|
August 18, 2009
Recall Effectiveness: A Review and Analysis of Current Issues
By Dr. Edward J. Heiden
Introduction
Recall effectiveness is one of the most discussed and written about of all product safety topics and issues. Seminars, workshops, meetings, and hearings have been held, experiments in recall design conducted, and research papers written (including one by this author under contract to CPSC). Numerous approaches have been suggested and tried at CPSC and elsewhere to improve recall effectiveness. Recently, recall effectiveness has surfaced again as a "hot" topic as CPSC began implementation of a product registration program for children's durable goods, as mandated by the 2008 Consumer Product Safety Improvement Act (CPSIA).
The opinions expressed in this piece are the author’s and do not necessarily represent those of Product Safety Forum.
Product Safety Forum is a free service of Product Safety Letter. Its mission is twofold: to promote frank and open airing of product safety issues and to provide advice and guidance from top experts in the field.
Do you have an idea or article you think would make a great piece for Product Safety Forum, including commentary or reaction to this one? Then email Publisher Sean Oberle to discuss it..
To read more stories, see the archives | Yet remarkably little decision-making attention in the safety community has been given in the last years to some of the most fundamental issues that characterize the recall effectiveness topic: (1) how effectiveness should be defined and measured and the serious problems and inconsistencies that arise from poor or imprecise definition; (2) what we do and do not know about actual recall effectiveness rates at the CPSC for the past several years; (3) what communication and notification techniques can best do the job of improving consumer response to recalls; (4) whether product registration cards, as called for in the case of children's durables under CPSIA, have any promise as an effectiveness-enhancing tool for these products; and (5) what new and innovative strategies are available for improving recall awareness, communication and response.
With a new Chairman and two new Commissioners taking office to join the two incumbents, and recall effectiveness surfacing again in the CPSIA context as a significant issue, it appears to be a good time to visit anew some of the fundamental issues and lessons learned with respect to the five fundamental topics mentioned above. This paper makes a preliminary attempt to do that.
I. Problems with CPSC Measurement of Recall Effectiveness
There are some fundamental measurement problems with the way in which effectiveness rate estimates are usually computed and reported by the CPSC and other interested parties. This causes effectiveness rates to be frequently misstated, sometimes dramatically. There are several major sources of such misstatement, and effectiveness is either over- or under-estimated depending on the significance of each misstatement source in a given recall.
Specifically, the measuring stick used most frequently by CPSC and others to calculate and report recall effectiveness rates is:
Number of products returned/repaired or refunds/credits requested
Number of product units originally distributed.
This particular statistic can be a relatively poor measure of the success of a particular corrective action for three major reasons.
First, the success of a recall should be defined in terms of how well consumers are protected from use of a product that is found to be potentially hazardous after they buy it. Thus the best benchmark for evaluating the correction of an alleged hazard for consumers should be the number of affected products actually in consumer use, rather than the number originally distributed. An adjustment to raw unit sales figures is therefore required to account for units that are still somewhere in the distribution chain (manufacturers, distributors, or retailers) at the time of the recall and have not yet been sold into consumers' hands. Capture of units still in the distribution chain is an important preventive part of a recall, but notification to relevant members of the distribution chain is accomplished through different channels than is notification to consumers of the product. Moreover, securing the remedy for affected products still in the distribution network does not involve any response or action on the part of consumers, and product return from the distribution chain is almost always much easier to obtain while held by known wholesale customers of the manufacturer and retailers than it is after it is sold into (usually) thousands or millions of individual households. In their early 1980s study of the determinants of recall effectiveness rates, Murphy and Rubin statistically demonstrated that for each ten percent of the affected units that exit the distribution network into consumers hands, the average effectiveness rate of a corrective action plan for that product falls by nine percent. 1
Thus, unless the percentages of products still under manufacturer, distributor, and retailer control are noted and counted separately, reported effectiveness rates that include products in the chain of distribution will usually be overstated, and will not provide meaningful information on whether or not a particular consumer corrective action plan was "successful".
CPSC itself has been involved in such overstatement of effectiveness levels. For example, it has stated that certain unspecified recalls of products such as catalogue items and major appliances have achieved effectiveness rates approaching in excess of 90 percent. As described below, studies of recalls by CPSC and others in the past have shown that high rates like these are typically achieved only by programs that captured larger-than-average shares of the total number of affected units before they were sold to consumers and for which the remaining units have been in the hands of consumers for only a short period of time. 2 Limited CPSC data available for the 1996-97 fiscal years on the experience of "fast track" recall effectiveness (53-60 percent participation achieved) relative to traditional recall effectiveness in that period (16-18 percent) may also reflect this overstatement. Manufacturers of products in current production, where many of the recalled units are still in identifiable distribution channels rather than in consumers' hands, typically select the fast track option. Some items in distributors' hands and still on retailer shelves are frequently counted as part of the success rate of the recall. Such counting of items in distributors' hands should be consistently separated from the counting of consumer returns if the focus of recall success is to stay on the consumer.
Second, on the other end of the product life cycle, effectiveness rates for relatively old or obsolete products will be underestimated by the formula stated above unless product "retirements" are properly taken into account. Product retirements are properly viewed as an effective alternative means of eliminating potential hazards. In many cases, the ongoing process of retirements and replacements actually provides a greater degree of hazard reduction than is generated by the corrective action program during the same time period.
An example is the 1999 recall by a major manufacturer of three million kitchen appliances sold between 1983 and 1989. Under reasonable assumptions about average product life, only about one-third of these products were still in use at the time the corrective action program was announced. Thus the degree of hazard reduction accomplished by this program was nearly three times the level that would be suggested by a standard computation of the recall effectiveness rate.
Recalls of products that are at the end of, or even well past, their average service lives are not uncommon. The correction rates reported for these actions should be based on a reasonable estimate of the population of products still in use, not the number of products that were originally distributed. The 1980 CPSC Recall Effectiveness Task Force report also makes this important point (Tab B, p. 2).
Third, the metric most commonly used as the top half of the reported effectiveness rate presented above—the number of products actually returned, repaired, or subject to a request for a refund or credit (the "3Rs" of recall remedies)—is also flawed. Conceptually, a more appropriate measure of recall "success" is straightforward: rather than counting only those products that have been corrected through repair, replacement, or refund/credit, what should be counted are the number of affected products that are no longer capable of causing harm after the recall is initiated. In many recall situations, hazard elimination can be accomplished by a number of means rather than just the 3Rs.
Of course, some of these cannot easily be tracked or added to the "success" statistics, but they are very real. For smaller, less durable, or lower value products such as many of those covered in toy recalls, simply discarding or removing the product from use is often the least burdensome and most effective way of eliminating the hazard. Toy recalls typically receive widespread attention from parents, but typically have very low reported effectiveness rates. CPSC's 1980 Task Force Report emphasizes that these other methods undertaken by consumers that eliminate the potential hazard, but which are not included in the number repaired or returned, may be particularly important in explaining low reported effectiveness rates for other lower value products as well, such as hand-held hairdryers.
For larger, higher value products that have been in consumers' hands for an extended period of time, the normal life-cycle of product retirements and replacements typically accomplishes a substantial measure of hazard reduction over the course of a recall. In many cases, the process of product retirements and replacements actually provides a far greater degree of hazard reduction than is generated by the corrective action program. In a 1998 recall of old baseboard heaters, for example, the manufacturer achieved an (expected) reported effectiveness rate in the low single-digit range. However, the normal life-cycle of these products meant that more than one-fifth (22 percent) of the units still in service at the time the recall was announced were removed from further use in each of the next three years. Three years later, fewer than half of the affected products were still in use, despite the low level of replacements under the program. In addition to these "natural" product retirements, there may be additional "induced" retirements, as consumers accelerate an impending replacement of the affected product with a newer model that effectively accomplishes the remedy.
These alternative routes of hazard elimination are quantitatively significant in many high-profile recalls where there has been extensive publicity. Irrespective of what additional response is undertaken, these consumers have been made aware of the hazard that exists and the options and incentives available for eliminating it. It would be reasonable to expect by the conclusion of a recall that a significant share of them will have taken action to eliminate the potential for harm, even if these actions are not tracked in a 3R-based measure of program participation.
Finally, it should be noted that manufacturers and retailers must ultimately rely on the good sense of consumers and on their willingness to accept personal responsibility for participating in a recall. Many consumers may have heard about and know about a particular recall program, but have deliberately chosen not to participate, performing – as the 1980 CPSC Recall Effectiveness Task Force report makes clear – their own personal cost-benefit analysis that weighs the time, effort, and cost of participating against the hazard removal benefit to them. Thus, in many instances, particularly for recalls involving low-value items where extensive media or other special notification measures have been used, a complete assessment of recall effectiveness should include measures of the level of awareness about the corrective action program (e.g., likely number of product users who were exposed to a magazine ad or video news release), as well as the number of actual product returns or corrections.
A More Appropriate Measure of Recall Success
To sum up, the recall effectiveness rates most commonly relied upon to evaluate the success of corrective actions are not usually based on appropriate measures of the actual size of the affected product population in consumers' hands and the extent of hazard elimination. Without more information about the circumstances of a specific recall, injudicious reliance on conventional recall effectiveness rate statistics can produce a distorted picture of the success and potential for further improvement of corrective action programs. This bias is particularly serious for recalls involving products that have been in the hands of consumers for a long period of time, relative to the average useful life of the product. The age and type of product, the length of time in distribution, and the average service life all strongly influence both the number of remaining products that are potentially affected and the likelihood that users of these products can be notified.
For these types of products, a more appropriate measure of success, in my view, would be:
Number of potential product hazards eliminated
Number of products still in use
On its own, however, even this statistic does not provide a wholly sufficient benchmark for evaluating the effectiveness of recall programs. A fair and complete assessment requires fuller examination of the nature of product-, hazard-, and program-specific factors that contribute to, or impede, the level of hazard reduction that can be reasonably expected from a specific corrective action program (e.g., degree of perceived utility of the recalled product). Specific factors that influence consumer response to recalls of a particular category of products (such as small toys or SCUBA gear) should also be noted.
II. Data on Actual Program Effectiveness Rates
Data on the effectiveness rates achieved by specific recalls in the past 25 years are relatively limited, both because of a lack of more recent CPSC studies and reports on the subject and because of manufacturers' concerns about the release of proprietary product distribution and marketing data. In a limited number of instances, information can be extracted from press releases or media discussions of the program. However, the recall-specific data do not cover a sufficiently wide range of hazards, product types, and distribution profiles to be useful in analyzing systematically the determinants of effectiveness for recalls since the early 1980s.
While effectiveness rate statistics for individual recalls have not been available for most corrective actions since the early 1980s, some summary measures have been reported by CPSC. 3 For instance, Commission staff indicated in a 2003 recall effectiveness workshop that the effectiveness rate of traditional CPSC product recalls in FY 1996-1997 was 16-18 percent, while a 53-60 percent participation rate was achieved in "fast track" recalls. CPSC has elsewhere cited average correction rates in the same time period of about 25 percent for programs involving juvenile products and counter-top appliances.
In the absence of more recent analyses, the current state of knowledge about factors that influence the effectiveness rate that can be achieved by a particular recall program is still based in large part on the two CPSC studies (the first directed by this author) conducted in the late 1970s and early 1980s respectively. Both of these studies show a high relationship between a product's distribution profile and the level of returns that was achieved.
The May 1978 CPSC recall effectiveness study concluded that the length of time a product was in distribution and the proportion of the recalled product that was actually in consumer hands at the time of the recall were the factors that played a critical role in determining recall effectiveness. For example, four of the 97 cases in the CPSC study involved products that had been in distribution for over five years between the first shipment and recall notification to CPSC. The median effectiveness rate for these cases was only six percent. There were 57 cases in the study where more than 66 percent of the affected units were actually in consumers' hands. In these cases, the average effectiveness rate was only 11 percent. More important, it was "significantly lower than 11 percent" (the specific percentage was not provided in the study) for recalls where all of the affected products were in consumers' hands.
The second study, a statistical model of recall effectiveness, was developed by then-CPSC chief economist Paul Rubin and co-researcher Dennis Murphy (the Murphy-Rubin model) and privately published by the authors. Murphy and Rubin constructed the model by using data from over 100 CPSC recalls conducted in the early 1980s. Recalled products that were entirely in the hands of consumers were characterized initially in the Murphy-Rubin model by a predicted level of effectiveness on the order of ten percent. A significant reduction from this ten percent baseline level was estimated in the model for any recall in which there was a significant lag between the last date that the recalled product was placed into the chain of distribution and the date of the initial recall notice. In the model that Murphy-Rubin used to predict the effectiveness rate for any given case, for each additional year's lag between distribution and recall, the expected return rate was reduced by another 1.3 percent from the ten-percent baseline.
More recent anecdotal data suggest that the basic conclusions reached in these older studies about the relationship between a product's distribution profile and the level of recall returns for that product are still valid. However, there are fairly stark differences between the effectiveness rates quoted in these old studies and the estimates cited above from the mid-1990s by CPSC (such as statistics from an undisclosed sample of recalls involving catalog sales and those of major appliances showing average effectiveness rates approaching in excess of 90 percent). While the CPSC offered these effectiveness levels as evidence that direct notification could achieve nearly complete returns of affected products, the distribution profiles of the affected products were not reported. An alternative, and, in my opinion more likely, explanation of these high effectiveness rates, consistent with the past CPSC studies of effectiveness, is that a very high proportion of products involved were still in the chain of distribution or had only recently passed into the hands of consumers. Retrieving a product that has been in use for a few months at the time of the recall is much easier than securing returns of items that have been in use for much longer periods of time. As the age of the product increases, the accuracy of the contact information provided at the time of purchase diminishes dramatically, as noted elsewhere below in this paper.
Citing summary recall effectiveness rate statistics without providing the supporting details as CPSC has done in the past does little to advance our understanding of what corrective action programs can be expected to (and do) achieve. To evaluate whether the factors identified in the previous CPSC studies are still significant determinants of recall effectiveness would require a new study. I offered to conduct a full review of all of the relevant factors—assuming CPSC cooperation in release of adequate data to do so—at the 1999 CPSC Product Registration Conference. Years later, it remains a critical need. CPSC has much relevant data for such a study in its files, and should develop an appropriate procedure for making adequate data available from its corrective action report files and other sources (e.g., the files of cooperative recalling firms) to facilitate this type of analysis.
However, this study should be undertaken with the understanding that there is not likely to be any "magic bullet" solution to achieve substantial increases in return rates for product recalls. On the contrary, evidence of consumer response from both past recall programs and from more general marketing and response studies discussed below strongly suggests that effectiveness rates are not likely to be significantly increased by corrective action programs involving products which have been in distribution, especially for long periods of time. Some anecdotal evidence shows that bounties can be effective in cases like this, but if set too high, they offer additional incentive to resurrect or retrieve scrapped units to take advantage of very generous incentives for returns, introducing additional motivations for participation in a recall that are unrelated to a real reduction in risk to consumers. 4
III. Consumer Response to Recall Programs and Other Rebate/Refund Offers
There are real difficulties in translating consumer awareness into actual consumer action to repair, replace, or get a refund for a product. These factors dragging recall effectiveness downward are extremely important. In some instances, difficulties may reflect widespread consumer disregard for the message being communicated. In others, consumers may simply have chosen not to undertake additional correction steps recommended by the manufacturer after having taken other effective actions of their own to eliminate the hazard. Data are available from a few high-profile corrective action programs that underscore the importance of taking such difficulties into account when evaluating the success of recall efforts.
Evidence from Recall Programs
One early example of this slippage between awareness and action is provided in the 1980 report of the CPSC recall effectiveness task force. According to the report, over half (55.6 percent) of consumers who knew about the asbestos issue in the widely publicized hair dryer recalls of the late 1970s were not motivated to check their own hair dryers to see if they had an affected product (p. C-21). Measured return rates were even lower than this statistic would suggest, amounting to only 4.5 percent of the total units sold. In many instances consumers were willing to discard the product (with or without checking it for the hazard) without applying for the refund that was offered.
A lower-than-expected level of product returns was also evident in a coffeemaker recall in the early 1990s. In this high-profile recall, which received national media attention, at least 30 percent of consumers who: 1) called the toll-free number, 2) indicated that they owned affected model coffeemakers; and 3) received postpaid shipping cartons, failed to return the product for a $25 refund. Because many of these coffeemakers were at or near the end of their expected useful life, many of these consumers may have simply discarded the product without seeking the refund.
Finally, consumers may elect not to participate in recall programs where the perceived risk of the hazard is very small. For example, many recalls involve potential misuse of items by children. Households that do not have children present but have the affected product (e.g., lead in paint) may reasonably decline to participate in the corrective action, but this lack of response does not have any effect on the level of hazard reduction achieved by the recall.
One important problem is that there is little quantitative evidence on consumer response available from comparatively recent recalls. An updated data collection and analysis effort is badly needed to assess issues such as the best channels for promoting consumer awareness of recalls and which media or programs are most likely to stimulate actual consumer response. Concern should be directed both to assessing the value of new channels for notifying consumers of recalls and to evaluating the effectiveness of more traditional approaches, such as point-of-sale notification, newspaper/magazine advertising, and direct mail, in recent programs. There is also a need to better understand not only the process by which consumers learn about recalls, but also what motivates them to respond to various types of appeals. CPSC undertook a contractor review of the consumer behavior and marketing literature and a series of Recall Workshops in 2003 to lay a foundation for addressing and providing insight into these issues, but to my knowledge has not reported publicly on any subsequent work on these questions.
As I noted in my remarks at the 1999 CPSC Product Registration Conference, the best available source of information that could be used to conduct some of this research is in the corrective action program files maintained by the CPSC. Other issues will require new data collection and analysis. In my view, mandating specific approaches to achieving more effective recalls is premature until this information can be obtained, reviewed, and analyzed.
IV. Consumer Response to Product Registration Requests
A few specific proposals have been made in the past several years to increase recall effectiveness. Much of this effort has focused on the use of product registration cards (PRCs) to be completed by consumers at the time of purchase and kept by manufacturers to facilitate consumer notification in the event of a recall. Under Section 104(b) of the 2008 CPSIA, CPSC has very recently set in place, as required by the statute, a final rulemaking proceeding regarding PRCs for durable infant and toddler goods. It should also be noted that a petition several years ago (2001) by the Consumer Federation of America (CFA) to require PRCs for children's products was rejected by CPSC in 2003 after extensive comments from interested parties, including some new analysis that examined consumer response rates to experimental notification materials.
I am not optimistic on the prospects for achieving significant increases in recall participation by imposing more stringent requirements for the provision and use of PRCs. The "best" case for them comes from a NHTSA study in the 1990s in which the mandatory inclusion of PRCs with child safety seats sold to consumers resulted in increased PRC returns from 3 percent to 27 percent, and an increase in car seats returned for repair from 13.8 percent to 21.5 percent for an expensive product with high safety-related content. 5 The likely response would almost certainly be significantly lower for less expensive, non-safety related products. Indeed, return rates for PRCs are typically quite low -- on the order of 10 percent or less -- for most common consumer products. Also, research experiments conducted by Equifax DMS Data Services in the late 1990s and early 2000s indicates that changing various format and design elements of a PRC will have only a modest impact on increasing the level of response - a maximum increase of 3.4 percent in response share, and less than 1 percent for low-value products (under $50).
Such examples of "incomplete" consumer responses to recall programs with PRCs are not surprising, but rather are entirely consistent with published experience that commercial fulfillment contractors have had with refund and rebate redemption rates. While these estimates are not specifically related to recalls, they share many of the same features that help to explain consumer response patterns. Research studies, again by Equifax DMS in the late 1990s and early 2000s, show that when consumers are offered significant financial incentives to participate in conventional marketing programs for newly purchased products, redemption rates reach the 50 percent mark only for generous offers associated with higher value products, and are much lower (in the teens or less) for lower-value products (under $50). Participation rates—at least as they are measured by returns of products and/or applications for incentive payments—achieved by corrective actions for products that have already been in the hands of consumers for an extended period of time, rather than newly purchased, will be much lower. Using the Equifax pattern of redemption rates as a guideline, achieving returns in the 10-15 percent range would represent the maximum consumer response that can be expected for products nearing or at the end of their useful service lives.
Likewise, two leading firms known to have conducted pilot experiments in response rates for product registration cards, Toro and Mattel, achieved disappointing results for pilot tests that they ran, according to CPSC comments in materials denying CFC's PRC petition in 2003. Toro had a low test result of 7 percent returns for a pilot PRC program for a leaf blower. Mattel found its results disappointing and discontinued a pilot program of requiring PRCs for a line of its motorized toys.
There are other reasons as well to question the value of a major effort to expand the provision and return of PRCs. Most importantly, the information collected by PRCs becomes out-of-date rather quickly, because of changes in both product ownership and household location. In the case of product registrations tied to consumer addresses at the time of purchase, it is important to note that nearly one out of every six people (16 percent) moves every year, and that the average person moves every 5-6 years (CPSC petition response package, 2003). NHTSA found that the usefulness of its PRC data base for child safety seats had declined to 10-13 percent of its original contact usefulness after only three years
Moreover, as discussed below, there are notification and publicity approaches that are promising alternatives to PRCs. The tools for notifying users of recalls include product registration and recall communication tools through electronic media venues, as well as video news releases (VNRs). Such tools provide the best avenue for improving the link between the initiators of recalls and the owners of affected products.
Taken as a whole, these results suggest that significant increases in PRC returns, such as that achieved by the NHTSA car seat campaign, are possible only for products that are strongly related to safety issues in the minds of consumers.
V. Electronic Channels for Increasing Recall Notification and Awareness
Several channels for increasing recall notification and awareness, though no longer new, may have yet to reach their full potential, based on their extremely rapid growth to date. Among these are video news releases and various web-based electronic media channels for notifying consumers.
Video News Releases
A significant number of corrective action programs have included production and distribution of video news releases (VNRs). The CPSC Web site lists 160 VNRs on its website (1998-2009). A total of 92 of these are for recalls, and 70 percent of those are for children's products. VNRs provide a potential means of notifying affected consumers who cannot be reached through conventional print media.
One problem faced by manufacturers, however, is getting television stations to use the VNRs that they distribute. In a paper presented at the 1999 CPSC Product Registration Conference, Dirk Gibson indicated that the distribution of these releases was not always followed by actual broadcast—of the 700 television stations which were sent a copy of the CPSC Recall Round-up VNR, only 140 stations (20 percent) broadcast it. 6
There is now more experience, and presumably data, available on the levels of exposure achieved by various VNRs. It is important for this information to be made available, on a "product de-identified" basis, to facilitate an assessment of VNRs' potential for achieving significant improvements in recall effectiveness.
The Internet
The emergence of both general and specialized internet safety and recall announcement sites also warrants discussion. The vast majority of consumer contacts with recall programs are through the web, rather than by phone or mail (CPSC 2003 Recall Effectiveness conference). CPSC's website and the associated website Recalls.gov, as well as manufacturer, distributor, and retailer websites, provide capabilities for consumers who are interested in individual products, as well as in many cases enhanced search capabilities for consumers who are interested in recalls of entire categories of products. These websites not only improve the publicity of recently announced recalls, but, perhaps more importantly, provide consumers with the ability to check on specific products long after the date when a particular corrective action was first announced. Efforts such as these are particularly important in helping consumers to identify older products, such as cribs and car seats, that do not meet current safety standards, but which cannot be located with any form of direct notification.
In addition to the CPSC sites and others that focus on publicizing recall announcements, online product registration is now more widespread for certain products beyond those in the computer and entertainment sectors in which they first appeared. Also appearing more frequently are centralized online product registration sites which would make it possible to combine purchase and registration activities in a single transaction. This approach may offer potential economies of scale to consumers who are interested in being contacted in the event of a recall, but who do not wish to take additional time and effort to complete and return paper registration cards for each individual product purchased.
Means of promoting recall awareness using electronic venues remain to be fully developed. Sponsored searches, banner ads and links at portal sites, and opt-in e-mail lists are all approaches now being used that also have the developmental potential to expand the level of consumer awareness about recall programs.
VI. Conclusion
As has been detailed in this paper, the conventional measurement of recall effectiveness typically used by the CPSC has conceptual problems that seriously limit its usefulness as a benchmark to evaluate the success of various corrective action programs. In addition, there is a scarcity of recent data on both recall effectiveness (however measured), and on consumer notification and response behavior for various recall techniques used. Addressing these serious limitations in a systematic manner through a new research effort that would take advantage of existing CPSC recall files, as well as collecting new data, are essential to providing a significantly improved foundation for answering the question of what additional steps would be most likely to improve recall effectiveness rates significantly.
Dr. Edward J. Heiden has been president of Heiden Associates, Inc., since 1981. He has directed and published numerous studies in the areas of product safety and micro-economics, and has presented analysis and testimony in many federal and state court cases, as well as regulatory proceedings and negotiations. Prior to formation of his firm, he was chief planning economist for the Federal Trade Commission and the U.S. Consumer Product Safety Commission (CPSC), as well as the White House Office of Consumer Affairs. He has also served on the Economics faculty at the University of Wisconsin (Madison) and Director of its Center for Firm and Market Behavior. While at CPSC he directed the agency's first major study of product recall effectiveness, and in 2003 as a CPSC contractor co-directed a comprehensive search of the product recall literature on consumer recall awareness and notification. He holds a doctorate in Economics from Washington University (St. Louis) and a B.A. summa cum laude from St. Louis University. Contact him at (202) 463-8171, eheiden@heideninc.com
----------------------------------------
|
E-Mail This Article
Printer-Friendly Version