Justice BAKER
delivered the opinion of the Court,
in which Chief Justice PHILLIPS, Justice HECHT, Justice OWEN, Justice ABBOTT, Justice HANKINSON, and Justice GONZALES joined.
These consolidated cases present two issues: (1) whether the good faith factors in Wadewitz v. Montgomery, 951 S.W.2d 464 (Tex.1997), an emergency response case, apply in a police pursuit case; and (2) whether the defendants in these consolidated cases have established good faith as a matter of law.
We hold that the Wadewitz factors apply to good faith determinations in police pursuit cases. In University of Houston v. Clark, we conclude that Sergeant Williams and Officer Stewart established good faith as a matter of law. Therefore, we reverse the court of appeal’s judgment on the good faith issue. Because the court of appeals did not consider whether Officer Stewart was performing a ministerial or discretionary function, we remand the case to the court of appeals to determine that issue. In Ener v. Thomas, we conclude that Deputy Constable Ener did not establish good faith as a matter of law, and therefore we affirm the court of appeals’ judgment.
I. BACKGROUND
A Clark
At about 2:00 a.m. one morning, Kevin Thomas was involved in a fist fight on the University of Houston campus. University of Houston police officers arrived at the scene. Thomas got into his jeep and the officers told Thomas not to leave the scene. But before the officers could identify him, Thomas fled. As Thomas left campus, Sergeant Jon Williams turned his patrol car sideways to try to stop Thomas, but Thomas drove his jeep around Williams’ car. Williams pursued Thomas. Williams then saw another University of Houston patrol car in position to intercept Thomas. Williams radioed that car’s driver, Officer Matthew Stewart, to stop [580]*580Thomas. Stewart activated his lights and siren and pursued Thomas. During the chase, Thomas ran a red light and collided with another vehicle, injuring Demetria Clark.
Clark sued Williams, Stewart, and the University of Houston for negligence, negligence per se, and civil rights violations.1 Williams and Stewart moved for summary judgment based on official immunity. The University of Houston moved for summary judgment based on sovereign immunity, relying on Williams’ and Stewart’s official immunity. The trial court granted summary judgment for the defendants, and Clark appealed.
On appeal, Clark did not dispute that Williams and Stewart were acting within the scope of their authority, but claimed that they were not acting in good faith and that Stewart was performing a ministerial rather than discretionary task in pursuing Thomas. The court of appeals reversed the trial court’s judgment on the ground that Williams and Stewart did not conclusively establish that they pm-sued Thomas in good faith. 979 S.W.2d 707. Accordingly, the court of appeals did not consider whether Stewart was functioning in a ministerial or discretionary capacity, and we’ do not consider that issue here.
B. Ener
Ener also involves an early morning police pursuit. Steven Gary Ener, a Harris County Deputy Constable, was on patrol. According to the plaintiffs’ response to the motion for summary judgment, Ener had a civilian passenger, Francis G. Thomas, Jr., riding with him, in violation of Harris County Constable policies. Ener observed a speeding car, which Ener’s radar indicated was going 71 miles per hour in a 55 mile-per-hour zone. Ener pursued the car and, at an intersection, Ener collided with another car. As a result, Thomas and John Alexander Roarke suffered injuries. According to the plaintiffs’ summary judgment response, Ener does not remember anything from the time he was about a block from the intersection where the collision occurred until he woke up in the hospital. Thomas and Roarke sued Ener and Harris County for negligence and gross negligence. Ener moved for summary judgment based on official immunity and Harris County moved for summary judgment based on sovereign immunity, relying on Ener’s official immunity. The trial court denied summary judgment. The court of appeals affirmed, holding that Ener did not establish good faith as a matter of law. 20 S.W.3d 712.
II. APPLICABLE LAW
Official immunity is an affirmative defense that protects government employees from personal liability. See City of Lancaster v. Chambers, 883 S.W.2d 650, 653 (Tex.1994). When official immunity shields a governmental employee from liability, sovereign immunity shields the governmental employer from vicarious liability. See DeWitt v. Harris County, 904 S.W.2d 650, 653 (Tex.1995). A governmental employee is entitled to official immunity: (1) for the performance of discretionary duties; (2) within the scope of the employee’s authority; (3) provided the employee acts in good faith. See Chambers, 883 S.W.2d at 653. Because official immunity is an affirmative defense, to obtain summary judgment on official immunity, the governmental employee must conclusively prove each element of the defense. See Kassen v. Hatley, 887 S.W.2d 4, 8-9 (Tex.1994).
In Chambers, we formulated a test for official immunity’s good faith element in police pursuit cases. See Chambers, 883 S.W.2d at 656. In doing so, we recognized the competing interests involved in good faith cases: (1) the injustice of imposing liability on an officer whose job requires him to exercise discretion and the danger that such liability will deter his willingness to exercise that discretion for the public [581]*581good; and (2) the rights of the public who are affected by an officer’s bad faith acts. See Chambers, 883 S.W.2d at 656. To accommodate these competing interests, we constructed a test for good faith that is analogous to the abuse of discretion standard of review. See Chambers, 883 S.W.2d at 657, n. 7. Under this test, an officer acts in bad faith only if he could not have reasonably reached the decision in question. See Chambers, 883 S.W.2d at 657 n. 7 (citing Landry v. Travelers Ins. Co., 458 S.W.2d 649, 651 (Tex.1970) (holding that abuse of discretion is shown only if a trial court could not have reasonably reached the decision in question)).
Thus, we held that to obtain summary judgment on good faith in a pursuit case, a police officer must prove that a reasonably prudent officer, under the same or similar circumstances, could have believed that the need to immediately apprehend the suspect outweighed a clear risk of harm to the public in continuing the pursuit. See Chambers, 883 S.W.2d at 656-57. An officer does not have to prove that it would have been unreasonable to stop the pursuit or that all reasonably prudent officers would have continued the pursuit. See Chambers, 883 S.W.2d at 656. Instead, the officer must prove only that a reasonably prudent officer might have believed that he should have continued the pursuit. See Chambers, 883 S.W.2d at 656-57.
But to controvert a police officer’s summary judgment proof on good faith, the nonmovant must do more than show that a reasonably prudent officer could have decided to stop the pursuit. See Chambers, 883 S.W.2d at 657. The nonmovant must show that no reasonable person in the officer’s position could have thought that the facts justified the officer’s acts. See Chambers, 883 S.W.2d at 656. Because this good faith test was a new legal standard, we did not apply it in Chambers, but remanded the case to the trial court for further proceedings consistent with our opinion. See Chambers, 883 S.W.2d at 657.
In Wadewitz, we applied the Chambers good faith test for the first time and elaborated on the test’s risk and need elements in the context of an emergency response. See Wadewitz, 951 S.W.2d at 467. The need element refers to the “urgency of the circumstances requiring police intervention,” or “the seriousness of the crime or accident to which the officer responds, whether the officer’s immediate presence is necessary to prevent injury or loss of life or to apprehend a suspect, and what alternative courses of action, if any, are available to achieve a comparable result.” Wadewitz, 951 S.W.2d at 467. The risk element of good faith refers to “the countervailing public safety concerns,” or “the nature and severity of harm that the officer’s actions could cause (including injuries to bystanders as well as the possibility that an accident would prevent the officer from reaching the scene of the emergency), the likelihood that any harm would occur, and whether any risk of harm would be clear to a reasonably prudent officer.” Wadewitz, 951 S.W.2d at 467.
We also emphasized that conclusory statements that a reasonable officer could or could not have taken some action will neither establish good faith on summary judgment nor raise a fact issue to defeat summary judgment. See Wadewitz, 951 S.W.2d at 467. Instead, testimony on good faith must discuss what a reasonable officer could have believed under the circumstances, and must be substantiated with facts showing that the officer assessed both the need to apprehend the suspect and the risk of harm to the public. See Wadewitz, 951 S.W.2d at 467.
III. ANALYSIS
A. Does Wadewitz apply to
A POLICE PURSUIT CASE?
In both Clark and Ener, the court of appeals analyzed the defendants’ summary judgment evidence under the Wa-dewitz factors. 979 S.W.2d 707, 713; 20 S.W.3d 712, 715. The University of Hous[582]*582ton argues that the Wadewitz factors do not apply in a police pursuit case. It contends that the risk considerations in emergency response cases and police pursuit cases are fundamentally different. The primary difference, the University argues, is that in an emergency response case, the officer has more control over the consequences of his actions in responding to an emergency and therefore can more easily assess the risks involved in his actions. The typical risk that an officer faces in responding to an emergency is that he will collide with a third party on his way to the emergency. But, the University argues, in a police pursuit case, the typical risk is that the criminal suspect being pursued will injure a third party. Thus, the University contends, it is more difficult to control the consequences of a police pursuit and evaluate its risks. The University also argues that, in responding to an emergency, an officer has the benefit of a defined point in time and space to balance the need to respond and risks of responding and then make a decision. But in police pursuits, it argues, officers must respond to an ongoing and evolving situation.
Clark counters that any distinction between emergency responses and police pursuits does not necessitate different risk and need assessments. Clark notes that whether the risk is a criminal suspect colliding with a third party or the officer colliding with a third party, the officer retains the same amount of discretion to initiate or continue his pursuit or emergency response. In both scenarios, his decisions impact the risks. Further, Clark points out that the risk analysis in both emergency response and police pursuit cases does not focus only on a specific point in time. Rather, officers in both situations will have to evaluate risks and make decisions in response to changing situations. Lastly, Clark points out that courts of appeals have applied Wadeimtz in police pursuit cases. See State v. Saenz, 967 S.W.2d 910, 914 (Tex.App.-Corpus Christi 1998, pet. denied); Chapa v. Aguilar, 962 S.W.2d 111, 114 (Tex.App.-Houston [1 st Dist.] 1997, no pet.).
We agree with Clark that the Wadewitz need and risk factors apply to good faith determinations in police pursuits as well as emergency responses. The Chambers balancing test inherently includes all the Wo-dewitz factors. Chambers requires, in the summary judgment context, that an officer in a police pursuit case conclusively prove that a reasonably prudent officer, under the same or similar circumstances, could have believed that the need to immediately apprehend the suspect outweighed a clear risk of harm to the public in continuing the pursuit. See Chambers, 883 S.W.2d at 656. The Chambers need element requires an officer to assess the need to immediately apprehend the suspect. In both an emergency response and a police pursuit, such an assessment necessarily includes the general considerations Wa-dewitz provides: the seriousness of the crime or accident to which the officer responds, whether the officer’s immediate presence is necessary to apprehend a suspect or to prevent injury or loss of life, and what alternative courses of action, if any are available to achieve a comparable result. See Wadewitz, 951 S.W.2d at 467. Similarly, the Chambers risk element — “a clear risk of harm to the public in continuing the pursuit” — necessarily includes considerations of the nature, severity, likelihood, and obviousness of the risks of the officer’s actions. Wadewitz, 951 S.W.2d at 467.
The University’s arguments do not persuade us otherwise. First, the University points out that, in an emergency response, the officer typically has a fixed destination and in a police pursuit, the officer does not know where the chase will ultimately end. But both emergency responses and police pursuits may involve rapidly changing circumstances. See, e.g., Wadewitz, 951 S.W.2d at 467. Therefore, contrary to the University’s arguments, both situations require a continuing assessment of need and [583]*583risk. See Wadewitz, 951 S.W.2d at 467 (holding that an officer responding to an emergency did not establish good faith because although his summary judgment evidence described why he chose to proceed on an emergency basis and why he chose a certain route to the scene, it did not include facts showing that he evaluated the risks created by his actions on the way to the emergency); Chambers, 883 S.W.2d at 656-57 (“[A]n officer must prove that a reasonably prudent officer might have believed that the pursuit should have been continued.”).
Second, despite their differences, pursuing a suspect and responding to an emergency involve the same general risk to the public — collision with a third party. The University distinguishes between emergency responses and police pursuits based on who is at risk of colliding with a third party. The University argues that in an emergency response, the risk is that an officer will collide with a third party, while in a police pursuit, the risk is that the suspect will collide with a third party. On the contrary, as the facts in Ener demonstrate, an officer in a police pursuit must assess both the risk that the suspect will injure a third party and the risk that the officer himself will injure a third party.
These multiple risks in a pursuit are almost always more difficult to assess than an emergency response’s risks. And although the general risks of a police pursuit are foreseeable, the fleeing suspect’s specific actions during pursuit are likely to be unpredictable and create unforeseeable risks. But both Wadewitz and Chambers acknowledge that the circumstances will differ in every case and that balancing risk and need must be done in light of those particular circumstances. See Chambers, 883 S.W.2d at 656; Wadewitz, 951 S.W.2d at 467. Indeed, one of Wadewitz ⅛ risk factors is “whether any risk of harm would be clear to a reasonably prudent officer.” Therefore, applying the Wadewitz factors does not penalize an officer for his inability to perceive or evaluate a risk due to circumstances beyond his control, such as the necessity to make a split-second decision, or the inability to foresee a particular risk caused by a fleeing suspect or otherwise. Nor does it prevent an officer from pursuing suspects for traffic violations or from pursuing suspects in residential or other populated or high traffic areas. The good faith test is whether, considering all the circumstances, a reasonably prudent officer could have believed that the need to immediately apprehend a suspect is outweighed by a clear risk of harm to the public in initiating or continuing pursuit. See Chambers, 883 S.W.2d at 656.
We recognize that, depending on the circumstances, an officer may not be able to thoroughly analyze each need or risk factor, and that this alone should not prevent the officer from establishing good faith. Police officers must make their decisions about pursuing a suspect rapidly and while under pressure. See Travis v. City of Mesquite, 830 S.W.2d 94, 98 (Tex.1992). At the same time, even if the circumstances require an officer’s immediate response, an officer pursuing a suspect in good faith should be able and should be required to quickly balance the need to immediately apprehend the suspect, based on the seriousness of the situation and the existence of any alternative actions, with the degree and likelihood of the risks of the officer’s actions. See Wadewitz, 951 S.W.2d at 467.
Accordingly, we conclude that to establish good faith in a police pursuit case, an officer must conclusively prove that a reasonably prudent officer in the same or similar circumstances could agree that the need to immediately apprehend the suspect outweighed the risk of harm to the public in continuing the pursuit, taking into account all the Wadewitz factors.
B. The Dissent
The dissent recognizes that police pursuits require a reassessment of whether to continue the pursuit in response to changing circumstances during the pursuit. 22 [584]*584S.W.3d 927. But the dissent asserts that entering a particular intersection or making a particular U-turn does not necessitate a reassessment of risk, because these circumstances do not significantly change the risks involved, but simply present the risks already contemplated in the initial decision to pursue. 22 S.W.3d 927. We disagree that these circumstances can never present a change in risks that would require a reassessment of need versus risk. See generally Wadewitz, 951 S.W.2d at 467 (holding that an officer did not establish good faith in an emergency response because he did not assess the risks created by his actions, including entering a blind intersection). For example, when an officer approaches a busy intersection with a red light and intersecting traffic is approaching the intersection on a green light and at a high rate of speed, the risk of collision significantly increases. Such an increase in risk could affect the balance between need and risk and thus, requires a reassessment.
The requirement to reassess does not, as the dissent argues, change the standard for liability from good faith to negligence. The officer’s actions are still judged under the good faith test — whether a reasonably prudent officer could have believed that the need to continue the pursuit outweighed a clear risk of harm to the public. See Chambers, 883 S.W.2d at 656-57. Again, the officer need not prove that it would have been unreasonable to stop the pursuit, nor must the officer prove that all reasonably prudent officers would have continued the pursuit. See Chambers, 883 S.W.2d at 656. As the dissent recognizes, if a reasonably prudent officer could have continued the pursuit despite the change in risk, the officer continues acting in good faith. 22 S.W.3d at 926-27.
Further, we disagree with the dissent that the availability of alternatives to pursuit is irrelevant in a police pursuit case. First, Chambers requires an officer to balance the “need to immediately apprehend the suspect” with “a clear risk of harm to the public in continuing the pursuit.” Chambers, 883 S.W.2d at 656. The need to immediately apprehend a suspect depends, among other things, on whether the suspect could be apprehended at a later time, i.e., whether there is an alternative to pursuing the suspect that would achieve a comparable result. Second, as a policy matter, given the inherent risk to the public in every pursuit, officers should be required to recognize reasonable alternatives to pursuit, and to justify their decisions to pursue when those alternatives exist by showing that the need to pursue nevertheless outweighed the risks. Requiring officers to assess alternatives insures that public safety is not “thrown to the winds in the heat of the chase.” Travis, 830 S.W.2d at 98.
C. Have the Defendants Here Established Good Faith as a Matter of Law?
We next consider whether the defendants have met their burden to conclusively prove good faith under Chambers and Wadewitz. We conclude that, in University of Houston, Sergeant Williams and Officer Stewart proved good faith as a matter of law. We conclude in Ener, however, that Deputy Ener did not conclusively establish good faith because his summary judgment evidence did not sufficiently substantiate his testimony that he acted in good faith.
1. Clark
The court of appeals held that Sergeant Williams and Officer Stewart did not establish good faith as a matter of law because their affidavits did not adequately discuss Wadewitz’s risk and need elements. Specifically on the need element, the court of appeals held that neither affidavit adequately discussed the seriousness of the suspect’s alleged assault or alternatives to immediately apprehending the suspect. 979 S.W.2d at 712. The court of appeals also held that neither affidavit adequately considered the risk to the public, [585]*585including the nature and severity of the risk to the public and whether a reasonably prudent officer would perceive a dear-risk of harm. 979 S.W.2d at 712.
We disagree with the court of appeals. Both Williams’ and Stewart’s affidavits sufficiently discuss both need and risk under Wadewitz and conclusively prove that a reasonable officer, under the same or similar circumstances, could have balanced need and risk as these officers did.
a. Sergeant Williams
Sergeant Williams’ affidavit discusses all Wadewitz'⅛ need factors. On the seriousness of the situation he was responding to, Williams stated:
When I made the decision to stop the ... jeep, I knew that the driver was a suspect in assaultive conduct and that he fled police officers who were attempting to detain him and had committed the criminal offense of evading detention which is a class B misdemeanor. Furthermore, the suspect had not been identified before he fled the foot patrol officers.... The manner in which the suspect operated his vehicle and the high rate of speed at which he traveled caused me to believe that the suspect posed a danger to the public.
Williams further stated that when he parked his patrol car in the suspect’s way to stop him, the suspect drove around Williams’ vehicle at a high rate of speed. Therefore, although Williams does not detail the seriousness of the assault allegedly committed on campus, he does discuss the seriousness of the situation as a whole, as Chambers and Wadewitz require. See Wadewitz, 951 S.W.2d at 467; Chambers, 883 S.W.2d at 656-57. On whether his presence was immediately necessary to apprehend the suspect and on the availability of alternatives to pursuit, Williams stated that he knew that the suspect had not been identified. Thus, Williams’ affidavit discussed the seriousness of the situation he was responding to, the need to immediately apprehend the suspect, and the availability of alternative actions to apprehend the suspect.
Williams also considered theWadewitz risk factors in assessing the risks to the public of his pursuit. He stated that when the pursuit occurred, around 2:00 a.m., “the weather was clear and the pavement was dry and the traffic was light.” He observed that the area was commercial and that there were “no cars other than the suspect vehicle and Officer Stewart’s on the roadway.” Williams further stated:
When the suspect vehicle proceeded onto M.L.K. Blvd., I knew that the street was divided by a median and that to the left was a green belt and to the right was a park; therefore, no traffic would be intersecting M.L.K. Blvd. There was no traffic from the point of entrance one (1) to M.L.K. Blvd. and no traffic from M.L.K. to Old Spanish Trail. From my professional experience, I was aware that suspect vehicles will often appear to flee but stop after having time to observe emergency lights and siren behind them. When I decided to pursue the suspect vehicle, I believed that at that time of night with the minimal traffic on the street and the nature of the area in which I traveled, diminished the risk from the pursuit itself, while the suspect and the manner in which he operated his vehicle posed a significant risk to the public. I believed, and in my opinion any reasonably prudent police officer in my position would also have believed, that during the pursuit, the need to stop the suspect vehicle was essential to insure the safety of the public by way of attempting to follow the suspect vehicle.
Although Williams did not use the exact language Wadewitz employs, we cannot agree with the court of appeals that Williams’ affidavit does not assess the nature, severity, likelihood, and obviousness of his pursuit’s risks to the public, namely the risk of a collision with a third party. See Wadewitz, 951 S.W.2d at 467. Sum[586]*586mary judgment requires that a movant establish facts upon which the court could báse its legal conclusion, not that the parties use particular words. See Wadewitz, 951 S.W.2d at 467; Siders v. State, 970 S.W.2d 189, 194 (Tex.App.-Dallas 1998, pet. denied); Pena v. Neal, 901 S.W.2d 663, 668 (Tex.App.-San Antonio 1995, writ denied). Just because Williams’s affidavit did not explicitly mention the risk of colliding with a third party does not mean that he did not assess this risk. This risk is present to some degree in every police pursuit. In assessing such facts as the time of day and traffic, weather and road conditions, Williams was assessing the specific circumstances present that affected this risk. And, the summary judgment record does not indicate that there were other circumstances that increased or decreased this risk or that presented a different risk to the public. We agree with Justice Edelmaris dissent in the court of appeals that an officer should not be required in his affidavit to affirmatively negate the existence of all circumstances or risks that did not actually exist. 979 S.W.2d at 719 (Edelman, J., dissenting).
Further, we conclude that Williams’ affidavit established that he acted in good faith. His affidavit includes facts upon which a reasonably prudent officer in the same or similar circumstances could agree that the need to pursue the suspect (considering that the suspect had allegedly committed an assault and was driving recklessly) outweighed the pursuit’s risk to the public (considering that there was very little traffic at that time of night, that the area was primarily commercial, that the weather and road conditions were good, and that there were few roads intersecting the road where the pursuit occurred).
b. Officer Stewart
Officer Stewart’s affidavit also established good faith as a matter of law. On the need element, Stewart stated: ‘While I was ordered by ... Sergeant Williams to stop the suspect vehicle, I was ... aware that the driver of the speeding vehicle was believed to have engaged in a fight on the University campus and evaded detention by University police officers.” He further stated:
After hearing Sergeant Williams’ radio transmission, I looked over my shoulder and saw a green Suzuki jeep rocket passed [sic] my patrol vehicle. I activated my emergency overhead lights and siren and attempted to follow the speeding vehicle. The suspect vehicle merged onto M.L.K. Boulevard and continued at a high rate of speed which I estimated to rise to 80 to 90 m.p.h. The posted speed limit was 35 m.p.h.
He thought that the suspect vehicle’s speed “was a clear indication that the driver of the suspect vehicle had no regard for the safety of the public and needed to be stopped.”
On alternatives to pursuit, Stewart stated that he “followed the suspect at a distance and was not able to get close enough to the suspect vehicle to obtain its license plate number. I had expected the suspect vehicle to stop when the driver observed my overhead lights and siren behind him,” but he did not.
On risk to the public, Stewart stated that he was familiar with the streets upon which he was traveling, that the area was primarily commercial, and that there was very light traffic at that time, about 2:00 a.m. He stated that he observed no car other than the suspect’s and Williams’ when he started the pursuit. He stated that he did not proceed through red lights or stop signs until it was safe to do so and that he did not go as fast as he could have to keep up with the suspect vehicle because he recognized the risk to the public. Further, when the suspect approached the intersection where the collision actually occurred, Stewart decided to end the pursuit and began slowing his vehicle. Stewart also stated:
When I decided to pursue the suspect vehicle, I believed that at that time of night with minimal traffic on the street [587]*587and the nature of the area in which I traveled, diminished the risk from the pursuit itself, while the suspect vehicle was operated by a person that engaged in an assault and fled the police and that posed a significant risk to the public. I believed, and in my opinion any reasonably prudent police officer in my position would also have believed, that during the pursuit the need to stop the suspect vehicle was essential to insure the safety of the public by way of attempting to follow the suspect vehicle.
Stewart’s affidavit shows that he considered all Wadewitz’s need and risk factors and supports his claim that a reasonably prudent officer could have assessed risk and need as he did. Therefore, Stewart met his summary judgment burden on good faith.
c. Clark’s Summary Judgment Evidence
In response to the defendants’ motions for summary judgment, Clark submitted the affidavit of expert Richard H. Turner, the chief executive officer of the National Academy for Professional Driving. Turner contends that Williams and Stewart were not acting in good faith and that “another reasonably prudent law enforcement officer, under the same or similar circumstances, could not have believed the need to immediately apprehend [the suspect] outweighed a clear risk of harm to other members of the public who may be using the highway during the pursuit.” Further, Turner stated that Williams’ and Stewart’s actions “did not meet industry standards for the University of Houston, City of Houston, Harris County, State of Texas, or the United States of America” and “violated the University of Houston policies regarding pursuit.” Turner bases his opinion on his review of “facts surrounding the pursuit,” “numerous documents including, but not limited to depositions taken, and policies and procedures presented in the discovery process,” and his visit to the pursuit scene.
Because Turner’s testimony on good faith is not substantiated with reference to each aspect of the need and risk balancing test, it is conclusory and is insufficient to controvert the defendant’s proof on good faith. See Wadewitz, 951 S.W.2d at 467.
Therefore, we conclude that Williams and Stewart established good faith as a matter of law.
2. Ener
In Ener, the court of appeals held that Deputy Constable Ener’s affidavit did not establish good faith because it did not substantiate Ener’s assessment of the risks of his pursuit, or of the need to pursue, specifically the availability of alternatives to pursuit. The court of appeals also held that Ener’s expert affidavits were conclusory and therefore were not competent summary judgment evidence.
We agree with the court of appeals that Ener’s affidavit does not contain facts showing that Ener adequately assessed the need to pursue, specifically, whether there were any alternatives to pursuit. See Wadewitz, 951 S.W.2d at 467.
On the need to pursue, Ener explained in his affidavit that while on patrol, he saw a vehicle speeding. His radar indicated that the vehicle was going 71 miles per hour in a 55 mile-per-hour zone. He put on his overhead lights and siren and pursued the vehicle for approximately one-fourth to one-half mile before approaching the intersection where the accident ultimately occurred. He further states: “At this time I believed the speeder was acting recklessly and was a danger and that I should definitely apprehend him.”
Nowhere in his affidavit does Ener mention facts to show that he assessed whether there were any alternatives to pursuit. Ener does not mention the viability of identifying the suspect and apprehending him later or any other alternative. To establish good faith under Wadewitz, Ener was required to show that he assessed the availability of alternatives to pursuit as [588]*588part of balancing the need to pursue and the pursuit’s risk to the public. See Wa-dewitz, 951 S.W.2d at 467. Without assessing the availability of alternatives, Ener did not have a suitable basis for concluding that a reasonable officer in his position “could have believed that the need to immediately apprehend the suspect outweighed a clear risk of harm to the public in continuing the pursuit.” Chambers, 883 S.W.2d at 656 (emphasis added); see Wa-dewitz, 951 S.W.2d at 467. Therefore, we agree with the court of appeals that Ener’s affidavit does not adequately substantiate his claim of good faith. We express no opinion on whether, if Ener had assessed the availability of alternatives to pursuit, his affidavit would have established good faith. Nor do we mean to imply that had Ener’s affidavit shown that he had a viable alternative to pursuit and decided to pursue the suspect anyway, we would conclude that he did not establish good faith. The availability of alternatives to pursuit is just one factor of the need/risk balancing test and the existence of alternatives to pursuit does not alone determine good faith. Therefore, just because an officer has identified a suspect and could apprehend him later but decides to pursue the suspect anyway cannot alone defeat good faith. We disagree with the dissent, therefore, that our opinion reheves an officer of the duty to apprehend a suspect whenever there is an alternative to pursuit. 22 S.W.3d at 925.
In addition to his affidavit, Ener also submitted the affidavits of experts Assistant Chief Rick Sadler and Captain Jim Sumner of the Harris County Constable’s office. But like Ener’s affidavit, these affidavits omit any reference to the availability of alternatives to pursuit. Therefore, they cannot establish good faith, either. Because we conclude that Ener did not meet his initial burden to prove good faith, we need not consider whether Thomas and Roarke controverted Ener’s summary judgment evidence.
IV. CONCLUSION
We hold that to conclusively prove good faith in a police pursuit case, an officer must prove that a reasonably prudent officer in the same or similar circumstances could have determined that the need to apprehend the suspect outweighed the risk to the public in continuing the pursuit and that the officer must substantiate his determination with facts showing that he assessed all Wadewitz’s need and risk factors.
In Clark, we conclude that Sergeant Williams and Officer Stewart conclusively proved good faith and that Clark’s summary judgment evidence did not successfully controvert their proof. Therefore, we reverse the court of appeals’ judgment on the good faith issue. Because the court of appeals did not determine whether Stewart was acting in a discretionary or ministerial capacity, we remand the case to the court of appeals for a determination on that issue.
In Ener, we conclude that Ener did not meet his burden to prove that he acted in good faith. Therefore, we affirm the court of appeals’ judgment denying summary judgment for Ener and Harris County.
Justice ENOCH filed a concurring and dissenting opinion.
Justice O’NEILL not sitting.