Ingersoll v. Palmer

Full title: WILLIAM INGERSOLL et al., Petitioners, v. ALFRED PALMER, as Chief of…

Court: Supreme Court of California

Date published: Oct 29, 1987

43 Cal.3d 1321 (Cal. 1987)

241 Cal. Rptr. 42

743 P.2d 1299

Opinion

Docket No. S.F. 25001.

October 29, 1987.

COUNSEL

Margaret C. Crosby, Alan L. Schlosser, Edward Chen, Amatai Schwartz, Donna J. Hitchens, Paul L. Hoffman, Mark D. Rosenbaum and Joan W. Howarth for Petitioners.

John K. Van de Kamp, Attorney General, Linda Ludlow, Thomas A. Brady, Martin S. Kaye, Dane R. Gillette and Ronald E. Niver, Deputy Attorneys General, James K. Hahn and Gary R. Netzer, City Attorneys, Frederick N. Merkin, Senior Assistant City Attorney, Lewis N. Unger, Assistant City Attorney, Donna Weisz and Pamela Victorine, Deputy City Attorneys, for Respondents.

Ira Reiner, District Attorney (Los Angeles), Harry B. Sondheim, Maurice H. Oppenheim and Richard Sullivan, Deputy District Attorneys, and Christopher N. Heard as Amici Curiae on behalf of Respondents.

OPINION

KAUFMAN, J.

This case presents the question whether sobriety checkpoints are permissible under the federal and state Constitutions. We conclude that within certain limitations a sobriety checkpoint may be operated in a manner consistent with the federal and state Constitutions.

FACTS

Petitioners are California taxpayers who seek to prohibit the operation of sobriety checkpoints in California. Respondents are chiefs of police of various California cities and the Commissioner of the California Highway Patrol. Petitioners alleged that the respondent law enforcement officers in the various jurisdictions around the state had begun or planned to begin using sobriety checkpoints.

In November 1984, in response to a request by the Commissioner of the California Highway Patrol, the Attorney General issued an opinion that roadblocks could constitutionally be used to detect and apprehend drunk drivers if certain safeguards were maintained to minimize the intrusion on motorists. (67 Ops.Cal.Atty.Gen. 471 (1984).)

That same month, the Burlingame Police Department (the Department) set up the first sobriety checkpoint program to operate in California, following the guidelines set forth in the Attorney General's opinion.¹ The Burlingame checkpoint was expected to serve as a model for others. We therefore examine the Burlingame checkpoint as illustrative of checkpoint operation procedures.

1 The California Highway Patrol shortly thereafter set up checkpoints at four test locations throughout the state. Other law enforcement agencies also announced or implemented sobriety checkpoint programs within a short time.

The Department promulgated a detailed manual to govern the checkpoint operations. The manual covered legal considerations, including the Attorney General's guidelines; a cost analysis; factors affecting location selection; required personnel and equipment; training and briefing of checkpoint personnel; press relations and publicity; as well as procedures for a follow-up evaluation.

The location for the Burlingame checkpoint was selected by taking into account frequency of drunk driving arrests and accidents, and safety factors such as traffic patterns and street layout. A suitable location was selected on El Camino Real for a checkpoint intercepting northbound traffic.² Warning signs (including a sign announcing a sobriety checkpoint) were posted. A cone taper diverted traffic to a single northbound lane. The signs and cone taper were set up according to Caltrans regulations for signing and lane closure.

2 At the location selected, El Camino Real was a divided road, providing safety and minimizing distraction to southbound traffic. There was a separate frontage road area which provided a safe place for directing motorists out of and back into the northbound traffic lanes. There was also a safe area in which to conduct field sobriety tests.

The checkpoint operation was supervised by a commander under whom two sergeants served. One sergeant supervised a team of traffic control and screening officers, and the second sergeant supervised the field sobriety test teams. Two traffic control officers, with support staff, set up the checkpoint and selected every fifth car for screening. There were one to four screening officers who contacted the motorists. Nonsworn reserve personnel were available for recording information and timing each contact. One to four officers, each with a nonsworn reserve assistant, were on duty to administer the field sobriety tests. There was also a booking officer, an officer to operate an intoxilizer, one for photographing and one alternate. There were also nonsworn personnel available for interpreting, transportation and booking assistance. All the officers chosen for checkpoint duty had a good record of "driving under the influence" (DUI) detection and arrest, all had recent refresher training on recognizing the symptoms of drug and alcohol use, and all had special training in checkpoint procedures, including conducting a simulated checkpoint. All officers on duty at the checkpoint were in full uniform.

On the night of the checkpoint operation, every fifth car was stopped and directed to a screening officer. The screening officer gave the driver a brief prescribed oral explanation of the checkpoint, and handed him or her an information flyer and a postage paid opinion survey card.³ During the contact, the screening officer observed the driver for bloodshot eyes, alcohol on the breath, and any other signs of impairment. The officer also shined a flashlight into the vehicle, looking for any open containers or other evidence of alcohol consumption. If no symptoms of impairment were observed, the driver was directed to continue into the northbound traffic lanes. If signs of impairment were observed, the driver was directed to a secondary testing area, where another officer would administer a field sobriety test. A sign announcing the checkpoint was posted sufficiently in advance of the checkpoint location to permit motorists to turn aside, and under the operational guidelines no motorist was to be stopped merely for choosing to avoid the checkpoint.

3 Approximately 29 percent of those stopped returned the postage paid survey cards. Of those responding, about 91 percent said they were not significantly delayed and 80 percent approved of drunk driving checkpoints.

The sobriety checkpoint was given advance publicity, including its date and general location. During the checkpoint operation, from 9:30 p.m. to 2:30 a.m. on November 16-17, 1984, 233 motorists were screened. Only 10 were asked to perform field sobriety tests, and all 10 passed. The checkpoint resulted in no arrests. The average detention periods for those cars stopped was 28 seconds. The average time for those who took the field sobriety tests was 6.13 minutes.

Petitioners filed an original petition for writ of mandate in this court within three days after Burlingame established its first sobriety checkpoint. We transferred the matter to the Court of Appeal. The First District, Division Three, denied petitioners' request for a stay and issued an alternative writ. The Court of Appeal issued an opinion in which the majority held sobriety checkpoints conducted in accordance with certain guidelines are permissible under the United States and California Constitutions. We granted the taxpayers' petition for review.⁴

4 Petitioners did not renew their request for a stay when they brought the matter before this court on the petition for review.

DISCUSSION

(1) Petitioners contend the validity of a sobriety checkpoint stop must be determined by the standard set forth in In re Tony C. (1978) 21 Cal.3d 888 [ 148 Cal.Rptr. 366, 582 P.2d 957], requiring an individualized suspicion of wrongdoing. If the primary purpose of the stop here were to detect crime or gather evidence of crime, we would agree with the contention that an individualized suspicion of wrongdoing is required. But, as we shall explain, the primary purpose of the stop here was not to discover evidence of crime or to make arrests of drunk drivers but to promote public safety by deterring intoxicated persons from driving on the public streets and highways. We therefore conclude the propriety of the sobriety checkpoint stops involved here is to be determined not by the standard pertinent to traditional criminal investigative stops, but rather by the standard applicable to investigative detentions and inspections conducted as part of a regulatory scheme in furtherance of an administrative purpose. (See People v. Hyde (1974) 12 Cal.3d 158, 165-166, 173 [ 115 Cal.Rptr. 358, 524 P.2d 830].)

In upholding airport screening searches, a majority of this court in Hyde applied the administrative search rationale. ( 12 Cal. 3d at p. 165 et seq.) The concurring minority, reaching the same result, preferred a more generic balancing test of reasonableness. (12 Cal.3d at p. 172 et seq.) But, verbal formulations aside, both the majority and the concurring minority in Hyde relied upon essentially the same principles and factors.

The majority noted: "Like all searches subject to the Fourth Amendment, an administrative screening must be measured against the constitutional mandate of reasonableness. In the case of administrative searches, however, 'there can be no ready test for determining reasonableness other than by balancing the need to search against the invasion which the search entails.' ( Camara v. Municipal Court (1967) supra, 387 U.S. 523, 536-537 [18 L.Ed.2d 930, 940, 87 S.Ct. 1727].) It is ironic, therefore, that by adopting the administrative search doctrine to evaluate the validity of airport screening procedures we must undertake a similar process of balancing to that which would have followed from a reliance upon Terry [v. Ohio (1968) 392 U.S. 1 (20 L.Ed.2d 889, 88 S.Ct. 1868)]." ( People v. Hyde, supra, 12 Cal.3d 158, 166, italics added.) The concurring minority reasoned: "It is now settled . . . that there is no fixed standard of reasonableness that applies to all types of governmental action which is subject to the mandates of the Fourth Amendment. Where, as here, we deal with a type of official conduct that (1) has objectives qualitatively different from those of the conventional search and seizure in the criminal context and (2) cannot feasibly be subjected to regulation through the traditional probable cause standard of justification, we may assess the reasonableness of the particular type of search and seizure by examining and balancing the governmental interest justifying the search and the invasion which the search entails. [Citations.]" ( Id., conc. opn. at p. 173. Italics added, fns. and original italics omitted.) We perceive no real inconsistency in the two analyses. They both employed a balancing test for reasonableness.

1. Reasonableness Standard Under the Fourth Amendment and the California Constitution

(2) The touchstone for all issues under the Fourth Amendment and article I, section 13 of the California Constitution is reasonableness. (See Terry v. Ohio, supra, 392 U.S. 1, 19 [20 L.Ed.2d 889, 904]; People v. Hyde, supra, 12 Cal.3d 158, 166, conc. opn. at pp. 172-173.)

The federal test for determining whether a detention or seizure is justified balances the public interest served by the seizure, the degree to which the seizure advances the public interest and the severity of the interference with individual liberty. ( Brown v. Texas (1979) 443 U.S. 47, 50-51 [61 L.Ed.2d 357, 361-362, 99 S.Ct. 2637].) In addition, federal constitutional principles require a showing of either the officer's reasonable suspicion that a crime has occurred or is occurring or, as an alternative, that the seizure is "carried out pursuant to a plan embodying explicit, neutral limitations on the conduct of individual officers." ( Brown v. Texas, supra, 443 U.S. at p. 51 [61 L.Ed.2d at p. 362], citing Delaware v. Prouse (1979) 440 U.S. 648, 663 [59 L.Ed.2d 660, 673-674] and United States v. Martinez-Fuerte (1976) 428 U.S. 543, 558-562 [49 L.Ed.2d 1116, 1128-1131].)

California constitutional principles are based on the same considerations, i.e., balancing the governmental interests served against the intrusiveness of the detention. (See People v. Hyde, supra, 12 Cal.3d 158, 166, also conc. opn. at pp. 172-173.) With respect to a seizure for conventional investigation of criminal activity, standards similar to federal standards have been articulated. (3) "[I]n order to justify an investigative stop or detention the circumstances known or apparent to the officer must include specific and articulable facts causing him to suspect that (1) some activity relating to crime has taken place or is occurring or about to occur, and (2) the person he intends to stop or detain is involved in that activity. Not only must he subjectively entertain such a suspicion, but it must be objectively reasonable for him to do so: the facts must be such as would cause any reasonable police officer in a like position, drawing when appropriate on his training and experience ( People v. Superior Court ( Kiefer) [1970] 3 Cal.3d [807,] at p. 827 [ 91 Cal.Rptr. 729, 478 P.2d 449]), to suspect the same criminal activity and the same involvement by the person in question." ( In re Tony C., supra, 21 Cal.3d 888, 893, fn. omitted.)

But Tony C. itself further pointed out that, for purposes of analysis under the Fourth Amendment and under California constitutional law, "[a] more fruitful approach focuses on the purpose of the intrusion itself. If the individual is stopped or detained because the officer suspects he may be personally involved in some criminal activity, his Fourth Amendment rights are implicated and he is entitled to the safeguards of the rules set forth above. But similar safeguards are not required if the officer acts for other proper reasons." ( In re Tony C., supra, 21 Cal.3d 888, at p. 895, italics added.) Thus, the court in Tony C., like the United States Supreme Court in Brown, supra, 443 U.S. 47, expressly recognized that individualized suspicion that the contactee is involved in criminal activity is not required in certain types of police-citizen contacts.

We therefore turn to a consideration of the kinds of stops permitted under federal and state law upon less than a reasonable suspicion of personal involvement in criminal wrongdoing.

2. Seizures Not Requiring a Reasonable Suspicion

In People v. Hyde, supra, 12 Cal.3d 158, this court considered the question of airport security screening searches. The majority in an opinion authored by Justice Mosk reasoned that airport searches could not be justified on the basis of Terry v. Ohio, supra, 392 U.S. 1, because Terry carefully limited the permissible search to a patdown necessary to discover weapons, and because, before even the limited patdown search could be conducted, Terry required there to be specific and articulable facts which would lead a reasonable officer to believe the safety of the officer was in danger.

"Nevertheless," we stated, "we do find support under the Fourth Amendment for the pre-departure screening of prospective passengers in the series of United States Supreme Court decisions relating to administrative searches. ( United States v. Biswell (1972) 406 U.S. 311 [32 L.Ed.2d 87, 92 S.Ct. 1593]; Wyman v. James (1971) 400 U.S. 309 [27 L.Ed.2d 408, 91 S.Ct. 381]; Colonnade Corp. v. United States (1970) 397 U.S. 72 [25 L.Ed.2d 60, 90 S.Ct. 774]; See v. City of Seattle (1967) 387 U.S. 541 [18 L.Ed.2d 943, 87 S.Ct. 1737]; Camara v. Municipal Court (1967) 387 U.S. 523 [18 L.Ed.2d 930, 87 S.Ct. 1727]; see also United States v. Davis (9th Cir. 1973) 482 F.2d 893; United States v. Schafer (9th Cir. 1972) 461 F.2d 856; Downing v. Kunzig (6th Cir. 1972) 454 F.2d 1230 [15 A.L.R.Fed. 926].) (4a) These cases recognize that 'searches conducted as part of a general regulatory scheme in furtherance of an administrative purpose, rather than as part of a criminal investigation to secure evidence of crime, may be permissible under the Fourth Amendment though not supported by a showing of probable cause directed to a particular place or person to be searched.' [Citation.]" ( People v. Hyde, supra, 12 Cal.3d 158, 165.)

We pointed out that the purpose of the airport search is not to ferret out contraband or preserve for trial evidence of criminal activity, although the mechanics of the search itself take the form of a search to detect criminal activity (carrying weapons or explosives aboard an aircraft). Rather, we characterized the search as "a central phase of a comprehensive regulatory program designed to insure that dangerous weapons will not be carried onto an airplane and to deter potential hijackers from attempting to board. [Citations.]" ( People v. Hyde, supra, 12 Cal.3d 158, 166.) In the reasonableness analysis under the Fourth Amendment, we found the governmental interest substantial, the intrusion minimal, and the method effective for its purpose (in fact, we found in that case that there was no other effective means of achieving the purpose). We pointed out it was possible for a traveler to avoid the intrusion by either checking his or her hand luggage or foregoing air travel and opting for alternate means of transportation. Finally, we pointed out that airport searches were singularly unsuited to the warrant procedure because of the extremely high volume of air passenger traffic, rendering it impractical if not impossible to issue a warrant for any individual passenger. In addition, the consequences of not having a warrant were found mitigated by (1) neutral application of the screening process to all air passengers, minimizing the discretion of the officials in the field, and (2) limiting the intrusiveness of the search to those actions strictly necessary to disclose the presence of weapons or explosives.

The three concurring justices in Hyde agreed that the airport screening procedures were constitutionally permissible but questioned whether the airport search could properly be labelled an "administrative search" like the building inspection in Camara v. Municipal Court (1967) 387 U.S. 523 [18 L.Ed.2d 930, 87 S.Ct. 1727]. In the view of the concurring justices in Hyde, supra, 12 Cal.3d 158, the Fourth Amendment considerations should simply be evaluated pursuant to a balancing test of reasonableness, consisting of an assessment of the governmental interest justifying the search and the intrusiveness entailed in the search. The concurring minority had no difficulty in concluding the governmental interest was compelling and the intrusion resulting from the search was minimal. Thus, the airport searches were concluded to be reasonable. No warrant was required because compliance with the warrant procedure, as the majority had also pointed out, would completely frustrate the legitimate governmental purpose.

(5) The sobriety checkpoint presents a compelling parallel to the airport screening search. While the label "administrative search" is open to some criticism in application to either the airport search or the sobriety checkpoint stop, both, although they operate mechanically as a search or inspection for the violation of law, actually serve a primary and overriding regulatory purpose of promoting public safety. Their primary purpose is to prevent and deter conduct injurious to persons and property; they are not conventional criminal searches and seizures. The fact that sobriety checkpoint stops will lead to the detection of some individuals involved in criminal conduct does not alter the fundamental regulatory character of the screening procedure. (See People v. Hyde, supra, 12 Cal.3d 158, at p. 166; see also New York v. Burger (1987) 482 U.S. ___, ___ [96 L.Ed.2d 601, 622-623, 107 S.Ct. 2636, 2651].)

Our analysis in Hyde is supported by decisions of the United States Supreme Court which have similarly approved regulatory searches in appropriate circumstances in the absence of any particularized suspicion of wrongdoing. Camara v. Municipal Court, supra, 387 U.S. 523 and Marshall v. Barlow's, Inc. (1978) 436 U.S. 307 [56 L.Ed.2d 305, 98 S.Ct. 1816] are examples.

In Camara, a city ordinance gave authorized city employees, upon presentation of credentials, the right to enter buildings or structures to perform necessary duties. The United States Supreme Court recognized that in performing a function such as building inspections, the governmental entity will rarely have knowledge of conditions in a particular building, but must necessarily rely on general conditions in an area. The court held that a warrant for building inspections based on area conditions, rather than upon probable cause to believe violations exist in a particular dwelling, was reasonable. The Supreme Court stated, "In determining whether a particular inspection is reasonable — and thus in determining whether there is probable cause to issue a warrant for that inspection — the need for the inspection must be weighed in terms of [the] reasonable goals of code enforcement . . . . [¶] . . . [¶] . . . [T]here can be no ready test for determining reasonableness other than by balancing the need to search against the invasion which the search entails." ( Camara v. Municipal Court, supra, 387 U.S. 523, 535-537 [18 L.Ed.2d 930, 939-940].)

In Marshall v. Barlow's, Inc., supra, 436 U.S. 307, the court examined the regulatory scheme for administrative inspections of business premises under the Occupational Safety and Health Act of 1970 (OSHA) (29 U.S.C.A. § 657(a)). The court held that the OSHA inspections should be subject to a warrant requirement, but significantly did not require an individualized suspicion of violation of OSHA regulations before the warrant could be issued.

Some industries are so heavily regulated that government inspections are held constitutionally permissible, without notice, warrant, or individualized suspicion of wrongdoing. ( Donovan v. Dewey (1981) 452 U.S. 594 [69 L.Ed.2d 262, 101 S.Ct. 2534] [mines]; United States v. Biswell (1972) 406 U.S. 311 [32 L.Ed.2d 87, 92 S.Ct. 1593] [firearms]; Colonnade Corp. v. United States (1970) 397 U.S. 72 [25 L.Ed.2d 60, 90 S.Ct. 774] [liquor].) Business owners in the heavily regulated industries are presumed to know that they are subject to the periodic inspections which are specified by and regularly carried out pursuant to enabling legislation.

Regulatory inspections and stops have also been permitted under decisions of the United States Supreme Court and the California courts in the absence of an individualized suspicion of wrongdoing in border patrol checkpoint inspections ( United States v. Martinez-Fuerte, supra, 428 U.S. 543), agricultural inspection checkpoints ( People v. Dickinson (1980) 104 Cal.App.3d 505 [ 163 Cal.Rptr. 575]), vehicle mechanical inspection checkpoints ( People v. De La Torre (1967) 257 Cal.App.2d 162 [ 64 Cal.Rptr. 804]), and license and registration inspection checkpoints ( People v. Washburn (1968) 265 Cal.App.2d 665 [ 71 Cal.Rptr. 577]).

The United States Supreme Court in United States v. Martinez-Fuerte, supra, 428 U.S. 543, held with respect to immigration checkpoints that neither a warrant nor particularized suspicion is required. The court upheld the constitutionality of an immigration stop without particularized suspicion at a checkpoint away from the international border by balancing the governmental interests served against the intrusion on Fourth Amendment interests. The court concluded the need for routine checkpoint stops was great because the flow of illegal aliens cannot be controlled effectively at the border. ( Martinez-Fuerte, supra, 428 U.S. 543, at pp. 556-557 [49 L.Ed.2d 1116 at pp. 1127-1128].) By contrast, the checkpoint stop was a "quite limited intrusion" on Fourth Amendment interests. Such a stop entailed only a brief detention, requiring no more than a response to a question or two and possible production of a document. Neither the vehicle nor the occupant was searched. The court also concluded the "subjective intrusion" of a fixed checkpoint stop was minimal, unlike a random or roving stop, because motorists could see that other vehicles were being stopped, could see visible signs of the officers' authority, and were much less likely to be frightened or annoyed by the intrusion. ( Id., at pp. 557-558 [49 L.Ed.2d at p. 1128].)

Moreover, the court found an area warrant was not required, and distinguished Camara, supra, 387 U.S. 523, both on the ground the checkpoint seizure of an automobile involves significantly different expectations of privacy from the traditional expectations of privacy in one's residence, as to which a warrant traditionally has been required, and on the ground the warrant requirement in Camara served purposes under the Fourth Amendment which were not relevant to a checkpoint operation.

The need to provide an assurance of legitimacy of the search/seizure required a warrant in the building inspection context, but that need was served alternatively in the checkpoint operation by the visible manifestations of authorization in the form of signs announcing the roadblock, official insignia and vehicles, and fully uniformed personnel. Another purpose of the warrant requirement in Camara was to prevent hindsight from coloring the evaluation of the reasonableness of a search or seizure. In the checkpoint operation, however, "The reasonableness of checkpoint stops . . . turns on factors such as the location and method of operation of the checkpoint, factors that are not susceptible to the distortion of hindsight, and therefore will be open to post-stop review notwithstanding the absence of a warrant. Another purpose for a warrant requirement is to substitute the judgment of the magistrate for that of the searching or seizing officer. [Citation.] But the need for this is reduced when the decision to 'seize' is not entirely in the hands of the officer in the field, and deference is to be given to the administrative decisions of higher ranking officials." ( United States v. Martinez-Fuerte, supra, 428 U.S. 543, 565-566 [49 L.Ed.2d 1116, 1133].)

The United States Supreme Court also strongly suggested that other checkpoint type stops would be viewed similarly. "Stops for questioning, not dissimilar to those involved here, are used widely at state and local levels to enforce laws regarding drivers' licenses, safety requirements, weight limits, and similar matters. The fact that the purpose of such laws is said to be administrative is of limited relevance in weighing their intrusiveness on one's right to travel; and the logic of the defendants' position, if realistically pursued, might prevent enforcement officials from stopping motorists for questioning on these matters in the absence of reasonable suspicion that a law was being violated. As such laws are not before us, we intimate no view respecting them other than to note that this practice of stopping automobiles briefly for questioning has a long history evidencing its utility and is accepted by motorists as incident to highway use." ( United States v. Martinez-Fuerte, supra, 428 U.S. 543, 560, fn. 14 [49 L.Ed.2d 1116, 1130].)

The intimation that neutrally operated checkpoint stops are permissible was reiterated in dictum in Delaware v. Prouse, supra, 440 U.S. 648. In that case, a single patrol officer decided to make a roving stop for the purpose of a license or registration "spot check," but he had no information or reasonable suspicion either that the driver was unlicensed or that the vehicle was improperly registered. The Supreme Court held that such a random roving stop made without a reasonable suspicion of law violation was contrary to the Fourth Amendment. However, the court was careful to state that "This holding does not preclude the State of Delaware or other States from developing methods for spot checks that involve less intrusion or that do not involve the unconstrained exercise of discretion. Questioning of all oncoming traffic at roadblock-type stops is one possible alternative." ( Id., at p. 663, fn. omitted [59 L.Ed.2d at pp. 673-674].) This dictum was not mere rhetoric, however. It is analytically consistent with the court's holdings in other cases. Standardless and unconstrained discretion on the part of government officers is what the court sought to circumscribe in the regulatory inspection and stop cases. ( Almeida-Sanchez v. United States (1973) 413 U.S. 266, 270 [37 L.Ed.2d 596, 601, 93 S.Ct. 2535]; Camara v. Municipal Court, supra, 387 U.S. 523, 532-533 [18 L.Ed.2d 930, 937-938].) (4b) Accordingly, such stops and inspections for regulatory purposes may be permitted if undertaken pursuant to predetermined specified neutral criteria ( Delaware v. Prouse, supra, 440 U.S. 648, 662 [59 L.Ed.2d 660, 673]) such as the criteria articulated for a checkpoint stop ( United States v. Martinez-Fuerte, supra, 428 U.S. 543, 553-554, 556-564 [49 L.Ed.2d 1116, 1126, 1127-1132]).

3. Regulatory Purpose

(6) Petitioners argue the sobriety checkpoint stop we examine here is a criminal investigation roadblock, subject not only to Tony C., supra, 21 Cal.3d 888, but barred by the Fourth Amendment under this court's decision in People v. Gale (1956) 46 Cal.2d 253 [ 294 P.2d 13]. (See also Wirin v. Horrall (1948) 85 Cal.App.2d 497 [ 193 P.2d 470].) In Gale, sheriff's officers stopped and searched cars at a roadblock explicitly for the purpose of "'curb[ing] the juvenile problem and also check for, well, anything that we might find, anything that looked suspicious.'" ( People v. Gale, supra, 46 Cal.2d 253, 255.) We do not agree.

Dragnet searches, explicitly undertaken for the purpose of uncovering evidence of crime but without any reason to believe any criminal activity has taken place, are unreasonable. ( People v. Gale, supra, 46 Cal.2d 253, 256; Wirin v. Horrall, supra, 85 Cal.App.2d 497, 504.) However, the sobriety checkpoint here was operated not for the primary purpose of discovering or preserving evidence of crime or arresting lawbreakers, but primarily for the regulatory purpose of keeping intoxicated drivers off the highways to the end of enhancing public safety. Analytically it is much the same as an immigration checkpoint or a checkpoint to inspect for the safety of equipment or compliance with agricultural regulations. The threat to public safety is at least as great and the intrusion into Fourth Amendment interests is no greater here than in those other regulatory checkpoint inspections which have invariably been held constitutionally permissible.

Our conclusion in this regard is based on factors related to the operation of the checkpoint in this case, on the stated goals of law enforcement agencies in implementing sobriety checkpoint programs, on the observable, albeit limited, experience with checkpoint operations in this and other states, as well as common sense.

In the Burlingame program, the sobriety checkpoints received substantial advance publicity, which was clearly designed both to inform the public of the serious problem of drunk driving and to deter potential drinking drivers before they decided to drink and drive. An important part of the Burlingame procedure was to educate by giving each stopped driver a leaflet about the checkpoint program, as well as a survey postcard. In addition, the checkpoint was not conducted as a criminal dragnet. Checkpoint personnel were specifically instructed that drivers were not to be stopped merely for avoiding the checkpoint.⁵ The road sign announcing the checkpoint was placed sufficiently in advance of the checkpoint that motorists could choose to avoid the checkpoint.

5 Cars avoiding the checkpoint would be stopped, however, if in avoiding the checkpoint the driver did anything unlawful, or exhibited obvious signs of impairment.

The stated goals of several law enforcement agencies explicitly point to deterrence as a primary objective of the checkpoint program. The Burlingame manual described the objectives of its program, noting the historical use of roving patrols as the principal law enforcement response to the drunk driving problem. Despite increased patrols, public awareness campaigns, stiffer drunk driving penalties, and increased arrests, the Burlingame Police Department found the major problem was that the public's perceived (and actual) risk of apprehension was very low. Two major goals of the checkpoint as stated in the manual were to increase public awareness of the seriousness of the problem and to increase the perceived risk of apprehension.

The evaluation report on the pilot project carried out by the California Highway Patrol (CHP) stated that, although a project of stepped up roving patrols in 1980 had resulted in approximately twice the number of arrests per work hour, "it must be remembered that accomplishing more arrests is not the intent of sobriety checkpoints. Rather, they are intended to deter persons who have been drinking from driving for fear of encountering a checkpoint. If checkpoints are truly accomplishing their purpose, DUI arrests, as well as DUI accidents, should decrease." (Italics added.) In addition, the report recommended a six-month long-term study to be carried out in two CHP test areas. The report recommended using two different patterns of roadblock implementation — employing sobriety checkpoints during major holiday seasons at one test location, and using twice monthly checkpoints at the other location. Significantly, the recommendation report stated that "This dual study method will not only permit long term evaluation of checkpoint deterrence, but may also identify the frequency necessary to produce deterrence."

A sobriety checkpoint program operated by the Arizona Highway Patrol is assertedly designed "to develop a public perception of the high risk of apprehension of drinking drivers," and the program abstract for the Maryland sobriety checkpoint project stated it was intended to function as a general deterrent to drinking drivers by instilling the perception that there was an increased likelihood of detection and arrest. An integral aspect of the Maryland program was publicity, to attain maximum public awareness and voluntary compliance with DUI laws.

Not only is deterrence the stated objective of DUI roadblock programs, but actual, though admittedly limited, experience with checkpoint programs indicates deterrence is in fact a significant result of such programs. In written responses to interrogatories posed by the Court of Appeal in the instant case, Burlingame Police Chief Alfred Palmer pointed out that deterrent value was demonstrated in two test areas of the Maryland program: incidence of alcohol related traffic accidents was reduced by 71 percent in Prince Georges County and fatalities were reduced 75 percent in Montgomery County in 1981. The follow-up report relating to the Burlingame checkpoint noted that some level of deterrence was indicated by the facts that traffic volume fell considerably below normal during the last two hours of their checkpoint operation, that the volume of business in Burlingame bars was also significantly below normal after 10 p.m., that calls for taxicabs were 12 percent above normal, and that, as officers on duty at the checkpoint noticed, several cars with sober drivers but intoxicated passengers proceeded through the checkpoint (the "designated driver" phenomenon). In New York, the Governor's Alcohol and Highway Safety Task Force found "'that the systematic . . . traffic checkpoint is the single most effective action in raising the community's perception of the risk of being detected and apprehended for drunk driving' (Report, at p. 103)." ( People v. Scott (1984) 63 N.Y.2d 518 [483 N.Y.S.2d 649, 473 N.E.2d 1, 4-5].)

Petitioners argue in their discussion of the balancing test that roadblocks are not effective for apprehending DUI violators, and point out that the CHP experience showed that roving patrols were over twice as effective as roadblocks per work hour in producing drunk driving arrests, and that the Burlingame checkpoint in fact resulted in no arrests. The absence of arrests, however, is both explained by and affords substantial support for the conclusion that increasing drunk driving arrests — i.e., conducting investigations for the purpose of gathering evidence of criminal activity — is not the primary purpose of sobriety checkpoints. An absence of arrests does not indicate a sobriety checkpoint is a futile exercise. It more likely indicates that the existence of the checkpoint program has succeeded in inducing voluntary compliance with the law, thus fulfilling the program's primary objective of keeping automobiles operated by impaired drivers off the roads. Drunk driving is not merely a crime, it is a serious public safety problem. A vehicle driven by an intoxicated person is as much a road hazard as a vehicle with defective brakes or a defective steering mechanism. Sobriety checkpoints serve to keep such hazardous instrumentalities off the road in the first instance. If checkpoints perform a significant deterrent function, it follows that drunk driving arrests would decrease in areas of checkpoint operation.

It is perhaps the characteristic of the automobile as a hazardous instrumentality that affords the greatest distinction between the sobriety checkpoint and an impermissible criminal dragnet. While the sobriety checkpoint differs from, e.g., an agricultural inspection, because the item to be examined is a person and not an offending plant, nevertheless, the automobile is stopped for reasons directly related to public safety, and not for purposes of criminal investigation. In this sense it is as we have said analogous to a permissible equipment inspection checkpoint. The fact that the officer's observations of a driver's demeanor have the potential to result in criminal sanctions is not determinative. Just as an airport screening search may result in criminal arrests and prosecutions, but is nevertheless not a criminal investigative search, the sobriety checkpoint inspection primarily serves the proper regulatory purpose of deterring intoxicated persons from driving and thus endangering the public.

4. The Balancing Test

(7) As we have explained, both the majority and concurring minority in Hyde, supra, 12 Cal.3d 158, and, ultimately, all other pertinent authorities determine the constitutional reasonableness of searches and seizures by a balancing test: weighing the gravity of the governmental interest or public concern served and the degree to which the program advances that concern against the intrusiveness of the interference with individual liberty. (See, e.g., Brown v. Texas, supra, 443 U.S. 47, at pp. 50-51 [61 L.Ed.2d 357, at pp. 361-362]; People v. Hyde, supra, 12 Cal.3d 158, at pp. 166-169, conc. opn. at pp. 172-178.)

Deterring drunk driving and identifying and removing drunk drivers from the roadways undeniably serves a highly important governmental interest. As we noted in Burg v. Municipal Court (1983) 35 Cal.3d 257, at page 262 [ 198 Cal.Rptr. 145, 673 P.2d 732], "The drunk driver cuts a wide swath of death, pain, grief, and untold physical and emotional injury across the roads of California and the nation. The monstrous proportions of the problem have often been lamented in graphic terms by this court and the United States Supreme Court. [Citations.] . . . [I]n the years 1976 to 1980 there were many more injuries to California residents in alcohol-related traffic accidents than were suffered by the entire Union Army during the Civil War, and more were killed than in the bloodiest year of the Vietnam War. [Citations.] Given this setting, our observation that '[d]runken drivers are extremely dangerous people' [citation] seems almost to understate the horrific risk posed by those who drink and drive." Stopping the carnage wrought on California highways by drunk drivers is a concern the importance of which is difficult to overestimate.

While it may be less self evident, the record here also supports a reasonable inference sobriety checkpoints of the sort here described do advance this important public goal. Petitioners contend that sobriety checkpoints are not as effective in detecting drunk drivers as other less intrusive alternatives, such as roving patrols. However, officers on a roving patrol can effect a stop only upon observable indications of impairment (i.e., reasonable suspicion). Petitioners point to the observation in the CHP report that a CHP project in 1980 utilizing stepped up patrols resulted in an arrest rate per work hour over twice that resulting from use of the roadblocks. But, as we have pointed out, the number of arrests does not necessarily measure the effectiveness of the sobriety checkpoint. If the checkpoint is properly serving its function — deterrence — it may result in no arrests at all. An Arizona court considering the question concluded that although a sobriety checkpoint may be no more efficient than a roving patrol in detecting, drunk drivers it is more effective in deterring drunk driving. ( State v. Super. Ct. in for County of Pima (1984) 143 Ariz. 45 [ 691 P.2d 1073, 1076-1077].) This is consonant with our conclusion that the primary purpose of sobriety checkpoints is deterrence.

Petitioners argue respondents have not made a sufficient showing of the effectiveness of sobriety checkpoints. However, such effectiveness is difficult to quantify. The experience both in California and in other states with sobriety checkpoints has been very limited, and no definitive statistics are yet available. It would be presumptuous in the extreme for this court to prohibit the use of an otherwise permissible and potentially effective procedure merely because its effectiveness is at the present time largely untested. Indeed, to do so would prevent the compilation of any data to show its effectiveness.

Nevertheless, there are indications of the effectiveness of the roadblocks even in the absence of statistical evidence. For example, the Maryland court in Little v. State (1984) 300 Md. 485 [ 479 A.2d 903, 913], noted certain evidence in that record that on the night of the checkpoint operation many people who had been drinking asked a sober companion to drive instead, that calls for taxi service by drunk individuals increased, and that certain groups anticipating consumption of alcohol at social events chartered vehicles instead of driving. "The prospect of being stopped at a roadblock thus convinced some intoxicated individuals to find alternate means of transportation." ( Little v. State, supra, 479 A.2d 903, 913.) Similar results were observed in connection with the Burlingame checkpoint in the instant case, and at oral argument counsel for petitioners conceded the likely deterrent effect of the sobriety checkpoints involved here.

We further observe that roving stops may not be a more effective alternative means of enforcing drunk driving laws. Constitutionally permissible roving stops must be based on an articulable suspicion of law violation. With respect to drunk driving, this requires an officer's observation of some objectively manifested behavior indicating impairment. By this method, the number of drunk drivers detected and arrested is estimated between one in two hundred to one in two thousand. Stepped up holiday patrols, with attendant publicity, have been used in many jurisdictions for many years, without appreciable effect on the drunk driving toll to people and property. In addition, although stiffer penalties for drunk driving in California appeared to result in a decline in alcohol-related accident incidence in 1981 and 1982, thereafter alcohol-involved accidents and fatalities began to increase again, to nearly pre-1981 levels. As noted in the Burlingame manual, despite countermeasures consisting of publicity, heightened patrol efforts and more severe penalties, an attitude of impunity continues to exist with respect to drinking and driving.

The failure of traditional methods of enforcement was commented on by Professor LaFave: "[A] rather strong argument can be made that mere patrol and stoppings based upon the Terry standard [of reasonable suspicion, supra, 391 U.S. 1] do not produce what the Camara Court [, supra, 387 U.S. 523] referred to as 'acceptable results.' For one thing, even if a patrolling officer is . . . in the vicinity where a drunk driver is operating his vehicle, it does not necessarily follow that the driver will at that particular time drive his car in such a fashion as to create a reasonable suspicion justifying a stop. And the chances of such observation in the first place are rather slight, given the substantial number of intoxicated drivers on the roads. . . . It is by no means surprising, therefore, that it has been reliably estimated that only one of every 2,000 drinking drivers is apprehended." (4 LaFave, Search and Seizure: A Treatise on the Fourth Amendment (2d ed. 1987) Vehicle Use Regulation, § 10.8(d), pp. 72-73.)

Justice Feldman, in a concurring opinion in State ex rel. Ekstrom v. Justice Ct. of State (1983) 136 Ariz. 1 [ 663 P.2d 992], observed that "The governmental interest sought to be protected by the roadblocks is greater than merely detecting and apprehending drunk drivers. Given the carnage on our highways, there is a unique societal interest in enforcing compliance with the law by deterring driving while under the influence of alcohol or other drugs. [¶] . . . [T]he state cannot satisfy this interest by traditional methods which satisfy the Terry test. The traditional system has left us far short of achieving the law's objective. . . . It is only fortuitous that an officer happens to be in a position to see a drunk entering the freeway on the off-ramp [ sic] before that drunk happens to kill some innocent person. . . . [¶] . . . [It is] obvious that traditional law enforcement methods, involving the arrest by roving officers of only those whom they can stop upon a founded suspicion of drunk driving, fall short of satisfying society's compelling interest in enforcing compliance with the laws prohibiting drunk driving." ( Id., 663 P.2d 992 at pp. 998-999, conc. opn. Feldman, J.)

Not only are roving patrol stops inadequate generally, but there are also indications that roving patrols are less effective than sobriety checkpoints in detecting lower but nonetheless dangerous levels of intoxication. The average person arrested for drunk driving by roving patrols tends to have a significantly higher blood-alcohol level than the average sobriety checkpoint drunk driving arrestee. The CHP evaluation report showed that in all four test areas, the blood-alcohol level of checkpoint arrestees was lower (though still above the presumptive drunk driving level) than the blood-alcohol level of roving patrol arrestees in the same area for the same period. Thus, there may in fact be no effective alternate means of detecting those drivers whose judgment has actually been seriously impaired by alcohol and whose blood-alcohol level is illegal, but who do not consistently manifest outwardly observable impaired driving behavior.⁶

6 It is also worthy of mention that in some cases stepped up roving patrols may not be a viable alternative for some law enforcement agencies for other reasons. The Riverside Police Department indicated, for example, that it was greatly handicapped in using roving patrols as an alternative because of a lack of patrol vehicles.

The third balancing factor is the intrusiveness on individual liberties engendered by the sobriety checkpoints. Upon examination of the record, we conclude that the programs at issue in this case have implemented procedures designed to provide minimal interference with individual liberties. The decisions of courts of other states and the California Attorney General's opinion which originally sanctioned the kind of checkpoints operated here have analyzed the issue of intrusiveness extensively and have identified a number of factors important in assessing intrusiveness. The standards articulated in these cases provide functional guidelines for minimizing the intrusiveness of the sobriety checkpoint stop.

A. Decisionmaking at the Supervisory Level

The decision to establish a sobriety checkpoint, the selection of the site and the procedures for the checkpoint operation should be made and established by supervisory law enforcement personnel, and not by an officer in the field. This requirement is important to reduce the potential for arbitrary and capricious enforcement. (See United States v. Martinez-Fuerte, supra, 428 U.S. at p. 559 [49 L.Ed.2d at p. 1129].)

Several out-of-state decisions are in accord on this point. Sobriety checkpoints have been upheld in a variety of situations in which the chief commanding officer of a law enforcement agency has drawn up a comprehensive procedures document (in some cases reviewed by other officials) or where the regulations were promulgated by supervisory personnel. ( People v. Scott, supra, 473 N.E.2d 1 [program set up by county sheriff]; State v. Super. Ct. in for County of Pima, supra, 691 P.2d 1073 [commander of traffic enforcement division issued detailed command directive]; Little v. State, supra, 479 A.2d 903 [regulations reviewed by Superintendent of State Police, the Governor and the Attorney General]; State v. Coccomo (1980) 177 N.J. Super. 575 [ 427 A.2d 131] [township police chief adopted regulations approved by state Attorney General]; State v. Golden (1984) 171 Ga. App. 27 [ 318 S.E.2d 693] [roadblock set up by supervising DUI task force project coordinator]; State v. Deskins (1983) 234 Kan. 529 [ 673 P.2d 1174] [roadblock a joint effort of several law enforcement agencies, and all personnel briefed by supervisory officers].)

In each of the sobriety checkpoint projects here, the decision to implement the checkpoints, the site selection and the establishment of operational procedures were made or done by command level personnel, and detailed program regulations were promulgated.

B. Limits on Discretion of Field Officers

A related concern is that motorists should not be subject to the unbridled discretion of the officer in the field as to who is to be stopped. Instead, a neutral formula such as every driver or every third, fifth or tenth driver should be employed. To permit an officer to determine to stop any particular driver or car when there is no legitimate basis for the determination would be to sanction the kind of unconstrained and standardless discretion which the United States Supreme Court sought to circumscribe in its decisions in Prouse, supra, 440 U.S. 648, Almeida-Sanchez, supra, 413 U.S. 266, and Camara, supra, 387 U.S. 523. In all the checkpoint programs at issue here, neutral mathematical selection criteria were used.

C. Maintenance of Safety Conditions

Primary consideration must be given to maintaining safety for motorists and officers. Proper lighting, warning signs and signals, and clearly identifiable official vehicles and personnel are necessary to minimize the risk of danger to motorists and police. (Cf. Jones v. State (Fla. Dist. Ct. App. 1984) 459 So.2d 1068, 1079.) The checkpoint should be operated only when traffic volume allows the operation to be conducted safely. Screening procedures may at times be altered consistent with traffic volume, such that, for example, every car might be stopped when traffic is light, but if traffic began to back up, a different neutral formula might be applied, such as every fifth or tenth car, or operations might be temporarily suspended until traffic volume permitted resumption of safe checkpoint operation.

The Burlingame and CHP checkpoints were operated with a very high degree of safety assurance. The sites of the checkpoint operations were carefully selected with safety considerations in mind, including ample offroad or shoulder area for screening or field sobriety test procedures. The lane closures and road signs complied with all Caltrans safety guidelines. During operations, designated officers were responsible for maintaining the safety of the traffic lanes and cone patterns. There were no safety problems with respect to traffic backups.

D. Reasonable Location

The location of checkpoints should be determined by policy-making officials rather than by officers in the field. The sites chosen should be those which will be most effective in achieving the governmental interest; i.e., on roads having a high incidence of alcohol related accidents and/or arrests. (See State v. Coccomo, supra, 427 A.2d 131, 134.) Safety factors must also be considered in choosing an appropriate location.

One state court has found a sobriety checkpoint unconstitutional largely because it was not at a permanent location. ( State v. Olgaard (S.D. 1976) 248 N.W.2d 392.) A decision of the Ninth Circuit also held that a border patrol stop at a temporary checkpoint was unlawful. ( United States v. Maxwell (9th Cir. 1977) 565 F.2d 596.) We believe, however, that the temporary nature of sobriety checkpoints does not affect their constitutionality. The Olgaard court's concern with lack of permanency was solely based on its worry about surprise and lack of publicity in connection with the checkpoint. Although it is not precisely clear from the record in Olgaard, it is inferrable from the circumstances that the Olgaard checkpoint was set up on a surprise basis. The checkpoint was operated by only four officers utilizing nothing but the red flashing lights on several patrol cars. They stopped all traffic in both directions. No lights or signs were used that would have given advance notice of the checkpoint. There was no advance publicity about the checkpoint. The checkpoint plainly also lacked sufficient indicia of legitimacy in terms of staffing strength. In addition, there was no showing who made the decision to set up the checkpoint, or how the location was selected. Thus the Olgaard court appears to have acted with propriety in holding the checkpoint unlawful.

Similarly, the "temporary" border patrol checkpoint at issue in Maxwell, supra, 565 F.2d 596, was deficient with respect to notice and indicia of legitimacy. The checkpoint was marked only by a "stop ahead" sign with battery operated blinking yellow lights, half a dozen traffic cones, one ordinary stop sign, and a border patrol car with a flashing red light. Whereas motorists know or may learn of a permanent immigration checkpoint, the checkpoint in Maxwell was in operation on an intermittent basis without advance notice. There were no structures or electrical equipment connections. So far as the motorist was concerned, he was called to a halt on a lonely road by a blinking red light which could belong to anybody. In addition, the location of the checkpoint may have been inappropriate for an immigration checkpoint. The immigration checkpoint in Martinez-Fuerte, supra, 428 U.S. 543, was justified in part by its being placed on a major highway to prevent easy access by illegal aliens into the interior. Just as a sobriety checkpoint would be improper at a location without any significant traffic or incidence of drunk driving, the location of the Maxwell checkpoint on a route without any significant traffic, by illegal aliens or otherwise, may have been improper. ( United States v. Maxwell supra, 565 F.2d 596, 597-598.)

As was pointed out in People v. Scott, supra, 473 N.E.2d 1, at page 5, "The fact that the [United States] Supreme Court has approved permanent roadblocks but disapproved roving patrol stops is not determinative. What is critical is the intrusiveness of the checkpoint in relation to the governmental purpose involved. The subjective effect upon a vehicle driver approaching a roadblock is unrelated to whether it is permanent or was established but a few minutes before the driver approached it; in either instance his or her observation of it will be measured in minutes if not seconds. The likelihood of there being the kind of fright or annoyance that invalidates a random stop made by a roving patrol is obviated in the case of a temporary checkpoint by the visible signs of authority which the checkpoint entails — signs announcing the purpose, lighting, and identifiable police vehicles and the observable fact that there is a uniform system for stopping cars [citations]." (Accord, Little v. State, supra, 479 A.2d 903, 914.)

With respect to the Burlingame checkpoint, the lighting, signing, substantial uniformed police presence, official vehicles, etc., provided advance notice to the motorist sufficient to ward off surprise and fright. In fact, sufficient advance notice was provided so a motorist could choose to avoid the checkpoint altogether. The objective and subjective intrusion into Fourth Amendment rights was no greater than that resulting from a permanent checkpoint. The checkpoints at issue here were reasonable as to location.

E. Time and Duration

The time of day that a checkpoint is established and how long it lasts also bear on its intrusiveness as well as its effectiveness. For example, a nighttime stop may be more hazardous and possibly more frightening to motorists, but it will also probably prove more effective. While mentioned as a factor in State v. Deskins, supra, 673 P.2d 1174, time and duration have received little attention in the decisions addressing sobriety checkpoints, although most of the checkpoints approved have been operated in the late evening and early morning hours. ( People v. Scott, supra, 473 N.E.2d 1; Little v. State, supra, 479 A.2d 903; State v. Coccomo, supra, 427 A.2d 131; State v. Golden, supra, 318 S.E.2d 693; State v. Deskins, supra, 673 P.2d 1174.) We agree with the assessment of the Court of Appeal that no hard and fast rules as to timing or duration can be laid down, but law enforcement officials will be expected to exercise good judgment in setting times and durations, with an eye to effectiveness of the operation, and with the safety of motorists a coordinate consideration.

F. Indicia of Official Nature of Roadblock

Those aspects of a sobriety roadblock which evidence its official nature are critical in minimizing its intrusiveness. The roadblock should be established with high visibility, including warning signs, flashing lights, adequate lighting, police vehicles and the presence of uniformed officers. Not only are such factors important for safety reasons, advance warning will reassure motorists that the stop is duly authorized.

Clearly visible warning lights and other signs of authority have been present in most of the checkpoints upheld by the courts of other states. (See People v. Scott, supra, 473 N.E.2d 1, 3; Little v. State, supra, 479 A.2d 903, 905-906; State v. Golden, supra, 318 S.E.2d 693, 694.) In contrast, most of the checkpoints found unlawful have not provided adequate warning to motorists. (See State v. McLaughlin (Ind. Ct. App. 1984) 471 N.E.2d 1125, overruled in State v. Garcia (Ind. 1986) 500 N.E.2d 158, 162 [holding checkpoints lawful]; Com. v. McGeoghegan (1983) 389 Mass. 137 [ 449 N.E.2d 349, 353]; State v. Olgaard, supra, 248 N.W.2d 392, 394; State ex rel. Ekstrom v. Justice Ct. of State, supra, 663 P.2d 992, 993; State v. Hilleshiem (Iowa 1980) 291 N.W.2d 314 [vandalism roadblock]; cf. State v. Smith (Okla. Crim. App. 1984) 674 P.2d 562, 564.)

The checkpoints at issue here clearly complied with requirements for proper lighting, signing, and official presence, both in the comprehensive regulations developed for the checkpoint operation and in actual practice.

G. Length and Nature of Detention

Minimizing the average time each motorist is detained is critical both to reducing the intrusiveness of the stop on the individual driver and to maintaining safety by avoiding traffic tie-ups. As occurred in the Burlingame and CHP checkpoints, each motorist stopped should be detained only long enough for the officer to question the driver briefly and to look for signs of intoxication, such as alcohol on the breath, slurred speech, and glassy or bloodshot eyes. If the driver does not display signs of impairment, he or she should be permitted to drive on without further delay. If the officer does observe symptoms of impairment, the driver may be directed to a separate area for a roadside sobriety test. At that point, further investigation would of course be based on probable cause, and general principles of detention and arrest would apply.

H. Advance Publicity

Advance publicity is important to the maintenance of a constitutionally permissible sobriety checkpoint. Publicity both reduces the intrusiveness of the stop and increases the deterrent effect of the roadblock.

The concurring opinion in State ex rel. Ekstrom v. Justice Ct. of State, supra, 663 P.2d 992, at page 1001 explained the value of advance publicity: "Such publicity would warn those using the highways that they might expect to find roadblocks designed to check for sobriety; the warning may well decrease the chance of apprehending 'ordinary' criminals, but should certainly have a considerable deterring effect by either dissuading people from taking 'one more for the road,' persuading them to drink at home, or inducing them to take taxicabs. Any one of these goals, if achieved, would have the salutary effect of interfering with the lethal combination of alcohol and gasoline. Advance notice would limit intrusion upon personal dignity and security because those being stopped would anticipate and understand what was happening." ( 663 P.2d 992, 1001, conc. opn. Feldman, J.; see also State v. Deskins, supra, 673 P.2d 1174, 1182.)

Publicity also serves to establish the legitimacy of sobriety checkpoints in the minds of motorists. Although the court in Jones v. State, supra, 459 So.2d 1068, found that advance publicity was not constitutionally mandated for all sobriety roadblocks, nevertheless the court offered the observation, consistent with finding reasonableness under the Fourth Amendment, that "'[A]dvance publication of the date of an intended roadblock, even without announcing its precise location, would have the virtue of reducing surprise, fear, and inconvenience.' [Citation.]" ( Id., at p. 1080.)

In the instant case, substantial advance publicity accompanied each sobriety checkpoint instituted.

I. Conclusions as to Intrusiveness

We conclude that, while the intrusiveness of a sobriety checkpoint stop is not trivial, the enumerated safeguards operate to minimize the intrusiveness to the extent possible. The fright or annoyance to motorists condemned in connection with roving stops is absent when the checkpoint is operated according to the guidelines followed here.

On balance, the intrusion on Fourth Amendment interests is sufficiently circumscribed so that it is easily outweighed and justified by the magnitude of the drunk driving menace and the potential for deterrence.

5. Statutory Authorization

Petitioners contend that sobriety roadblocks are impermissible in the absence of specific authorizing legislation. Petitioners make three points. (8a) The first and broadest argument is that the police may not enforce traffic laws in any manner not specifically authorized by statute. (9a) The second is that the Vehicle Code provides for uniform statewide rules governing vehicle use and police regulation of that use, so that to allow municipalities to set up roadblocks on an ad hoc basis would allow a balkanization of vehicle use regulation that various sections of the Vehicle Code show the Legislature did not intend. (10a) The third point is that the Vehicle Code specifically permits police officers to use roadblocks in limited circumstances. Pointing out that two recent bills to amend the Vehicle Code to permit drunk driving roadblocks have died in committee, petitioners advance the proposition that since the Legislature has only permitted roadblocks in limited circumstances, roadblocks in any other circumstances are impermissible.

(8b) Petitioners cite no persuasive authority for the proposition that police officers may not enforce the Vehicle Code in any manner not specifically provided for by statute. Citing People v. One 1960 Cadillac Coupe (1964) 62 Cal.2d 92, 95-96 [ 41 Cal.Rptr. 290, 396 P.2d 706], they claim that the general police power only permits detention on reasonable suspicion when a motorist is engaged in wrongdoing unless there is statutory authority for other police action. The cited case is inapposite; it merely applies familiar principles as to the circumstances necessary to justify a detention, and establishes that the exclusionary rule applies in a civil action for forfeiture of a car believed to be involved in drug trafficking.

Petitioners also cite People v. McGaughran (1979) 25 Cal.3d 577, 583-584 [ 159 Cal.Rptr. 191, 601 P.2d 207] (warrant checks during traffic stops); People v. Superior Court ( Simon) (1972) 7 Cal.3d 186, 199-200 [ 101 Cal.Rptr. 837, 496 P.2d 1205] (search incident to arrest for violations for which accused would not be booked) and People v. Franklin (1968) 261 Cal.App.2d 703, 707 [ 68 Cal.Rptr. 231] (scope of statutes allowing stops for vehicle safety and registration inspection) for the proposition that the Vehicle Code is comprehensive and controls methods of enforcement of its provisions. None of these cases, however, establishes that unless a method of law enforcement is specifically authorized in the Vehicle Code, it is prohibited. Rather, these cases interpret the limits on officers' authority which have been expressly established by statute.

For similar reasons, petitioners' position is not aided by their citation to People v. Welsch (1984) 151 Cal.App.3d 1038 [ 199 Cal.Rptr. 87] (warrantless arrest for hit and run outside officer's presence not authorized by statute); People v. Horvath (1982) 127 Cal.App.3d 398 [ 179 Cal.Rptr. 577] (neither Pen. Code nor Pub. Util. Code authorized arrest of pilot who flew while intoxicated outside officer's presence); or People v. Aldapa (1971) 17 Cal.App.3d 184 [ 94 Cal.Rptr. 579] (arrest outside jurisdiction not authorized by former Pen. Code, § 817). In each case, the officer breached a statutory limitation on his authority; none of these decisions holds that methods of law enforcement not specifically authorized are prohibited.

It is illogical to suggest that an officer who has a reasonable suspicion an individual is driving under the influence of intoxicants and thus endangering the public may take corrective action, but that a law enforcement agency having knowledge that on any given night hundreds of drivers will be under the influence of intoxicants and thus endangering the public may not. The threat to public safety in the second instance is immeasurably greater than in the first. We conclude, as did the Court of Appeal majority, that the requisite authority is implicit in law enforcement's statutory authority to enforce criminal laws generally or traffic laws specifically. (See, e.g., Veh. Code, § 2400; Gov. Code, §§ 26600, 26601.)

(9b) With respect to the second point, it is true that the Vehicle Code generally preempts the field of traffic regulation vis-a-vis local ordinances. (See Veh. Code, § 21)⁷ We have observed that unless the Legislature so provides, a city has no authority over traffic control. (See Rumford v. City of Berkeley (1982) 31 Cal.3d 545, 550 [ 183 Cal.Rptr. 73, 645 P.2d 124] [city has no authority to erect traffic barriers not qualifying as traffic control devices under Veh. Code].) While this rule of preemption might conceivably prevent municipalities from establishing permanent drunk driving roadblocks that might in effect regulate traffic, it does not affect the statutory authority of the CHP and local police to enforce the Vehicle Code and other laws with checkpoints at more temporary locations. (See, e.g., Veh. Code, § 2400; Gov. Code, §§ 26600, 26601.)

7 Vehicle Code section 21 provides: "Except as otherwise expressly provided, the provisions of this code are applicable and uniform throughout the State and in all counties and municipalities therein, and no local authority shall enact or enforce any ordinance on the matters covered by this code unless expressly authorized herein."

(10b) Petitioners' arguments as to their third point, again go far beyond the authority they cite. Petitioners point out examples in which the Legislature has permitted police to stop or inspect cars. The Vehicle Code authorizes police officers to require motorists to stop and submit their vehicles for safety inspections upon reasonable cause to believe that the vehicle is in violation of the code. (Veh. Code, § 2806.) CHP and law enforcement officers "whose primary responsibility is to conduct vehicle theft investigations" may make warrantless inspections for vehicle registration. (Veh. Code, § 2805.) The CHP is authorized to run mechanical inspection stations. (Veh. Code, § 2814.) And the Legislature has provided for agricultural inspection stations at state borders. (Food Agr. Code, § 5341 et seq.) But it does not follow that because the Legislature has specifically authorized these inspections, no other inspections are permissible under the general police power. Indeed, it may be more reasonable to assume the Legislature would not feel obliged to enact specific legislation authorizing conduct it deemed to be constitutional and appropriate within the scope of existing police power. Legislative silence is an unreliable indicator of legislative intent in the absence of other indicia. We can rarely determine from the failure of the Legislature to pass a particular bill what the intent of the Legislature is with respect to existing law.⁸ "As evidences of legislative intent they [unpassed bills] have little value." ( Sacramento Newspaper Guild v. Sacramento County Bd. of Suprs. (1968) 263 Cal.App.2d 41, 58 [ 69 Cal.Rptr. 480]; see Marina Point, Ltd. v. Wolfson (1982) 30 Cal.3d 721, 735, fn. 7 [ 180 Cal.Rptr. 496, 640 P.2d 115, 30 A.L.R.4th 1161]; Miles v. Workers' Comp. Appeals Bd. (1977) 67 Cal.App.3d 243, 248, fn. 4 [ 136 Cal.Rptr. 508 ]; see also United States v. Wise (1962) 370 U.S. 405, 411 [8 L.Ed.2d 590, 594-595, 82 S.Ct. 1354]; Gregory v. City of San Juan Capistrano (1983) 142 Cal.App.3d 72, 84 [ 191 Cal.Rptr. 47]; cf. 2A Sutherland, Statutory Construction (4th ed. 1984) § 49.10, pp. 407-408.)

8 Petitioners point out the failure of Senate Bill No. 5 (1985-1986 Reg. Sess.) (see Sen. Weekly Hist., No. 134 (1985-1986 Reg. Sess.) Sept. 13, 1985) and Assembly Bill No. 14 (1985-1986 Reg. Sess.) (see Assem. Weekly Hist., No. 128 (1985-1986 Reg. Sess.) Sept. 13, 1985). They also cite the failure of Assembly Bill No. 104 (1983-1984 Reg. Sess.) and Assembly Bill No. 3604 (1983-1984 Reg. Sess.).

DISPOSITION

For the reasons stated, the judgment of the Court of Appeal denying the writ of mandate is affirmed.

Lucas, C.J., Arguelles, J., and Eagleson, J., concurred.

BROUSSARD, J.

I dissent. The majority uphold drunk driving¹ roadblocks on the theory that they are administrative inspections, not subject to the usual rule that any detention be justified by reasonable suspicion of individual wrongdoing. In my opinion, when uniformed law enforcement officers stop motorists to check them for intoxication, shine a light in the car to look for open containers of alcohol or other evidence of intoxication, with special officers ready to administer blood-alcohol tests and booking officers and police vans ready to take offenders to jail, it is not an administrative inspection but an ordinary police detention, which must be justified on the same grounds as any other detention for the purpose of law enforcement.

1 For the purpose of this opinion, the term "drunk driving" includes driving under the influence of alcohol or drugs. (See Veh. Code, § 23152 et seq.)

ADMINISTRATIVE SEARCH DOCTRINE

The majority concede that if the primary purpose of the roadblock were to detect crime, the detention of a driver without individualized suspicion that the driver had engaged in criminal activity would be unconstitutional. In fact, the roadblock has two purposes: detection of drunk drivers and collection of evidence. The majority maintain, however, that the primary purpose of these roadblocks is to promote public safety by deterring drunk driving. The majority assert that this is a regulatory or administrative purpose, and conclude that detention without individualized suspicion is permissible by analogy to the administrative search doctrine we adopted in People v. Hyde (1974) 12 Cal.3d 158 [ 115 Cal.Rptr. 358, 524 P.2d 830].

In Hyde, we permitted predeparture screening of airline passengers, without individualized suspicion of wrongdoing. Our theory was that the screening was a central part of a general regulatory scheme in furtherance of an administrative purpose, not an effort to seize contraband or evidence of crime. ( People v. Hyde, supra, 12 Cal.3d at p. 166.)

Hyde does not bring the drunk driving roadblock into the administrative search doctrine. First of all, the Vehicle Code provisions prohibiting drunk driving are not a "regulatory scheme." In Hyde, we used federal cases approving warrantless inspection of the firearms and liquor industry as examples of pervasively regulated activities in which a warrantless inspection was permissible. ( People v. Hyde, supra, 12 Cal.3d 158, 165, citing United States v. Biswell (1972) 406 U.S. 311 [32 L.Ed.2d 87, 92 S.Ct. 1593]; Colonnade Corp. v. United States (1970) 397 U.S. 72 [25 L.Ed.2d 60, 90 S.Ct. 774].) The rationale of those cases is that a person engaging in the pervasively regulated industry is on notice that he has a limited expectation of privacy because the regulations provide for effective inspection. ( United States v. Biswell, supra, 406 U.S. 311, 316 [32 L.Ed.2d 87, 92], see also Marshall v. Barlow's, Inc. (1978) 436 U.S. 307, 313 [56 L.Ed.2d 305, 311-312, 98 S.Ct. 1816].) No such "regulatory scheme" puts California drivers on notice that they are subject to detention without reasonable suspicion to determine whether they are driving under the influence of alcohol or drugs. Drivers do not "impliedly consent" to being inspected for alcohol on their breath.

The United States Supreme Court has rejected the Colonnade/Biswell analogy for automobile inspections on the ground that motorists have a considerable and legitimate expectation of privacy in the automobile, including an expectation of freedom of movement. ( Delaware v. Prouse (1979) 440 U.S. 648, 662-663 [59 L.Ed.2d 660, 673, 99 S.Ct. 1391].) The Supreme Court also has rejected the argument that driving is a pervasively regulated activity subjecting motorists to suspicionless roving immigration stops. ( Almeida-Sanchez v. United States (1973) 413 U.S. 266 [37 L.Ed.2d 596, 93 S.Ct. 2535].) The court has explained that a roving stop of a motorist to check for illegal aliens was unreasonable; the driver was not in the same position as the gun manufacturer or liquor distributor who had in effect consented to inspection by entering a heavily regulated industry. ( Id. at pp. 271-272 [37 L.Ed.2d at p. 602].)

A drunk driving roadblock also differs from the usual administrative or regulatory inspection because there is no "regulatory" agency to enforce the drunk driving prohibitions other than the police and the criminal courts.² The clear purpose of these laws is not to regulate, but to detect and punish criminal drunk driving. Nothing distinguishes this crime from any other serious one.

2 Although the Department of Motor Vehicles administers the license and registration provisions of the Vehicle Code, it has no agents enforcing the prohibition against drunk driving.

The majority suggest that as long as the purpose of a drunk driving roadblock is to deter rather than detect crime, the roadblock is "regulatory." But we certainly did not hold in Hyde, supra, 12 Cal.3d 158, as the majority suggest, that if the purpose of a detention is to deter rather than detect crime, it may be justified as an administrative search. Criminal law enforcement encompasses both detection and deterrence. If we allowed detentions without individualized suspicion to deter crime, we would allow preventive detentions in high crime areas. But we do not allow such practices. (See People v. Loewen (1983) 35 Cal.3d 117, 124 [ 196 Cal.Rptr. 846, 672 P.2d 436].) What distinguishes the permissible administrative inspection from other searches is not that they are only intended to deter, but that they carry out an administrative scheme that is not part of the penal system. There is no such administrative scheme here. In fact, the majority would permit roadblocks carried out without uniform regulation, without statewide oversight, in a Balkanized system varying from jurisdiction to jurisdiction.

The majority also rely on dictum in Delaware v. Prouse, supra, 440 U.S. 648, another case involving random stops of automobiles. There the high court disapproved a roving patrol stop conducted without individualized suspicion to check for current license and vehicle registration. The court said that its holding did not mean that police could not try other methods to enforce license and registration laws, such as a permanent roadblock to inspect for license and registration violations. But the Prouse dictum is inapposite. A roadblock inspection for license and vehicle registration is an administrative inspection, since these aspects of motoring are closely regulated. Since license and registration violations do not involve criminal sanctions primarily, the inspections themselves are less intrusive for the average motorist. A request to look at one's license is far less accusatory than an inspection for red, watery eyes, slurred speech, alcohol on the breath, open containers in the car, and the other signs of intoxication. It does not follow that, because a roadblock may be permissible to check for drivers' licenses, it must be permissible to check for drunk driving.

To call a drunk driving roadblock an administrative inspection ignores its true purpose — apprehension of drunk drivers. The fact is that the apparatus of the law enforcement system is moved to the scene of the roadblock — with breathalyzers ready to take evidence for introduction at a criminal trial, police officers ready to arrest offenders, and police vans ready to take suspects away. If we call the Burlingame roadblock an administrative inspection, then a detention to investigate any crime could be deemed an administrative inspection. The Constitution cannot, and should not, be stretched so far.

APPLICATION OF THE BALANCING TEST

The propriety of an administrative search is judged under a balancing test in which the invasion of individual liberty is weighed against the necessity for the invasion and its effectiveness in achieving the state's goal. (See Camara v. Municipal Court (1967) 387 U.S. 523, 536-537 [18 L.Ed.2d 930, 940, 87 S.Ct. 1727]; People v. Hyde, supra, 12 Cal.3d 158, 166.) Even assuming that the analogy to administrative searches is proper, and that we should abandon individualized suspicion in favor of a balancing test, I would conclude that roadblocks are neither necessary nor effective enough to warrant the intrusion on the individual that they cause.

We all agree that the government has a profound interest in deterring and punishing drunk driving. We have recently lamented the ". . . horrific risk posed [to public safety] by those who drink and drive." ( Burg v. Municipal Court (1983) 35 Cal.3d 257, 262 [ 198 Cal.Rptr. 145, 673 P.2d 732].) Yet the necessity for and effectiveness of drunk driving roadblocks remains to be demonstrated. And the intrusion is far from minimal.

In the federal cases allowing detentions and other intrusions without individualized reasonable suspicion that wrongdoing was taking place, there was little alternative available to the state, and this entered into the balance in determining whether the stop was reasonable. In those cases, the suspicionless intrusions were literally necessary, since the transgressions to be detected could not be observed unless the inspectors entered the premises; there were no objective indicators visible from the outside upon which an official could form a reasonable suspicion. (See United States v. Biswell, supra, 406 U.S. 311, 316 [32 L.Ed.2d 87, 92]; Colonnade Corp. v. United States, supra, 397 U.S. 72, 74, 76-77 [25 L.Ed.2d 60, 64]; Camara v. Municipal Court, supra, 387 U.S. 523, 537 [18 L.Ed.2d 930, 940]; cf. United States v. Martinez-Fuerte (1976) 428 U.S. 543, 557 [49 L.Ed.2d 1116, 1128, 96 S.Ct. 3074].) By contrast, drunk drivers are conspicuous. We have all observed drunks weaving down the road, speeding up and slowing down, straddling lanes, and ignoring traffic and traffic signs. It is preposterous to claim that police have no way other than a roadblock to detect or deter drunk drivers.

The majority suggest that roadblocks are necessary because existing enforcement techniques have not eradicated the problem of drunk driving. If this were a proper consideration, the Fourth Amendment would have little meaning. Existing enforcement techniques have not eradicated the scourge of crime in our society, yet no one would seriously propose that the Constitution therefore permits the police to make unprecedented invasions of personal liberty. If we allow mass detentions through the means of roadblocks merely because the police claim that they may be more effective and helpful to law enforcement, we have gone a long way towards abandoning the protection of the Fourth Amendment.

The majority find that the deterrent effect of drunk driving roadblocks weighs heavily in the balance. (The majority concede that roadblocks do not produce nearly as many arrests per officer hour as patrols in which drivers are stopped for cause.)³ This assertion is based on anecdotal evidence and flawed logic. Some states justify roadblocks by comparing accident rates in counties having roadblocks with others having none, but to conclude that it was the roadblock that caused the difference is the rankest speculation.⁴ The California Highway Patrol concedes that such evidence is inconclusive. In fact, some studies indicate that whatever deterrent effect a roadblock may have is entirely the result of its novelty and the waywardness of publicity. For example, as European drivers became accustomed to roadblocks and the publicity about them died down, their deterrent effect disappeared. (See ABA, Assessment of Effectiveness, supra, at p. 3.)

3 The majority do maintain that roadblocks may be effective in detecting the drunk driver with a low blood-alcohol level whose driving would not give objective signs that he is drunk. While this may be true, I fail to see the point of dedicating twice as many officer hours to arrest a mildly intoxicated driver as would be employed to arrest a seriously intoxicated driver. Police resources being limited, it is obviously more effective to use them to apprehend the more dangerous offender.

4 This point is made in great detail in Grossman, Sobriety Checkpoints: Ineffective and Intrusive in American Bar Association, Criminal Justice Section, Drunk Driving Laws and Enforcement, an Assessment of Effectiveness (1986) 15, 17 (hereafter ABA, Assessment of Effectiveness), and in Grossman, Sobriety Checkpoints: Roadblocks to Fourth Amendment Protections (1984) 12 Am. J. Crim. L. 123, 162-165. See also Jacobs Strossen, Mass. Investigations Without Individualized Suspicion: A Constitutional and Policy Critique of Drunk Driving Roadblocks (1985) 18 U.C.Davis L.Rev. 595, 640-641.

The majority admit that the deterrent effect of drunk driving roadblocks is not established and that "[t]he experience both in California and in other states with sobriety checkpoints has been very limited, and no definitive statistics are yet available." (Maj. opn., ante, at p. 1339.) Yet, the majority insist that "[i]t would be presumptuous in the extreme for this court to prohibit the use of an otherwise permissible and potentially effective procedure merely because its effectiveness is at the present time largely untested." ( Ibid.) This distorts the balancing test and makes it possible for any law enforcement method to pass constitutional muster as long as a plausible argument can be made that it might turn out to be effective. If this is the balancing test, it is not a test but a rubber stamp.

We also must weigh the intrusion of the roadblock on the individual. There can be no question of the reasonableness of the motorist's expectation of privacy. Though the expectation of privacy in the automobile is not as great as in the home, it is clear from Almeida-Sanchez, supra, 413 U.S. 266, and Delaware v. Prouse, supra, 440 U.S. 648, that motorists do retain a reasonable expectation of considerable privacy in the automobile. The invasiveness of a drunk driving roadblock is far greater than the invasion that the high court has characterized as minimal in the immigration checkpoint. (See United States v. Martinez-Fuerte, supra, 448 U.S. 543, 559 [49 L.Ed.2d 1116, 1129].) In the immigration checkpoint, the immigration agent's primary purpose is not to make arrests. But at a drunk driving roadblock, officers stop individuals with the purpose of determining if they are then committing the crime of drunk driving — a crime now involving considerable public stigma, to say nothing of the substantial criminal penalties that now result from a drunk driving conviction. The United States Supreme Court has repeatedly distinguished the minimal invasion of the administrative inspection from the necessarily hostile, threatening, and frightening intrusion of an investigation for crime. (See, e.g., Camara v. Municipal Court, supra, 387 U.S. 523 at pp. 530, 537 [18 L.Ed.2d 930 at pp. 936, 940].) Moreover, the detention at a drunk driving roadblock is necessarily experienced as personally intrusive, since unlike in the license inspection or immigration checkpoint, the officer's object is to inspect the interior of the vehicle for evidence of crime and to examine the present mental and physical condition of the driver to determine if he or she should be arrested.

The majority seem to suggest that as long as a neutral plan assures that the roadblock is run safely and without arbitrariness, the individual's interest in being free from police detention does not weigh in the balance at all. This antiseptic approach denies the unavoidable invasion of privacy which occurs when a citizen is confronted by the police and his demeanor inspected for evidence that he is committing a crime. Furthermore, the protection of the neutral plan is illusory. What recourse does any driver have if the neutral plan was not being followed when he or she was stopped? In the Burlingame example, the plan provided that motorists who refused to stop would be allowed to proceed. Yet one of the participating officers said he would have pursued any motorist who refused to stop. As there is apparently no remedy for violations of the neutral plan, the plan is no protection against arbitrariness.

The pervasiveness of the invasion also must be considered. Take one example. The New York City police used 100 officers to operate a series of drunk driving roadblocks from May 27 to June 26, 1983. The police stopped 184,828 cars. There were 222 arrests for drunk driving. (N.Y. Times (June 27, 1983) at p.B1, col. 2, described in Grossman, Sobriety Checkpoints: Roadblocks to Fourth Amendment Protections, supra, 12 Am. J. Crim. L. 123, 157.)⁵ During a one-month period, 184,606 people who turned out to be innocent were detained by the police. For every arrest there were 831 innocent drivers whose privacy was infringed. We certainly would be concerned about the propriety of detaining the same number of citizens on our streets for "inspection" for drug abuse or other crimes. It is one thing to invade personal privacy in order to apprehend dangerous criminals, but when the purported object is deterrence, such mass detentions are a very high price to pay when the effectiveness of such detentions is questionable at best.

5 A similar example is the experience of Missouri. In a 12-month period, there were 83 roadblocks; 23,934 cars were stopped. There were 181 arrests for drunk driving and 34 for drug-related offenses. (See ABA, Assessment of Effectiveness, supra, at p. 9.)

The invasion of privacy occasioned by these roadblocks also may become pervasive in the sense that the roadblocks will be everywhere. If we approve drunk driving roadblocks, they may appear in every community. This could mean 20 or 30 or more roadblocks in any urban area on any given night. Omnipresent police blockades at each community's border would be not only inconvenient for motorists, but also would be a contradiction of our values as an open and free society.

The Fourth Amendment is highly inexpedient to law enforcement, yet to date we have not allowed mass detentions on the theory that these might prove useful in combatting crime. I see no basis for distinguishing a drunk driving roadblock from any other mass detention established to prevent crime or apprehend wrongdoers. While drunk driving is a revolting crime, it is not the only one which the community abhors. If we abandon constitutional protections to combat every abhorrent crime which has captured the public's attention, we will find ourselves naked and unprotected in a hurry.

CONCLUSION

Since I regard a drunk driving roadblock under which a motorist is stopped with no reasonable suspicion that he is intoxicated inconsistent with the federal and state Constitutions, I would reverse the decision of the Court of Appeal and order the issuance of a peremptory writ.

Mosk, J., and Panelli, J., concurred.