Accident Analysis and Prevention 32 (2000) 383–388

What you don’t know can hurt you: household products andevents

S. David Leonard a,*, Michael S. Wogalter b

a Department of Psychology, Uni6ersity of Georgia, Athens, GA 30602-3013, USAb Department of Psychology, 640 Poe Hall, North Carolina State Uni6ersity, Raleigh, NC 27695-7801, USA

Accepted 24 May 1999

Abstract

Product safety is affected by product design and by the knowledge of the user, either through the user’s own background orthrough instructions and warnings presented with the product. Given adequate knowledge, warnings can serve primarily to remindindividuals of the hazards and precautions that can be taken. This study examined people in the USA (represented by two diversesamples) to evaluate their knowledge about the hazards associated with common household products and situations using bothmultiple choice and open-ended surveys. The results indicated that the respondents were aware of a substantial number ofhazards, but their knowledge often did not extend to the specific circumstances that could produce personal injury and propertydamage. Further, comparisons of cued and non-cued responses suggested some hazards are not well recognized without the cue.The results indicate warnings are needed both as reminders and to provide safety information. © 2000 Elsevier Science Ltd. Allrights reserved.

Keywords: Household products; Hazard identification; Risk perception

www.elsevier.com/locate/aap

1. Introduction

This paper is primarily concerned with safety in thehome environment. Specifically, it will examine knowl-edge about hazards often encountered in the home. Tothe extent that lack of knowledge influences the safetyof people it is important to know what they know anddon’t know about objects and events in their environ-ments in order to provide adequate warnings. Manyphysical events are not well understood by the generalpublic (cf. McCloskey et al., 1983). Some of these arerelevant for understanding the seriousness of hazardsand their consequences. Ergonomic development ofproducts commonly used by consumers includes thenecessity for users to be informed about hazards associ-ated with those products. Understandably, the extent towhich one perceives the risk in a hazard is based onunderstanding its consequences. Two factors importantfor safety are the physical characteristics of the prod-

ucts and the extent to which the limitations on use andthe hazards associated with the products are recognizedby the users. In industrial settings it may be possible toprovide safety training regarding hazards of a specificapparatus when it is introduced. Thus, as new anddifferent hazards occur, information about them can beintroduced with them and transmitted rapidly, althoughthis may not occur as quickly as it should in someinstances. Unfortunately, for products used in thehome, formal training in safe use is generally notpossible.

Manufacturers may assume that the public at large isaware of the hazards, because those who develop andproduce the products are experts who know the generalphysical principles and how they may produce thosehazards. However, the small number of incidents re-ported may simply reflect the fact that many individualshave been fortunate and did not receive serious injurieswhen the hazard was experienced. Many reporting sys-tems for injuries (e.g., the National Electronic InjurySurveillance System in the USA) do not include rela-tively minor events. Some products may become morewidely available and potentially increase the number of

* Corresponding author. Tel.: +1-706-542-3041; fax: +1-706-542-3275.

E-mail address: dleonard@egon.psy.uga.edu (S.D. Leonard)

0001-4575/00/$ - see front matter © 2000 Elsevier Science Ltd. All rights reserved.PII: S 0 0 0 1 -4575 (99 )00072 -X

S.D. Leonard, M.S. Wogalter / Accident Analysis and Pre6ention 32 (2000) 383–388384

accidents (Godfrey and Laughery, 1984). Further, be-cause of prior use of earlier similar but less hazardousproducts, later more hazardous products may producemore injuries in the future. Experiences with hazardousmaterials that have not affected one adversely mayreduce the perception of risk (Karnes et al., 1986). Assuggested by the data of Wright et al. (1982), concernfor reading instructions may be a result of how seri-ously one views the hazard. In some instances efforts atguarding may prove counterproductive, as in the caseof motorized belt systems in automobiles. Because ap-preciable numbers of drivers do not understand thehazard involved if one fails to use the lap belt when themotorized shoulder belt is in place (Lehto and James,1997; Leonard and Karnes, 1998), they may be uncon-cerned with the necessity for making the effort toprotect themselves.

It was the specific concern of this research to evaluatepeople’s knowledge about hazards posed by commonhousehold products and events rather than the informa-tion provided by warnings. If everyone understands ahazard, all that need be done is to provide a warning asa reminder. If the public is unaware of a hazard,additional effort by manufacturers in terms of design,guarding, or warning needs to be expended in order tocounteract behavior that might result in injuries. Some-times the best a manufacturer can do is to warn abouta hazard. Incorporating awareness of hazards into theeducational process may be possible, as has been donewith safety training in the schools, but for hazards ofnewly developed products or new hazards incorporatedin later versions of old products, a safety campaign mayneed to be mounted to inform those individuals whoseschool days are well behind them.

Warnings must adequately inform individuals whoare not knowledgeable about the topic. Warnings gen-erally include hazard information, information aboutpossible consequences and safety instructions for avoid-ing those consequences. A problem with many warn-ings is that they are not well presented (e.g. in smallprint on the back of the container), and their constantpresence can produce habituation even to the reminderfunction that they might serve. Thus, elimination ormitigation of the hazard by design or guarding isgenerally the best solution to the problem. Previousresearch has shown terms used in warnings on commonconsumer products are not understood by many indi-viduals in the US population (e.g. Leonard et al., 1991;Leonard and Digby, 1992). This lack of understandinghas resulted in inappropriate use of products withdisastrous consequences in some cases. For example, ifone is aware that flammable materials can cause explo-sions, the term flammable in a warning may influencehow, or whether, one uses a product. Lacking thisknowledge, misuse may occur, possibly producing in-jury. For example, Leonard et al. (1997) found consid-

erable misunderstanding of common electrical termsand concepts which could lead to injuries and possiblyeven death.

Effective warnings include information about activi-ties that might be associated with the hazards. Thus, itis important for the manufacturer to evaluate the possi-ble uses and misuses of the product and determinewhich hazards need to be warned against. Those haz-ards would then be described and the appropriate coun-termeasures presented. Some hazards may be open andobvious to users, while others may be ‘hidden’ andrequire explanation and warning. Automobile accidentsusually receive publicity if someone is killed or a num-ber of people are injured, and an unusual incident suchas a tree falling on a person may also be publicized.However, the ingestion of an overdose of iron tabletsby a young child (possibly an unusual event) is unlikelyto be reported in the news media, despite the fact thechild may need treatment by professionals. A parentwho is aware of the possibility of such an event and thedanger it presents will be more likely to take appropri-ate precautions than one who isn’t. The hidden or lesserknown hazards are most important to warn against.The present research was done to examine people’scognizance about hazards in the home for which therewas some question regarding the general public’sknowledge, at least in the USA.

In contrast with previous studies that have examinedunderstanding of warnings, this research focused onpeople’s knowledge about certain hazards. Two typesof questionnaires were used to examine the ability ofrelatively untrained people to recognize what possibleill effects might from various actions with severalhousehold products and some events that might occurin the home. One questionnaire was in a multiple choice(MC) format and administered in groups, while theother was an open-ended (OE) questionnaire that wasadministered individually. It was assumed that the cuesprovided in the MC condition would generate higherrecognition of hazards than in the OE condition. Asubstantial difference between the two might indicate adifference in the extensiveness of the warning required.

2. Methods

A total of 62 students from two geographicallywidely separated colleges within the USA (Metropoli-tan State College of Denver in Colorado, MSCD andNorth Carolina State University, NCSU) completed asurvey in the MC format. Of these 25 were men and 39were women. The ages ranged from 18 to 55, but only12 were 25 or older. Some respondents received coursecredit as incentive to participate. Two groups compris-ing a total of 115 individuals participated in the OEsurvey. One group completing the OE survey included

S.D. Leonard, M.S. Wogalter / Accident Analysis and Pre6ention 32 (2000) 383–388 385

Table 1Percentages of respondents selecting possible consequences

Injury or illness Selecting itemObject or eventpossibility (%)

a. Cancer develop-Hair spray 11ment

48b. Respiratory disease89c. Explosive burning42d. Asphyxiation

Removing old paint 92a. Lead sickness6b. Difficult to repaint

c. Toxic fumes from 84solvent

a. HeatingHow to remove paint 39around children b. Use of chemicals 27

45c. Chipping andscraping

a. Appreciation of 45Effect of age on toypurchase toy by child

b. Choking from 97small partsc. Sharp edges 90

a. Vitamin C 40Food supplementsthat may poison a b. Iron tablets 89

c. Multi-vitamin/min- 45childeral tablets

50d. Vitamin B1

a. Flames glow moreEnvironmental signs 16of carbon monoxide brightly

63b. Pets lethargic95c. May feel drowsy

6d. See carbonmonoxide vapor

a. NauseaSymptoms of carbon 85b. Dizziness 97monoxide

87c. Headache0d. Excitability

Object or event Injury or illness Selecting item(%)possibility

Indoor use of gas or 94a. Possibility of fireb. Buildup of carbonkerosene heaters 68monoxidec. Catch hepatitis 2d. Respiratory prob- 34lems from spill

a. Turn on water 5What not to do withb. Lighting any flame 100gas leak

85c. Use any electricalequipmentd. Opening doors 5

a. Leave immediately 77What to do if you40suspect a gas leak b. Call gas company

and then leave0c. Call gas company

and stay0d. Don’t worry about

it

Use of product that a. Hot but not flaming 73says do not use charcoal grill

b. Gas furnace that is 23around an opennot runningflamec. Lit cigarette 92d. Cigarette lighter 77with shield

69e. Glass-enclosed fire-place

a. Use only outdoors 26What does message‘use in a well venti- b. Use fan while using 58lated area’ mean to product

79c. Be sure door is openyouto area

56d. Window needs to becracked open

84e. Not in small en-closed space like closet

65 students from NCSU. Of these 36 were men and 29were women. The ages ranged from 18 to 35 years, butvery few were older than 23 years. A second group withan additional 50 individuals, 27 women and 23 men,were sampled from the nearby community. Their agesranged from 21 to 70 years with a median of 32.5 years.The community group were roughly comparable to thestudent population in intellectual capabilities as indi-cated by the fact that 25 of them indicated they haddegrees and an additional 15 stated they had from oneto three years of college training. They were recruited ata flea market and received small prizes for their partic-ipation. An attempt was made to avoid overweightingthe samples with persons having technical jobs. Mostwere employed in business capacities.

The survey forms were similar in terms of the ques-tions, but the MC forms included several possible re-sponses for the items. Each item consisted of a shortsentence describing the hazardous object or event withdescriptions of a hazard or hazards associated with it aswell as some hazards not related to it in most items.The respondents were instructed to check all items theythought were correct and were allowed to select asmany of them as they wished. Abbreviated versions ofthe items and possible responses are included in Table1. In the OE group the survey form did not include anypossible responses, but participants were asked to givetheir ideas about the items. Each item consisted of ashort question about an object or event that has haz-ards associated with it. In general, the items examined

S.D. Leonard, M.S. Wogalter / Accident Analysis and Pre6ention 32 (2000) 383–388386

dealt with problems of home activities that are rela-tively common. Each participant was presented withone of the safety surveys containing 23 items. The itemsshown in abbreviated form in Table 1 were interspersedwith additional items regarding electrical hazardswhich have been reported elsewhere (Leonard et al.,1997).

3. Results and discussion

As shown in Table 1, the level of understandingvaries among topics. To simplify matters we have pre-sented the data in content categories related to theunderstanding involved. The topics will be consideredin their order in the table. The presentation order in theoriginal survey was somewhat different in order toobviate possible effects of juxtaposition of similaritems. One concern about the procedure was that be-cause of usual test procedures individuals might selectonly one of the possible hazards despite the instructionsallowing them to select more than one. The responsepercentages displayed in Table 1 indicate this fear wasunfounded. Only the items involving gas leaks andpaint removal produced minimal responding, and forboth of those items the respondents could have consid-ered some of those responses to be mutually exclusive.

It appears that most of the respondents to the MCform were aware of the explosiveness of hair sprays inthat 89% selected that hazard. However, the percent-ages associated with some of the other responses maysuggest a bit of a positive response bias. For example,asphyxiation was listed by 42% of the respondents.Further, without the cue only 65% of the respondentsprovided a response of flammability. The differencebetween cued and non-cued percentages were found tobe significant at PB0.001 by the chi-squared test ofindependence. All differences between the OE and MCconditions described in these results as significantreached that probability level. The percentages of re-sponses provided by the different subgroups on the OEquestions are provided in Table 2. It should be notedthat responses to the MC items were tied to the termsused in the survey form, while responses to the OEquestions were stated in the words used by the respon-dents. Thus, as seen in Table 2 the respondents tendedto use the word flammability. These data do not allowdetermination of whether or not use of the term explo-siveness by the MC respondents was influenced by itspresence on the form.

The possibility of lead poisoning from paints wasrecognized by 92% of the MC group, although againthe percentage of those making the response was signifi-cantly less (70%) when no cue was provided. However,

fact that 45% would adopt the technique of chippingand scraping off the old paint as the solution to itsremoval suggests that they may not recognize thatyoung children might ingest the resulting chips. It isalso possible that some respondents thought that theywere required to make some response to each item andconsidered this to be the best alternative. This possibil-ity is considered because relatively fewer multiple re-sponses were made to this item than to most of theothers. Because there was almost universal recognition(97%) that small children can choke on small toys, onecan assume the respondents recognized that young chil-dren often put things in their mouths. However, theconnection between ingestion of paint chips and leadsickness may not have been made. The problem ofrecognizing that an object may have the characteristicthat is hazardous is also seen in the responses to theitems regarding food supplements and use of productsaround an open flame. Clearly, a very high percentage(90%) of individuals can recognize iron tablets pose adanger of overdose among young children, but thepossibility that a multiple vitamin and mineral tabletmight contain iron is not considered by many people.Only about half as many (45%) respondents selectedthat particular response. The effect of cues on responsesis perhaps most evident in responses to this item. Only26% of the respondents in the OE condition sponta-neously mentioned iron tablets as dangerous.

While awareness of symptoms of carbon monoxide(CO) was high among those in the MC condition with95% selecting drowsiness as a symptom, only 12% inthe OE condition responded in terms of physical symp-toms. Only 68% recognized that using a gas or keroseneheater indoors could produce CO buildup. Again therewas a substantial difference as a function of cueavailability with only one-third of the respondents inthe OE condition mentioning the CO hazard in con-junction with using a grill indoors. In a similar fashionthere was universal awareness that one should not lighta flame if a gas leak exists, and 85% responded that oneshould not use electrical appliances. Again significantlyfewer respondents in the OE condition (only 21%) whomentioned using electrical equipment. Moreover, in theMC condition 40% chose the option of calling the gascompany and then leaving the house. In the OE condi-tion a similar percentage (47%) indicated that theywould call an authority. A logical inference is thatmany people do not recognize the possibility of sparkproduction by the telephone.

Many products warn against their use around anopen flame. However, the concept of an open flame isapparently quite varied in the thinking of these respon-dents, as shown by the percentages of individuals notresponding to many of the options in the MC question.The small percentage that recognized pilot lights onfurnaces suggests warnings about open flames may not

S.D. Leonard, M.S. Wogalter / Accident Analysis and Pre6ention 32 (2000) 383–388 387

be adequate without some description of what is meant.While this specific problem is being partially amelio-rated by the replacement of pilot lights by igniters innewer furnaces, many older pilot lit furnaces will con-tinue to be used for years to come. More troublesome isthe possibility that in moderately cool weather a fur-nace may be cycling on aperiodically with extended offperiods, thus providing a source of ignition for anyexplosive gasses. In such circumstances one may beunaware of the furnace’s operation. The fact that only69% indicated that they would consider a glass enclosedfireplace to be an open flame suggests that some peopleconsider only the fire moving out rather than thevapors moving into the flames. Therefore, it appearswarnings concerning the use of highly flammable andexplosive materials will not be adequate if they includeonly a statement to avoid open flames.

From both a safety standpoint and from the stand-point of efficiency of use the presentation of warningsor instructions in vague terms has plagued the con-sumer. In an effort to evaluate whether or not people ingeneral are likely to interpret vague instructions in thesame way the respondents were asked to indicate what,

‘Use in a well ventilated area’, meant to them. As seenin Table 1 the responses were very variable. Somewhatsurprisingly only 84% checked the choice of not usingthe substance in a closet. Although instructions regard-ing cubic feet per minute of airflow would likely be nomore instructive to the average user than ‘well venti-lated’, it should be possible to provide some sort ofbenchmarks for size of the area and possible ventilatingtechniques that would be less vague.

4. Conclusions

If a hazard is well understood by the general public,its consequences also might be expected to be under-stood. In such cases instructions are not necessary, butthe results obtained in this study indicate warnings maybe necessary to remind users of the hazards. The impor-tance of warnings as reminders, even if not necessary asinformation providers, is clear from the differencesbetween the MC and OE conditions. The cues providedby the alternative responses given in the MC conditionproduced substantially more appropriate responses

Table 2Number of respondents specifying possible consequences

ConsequenceObject or event Percent listingNumber listing

Student Non-student

F M F M

Flammability 14 22Hair spray 16 16 6519Eye irritation 6 6 5 3

17Lead sickness 7015 21 20Removing old paint

111826 7625Choking on partsEffect of age on toy purchase3 3 14Other safety factor 4 5

Age appropriateness 2 6 4 5 16

Iron tablets 7 7Food supplements that may poison child 9 4 265 2 10Vitamin 1 2

Vitamin C 5 4 1 3 12Other 14 13 5 6 36

171521 6515Use detectorDetection of carbon monoxideOdor 5 5 4 3 16Physical symptoms 5 5 2 1 12

Nausea 10 11Symptoms of carbon monoxide 8 6 33Lightheaded, dizzy 11 15 9 7 40Headache 3 3 5 4 14

Fire 21 31 21 20 89Indoor use of grillSmoke 19 26 10 10 62Carbon monoxide 331112 111

Lighting any flame 14What not to do with gas leak 28 17 13 696565Use electrical equip. 21

Use telephone 0 00 33

What to do if suspect a gas leak Leave immediately 14 12 1012 46Call authority 16 14 13 6 47

S.D. Leonard, M.S. Wogalter / Accident Analysis and Pre6ention 32 (2000) 383–388388

than were obtained without the cues. Of course, it ispossible that the warnings may simply not be noticed.Prior research on label preferences in the senior au-thor’s laboratory indicated this is often the case. Thepresent studies were designed only to determine under-standing of the problem. Other research is necessary toevaluate how warnings in some of these cases can beaccomplished effectively. Presumably, if warnings aresimply reminders the amount of information presentedcould be limited, but in the cases where understandingis generally poor, more information about avoiding theconsequences would be included.

The results indicate that the US public, at least asrepresented in these samples, is aware of many hazards,but that awareness does not necessarily translate intounderstanding the sorts of actions necessary to avoidthe hazards. This is particularly true when more thanone sort of knowledge is required to obviate a hazard.The knowledge that an electric spark can set off a gasexplosion may be negated if one is unaware that atelephone connection is electrical. The responses toquestions dealing with CO are very revealing in termsof how lack of specific knowledge can place one at risk.Although there were other questions in the surveyabout CO, only 68% of the respondents in the MCcondition recognized it as a potential threat from in-door combustion, and only 30% noted it when no cuewas provided.

Generalization of these results is difficult to evaluate.Because college students and others who were primarilycollege educated were used, we can assume our sampleto be better educated and perhaps more able to recog-nize problems and combine separate components thanthe population at large. The relatively skewed distribu-tion of ages, however, suggests that the life experiencesmay be less extensive. Owing to the relatively smallnumber of older respondents, no statistical analysis ofdifferences was attempted, but examination of the rawdata suggested that differences between older andyounger subjects were minimal. There were also few, ifany, differences between the samples from North Caro-lina and from Colorado. It might be expected that someknowledge, for example, about chemicals would bebetter understood by well educated individuals, andother knowledge might be likely to be accrued with age.Determination of this will take more extensive research.In any case, it is a fact that younger people and poorlyeducated people do use the products and engage in theactivities considered in this survey and may be placed insituations where gas leaks and CO buildup can occur.Thus, it is important to provide them with appropriateinformation. Sometimes the best a manufacturer can dois to warn about a hazard. In some cases the informa-tion may be adequately presented by on product in-structions, but in others an educational effort may be

needed, such as has been done with other safety train-ing in the schools.

Although the percentages obtained in this survey aresample values, the upper end of a 99.9% confidenceinterval for an appropriate response made 70% of thetime would be approximately 90%. Thus, in a countryof 250 million people, such as the USA, it is likely thatat least 25 million people would be at risk. If a similarpercentage of the European population follows thepattern, the individuals at risk would be more thandoubled. Obviously, these problems are matters of con-cern. Manufacturers of products involving new hazardsor new forms of older hazards should be encouraged toinform their clientele because of safety factors. It wouldalso be important from a business standpoint to avoidthe publicity that might discourage sale of productsthat would be safe if used appropriately. Safety cam-paigns may need to be mounted to aid in dealing withhazards of newly developed products. Further researchis needed to determine what sorts of information andeducational procedures will be most effective in bring-ing the hazard to the attention of the public and inproviding adequate understanding of the characteristicsof the hazard.

References

Godfrey, Laughery, K.R., 1984. The biasing effects of product famil-iarity on consumer’s awareness of hazard. In: Proceedings of theHuman Factors Society 28th Annual Meeting 1984, Human Fac-tors Society, Santa Monica, CA, 483–486.

Karnes, E.W., Leonard, S.D., Rachwal, G., 1986. Effects of benignexperiences on the perception of risk. In: Proceedings of theHuman Factors Society 30th Annual Meeting 1984, Human Fac-tors Society, Santa Monica, CA, 121–125.

Lehto, M.R., James, D.S., 1997. Safety knowledge of users andnon-users of the lap belt on two-point motorized belt systems.Accident Analysis and Prevention 29, 739–744.

Leonard, S.D., Creel, E., Karnes, E.W., 1991. Commonly usedhazard descriptors are not well understood. In: Karwowski, W.,Yates, J.W. (Eds.), Advances in Industrial Ergonomics andSafety, III. Taylor and Francis, London, pp. 731–738.

Leonard, S.D., Digby, S.E., 1992. Consumer perceptions of safety ofconsumer products. In: Kumar, S. (Ed.), Advances in IndustrialErgonomics and Safety, IV. Taylor and Francis, London, pp.169–176.

Leonard, S.D., Griffin, R.S., Wogalter, M.S., 1997. Electrical hazardsin the home: what do people know? In: Proceedings of the HumanFactors Society 41st Annual Meeting 1997, Human Factors Soci-ety, Santa Monica, CA, 840–843.

Leonard, S.D., Karnes, E.W., 1998. Perception of risk in automo-biles: is it accurate? In: Proceedings of the Human Factors Society42nd Annual Meeting 1998, Human Factors Society, Santa Mon-ica, CA, 1083–1087.

McCloskey, M., Washburn, A., Felch, L., 1983. Intuitive physics: thestraight down belief and its origin. Journal of ExperimentalPsychology: Learning, Memory and Cognition 9, 636–649.

Wright, P., Creighton, P., Trelfall, F.M., 1982. Some factors deter-mining when instructions will be read. Ergonomics 25, 225–237.

.