expected to testify that ... the mound significantly reduces the safe speed at which an ATC can traverse the area of the accident in the Pantano Wash. He will express an opinion that MITCHELL GOSEWISCH was traveling at a nominal speed of 30-35 miles per hour at the time of the accident, but that a safe speed is no more than 15 miles per hour for the mound in question.”

At trial Dr. Packer testified that plaintiff’s accident was the result of going too fast when he struck the mound:

“Well, it’s my opinion that the accident happened as the result of Mr. Gosewisch *396 traveling down the wash at a high speed, down the area approaching a highly visible obstacle, and approaching it at far too fast a speed, far too fast a speed to negotiate that obstacle safely, and neither slowing down for (sic) avoiding the obstacle nor bringing his bike into a safe operating mode before he passed over that obstacle, sir.

Q. Is there any way for you to be precise for the jury as to when, between the striking of the mound and the coming to rest of the machine, as to when in that point of time Mr. Gosewisch was ejected from the machine?

A. No. He would start—he’s going—the bike’s going to decelerate, he’s going to go on and he’ll start to leave the machine at the time of the contact. But to give you a precise scenario, no, I cannot.”

Plaintiffs here contend that the trial court erred in not granting their motion to exclude the testimony of Dr. Packer since his testimony at trial was that of a “crash theory” whereas the supplemental answers to interrogatories indicated that Dr. Packer-was going to testify to an “airborne theory.” We do not agree. We perceive no material difference between the supplemental answers to the interrogatories and Dr. Packer’s testimony at trial. There was no failure here to divulge an entire defense. See e.g., Smith v. Ford Motor Company, 626 F.2d 784 (10th Cir.1980), cert. denied 450 U.S. 918, 101 S.Ct. 1363, 67 L.Ed.2d 344 (1981). Assuming arguendo that there was some variation between the supplemental answers to the interrogatories and Dr. Packer’s testimony at trial, such variations are not uncommon with witnesses, especially expert witnesses, and the usual way to handle such variations is by means of cross-examination. The imposition of sanctions to make discovery is within the sound discretion of the trial court. Fleitz v. Van Westrienen, 114 Ariz. 246, 560 P.2d 430 (App.1977). The record does not support plaintiffs’ contention that Dr. Packer presented a new defense at trial. Nor have plaintiffs demonstrated that they were prejudiced by this testimony at trial. The trial court did not err in refusing to exclude Dr. Packer as a witness.

VIII. EXCLUSION OF EVIDENCE OF OTHER ACCIDENTS

The trial court refused to allow into evidence a “study” done by Dr. Michael Rieser, an emergency room physician affiliated with the Yuma Regional Medical Center in Yuma, Arizona, and refused to allow the doctor to give an opinion concerning the results of his study. In the course of his duties, Dr. Rieser had seen many ATC accident victims, injured while riding ATCs recreationally in the surrounding areas, and he had conducted this study to discover whether there was a pattern to the injuries. Plaintiffs contend that the trial court erred in rejecting this evidence because the study was admissible under Rules 803(4) and 803(24), of the Rules of Evidence, 17A A.R.S., and because the doctor was qualified to give his opinion as an expert witness.

Rule 803(4) provides that the following are exceptions to the hearsay rule:

“Statements made for purposes of medical diagnosis or treatment and describing medical history, or past or present symptoms, pain, or sensations, or the inception or general character of the cause or external source thereof insofar as reasonably pertinent to diagnosis or treatment.”

Since plaintiffs provided no foundation showing that the statements were made for medical diagnosis or treatment, the court did not err in rejecting the admissibility of the evidence under Rule 803(4).

Rule 803(24) provides another exception to the hearsay rule and allows admissibility of:

“A statement not specifically covered by any of the foregoing exceptions but having equivalent circumstantial guarantees of trustworthiness, if the court determines that (A) the statement is offered as evidence of a material fact; (B) the statement is more probative on the point for which it is offered than any other evidence which the proponent can procure through reasonable efforts; and *397 (C) the general purposes of these rules and the interests of justice will best be served by the admission of the statement into evidence____”

Case-specific surveys are generally admissible if they are conducted according to the principles accepted by social scientists and statisticians for gathering and analyzing survey data. C.A. May Marine Supply Company v. Brunswick Corporation, 649 F.2d 1049 (5th Cir.1981), cert. denied 454 U.S. 1125, 102 S.Ct. 974, 71 L.Ed.2d 112; Baumholser v. Amax Coal Company, 680 F.2d 550 (7th Cir.1980); McCormick on Evidence, 3rd Ed., § 208. Therefore, if the proper foundation had been laid, the study or survey would have been admissible as an exception to the hearsay rule under Rule 803(24). The plaintiffs’ foundation for the admissibility of the study consisted of the testimony of Dr. Rieser and the affidavit of Dr. Thomas Moon, who had a doctorate in biostatistics and epidemiology from the University of California at Berkley. He was a professor in the Department of Medicine at the University of Arizona Medical School and served as assistant director at the Cancer Center and as head of the Biometry, Computing and Epidemiology Unit at the Cancer Center. As for Dr. Rieser, his knowledge of statistics consisted only of a few college math courses and one in medical school that he could not remember having completed. He did not consult any sources or textbooks when he designed his “study.” The affidavit of Dr. Moon shows that he has training and experience in the statistical processes and methods applied to the analysis of biological phenomena. However, it does not show that he has any experience in conducting surveys. His affidavit stated, inter alia:

3. I have conferred with Dr. Michael Rieser regarding his ATC accident study in Yuma, Arizona, and am aware of the methodology employed in his study. The approach taken by him is one that is generally accepted in the scientific community as a standard and valid method of research;

4. The method employed by Dr. Rieser is classified as a near-complete census of the population of ATC accident victims in the Yuma area who came to the Yuma Regional Medical Center for emergency medical treatment between April 15,1983 and to current date. This method is reliable and accurate, and is a variation of the survey method known as ‘sampling’, which has furnished statistical data for many articles published in professional journals known to me. * * * ”

Neither Dr. Rieser nor Dr. Moon had any training in the collection and studying of accident data.

In opposition, the defendants provided the court with the affidavit of Roger L. McCarthy. Among his degrees are a bachelor of arts with high distinction from the University of Michigan; a bachelor’s degree in mechanical engineering with high distinction from the University of Michigan, two degrees in mechanical engineering from the Massachusetts Institute of Technology, and a doctorate in mechanical engineering from the Massachusetts Institute of Technology. He is a member of the American Society of Mechanical Engineers, Society of Automotive Engineers, American Society of Metals, System Safety Society, Human Factors Society, the National Safety Council, the National Society of Professional Engineers and the American Society for Testing and Materials. He is also a registered mechanical engineer in the states of California and Arizona. He is the president of Failure Analysis Associates, allegedly the largest single repository of accident and failure data in the nation. He has personally directed the construction of accident report data bases and telephone questionnaires. He regularly assesses data groups and reporting samples to appraise the utility and evaluate the safety of various products. He reviewed the deposition of Dr. Rieser, Dr. Rieser’s affidavits, the results of the “survey”, the questionnaire that was provided to the persons who reported the accidents, the instructions or letter that accompanied the questionnaires, and the affidavit of Dr. Moon. He stated that as a result of his review he concluded that Dr. Rieser’s survey was so technically *398 flawed and almost devoid of scientific controls that it could not be used to come to, or support any valid conclusions about, the safety, design or accident propensities of ATVs (all-terrain-vehicles), including ATCs. He also concluded that the confounding bias in Dr. Rieser’s study was so great that there was no method, technique, or statistical procedure that could be applied to Dr. Rieser’s data that would produce valid conclusions relative to ATV or ATC safety. In his affidavit Mr. McCarthy pointed out the error in the selection process, found errors in the instructions given to the medical personnel who filled out the questionnaires, and found the survey form poorly designed to the extent that it would not elicit proper answers. The affidavit of Dr. McCarthy also states:

it * * *

16. The data that Dr. Rieser has collected cannot be related in any statistical manner to the ATV or ACT user population as a whole, the ATV or ATC accident population as a whole, or ATV or ATC safety. The relationship of Dr. Rieser’s sample to ATV or ATC riders as a whole is completely unknown. The ‘exposure’ or vehicle usage that produced these fifty-six accidents is an even greater mystery. Important variables, such as customer/user vehicle modifications, were not controlled for at all because of Dr. Rieser’s disinclination to understand or classify them.

17. The goal of determining ATV or ATC accident causation, ostensively [sic] one of the goals of Dr. Rieser’s survey, could never be achieved by the method Dr. Rieser employed. Numerous studies of accidents involving other motorized vehicles typically find 70-90% of the accidents caused by ‘human error.’ There is no a priori reason to assume this factor would not be a very significant, if not the most significant, cause of ATV or ATC accidents. Yet, such a contribution would be vastly understated by any survey technique that relied upon accident description and analysis by the potentially at-fault operator. To determine accident causation has historically required in depth [sic], accident-by-accident, investigation by skilled specialists. Dr. Rieser clearly recognized the propensity of accident-involved drivers to misstate alcohol consumption, but completely ignored this same effect among such drivers in describing accident causation.”

Surveys, since they involve hearsay, must be conducted with proper safeguards to ensure accuracy and reliability. Several factors must be examined when determining whether a poll meets generally accepted principles. A proper universe must be examined and a representative sample must be chosen; the persons conducting the survey must be experts; the data must be properly gathered and accurately reported. It is essential that the sample design, the questionnaires and the manner of interviewing meet the standards of objective survey and statistical techniques. Just as important, the survey must be conducted independently of the attorneys involved in the litigation. Here, 18 of the persons interviewed were contacted a second time by Dr. Rieser, at the request of plaintiffs’ counsel, to supplement or clarify answers to the questionnaires. The interviewers or sample designers should, of course, be trained, and ideally should be unaware of the purposes of the survey or the litigation. Pittsburgh Press Club v. United States, 579 F.2d 751 (3rd Cir.1978).

The record here justifies the trial court’s conclusion that the study did not meet the criteria of trustworthiness and reliability and justifies the trial court’s refusal to allow it to be admitted into evidence.

As far as Dr. Rieser himself testifying as to his conclusions from the poll, we first note that Dr. Rieser had no special knowledge in the areas in which the plaintiffs hoped to elicit his testimony. He is a medical doctor. He is not a statistician, an engineer, an expert on vehicle design or a trained accident investigator.

Rule 703, Rules of Evidence, provides that an expert may rely upon evidence that is not admissible only if it is of a type reasonably relied upon by experts in a *399 particular field. State v. Clark, 112 Ariz. 493, 543 P.2d 1122 (1975). There was no foundation here to show that experts in the accident reconstruction field use the type of survey produced by Dr. Rieser. Furthermore, once the survey was determined to be untrustworthy and unreliable, and therefore inadmissible under the hearsay exception, it could not come into evidence under the guise of Dr. Rieser’s testimony. See Pittsburgh Press Club v. United States, supra.

Plaintiffs next contend the trial court erroneously sustained objections to the following questions asked of Mr. Takeuchi, a representative for Honda, during cross-examination:

“Do you know how many people have been crippled or killed with a forward flip of these vehicles between 1970 and 1980? ...

Have you or Honda made any effort to find out how many people are being injured in the field with your invention?”

We do not agree. Discretion must be exercised to prevent cross-examination that is argumentative, irrelevant, without foundation, and unfairly prejudicial. See Poole v. Superior Court, 139 Ariz. 98, 677 P.2d 261 (1984); State v. O’Brien, 123 Ariz. 578, 601 P.2d 341 (App.1979). An argumentative question seeks no factual testimony, but requires the witness to acquiesce in inferences drawn by counsel from prior testimony. M. Udall and J. Livermore, Arizona Practice: Law of Evidence § 33 (2d ed. 1982). The thrust and implication of both questions is that Honda has loosed upon the public a vehicle which is maiming and killing people. To this extent the questions are argumentative. Furthermore, the trial court had already ruled, correctly, that only evidence of similar accidents was admissible. Thus, the questions were improper and were an attempt to circumvent the ruling of the trial judge.

Finally, the trial court refused to allow into evidence testimony by Sheriff Sgt. Bowman. He was going to testify about three or four ATC accidents which he had investigated. Plaintiffs’ counsel advised the court that Sgt. Bowman:

“... Would testify that he’s investigated three or four other accidents similar to this one where there has been a bounce-out, or a flip-out where a cycle has hit a series of bumps and hit another bump or depression and has flipped over, while he has been deputy sheriff here in Pima County.”

The court responded to the offer as follows:

“If you want to go into further matters such as what makes you say similar accidents, why, I’ll allow you to question the witness along those lines.

But just to draw the conclusion that it’s similar accidents, I don’t think there’s anything before the Court that would— at this point that would justify my being able to come to that conclusion, which I think is necessary before- we can have admissibility of other accidents.”

Plaintiffs never pursued the matter further by showing, for example, that the accidents involved a similar speed, that the vehicle flipped over forward as occurred in plaintiff's accident, the nature of the terrain, whether the accident occurred in daytime or nighttime, how many people were on the vehicle at the time, and whether the driver of the vehicle had been drinking at the time or not. Having failed to pursue the matter, the plaintiffs cannot argue that the trial court was in error.

Affirmed.

BIRDSALL, P.J., and FERNANDEZ, J., concur.

. But see Brown v. Superior Court, supra, where the court held in a bad-faith failure-to-settle case that such mental impressions, conclusions, opinions or legal theories of an attorney or other representative of a party concerning the litigation were discoverable since they were directly at issue in the case.