Citations
- 931 F. Supp. 2d 409
Full opinion text
MEMORANDUM & ORDER
MARGO K. BRODIE, District Judge.
Plaintiffs Matthew Valente and James Valente filed the instant products liability action in New York state court against Defendants Textron and the E-Z Go Division of Textron. Matthew Valente was seriously injured while operating a golf cart made by Defendants and alleges that Defendants are liable under a theory of strict liability, negligence, breach of implied warranty and failure to warn. Matthew Valente’s father, James Valente, brings a claim for loss of consortium. Plaintiffs allege that the golf car at issue was defectively designed because it only had a rear-wheel braking system and did not have a seatbelt restraint system. Defendants removed the action to this Court and now move to preclude the testimony of Plaintiffs’ experts, Kristopher Seluga and Bruce Gorsak, and for summary judgment. Plaintiffs also move to preclude the testimony of Defendants’ experts, Matthew Schwall and David Bizzak, and for summary judgment. The Court held a Daubert hearing on January 31 and February 1, 2013 with respect to Plaintiffs’ experts. The Court heard argument on the parties’ motions on February 28, 2013. For the reasons set forth below, the Court grants Defendants’ motions to preclude the testimony of Plaintiffs’ experts and for summary judgment. Plaintiffs’ motions to preclude the testimony of Defendants’ experts and for summary judgment are denied.
I. Background
In August of 2007, plaintiff Matthew Valente (“Matthew”) was working as a cart and range attendant at La Tourette Golf Course (“La Tourette”) on Staten Island. (Def. 56.1 ¶¶ 1, 4.) Matthew was 18 years old at the time. (PI. 56.1 ¶ 4.) As an attendant, Matthew was responsible for, among other things, driving golf cars between the pen and the area where golfers picked them up. (Def. 56.1 ¶ 5; PI. 56.1 Reply ¶ 5.) Matthew was trained at La Tourette regarding how to operate a golf car. (Def. 56.1 ¶ 6; PI. 56.1 Reply ¶ 6.)
On August 18, 2007, the day of the accident, Matthew was driving an E-Z-Go golf car. (Def. 56.1 ¶ 3; PL 56.1 Reply ¶3.) Textron manufactures E-Z-Go golf cars. (Def. 56.1 ¶ 2.) Matthew was driving on the path to the 10th hole, and his hat blew off. (Def. 56.1 ¶ 10.) Along the left side of the path, there was a series of posts connecting a rope. (Deposition of Matthew Valente (“M. Valente Dep.”) 88:21-89:11.) Defendants claim that Matthew reached back to retrieve his hat, removing his foot from the accelerator and stepping on the brake in an effort to stop the car. (Def. 56.1 ¶ 11.) Matthew claims that, when his hat blew off, he did not attempt to reach for his hat or turn his body. (Pl. 56.1 Reply ¶ 11.) According to Matthew, the path turned slightly to the left and the only adjustment to the steering that he made was to turn the steering wheel slightly “to maintain a straight course on the path.” (PL 56.1 Reply ¶ 14.) He simply applied the brakes, and the golf car yawed, or fishtailed. (PL 56.1 Reply ¶¶ 17-18.) Matthew remembers the car sliding, coming out of his seat and hitting his head. (Def. 56.1 ¶ 12; Pl. 56.1 Reply ¶ 12.) The golf car rolled over onto its passenger side. (Def. 56.1 ¶ 13; Tr. 25:2-8.) The parties agree that the golf car did not have any mechanical difficulties the day of the accident. (Def. 56.1 ¶ 8; Pl. 56.1 Reply ¶ 8.) Matthew suffered serious injuries, including a spinal fracture, and is paralyzed below the waist with partial paralysis in his upper body. (PL 56.1 ¶ 13.)
II. Admissibility of Expert Testimony
Rule 702 of the Federal Rules of Evidence provides that “[a] witness who is qualified as an expert by knowledge, skill, experience, training, or education may testify in the form of an opinion or otherwise if: (a) the expert’s scientific, technical, or other specialized knowledge will help the trier of fact to understand the evidence or to determine a fact in issue; (b) the testimony is based on sufficient facts or data; (c) the testimony is the product of reliable principles and methods; and (d) the expert has reliably applied the principles and methods to the facts of the case.” Fed. R. Evid. 702. The proponent of the expert testimony bears the burden of “establishing by a preponderance of the evidence that the admissibility requirements of Rule 702 are satisfied.” United States v. Williams, 506 F.3d 151, 160 (2d Cir.2007) (citing Daubert v. Merrell Dow Pharmaceuticals, Inc., 509 U.S. 579, 592 n. 10, 113 S. Ct. 2786, 125 L.Ed.2d 469 (1993)). However, “the district court is the ultimate ‘gatekeeper.’ ” Id. (citations omitted); see also United States v. Farhane, 634 F.3d 127, 158 (2d Cir.2011), cert. denied, — U.S.-, 132 S.Ct. 833, 181 L.Ed.2d 542 (2011) (“The law assigns district courts a ‘gatekeeping’ role in ensuring that expert testimony satisfies the requirements of Rule 702.” (citation omitted)).
Before permitting a person to testify as an expert under Rule 702, the court must make the following findings: (1) the witness is qualified to be an expert; (2) the opinion is based upon reliable data and methodology; and (3) the expert’s testimony on a particular issue will “assist the trier of fact.” Nimely v. City of New York, 414 F.3d 381, 396-97 (2d Cir.2005); see also United States v. Cruz, 363 F.3d 187, 192 (2d Cir.2004) (the court is tasked with “ensuring that an expert’s testimony both rests on a reliable foundation and is relevant to the task at hand” (quoting Daubert, 509 U.S. at 597, 113 S.Ct. 2786)). In Daubert v. Merrell Dow Pharmaceuticals, the Supreme Court set forth a list of factors, in addition to the criteria set forth in Rule 702, that bear on the determination of reliability: “(1) whether a theory or technique has been or can be tested; (2) “whether the theory or technique has been subjected to peer review and publication;’ (3) the technique’s ‘known or potential rate of error’ and ‘the existence and maintenance of standards controlling the technique’s operation;’ and (4) whether a particular technique or theory has gained general acceptance in the relevant scientific community.” Williams, 506 F.3d at 160 (quoting Daubert, 509 U.S. at 593-94, 113 S.Ct. 2786); see also Zaremba v. Gen. Motors Corp., 360 F.3d 355, 358 (2d Cir. 2004) (same). The Daubert inquiry for reliability is a “flexible one” and does not “constitute a definitive checklist or test,” Kumho Tire Co., Ltd. v. Carmichael, 526 U.S. 137, 150, 119 S.Ct. 1167, 143 L.Ed.2d 238 (1999) (citation omitted), and, thus, the Daubert factors “neither necessarily nor exclusively appl[y] to all experts or in every case,” id. at 141,119 S.Ct. 1167.
The district court is afforded “broad latitude when it decides how to determine reliability as it enjoys [with] respect to its ultimate reliability determination.” Kumho Tire, 526 U.S. at 142, 119 S.Ct. 1167 (emphasis in original). Expert testimony should be excluded if it is “speculative or conjectural.” Major League Baseball Prop., Inc. v. Salvino, Inc., 542 F.3d 290, 311 (2d Cir.2008) (quoting Boucher v. U.S. Suzuki Motor Corp., 73 F.3d 18, 21 (2d Cir.1996)). When an expert’s opinion is based on data or methodologies “that are simply inadequate to support the conclusions reached, Daubert and Rule 702 mandate the exclusion of that unreliable opinion testimony.” Ruggiero v. Warner-Lambert Co., 424 F.3d 249, 253 (2d Cir.2005) (citation omitted); see also Nimely, 414 F.3d at 396 (“[N]othing in either Daubert or the Federal Rules of Evidence requires a district court to admit opinion evidence which is connected to existing data only by the ipse dixit of the expert. A court may conclude that there is simply too great an analytical gap between the data and the opinion proffered.” (alteration in original) (quoting Gen. Elec. Co. v. Joiner, 522 U.S. 136, 146, 118 S.Ct. 512, 139 L.Ed.2d 508 (1997))). Nevertheless, “in accordance with the liberal admissibility standards of the Federal Rules of Evidence, only serious flaws in reasoning or methodology will warrant exclusion.” In re Fosamax Products Liab. Litig., 645 F.Supp.2d 164, 173 (S.D.N.Y.2009) (citing Amorgianos v. Nat’l R.R. Passenger Corp., 303 F.3d 256, 267 (2d Cir.2002)).
a. Defendants’ Motion to Preclude
Defendants seek to exclude the testimony of Plaintiffs’ experts, Kristopher Seluga and Bruce Gorsak. Defendants do not dispute the expert qualifications of Seluga but contend that Seluga’s computer simulation model is not reliable because it has not been validated and uses flawed input values. (Def. Mem. 5-18.) With respect to Gorsak, Defendants argue that he lacks the requisite qualifications to testify as an expert and that his opinions are unreliable and manufactured for the purpose of litigation. (Def. Mem. 19-22.) For the following reasons, the Court finds that neither Seluga nor Gorsak meet the requirements of Rule 702, and, therefore, Defendants’ motion to preclude the testimony of Seluga and Gorsak is granted.
i. Kristopher Seluga
Kristopher Seluga is a forensic engineer and accident reconstructionist. (Tr. 4:2-5.) Seluga has a master’s degree and a bachelor’s degree in mechanical engineering from Massachusetts Institute of Technology. (Tr. 4:21-23.) He is a licensed professional engineer in Connecticut and New York. (Tr. 5:5-7.) Seluga has previously published four articles related to golf car safety, and in a 2006 article he addressed the yaw instability of a golf car during rear-wheel braking. (Tr. 8:2-5, 9:24-10:6.) Seluga testified that, based on his prior research and study of braking systems, he believes that rear-wheel brakes are unstable and “cause the car to lose its directional stability, in other words, spin around.” (Tr. 13:3-10.)
In October of 2007, Seluga went to La Tourette to inspect the accident site. (Tr. 23:24-24:2.) Seluga took photographs and measurements of the golf car that Matthew was driving the day of the accident. (Tr. 24:18-22.) Upon inspection, Seluga saw evidence, including a broken hip restraint and grass in one of the wheels, consistent with a rollover onto the passenger side of the golf car. (Tr. 25:2-8.) Seluga measured the slope where the accident occurred and found that the hill was between seven and nine degrees. (Tr. 25:13-16.) Seluga performed speed tests with the actual golf car, using a hand-held GPS, in two locations — a relatively flat portion of the path and the hill where the accident occurred. (Tr. 34:23-35:2.) The maximum speed he obtained was ten miles per hour on the flat portion and ten and a half miles per hour on the hill. (Tr. 35:3-8.) Seluga testified that he noticed that the golf car was “running very rough,” and, as a result, he believed that the golf car was running slower than it generally would. (Tr. 35:9-19.) Seluga did not calculate the center of gravity of the subject golf car as part of his inspection. (Tr. 32:13-17.) Instead, he calculated the center of gravity based on data that he was later provided. (Tr. 32:16-33:10; PI. Ex. 4.)
While at the golf course, Seluga conducted various tests to determine the coefficient of friction. (Tr. 28:5-9, 29:9-30:4.) The coefficient of friction is the “ratio between the amount of force it takes to slide an object across another surface divided by the amount of weight that is pushing the two objects together.” (Tr. 20:4-10.) With a lower coefficient of friction, there is less braking force, and, as a result, an object will take longer to come to a stop. (Tr. 21:1-7.) There are two different forms of friction that can be tested: static friction and dynamic friction. (Tr. 20:16-25.) Static friction is the amount of friction that it takes for an object to start moving, and dynamic friction is the amount of friction that is generated once the object is moving. (Tr. 20:22-25.) Both methods for testing the coefficient of friction are accepted, but Seluga testified that the static friction test is not as precise as the dynamic friction test. (Tr. 89:4-17.) According to Seluga, the static test, when performed by the same individual, has a margin of error of approximately ten percent. (Tr. 90:4-11.)
Seluga first performed braking tests of the golf car by the maintenance shed in order to determine the coefficient of friction. (Tr. 28:4-9.) Seluga testified that he did not conduct the test for the dynamic coefficient of friction on the actual path because he was “aware of the instability problem and ... was unsure if [conducting the test] would produce a dangerous effect, in other words, it would do what Mr. Valente claimed it did at the time of his accident.” (Tr. 30:13-18.) Seluga claims that he chose the area by the maintenance shed because it was a “relatively flat area,” which “makes it a little easier to do the calculations.” (Tr. 28:10-14.) As can be seen in the photograph Seluga took that day, the path by the maintenance shed had a significant amount of sand on it. (Def. Ex. F.) This area is next to a dumpster and a sandy area, unlike the actual path where the accident occurred, which is bordered by grass. (Tr. 92:22-93:3.) Moreover, the maintenance shed is where vehicles are refueled, maintenance work is performed and a pressure washer is used to wash the “grime, oil, [and] dirt off the vehicles.” (Tr. 236:19-24.) Contaminants such as sand, gas, cleaning fluids and oil can affect the coefficient of friction. (Tr. 236:24-237:1.)
Seluga described how he performed the test for the dynamic coefficient of friction at the maintenance shed path as follows:
I attached an accelerometer to the vehicle and then I recorded the acceleration on the accelerometer as I drove the vehicle up to its top speed. I slammed on the brakes to get the rear wheels to slide. Then I would turn the car around and do that again, and I did that four times.... I obtained measurements of the acceleration of the car. Then I used those to determine what the coefficient of friction between the tires and road must have been in order to produce that amount of deceleration.
(Tr. 28:15-29:1.) Based on these tests, Seluga found that the dynamic coefficient of friction by the maintenance shed was between 0.53 and 0.57. (Tr. 102:25-103:9.) However, in conducting these tests, Seluga failed to account properly for the accelerometer’s gravity bias. (Tr. 51:20-22, 99:4-14.) The gravity bias is the effect that gravity has on the accelerometer. (Tr. 51:10-19.) Accelerometers read the acceleration caused by gravity in addition to the acceleration caused by the golf car’s movement. (Tr. 227:25-228:1.) If an accelerometer is on a slope as opposed to a flat surface, the accelerometer will not read zero, even if the golf car is not moving. (Tr. 51:10-19, 228:2-7.) Here, the path by the shed contained a 1.7 degree slope at the beginning and a 3 degree slope at the end, with a depression between the two rising slopes. (Tr. 94:13-22.) Seluga set his accelerometer to zero at the beginning of his run and, thus, only accounted for the 1.7 degree slope at the beginning of the run. (Tr. 51:20-22, 99:4-14.) He did not correct his results for the 3 degree slope at the end of the run. (Tr. 51:20-22, 99:4-14.) After reviewing the criticisms of Matthew Schwall, one of Defendants’ experts, Seluga acknowledged that he should have measured the bias at the end and recalculated the coefficient of friction. (Tr. 51:23— 52:22.) According to Seluga, after he corrected his calculations, the coefficient of friction for the dynamic tests by the maintenance shed increased between 0.54 and 0.58. (Tr. 52:20-22,102:18-24.)
Seluga also conducted a static test, also known as a drag sled test, on the path by the maintenance shed and the actual slope. (Tr. 29:11-30:4, 83:23-84:2.) The drag sled test involves pulling an exemplar tire and some known amount of weight across the path in order to measure the static coefficient of friction, i.e. the friction required to start moving an object. (Tr. 29:11-30:4.) When Seluga performed the drag sled test by the maintenance shed, he found coefficient of friction values between 0.69 and 0.73. (Tr. 96:15-22.) Seluga then performed the drag sled test on the actual slope, and found the coefficient of friction to be between 0.83 and 0.90. (Tr. 90:17-22.) Despite these various calculations, Seluga used 0.53 exclusively as the coefficient of friction in all of the simulations he ran in preparing his report. (Tr. 103:11— 16.) Seluga testified at his deposition that he used 0.53 because he “wanted to see if this accident could happen a certain way and the lowest friction value that was measurefd] is the most likely to see if it could happen.” (Deposition of Kristopher Seluga (“Seluga Dep.”) 134:17-135:10.)
After collecting the relevant data, Seluga created a computer simulation in order to determine what conditions would “produce a significant yaw instability that would cause the car to rotate sideways.” (Tr. 36:8-18.) Seluga used a software program called Matlab for his simulations. (Tr. 37:18-38:2.) Matlab is analogous to Microsoft Excel in that the user creates the formulas and equations, and the program merely solves the user-created model. (Tr. 41:5-18, 67:1-4.) Accordingly, Seluga wrote the computer code for the model, including the relevant algorithms and formulas. (Tr. 39:5-9.) Seluga conceded at the Dauberb hearing that it was not the general practice of accident reconstructionists to write their own code. (Tr. 38:22-39:2.) Seluga testified, however, that he used algorithms and formulas in writing the code for the simulation that he had previously used in his published, peer-reviewed articles. (Tr. 37:16-20.) However, the formulas and equations had generally only been used in the context of automobiles and large vehicles. (Tr. 41:14-42:5, 45:18-23.) The underlying computer code, i.e. Seluga’s simulation model, has not ever been used or validated by anyone outside of Seluga’s company. (Tr. 68:17-69:2.) The model is a proprietary model and is not commercially available. (Tr. 67:16-20.)
Seluga used a tire model for his simulation that had previously been used for automobiles. (Tr. 38:15-18.) Seluga testified that the particular tire model is not important in this action. (Tr. 47:15-22.) In contrast, Schwall claims that the tire model “is probably the most important part because all forces on a vehicle are coming through the tires.” (Tr. 211:17-20.) Although Seluga maintains that the laws of physics apply identically to automobiles and golf cars, he had to change at least one of the equations — the yaw moment of inertia — used in his model, after he discovered the correction factor used in the equation is significantly different for golf cars than it is for automobiles. (Tr. 57:3-19,117:2-24.)
Based on his simulations, Seluga concluded that, at a speed of 14 miles per hour and on a slope of less than 10 degrees, a golf car would “yaw significantly enough to produce a rollover if the friction of the tires on the road were in the vicinity of about 0.55.” (Tr. 50:4-10.) Seluga testilled that yaw instability “was the best explanation for [Matthew’s] accident.” (Tr. 23:18-23.) Seluga determined that “[t]his instability does not require a major steering input by the driver other than the small steering angle require[d] to follow the gentle left curve of the path and results entirely from the inherent instability of a vehicle with skidding rear wheels and rolling front wheels.” (Seluga Report 11.) Seluga further opined that the simulations “demonstrated that the yaw instability that caused the fishtail would have been avoided if the subject car had been equipped with either front only or 4-wheel brakes.” Id. Seluga concluded that “[i]t may be stated to a reasonable degree of engineering certainty that the subject rollover accident was the result of providing a golf car equipped with brakes on the rear wheels only.” Id. at 14. This conclusion was based on “multiple dynamic computer simulations,” which Seluga conducted in order to evaluate the “yaw instability of the golf car.” Id. at 11.
1. Reliability of Seluga’s Simulation Model
Defendants concede that Seluga is qualified but argue that his simulation model is not reliable. (Def. Mem. 5-18.) As an initial matter, Plaintiffs contend that Seluga’s simulation model does not involve a “novel scientific procedure” and, therefore, “Daubert should be generally inapplicable, or applied in the context of common sense and logic.” (PL Opp’n 23.) As the Supreme Court held in Kumho Tire, “Daubert’s general holding — setting forth the trial judge’s general ‘gatekeeping’ obligation — applies not only to testimony based on ‘scientific’ knowledge, but also to testimony based on ‘technical’ and ‘other specialized’ knowledge.” 526 U.S. at 141, 119 S.Ct. 1167 (quoting Fed.R.Evid. 702). As a result, courts have repeatedly applied Daubert in determining whether a specific accident reconstruction software program is sufficiently reliable to meet the requirements of Rule 702. See, e.g., Royal & Sun Alliance Ins. PLC v. UPS Supply Chain Solutions, Inc., No. 09 Civ. 5935, 2011 WL 3874878, at *8-9 (S.D.N.Y. Aug. 31, 2011) (conducting an analysis regarding whether the expert’s methodology to reconstruct the accident in question met Rule 702’s reliability requirements); Moon v. United States, No. 08 Civ.1990, 2011 WL 181741, at *5 (S.D.N.Y. Jan. 13, 2011) (“As have other courts, I find that the use of PC-Crash to analyze the crash is a reliable methodology.” (collecting cases)); Turner v. Liberty Mut. Fire Ins. Co., No. 07 Civ. 163, 2007 WL 2713062, at *3-4 (N.D.Ohio Sept. 14, 2007) (finding the computer simulation program validated where it was subject to peer review and publication, has known error rates and is generally accepted by the relevant scientific community). Accordingly, the Court must determine whether Seluga’s simulation model is reliable. Plaintiffs bear the burden of proving by a preponderance of the evidence that Seluga’s simulation model meets the requirements of Rule 702. Williams, 506 F.3d at 160. In determining whether a computer simulation is reliable, the court may consider whether the program has been or can be tested, has been subjected to peer review and publication, has a known or potential rate of error and has gained general acceptance in the relevant scientific community. Id. (citing Daubert, 509 U.S. at 593-94, 113 S.Ct. 2786).
a. Validation
Defendants do not dispute that Matlab is an acceptable computer program to use in creating a simulation. (Def. Mem. 8-9; Tr. 243:12-20.) However, Matlab, unlike other generally accepted computer software programs, allows the user to create the underlying mathematical model. (Tr. 41:14-18, 67:1-4; Declaration of David Osterman (“Osterman Deck”) Ex. G (“Schwall Aff.”) 6.) As a result, Seluga’s simulation model must be validated. (Schwall Report 12 (“Only when the underlying physical model has been properly validated and the inputs to the model are accurate will the resulting output of the model be reliable.”).) Plaintiffs contend that Seluga’s simulation model is reliable because (1) the model uses accepted models and laws of physics and, therefore, need not be validated; (2) the model was properly validated using test data; and (3) the model was validated through real-world testing conducted by Defendants’ experts. (PI. Opp’n 25-26; Oral Arg. Tr.42:9-44:16.)
Plaintiffs maintain that “[w]here a simulation utilizes a well-understood and accepted model and laws of physics, and the utilized techniques, models and equations have previously been validated for vehicle simulations, no further validation is necessary for the results to be considered reliable.” (PI. Opp’n 25.) Plaintiffs contend, therefore, that Seluga did not need to perform any tests to validate his simulation. Id. at 25-26. “Engineering testimony rests upon scientific foundations, the reliability of which will be at issue in some cases.” Kumho Tire, 526 U.S. at 150, 119 S.Ct. 1167. A district court must determine whether a methodology, even one based on established scientific foundations, is reliable for the factual issues raised in a particular case. See Dreyer v. Ryder Auto. Carrier Group, Inc., 367 F.Supp.2d 413, 434 (W.D.N.Y.2005) (“The reliability of the expert’s methodology in reaching his conclusions must ... be evaluated against the specific facts at issue, not generalized theories.” (citing Kumho, 526 U.S. at 154, 119 S.Ct. 1167)).
As Schwall testified at the Daubert hearing, the first step in creating a model is to determine the purpose for which the model is being created:
First, in creating a model, what is important is to understand what the purpose of the model will be. And the reason for that is that nature is very complex. In order to create a model, we have to make simplifications. So if we’re creating a vehicle model, we need to know what it will be used for. If it’s used for determining how vehicles behave in a crash, then we are going to be interested in modeling the crush structure of the vehicle whereas if we’re creating a model for vehicle dynamics and handling, we are interested in things such as the tire and suspension but not interested in how it behaves in a crash. So in order to make the simplifications, we need to know how the model will be used and that is the first step.
(Tr. 210:20-211:8.) In other words, simplifications that are permissible to recreate a collision may not be appropriate to recreate a non-impact rollover. See Kumho Tire, 526 U.S. at 154, 119 S.Ct. 1167 (noting that “the reasonableness in general” of using a specific methodology does not mean that using that methodology is reasonable “to draw a conclusion regarding the particular matter to which the expert testimony was directly relevant” (emphasis in original)). “[VJalidation is the process of making sure that despite the simplifications that have been made, the model still accurately predicts the way the object [the expert is] trying to simulate behaves.” (Tr. 212:5-8.) Even a generally accepted computer simulation program, like PC-Crash, which is “based on the laws of physics and accepted principles of accident reconstruction,” Moon, 2011 WL 181741, at *5 (collecting cases), is not a reliable methodology in all factual circumstances, see, e.g., Fairley v. Clarke, No. 02 Civ. 2219, 2004 WL 877102, at *5-6 (E.D.La. Apr. 22, 2004) (holding that PC-Crash, as well as another computer simulation program, was “not the product of reliable principles and methods with regard to the underride accident in question, and the Court is not convinced that the witness has applied the principles and methods reliably to the facts of the case”); State v. Sipin, 130 Wash.App. 403, 123 P.3d 862, 869 (2005) (finding that PC-Crash is not a reliable methodology for simulating multi-impact accidents). Without validation, the Court cannot determine whether Seluga’s simulation model, reliably simulates an accident involving a vehicle rollover.
Seluga’s reliance on equations created and verified for automobiles, as opposed to golf cars, further demonstrates the necessity of validation. Seluga claims that the laws of physics are the same for golf cars and automobiles. (Tr. 13:28-14:1.) However, Seluga concedes that at least one of the automobile-based assumptions used in the original model, the yaw moment of inertia, had to be modified after he learned that a portion of the calculation was different for golf cars. (Tr. 116:21-117:19.) Seluga testified that yaw inertia is a law of physics — “just like force is equal to mass times acceleration, the torque applied to an object is equal to its inertia in the rotational axis times its angular acceleration.” (Tr. 116:21-117:1.) As with the majority of Seluga’s model, he used an equation to calculate the yaw moment of inertia based on automobiles. (Tr. 117:2-10.) The equation involved a correction factor, which Seluga later learned is significantly lower for golf cars than for automobiles. (Tr. 117:5-19.) The original yaw moment of inertia that Seluga used was 0.95, but after he made the correction it was only 0.60. (Tr. 117:20-24.) Seluga testified that this adjustment was significant. (Tr. 117:17-19.) Although this portion of the model was presumably corrected, the Court cannot know whether the remaining automobile-based equations are, in fact, reliable when used in the context of a golf car rollover simulation. Accordingly, the Court finds that Plaintiffs’ claim that Seluga’s simulation model does not need to be validated because it relies on the laws of physics is without merit.
Next, Plaintiffs claim that Seluga adequately validated his model with test data that he received from a manufacturer. (PI. Opp’n 25-26; PI. 56.1 Reply ¶ 198.) Seluga no longer has the data in his possession, (Tr. 76:21-23), and does not recall who provided him with the data, (Tr. 76:24-77:1.) Seluga does not know if the data involved gas or electric golf cars, and he does not know who performed the tests reflected in the data. (Tr. 77:18-25.) The data is confidential and has never been published. (Tr. 77:6-9.) Seluga does not know the coefficient of friction of the test surfaces used in collecting the data. (Tr. 77:2-5.) Seluga does not believe that the data involved vehicles being driven on a hill while braking and steering simultaneously, or simultaneous hard braking and steering on any surface. (Tr. 78:17-23.) Plaintiffs essentially ask the Court to accept Seluga at his word, without any tangible evidence, that this confidential data from an unknown source sufficiently validates Seluga’s simulation model. The Court would not be performing its gatekeeping function, if it merely accepted, without any proof, a party’s contention that its expert’s opinion is reliable.
Even putting aside the various unknowns related to the confidential data, the data could not validate Seluga’s simulation model because, according to Seluga’s testimony, it does not involve conditions that are either similar to or encompass the conditions being simulated for this litigation. Seluga testified that this data does not involve combined braking and steering maneuvers on any surface, let alone a sloped surface. (Tr. 78:17-23.) Schwall contends that Seluga “would not have been able to sufficiently validate his model for the purposes for which he used it in this case,” (Schwall Report 13), without test data that, at a minimum, involved “some combined braking-steering maneuvers,” (Tr. 215:17-21). Similarly, Seluga conceded that he did not know the coefficient of friction involved in the test data that he used to validate his simulation. (Tr. 79:8-11.) The actual dynamics of a tire are different depending on the surface and involve different mechanisms of friction, and, therefore, the model must be validated with an analogous coefficient of friction. (Tr. 217:6-218:1.) In the instant action, where the central issue is the yaw instability of the golf caí-, and the determinative factor is the coefficient of friction, the use of confidential data that does not involve similar circumstances does not render the simulation reliable.
Finally, Plaintiffs argue that the real-world testing conducted by Defendants’ experts validates Seluga’s simulation model. Schwall and David Bizzak, another expert for Defendants, conducted real-world testing, using the same inputs in Seluga’s simulation, in order to determine whether the real-world results were consistent with the simulation. (Schwall Aff. 8.) The only input that was different was the coefficient of friction, which was only different because Schwall and Bizzak performed the tests on the actual slope where the accident occurred, and the slope, as discussed in more detail below, has a significantly higher coefficient of friction than the value used in Seluga’s simulations. Schwall found that Seluga’s simulation model “produce[d] dramatically different results than the vehicle’s actual behavior.” Id. During Schwall’s testing, the vehicle was directionally stable and quickly came to a stop on the path with minimal change in the heading angle. Id. In contrast, the simulations, using the recorded inputs, “predicted significant vehicle rotations before coming to a stop, often fully or partially off the path.” Id. Plaintiffs contend that this testing validates Seluga’s model because, after Defendants performed this testing, Seluga ran his simulation using a coefficient of friction similar to the coefficient of friction that existed on the day Defendants performed their tests and Seluga’s simulation produced results similar to the real-world testing. (Oral Arg. Tr. 43:2-8.) When Defendants’ experts conducted their tests, “the friction ... was on the order of 0.82.” (Tr. 61:6-13.) Seluga ran his simulation, assuming a coefficient of friction of 0.80 or higher, and “the result was that there was very little yaw and in fact it matched very well with their tests.” (Tr. 61:9-13.) As a result, Plaintiffs claim that Defendants’ experts’ tests validate Seluga’s simulation.
Putting aside the issue of whether Seluga’s artificially low coefficient of friction is, in and of itself, fatal to the admissibility of his simulation, the Court finds that the real-world testing of Defendants’ experts does not validate Seluga’s simulation model for the purposes of this case. As Schwall explained at the Daubert hearing, in order to validate a simulation through real-world testing, an individual must put certain inputs into both the simulation and the real-world system and compare the results to see if “they are similar enough within some desired degree of accuracy.” (Tr. 212:5-16.) However, “one needs to use inputs that are representative of the inputs that one would be using when they use the simulation.” (Tr. 212:18-19.) For example, if someone were to create a model for the purpose of calculating how long it takes an object to fall from a specified height, he or she would create the model based on the relevant laws of physics and then attempt to validate it. (Tr. 212:21-24.) If the individual ran tests, dropping a pen from heights up to ten feet and the results were consistent with the simulation, the simulation would be validated for heights up to ten feet. (Tr. 212:25-213:8.) The simulation would not be validated for all heights. As Schwall explained, if he dropped a pen from a height of several stories, the results would not necessarily be consistent with a simulation that was only validated for heights up to ten feet because there are additional factors that must be accounted for when an object falls from a greater height. (Tr. 213:9-13.) For instance, if the model failed to take into account air resistance, that omission would not have a significant effect on the drop time of a pen from heights up to ten feet but would have a significant effect on the drop time of a pen from several stories. (Tr. 213:9-16.)
Accordingly, a simulation must be “evaluated against the specific facts at issue” in order to ensure that the model can reliably recreate the relevant accident at issue. Dreyer, 367 F.Supp.2d at 434; see also Jarvis v. Ford Motor Co., No. 92 Civ. 2900, 1999 WL 461813, at *4-5 (S.D.N.Y. July 6, 1999) (finding that the expert’s methodology was reliable with respect to two of his opinions regarding the causes of sudden acceleration in a car but not the third because only the first two conditions had been “verified through repeated tests on a model that accurately reflects the relevant electrical components” (emphasis added)). Seluga utilized the computer simulations in order to determine the “yaw instability of the golf car.” (Seluga Report 11.) The real-world testing conducted by Defendants’ experts does not address nor validate the yaw instability portion of Seluga’s simulation model, and Seluga concedes that he did not conduct any real-world testing to validate the yaw instability portion of his model. (Tr. 69:21-70:2.) As Schwall explained at the Daubert hearing:
[W]hat Mr. Seluga described yesterday was that he took the data that I had taken, used a high coefficient of friction such as the one that I measured and then received similar results in that the car rapidly came to a stop. What that does is that validates that the model can show a vehicle rapidly coming to a stop. Again, as we discussed the type of inputs used for validation, it certainly doesn’t validate that on a low coefficient of friction, the vehicle has yaw instability. It’s much easier to model a vehicle rapidly coming to a stop than it is to model the complex dynamics of yaw instability.
(Tr. 253:22-254:9 (emphasis added).) Simply because the simulation can reliably predict a real world event does not prove that the simulation can reliably predict any real world event. Defendants argue that Seluga’s simulation model made the following simplifications, which affect its ability to reliably predict yaw instability: (1) use of an overly simplified tire model; (2) failure to include any simulation of the vehicle roll angle, vehicle suspension or tire deflection, or any difference between the path and the grass; and (3) failure to take into account the golf car suspension system. (Schwall Report 12-13; Schwall Aff. 7; Tr. 122:8-18, 211:17-24.) Seluga maintains that neither the specific tire model nor the golf car’s suspension system affect the reliability of the model. (Tr. 47:15-22, 122:8-18.) However, the Court cannot determine whether or not these simplifications affected the reliability of the model because Plaintiffs have not provided any evidence that the model has been validated. The Court finds that Plaintiffs have failed to demonstrate that Seluga’s simulation model has been validated or tested in any meaningful way.
b. Peer Review and Acceptance in the Scientific Community
Seluga concedes that his simulation model has never been subjected to peer review. (Tr. 68:23-69:2.) In fact, other than in litigation, no one outside of Seluga’s company has used or evaluated his simulation model. (Tr. 68:23-69:2, 238:6-12.) His model is not available to the public and, thus, does not have general acceptance in the scientific community. (Tr. 67:16-20.) Moreover, Plaintiffs have not provided the Court with any peer-reviewed literature supporting Seluga’s opinion that when a golf car is driven on a slope of less than ten degrees, at a speed of less than fifteen miles per hour with minimal steer input, a rear-wheel braking system will cause significant yaw instability. To the contrary, in Seluga’s 2006 article, “Braking Hazards of Golf Cars and Low Speed Vehicles,” Seluga states:
The simulation results consistently demonstrate that for speeds at or below the maximum flat ground speed of 24 km/h (15mph), and downhill slopes as high as 10°, the rear brakes configuration is not likely to cause significant yaw displacements before the vehicle comes to a stop. Therefore, for these conditions (ie. no large steering inputs), the rear wheel only braking configuration does not lead to large yaw instabilities.
(PI. Ex. N at 63 (emphasis added).) Plaintiffs argue that this article does not contradict his current opinion because “the presumed coefficient of friction in Seluga’s article ... is much higher than the coefficient of friction measured on areas [sic] cart path surfaces at La Tourette Golf Course.” (PI. Opp’n 43.) Seluga’s article assumes a coefficient of friction of 0.75, which is significantly higher than the coefficient of friction used in Seluga’s simulations but consistent with the coefficient of friction measured on the actual slope where the accident occurred. (Tr. 75:8-12.) In any event, even if the Court was persuaded by Plaintiffs’ attempts to distinguish Seluga’s 2006 article, Plaintiffs have only established that his opinion is not contradicted by peer-reviewed literature. Plaintiffs still have not provided any peer-reviewed literature supporting his model or his opinion or any other evidence that either his model or his opinion is generally accepted in the scientific community.
c. Error Rate
The Court also finds that Seluga’s simulation model is not reliable because its error rate is unknown and cannot be determined. Seluga testified that he ran the simulation, which forms the basis of his opinion, dozens of times. (Tr. 129:20-24.) Each time he ran it, he added a random noise component. (Tr. 59:17-60:3, 127:5-17.) Noise, according to Seluga, introduces random variations to the underlying inputs to “represent the actual ... randomness that occurs in the real world.” (Tr. 59:23-60:3.) The model automatically changes the steering input and coefficient of friction each time that it is run because of the random noise component. (Tr. 127:9-17.) Random noise is not a part of other, existing accident reconstruction models. (Tr. 128:24-129:1.) Seluga testified at his deposition that he determined the effect of the noise component on the steering input and coefficient of friction based on “[his] understanding of [how] the car actually works.” (Seluga Dep. 185:9-11.) Seluga does not offer any data to support this determination. (Seluga Dep. 185:17-186:5.)
As a result of random noise, a different expert cannot verify Seluga’s results. (Tr. 129:2-14, 237:21-22.) As Schwall testified, he cannot “quantify the error” because the input values change each time he runs the simulation. (Tr. 238:4-5.) Moreover, Seluga did not keep a record of the input values used in his simulations after the random noise adjustment; in fact, Seluga did not even keep a record of the number of times that his simulation actually predicted a rollover:
DEFENSE COUNSEL: Did you produce the data from the dozens of runs or did you only produce it from one run?
SELUGA: Only from one because it doesn’t get saved anywhere unless you actively save it.
DEFENSE COUNSEL: Did you write down in your notes anywhere how many times out of a hundred [the simulation] predicted a rollover?
SELUGA: No. I know it was definitely not all of the time. It was not a one in a hundred chance but it was something I think around half the time, maybe a quarter of the time.
DEFENSE COUNSEL: So the simulation that you say best represents the accident scenario in this case, the simulation upon which the animation is based is a simulation that predicted a rollover approximately 25 percent of the time?
SELUGA: Like I said, it might have been up to 50 percent. I don’t know. It was something like that.
(Tr. 130:4-14.) Schwab ran Seluga’s simulation with a 0.53 coefficient of friction and found that it predicts a rollover 25 percent of the time. (Tr. 235:17-20.) Seluga’s use of random noise and his failure to keep track of the simulation results undermine the reliability of both his model and his opinion.
2. Coefficient of Friction
The reliability of Seluga’s simulation model is further compromised by the unreasonably low coefficient of friction used in his simulations. Although challenges to an expert’s assumptions generally go to the weight, not the admissibility, of the testimony, “a trial judge should exclude expert testimony if it is speculative or conjectural or based on assumptions that are ‘so unrealistic and contradictory as to suggest bad faith’ or to be in essence ‘an apples and oranges comparison.’ ” Zerega Ave. Realty Corp. v. Hornbeck Offshore Transp., LLC, 571 F.3d 206, 213-14 (2d Cir.2009) (quoting Boucher v. U.S. Suzuki Motor Corp., 73 F.3d 18, 21 (2d Cir.1996)); see also Major League Baseball Prop., 542 F.3d at 311 (“An expert’s opinions that are without factual basis and are based on speculation or conjecture are ... inappropriate material for consideration on a motion for summary judgment.”).
All parties agree that the coefficient of friction is “the determining factor” in the instant action. (Tr. 60:18-23, 79:12-19, 236:4-10.) Seluga concedes that in circumstances where the coefficient of friction is 0.80 or higher, there would be no significant yaw. (Tr. 61:9-13, 62:17-21.) Seluga’s 2006 article states that there is no significant yaw, when the coefficient of friction is 0.75. (Tr. 75:8-12.) Seluga’s simulation model does not ever predict a rollover when the coefficient of friction is 0.57 or higher. (Tr. 235:17-20.) Here, Seluga used a coefficient of friction of 0.53. (Tr. 103:11-16.) In light of the existing literature, and the tests conducted by Seluga and Defendants’ experts, the Court finds that Seluga’s use of a 0.53 coefficient of friction is so unrealistic and speculative that it renders his simulations unreliable.
First, the Court finds that Seluga’s own testing methodology demonstrates that the 0.53 coefficient of friction is artificially low. Seluga conducted two types of friction tests — dynamic and static — in two different areas of the golf course. Seluga only conducted the dynamic test by the maintenance shed and found that the coefficient of friction was between 0.53 and 0.57. (Tr. 102:25-103:9.) He conducted the static tests in both locations and found the coefficient of friction on the path by the maintenance shed was between 0.69 and 0.73, (Tr. 96:15-22), and the coefficient of friction on the actual slope was between 0.83 and 0.90, (Tr. 90:17-22). After conducting these tests, Seluga knew that the “slope at the accident scene” had a higher coefficient of friction than the path by the maintenance shed. (Tr. 104:22-105:4.) Despite knowing this, Seluga used a 0.53 coefficient of friction, which is the lowest value that he received when conducting his tests by the maintenance shed. (Tr. 102:25-103:9.) Further undermining the reliability of Seluga’s coefficient of friction value are the corrections he made to adjust for the gravitational bias. (Tr. 102:18-24.) After Seluga’s calculations, he determined that for the dynamic test by the maintenance shed, his results were actually between 0.54 and 0.58. (Tr. 102:18-24.) Thus, Seluga’s own tests did not ever produce a coefficient of friction value as low as 0.53.
Plaintiffs contend that the coefficient of friction value that Seluga used is consistent with peer-reviewed literature. At the Daubert hearing, Plaintiffs presented a table, which appears to be excerpted from either a peer-reviewed paper or textbook entitled “Skidmarks Analysis,” listing the coefficient of friction values for an automobile tire on various surfaces. (PI. Ex. 6.) The table indicates that the coefficient of friction on sand is 0.55 and on dry asphalt is 0.80. Id. This article does not validate but rather further refutes Seluga’s use of 0.53. Plaintiffs have not presented any evidence that the path where the accident occurred was covered in sand, such that the coefficient of friction for sand would be a relevant number. Moreover, Seluga’s 0.53 coefficient of friction value is not only contradicted by peer-reviewed literature— both the “Skidmarks Analysis” table and Seluga’s 2006 article, which assumed a coefficient of friction of 0.75 — but also the tests conducted by Defendants’ experts.
Defendants’ experts performed dynamic tests on the actual slope in July of 2009 and October of 2011. (Tr. 219:20-220:6.) In July of 2009, Schwall measured the coefficient of friction on the slope where the accident occurred using the actual golf car and an exemplar golf car, and he determined that the coefficient of friction was between 0.85 and 1.0. (Tr. 221:21-222:8; Schwab Report 9.) In performing the tests, Schwab used both a GPS and an accelerometer to verify the accuracy of his results. (Tr. 222:1-8; Schwab Report 9.) In October of 2011, after the path had been repaved, Schwab conducted additional tests with the actual golf car on the actual path. (Tr. 222:20-223:2.) Schwab found the coefficient of friction to be in the range of 0.85 to 0.95. (Tr. 223:6-7; Schwab Report 9.) Schwab’s values were “consistent with published results for tires on dry asphalt.” (Schwab Report 10.) As the below chart emphasizes, Seluga’s use of a 0.53 coefficient of friction is completely unrealistic given the coefficient of friction values obtained by all of the experts:
Maintenance Shed Actual Slope
Static 0.69-0.73 0.53-0.57 (without bias correction) 0.54-0.58 Dynamic (Seluga’s bias correction) 0.61 (Schwab’s bias correction) 0.83-0.90 0.85-1.0 (old pavement) 0.86-0.95 (new pavement)
(Def. Ex. G.) Seluga concedes that he used 0.53 in order to get the desired result, a rollover, (Seluga Dep. 134:17-135:10), and that his opinion is dependent on an equal amount of sand having been present on the slope where the accident occurred on the day of the incident, (Tr. 108:17-109:1).
There is no evidence in the record to suggest that on the day of the accident the path that Matthew was driving on was covered in sand and, thus, no basis for this assumption. See Russo v. Keough’s Turn of the River Hardware, LLC, No. 11 Civ. 994, 2012 WL 4466626, at *3-4 (S.D.N.Y. Sept. 25, 2012) (finding no basis for the expert’s assumption that a lack of complete uniformity in the ladder’s thickness caused the accident in question, particularly where the expert never measured the thickness of the ladder in the actual area where it folded). Accordingly, the Court finds that Seluga’s use of an unrealistically low coefficient of friction value is further evidence that his simulations are not admissible under Rule 702. See id. at *4 (“There is simply too great an analytical gap between the data measured by Dr. Marietta and the assumption that the ladder was defectively manufactured.”); Baker v. Urban Outfitters, Inc., 254 F.Supp.2d 346, 354 (S.D.N.Y.2003) (excluding the expert’s testimony where she was “engaging in the sort of ‘apples and oranges’ comparison that has been rejected in the past as irrelevant”); see also Lynn ex rel. Lynn v. Yamaha Golf-Car Co., 894 F.Supp.2d 606, 618-19 (W.D.Pa.2012) (“[Cjourts ... have held that preclusion of an expert’s testimony is proper when necessary to screen inaccurate data used in computer modeling.”).
In conclusion, the Court finds that Seluga’s simulation model does not meet the reliability requirements of Rule 702. The underlying model has not been validated, does not have a known error rate, has not been subjected to peer review and does not have general acceptance in the scientific community. The input value for the coefficient of friction is so contradictory to the facts of this case that it renders the simulation model inadmissible. Moreover, the ultimate opinion reached by Seluga is contradicted by his own 2006 peer-reviewed article and by his own simulation, which only gets the desired outcome 25 percent of the time even after he assumes that a significant amount of sand was on the path the day of the accident. Simply put, Plaintiffs have not presented any evidence from which the Court could find that Seluga’s simulation model is reliable for the purposes of this litigation. Having found the simulations inadmissible, the Court concludes that Seluga does not have a proper basis for his opinion, and his testimony must be excluded. See Ruggiero v. Warner-Lambert Co., 424 F.3d 249, 253 (2d Cir.2005) (“[W]hen an expert opinion is based on data, a methodology, or studies that are simply inadequate to support the conclusions reached, Daubert and Rule 702 mandate the exclusion of that unreliable opinion testimony.” (citation omitted)). The Court grants Defendants’ motion to exclude the testimony of Seluga.
ii. Bruce Gorsak
Bruce Gorsak, Plaintiffs’ second expert, contends that the golf car was defective because it lacked seatbelt restraints, four-wheel brakes and an adequate warning. Defendants argue that Gorsak is unqualified and that his opinions are unreliable. (Def. Mem. 19.) Gorsak received a bachelor’s degree in mechanical engineering technology from the Milwaukee School of Engineering. (Tr. 137:Il-ls.) Whereas mechanical engineering involves the actual design of the mechanical aspects of a product, mechanical engineering technology involves the techniques and systems used for manufacturing a product in accordance with the mechanical engineer’s design. (Tr. 171:8-19.) Gorsak is not a licensed mechanical engineer and, therefore, by law is not permitted to design products or otherwise offer himself out as a mechanical engineer. (Tr. 171:18-172:24.) Gorsak previously worked at E-Z Go, but he worked in the maintenance department and was responsible for supervising the plant. (Tr. 139:17-19.) Gorsak was not involved with the manufacturing of E-Z Go’s golf cars. (Tr. 139:17-18.)
In preparing his report, Gorsak did not go to La Tourette, inspect the subject golf cart or conduct any independent testing. (Tr. 179:6-17.) At the time he prepared his report, Gorsak was not aware of the relevant American National Standards Institute (“ANSI”) standards or of any standards governing the golf car at issue. (Tr. 180:9-19.) Gorsak conceded at the Daubert hearing that he did not rely on or consider the ANSI standards specific to golf cars in forming his opinions. (Tr. 182:16-19.) Instead, Gorsak’s opinions were based on the federal standards for automobiles. (Tr. 185:17-20.) With respect to the rear-wheel brakes, Gorsak states that “[t]he 2-wheel brake application of the E-Z-GO TXT golf car does not allow for any form of failure, it is not fail-safe, is defective from the design standpoint, and is a use [sic] of Valente’s injuries.” (Gorsak Report 7.) In his report, Gorsak does not provide any underlying analysis or reasoning for this conclusion. At the Daubert hearing, Gorsak explained that he relied on his familiarity with the brake systems that are used in golf cars in forming his opinion. (Tr. 150:4-6.) When asked the basis of this familiarity, Gorsak responded: “We fixed them. I didn’t fix them. I had guys that would do that in my department if there were problems.” (Tr. 150:9-10.) Gorsak testified that he could not say to a reasonable degree of engineering certainty that Matthew’s accident could have been prevented by four-wheel brakes without conducting testing. (Tr. 194:9-195:10.)
Similarly, when Gorsak testified regarding the need for a restraint system in the subject golf ear, he stated that his opinion was based on his “use of the golf car every day for several years, [and] that you can feel the propensity to go out of it, especially if you are in a hurry.” (Tr. 158:7-9.) Gorsak opined that the lack of a restraint system was one of the reasons that Matthew had his accident. (Tr. 160:18-22.) However, Gorsak did not know whether the G forces involved in Matthew’s accident would have been sufficient to trigger the lock-up function of a retractable seat-belt, even if the golf car had been equipped with such a belt. (Tr. 198:16-22.) When asked whether he would suggest a three, five or six point harness system, Gorsak responded that he would have to conduct some testing or research in order to determine the best harness system. (Tr. 197:5-198:5.) Gorsak was not aware of any peer-reviewed publication that recommended a retractable seatbelt or a fixed-lap belt for golf cars intended for use under 15 miles per hour. (Tr. 199:20-23.)
Finally, Gorsak testified that the warnings on the E-Z Go golf car do not sufficiently advise the user of potential hazards and what the user must do in order to avoid those hazards. (Tr. 162:1-5.) At the Daubert hearing, Gorsak admitted that he did not “specifically state what the warning is” in his report. (Tr. 163:7-164:2.) Gorsak could not recall the specific language of the warning, but he testified that it “brings up [that] you could be seriously hurt but then it downsizes it, if you will to use a common term, to just drive safely.” (Tr. 164:5-8.) Gorsak stated that his opinion that the warning was insufficient was based on “the Peters book,” which says that a warning that “calls for safe use, avoidance of use and abuse, and to be careful” is not sufficient. (Tr. 164:9-14.) Gorsak could not offer an opinion to a reasonable degree of engineering certainty that a different warning would have prevented the accident. (Tr. 196:20-25.) Gorsak did not draft an alternative warning. (Tr. 195:20-23.)
1. Qualifications
Gorsak does not have the requisite “knowledge, skill, experience, training or education” in the area in which his testimony is offered. Fed.R.Evid. 702. Although Gorsak need not have an engineering license in order to testify regarding design defects, he must be “educated and experienced enough to make himself an expert regarding the [product] in question.” Russo v. Keough’s Turn of the River Hardware, LLC, No. 11 Civ. 994, 2012 WL 4466626, at *3 (S.D.N.Y. Sept. 25, 2012); see also Dreyer, 367 F.Supp.2d at 430 (“[W]hile experience can provide the basis to qualify a witness as an expert, the experience must be demonstrated and have direct relevance to the issues in the case.” (collecting cases)). Gorsak has a degree in mechanical engineering technology, but he does not have a degree in mechanical engineering and is not a licensed engineer. (Tr. 137:11-13,171:18-22.) He does not have any special training or expertise related to golf cars, braking systems, restraint systems or warnings. He has never designed a braking system, restraint system or written a warning for a commercially-available product. (Tr. 142:15-19, 178:8-23.) Gorsak was previously employed by E-Z Go, but he was a maintenance supervisor, whose job required him to fix conveyor belts and did not have any involvement in the design process. (Tr. 139:17-19.) He was not involved with the manufacturing of E-Z Go’s golf cars or any golf cars. (Tr. 139:17-18.) In fact, Gorsak has little experience with golf cars, claiming that he developed his expertise by driving a golf car “every day of the week for two and a half years.” (Tr. 142:19-20.)
Accordingly, the Court finds that Gorsak does not have the knowledge or experience required in order to qualify him as an expert in this action. See Fernandez v. Cent. Mine Equip. Co., 670 F.Supp.2d 178, 184 (E.D.N.Y.2009) (“Although Anderson is a licensed mechanical engineer, his work experience contains little to no involvement in the drilling industry. Nothing in his resume or his deposition testimony permits this Court to conclude that Anderson has any relevant experience in the field of geotechnical or water well drilling or with a manually operated cathead, such as the one at issue in this litigation.”); Solorio v. Asplundh Tree Expert Co., No. 02 Civ. 8035, 2009 WL 755362, at *2 (S.D.N.Y. Mar. 23, 2009) (noting the expert’s “general lack of qualifications,” where the expert had “a degree in ‘mechanical engineering technology’ not ‘mechanical engineering,’ ” was not a professional engineer and had not written any “peer-reviewed articles, participated in the drafting of any standards or regulations, designed a product that went to market, or received any patents” (emphasis in original)); Barban v. Rheem Textile Sys., Inc., No. 01 Civ. 8475, 2005 WL 387660, at *3-4 (E.D.N.Y. Feb. 11, 2005) (holding that the expert was not qualified to testify as an expert regarding the design of a laundry press machine because he “has never designed a machine of any kind, and has never worked with laundry machines in any capacity that bears on the conclusions he reaches in this case”), aff'd, 147 Fed.Appx. 222 (2d Cir. 2005).
2. Reliability and Relevance
Even assuming arguendo that Gorsak is qualified to testify as an expert, his proposed testimony is neither relevant nor reliable. In forming his opinion, Gorsak did not conduct any independent testing. (Tr. 179:6-17.) He did not examine the golf car at issue or an exemplar golf car. (Tr. 179:6-17; Deposition of Bruce Gorsak (“Gorsak Dep.”) 13:3-10.) Gorsak is not able to state with a reasonable degree of engineering certainty that either the absence of four-wheel brakes, seatbelts or an adequate warning caused Matthew’s accident. (Tr. 194:9-19, 196:20-25; Gorsak Dep. 69:2-11.) Gorsak testified that he could not make that determination without conducting tests, and he did not conduct any tests. (Tr. 194:20-195:10.) Gorsak did not even consider the governing standards for golf cars in reaching his opinions. (Def. Mem. 20.)
Gorsak merely reiterates, without any basis