The influence of safety at work on safety at home and during leisure time

The in?uence of safety at work on safety at

home and during leisure time

J.Lund a,*,J.Hovden b

a

Department of General Practice and Community Medicine,University of Oslo,PO Box 1130,

Blindern,N-0317,Oslo,Norway

b

Department of Industrial Economics and Technology Management,The Norwegian University of Science and Technology,Trondheim,Norway

Abstract

It has been argued that the e?ects of safety programmes extend to safety related behaviour outside the work environment.Data from studies of the behaviour of workers in three industries with a high focus on safety were examined in order to explore this argument.By means of data collected through a questionnaire,safe behaviour and emergency preparedness behaviour at home and during leisure time were measured in a case-control study.The ?nd-ings indicate that workers in two of the three studies did not transfer safe behaviour from workplaces with a high focus on safety to home and leisure arenas.The safety consciousness acquired by workers at their workplaces in the absence of specially designed home and leisure time intervention projects seems to be insu?ciently comprehensive or deep enough to in?u-ence safe behaviour in other arenas.While safe behaviour was not transferred,the emergency preparedness behaviour was.This indicates that the contextual and situational aspects of emergency preparedness are shared across risk arenas.#2003Elsevier Science Ltd.All rights reserved.

Keywords:Behaviour transfer;Safety behaviour;Emergency preparedness behaviour;Risk arenas;Home and leisure safety

1.Introduction

The 1980s represented a shift in focus in safety science theory and practice from tech-nical and human factors to management and organisational aspects of safety promotion (Hale and Hovden,1998).These changes were partly driven by lessons learned from major disasters and approaches to regulation,and partly by an international ‘‘revival’’

*Corresponding author.Tel.:+47-22-850550;fax:+47-22-850620.E-mail address:johan.lund@samfunnsmed.uio.no (J.Lund).

2J.Lund,J.Hovden/Safety Science41(2003)&–&

in industry of an integrated strategy of goals and means to achieve general loss prevention and business success through the prevention of occupational accidents. This‘‘revival’’was to a large extent based on the dogma of DuPont,which main-tained beliefs such as‘‘all injuries can be prevented;safety must be equal to quality, costs,and productivity;safety is a line organization responsibility;good safety is good business’’(Webber,1987,p.1).A central argument for the focus on Lost Time Injury(LTI)rates,near misses,and deviations was the triangle for ratios between minor,major,and fatal injuries in the‘‘iceberg theory’’(Heinrich,1931;Bird and Germain,1996).For a review,see Hale(2000).Emerging trends included arguments for more holistic approaches to safety,health,environment,and quality manage-ment in industry that were extended to include safety in tra?c,home,and during leisure time in the same framework.Beliefs evolved that synergy e?ects,experiences, and practices regarding attitudes and behaviour could successfully transfer from one risk arena to another.

The safety promotion programmes that were implemented successfully reduced the occupational accidents and injuries in many companies to a very low level (Kjelle n et al.,1997).Some ambitious companies then began to turn their attention to the safety of their employees and their families outside the workplace.Some big international companies as well as the companies studied in this article argued that safety programmes at work also in?uenced behaviour outside the workplace.These e?orts by industry to achieve‘‘24hours safety’’corresponded with initiatives from health authorities to promote and implement concepts such as‘‘Safe communities’’(Lund and Svanstro m,1995).

We were interested in exploring these beliefs about behavioural transfer from occupational safety to other arenas because we had not seen empirical evidence of similarities in behaviours across di?erent arenas.Data from three surveys were examined.

Hypotheses for testing:we assume that e?cient safety programmes,for example, education and training,propaganda,strict control regimes,legislation,and regula-tion at the workplace will a?ect not only safe behaviour at the workplace,but also safe behaviour in other arenas.As safety at work is normally directed towards both safe behaviour and emergency preparedness behaviour,we will separately test the in?uence of programmes on each of these behaviours.Generally,emergency pre-paredness behaviour is based more on practical training and exercises,whereas measures for safe behaviour represent a mixture of organisational and physical strategies directed towards the individual.The measures directed towards the indi-vidual are safety related job education and training,attitude in?uencing measures (campaigns),and behavioural modi?cation measures.The H1hypothesis above expresses common beliefs.A H0hypothesis,i.e.there is no transfer of safe beha-viours between arenas,is more challenging and interesting from a safety science point of view,because it contradicts mainstream thinking.

Our data are based on three independent surveys of self-reported behaviour in groups of employees heavily exposed to occupational safety and emergency pre-paredness propaganda.These groups were compared with employees not specially exposed to safety measures at work.Two of the surveys were MSc theses,giving us

second-hand analysis of the material without possibilities of exploring interactions with contingency factors such as gender beyond the analyses of the original reports (Bogen and Holmaas,1994;H?ilund,1995).Nevertheless,we found that the results from these MSc theses could be of interest to the safety science community both in terms of sources of hypotheses testing results that appear to con?ict with common beliefs,and as an explorative basis for further and more sophisticated research on the transfer of behaviour between risk arenas.

The hypotheses?t into a model on accident prevention developed by Lund and Aar?(submitted for publication),which places emphasis on human,structural and cultural factors(Fig.1).This model is proposed as a supplement to the KAP-model: Knowledge!Attitude!Practice(Behaviour)due to reviews of research on attitude changing interventions,giving conclusions which are rather pessimistic(Wilde,1993; OECD,1994;Aar?and Rise,1996).Changing of attitudes and behaviours seems to

4J.Lund,J.Hovden/Safety Science41(2003)&–&

follow a more complex pathway than in the simple KAP-model.Behavioural and structural modi?cations might also in?uence workers’attitudes and beliefs towards safety.Social norms and cultural factors regarding safety will be in?uenced,which in turn might change the safety climate(safety culture)in a particular setting. According to the H1hypothesis,these in?uences will also improve the behaviour of the worker at home and during leisure time,resulting in generally safer behaviour. From the theoretical perspectives of social and cognitive psychology(Ross and Nisbett,1991;Kimball and Holyoak,2000),similarities between the situational contexts of work and behaviour outside the workplace may be crucial for the success of knowledge and behavioural transfer between the arenas.

2.Material and methods

This article is based on three similar,but independent studies of safe behaviour and emergency preparedness behaviour in the home and during leisure time among groups of employees.Workers were studied in three Norwegian industries:oil-producing platforms in the North Sea(Bogen and Holmaas,1994),an oil re?nery (H?ilund,1995),and a smelting plant(Lund,2000).All of these industrial settings have potential for catastrophe.They have extensive safety systems and have worked with improving their safety records for a considerable time.Their safety pro-grammes consist of a combination of measures from the three types of prevention measures in Fig.1:attitude,behaviour,and structural modi?cation.All three industries have lowered their LTI(lost time injury)rates substantially.The LTI rates at the oil-producing platforms were reduced by approximately70%and in the smelting plant by approximately50%from1984to1994.The oil re?nery had achieved an LTI rate of zero in1994.

We were involved in the design and implementation of the?rst two studies,and had access to comprehensive and detailed reports with tables,?gures,and statistical analyses.The last study was carried out as part of an evaluation of a3-year inter-vention project commencing in1995,in a smelting plant that was the cornerstone enterprise of a small village in Norway(Lund,2000).The project objective was to reduce the number of occupational accidents in the smelting plant,since the LTI rates had slowly increased after1988.In addition,the project was also directed towards reducing the number of accidents in the home and during leisure time for the smelting plant workers and for the village population.

The three studies were based on self-reported behaviours by means of a questionnaire sent to workers.Employees in industry,municipality administration,and clerical work where safety issues were not a primary focus,were selected as control groups.

At the o?shore oil installations,about1500persons were employed,working in three shifts of2weeks’duration.They normally enjoyed their leisure time during o?-shifts with their families,which were located across Norway but concentrated along the west coast.In1994,workers from one of these shifts were selected as study participants for this investigation,and a questionnaire was distributed through the medical personnel of the platforms.Questionnaires were distributed through members

J.Lund,J.Hovden/Safety Science41(2003)&–&5 of a national union of industrial hygienists across Norway,to a control groupof 2000employees in enterprises that paid little attention to safety.

In1995,about250sta?of an oil re?nery in the county of Vestfold received a slightly modi?ed version of the same questionnaire(see Tables1and2)used in the previous studies.The sta?of210persons in the technical sector of the municipality administration of a nearby city were selected as the control group.

The third study was carried out in a small village at the northern side of Hard-anger fjord on the west coast of Norway,which included a smelting plant situated in the village.The village of304households and a population of about800has devel-oped around the smelting plant established by a multinational company to produce ferro-silicon by exploiting nearby waterfalls for energy.Of the sta?of about230 persons,about150live in the village in132households.In39other households, adults are employed in other industry/clerical occupations.Adults of the remaining 133households receive pensions.In1998,the132households with employees in the smelting plant were selected for this study.The39households with employees in other occupational branches were selected as a control group.All households received by mail a slightly modi?ed version of the same questionnaire used in the previous studies(see Tables1and2).The head of each household was requested to complete the questionnaire.

The2-year intervention programme of home and leisure safety for the workers at the plant consisted of three voluntarily family gatherings at the plant to attend small groupdemonstrations of safety for?re,tra?c,electricity,p oisonings,skiing and sledging,boating and?shing,and?rst aid.Between50%and almost100%of the workers with families attended these gatherings.

Booklets and brochures on home and leisure safety were distributed to households. Safety equipment was distributed as Christmas gifts,o?ered at subsidised prices,and also loaned to the workers.An attempt was made to implement a diary system for registering injuries at home and during leisure time.While the system did not function, mainly due to an underestimation of the e?ort required to keepit running,it p ossibly contributed to an increase in workers’awareness of home and leisure safety.

The2-year intervention programme towards home and leisure safety for the vil-lage population,which included both case and control groups,consisted of infor-mation about home and leisure safety distributed through the local newspaper,and safety meetings for the pensioners,and school and kindergarten teachers and per-sonnel.The pupils of the school distributed home safety checklists and ques-tionnaires to all households and analysed the data collected from the questionnaires.

A local hazards reporting system was established.Over a period of4months about 50hazards were reported,and many were dealt with by the plant and municipality.

A bicycle safety campaign at the school was carried out with bicycle safety checks, pedestrian training,free helmets to all pupils above second grade,and free re?ex vests for the pupils in grades1and2.The plant paid for the safety equipment.

The questionnaire was developed to measure safe and emergency preparedness behaviours at home and during leisure time.Behaviour is regarded as safe when a person follows recommended safety rules in a relatively normal situation,and adjusts their behaviour because of knowledge and consciousness of safety aspects of

Table1

Questions for measuring safe behaviour at home and during leisure time.Questions included in the safe behaviour index are marked with X.Mean values of the answers are given with the standard deviation. (Quite often=5,Often=4,Occasionally=3,Seldom=2,Never=1.Question is not asked:–) Questions for measuring safe behaviour O?shore oil inst.Oil re?nery Melting plant

Case

N=332Control

N=831

Case

N=128

Control

N=85

Case

N=82

Control

N=30

1.Do you keepto the30km/h sp eed X4.15 4.24X3.98 4.20X4.36 4.30 limit for cars?0.940.93 1.010.820.840.82

2.Do you keepto the80km/h sp eed X

3.53 3.72X3.82 3.93X

4.19 4.11 limit for cars?0.960.950.850.840.900.75

3.Do you turn o?the main switch on X3.96

4.15X4.13 4.19X4.41 4.10 the TV every night? 1.46 1.44 1.47 1.46 1.32 1.47 4.Do you always put tools away in a X3.85 3.97X3.99 3.94––safe place after using them at home? 1.03 1.05 1.01 1.17

5.Do you use a lifejacket when boating?X2.69 2.67X2.89 2.88X3.29 3.33

1.29 1.34 1.30 1.30 1.35 1.11

6.Do you warn friends/neighbours if they X3.10 2.86X2.76 2.54X2.96 2.63 are undertaking dangerous activities? 1.05 1.060.88 1.07 1.02 1.13

7.In hotels or on ships etc.do you read X3.93 3.74X4.11 3.61X3.93 3.43?re instructions and check the route to

the emergency exit?

1.08 1.230.93 1.25 1.19 1.33

8.In hotels or on ships etc.do you check X2.82 2.36––––the emergency exit,that it is open,and

leads out of the building?

1.32 1.37

9.When buying new products for the X3.56 3.65X3.78 3.79X3.79 3.68 household,do you check product

safety?

1.03 1.13 1.06 1.12 1.12 1.08

10.Do you pay extra for a safer product?X3.55 3.54X3.69 3.63X3.58 3.73

1.07 1.13 1.01 1.17 1.15 1.05 11.Do you pay extra for an environmentally X

2.93

3.04––––safe product? 1.08 1.28

12.When using a ladder,do you ensure it 3.98 4.18--X4.23 3.65 is securely fastened? 1.01 1.010.89 1.23 13.Do you use a bicycle helmet? 1.54 2.09 2.24 2.21 2.59 2.33

1.17 1.64 1.61 1.69 1.58 1.57 14.Do you use a light on a bicycle when 3.91 4.06 4.04 3.94 4.23 3.72 riding in the dark? 1.39 1.35 1.30 1.37 1.25 1.71 15.Do you smoke in bed? 1.27 1.39 1.33 1.18 1.32 1.00

0.72 1.04 1.030.780.990.00 6J.Lund,J.Hovden/Safety Science41(2003)&–&

the situation in order to avoid accidents.All the questions selected for measuring safe behaviour in the questionnaires are given in Table 1.

The questions selected for measuring emergency preparedness behaviour are given in Table 2.They are related to behaviour that reduces the consequences of a crisis such as an accident or a ?re.Questions 7and 8from Table 1are included here as they are measuring both safe behaviour and emergency preparedness behaviour.They are related to the consequences of relatively rare crisis situations such as a ?re

Table 2

Questions for measuring emergency preparedness behaviour at home and during leisure time.Percentage of groups with answers giving value 1in index (questions 1–10),and mean value of answers given with standard deviation (questions 11–12).(–:question is not asked.)Questions for measuring emergency preparedness behaviour

Answer which gives value 1in index

O?shore oil inst.Oil re?nery

Smelting plant

Case N =332Control N =831Case N =128Control N =85Case N =82Control N =30

1.How many smoke alarms do you have at home?

2,3or more 89%74%a

a

94%80%2.Have they been tested in the last six months?

Yes 92%83%91%

94%

95%93%3.How many ?re extinguishers do you have at home?

2,3or more 72%35%a

a

26%28%4.Have you checked their pressure in the last six months?

Yes

71%49%62%61%69%54%5.Have you turned them upside down in the last six months to loosen the powder?

Yes 51%

30%

38%

49%

81%

37%

6.Have you or your family members checked emergency escape routes in case of ?re?

Yes

68%54%70%60%70%69%

7.Have you tested the escape routes?

Yes 27%20%20%15%33%25%8.Do you have a ?rst aid kit at home?Yes 92%70%80%60%79%71%9.Do you have a ?rst aid kit in your car?

Yes 73%62%69%46%78%61%10.Do you have a ?re extinguisher in your car?

Yes ––19%17%25%21%11.In hotels or on ships,etc.do you b Quite often, 3.93 3.74 4.11 3.61 3.93 3.43read ?re instructions and check the route to the emergency exit?often 1.08 1.230.93 1.25 1.19 1.3312.In hotels or on ships etc.do you Quite often, 2.82 2.36–

–

–

–

check the emergency exit,that it is open,and leads out of the building?b

often

1.32

1.37

a Results not found in report from study.

b

Questions 11and 12are also used for measuring safe behaviour—ref.Table 1.

J.Lund,J.Hovden /Safety Science 41(2003)&–&

7

8J.Lund,J.Hovden/Safety Science41(2003)&–&

in a hotel or a ship,which require consciousness of safety procedures such as checking emergency routes.

In addition to questions on behaviours,the questionnaire collected personal data such as age,gender,education,income,company type,and length of employment,as well as safety and emergency activities in which the interviewee had been involved. Most of the questions and response choices were similar in all of the three studies. The?ve possible alternative choices for answering the questions in Table1were attributed values:never:1;seldom:2;occasionally:3;often:4;quite often:5.For each question,the mean value for the case groupand control groupwas calculated. An index was created to obtain an aggregate measure of safe behaviour for each subject.As some questions had been deleted in the second and third studies,the indexes are constructed somewhat di?erently.The questions marked with an X in Table1are included in the index.Questions12–15(lower part of Table1)were deleted in all studies due to low sample size,except for question12in the smelting plant study where the sample size was considered to be su?ciently large.Accord-ingly,in Table1the index in the?rst study consisted of11questions,while the other two studies contained nine questions.

If the worker answered quite often(5),or often(4)to the relevant questions in Table1,the behaviour of the worker was considered safe,and received a value of1 in the index.If the answers were occasionally(3),seldom(2),never(1),or not given (unknown),the behaviour was considered unsafe,and the worker received a value of 0in the index.For each worker,the sums of the0and1values were divided by the number of the question answered.Not all questions were relevant.For instance some workers did not drive.These workers had two questions less in the index than workers who drove cars.

An index of emergency preparedness behaviour was similarly constructed.All questions in Table2were included in the index,with the exceptions of no.10,which was not asked in the?rst study,and question12,which was asked only in the?rst study.Thus,the index consisted of11questions in all three studies.If the worker’s answer was none or one to questions1and3in Table2,no or not given(unknown) to questions2and4–10,and occasionally,seldom,never or not given(unknown)to questions11and12,the worker received0in the index.Otherwise they received the value of1.

To compare the level of safety and emergency activities of the respondents,they were asked if they had participated in any of the following safety and emergency activities: emergency exercises;

courses/lectures/seminars on safety and/or emergency preparedness;

safety meetings;

safety rounds/inspections;

the establishment of safety goals for the company;and

membershipin voluntary organisations,such as Red Cross.

All the workers at the o?shore oil installations had participated in special safety training courses of2weeks duration.They received questions about this safety

course,repeat courses,and the number of emergency activities they had participated in with regards to:?rst aid teams,evacuation,training with breathing apparatus,and use of lifeboats.

The data were coded and analysed in statistical programmes.Data from the study conducted at the o?shore oil installations was analysed using SAS,while data from the studies conducted at the oil re?nery and the smelting plant were analysed using SPSS.The signi?cance-test for the o?shore oil installation study was made with the Ryan–Einot–Gabriel–Welsch Multiple Test (REGWQ-test),while the Students t -test was used for the oil re?nery and smelting plant studies.

3.Quality of data

In Table 3the response rates and characteristics of the respondents in the case and control groups are given.The response rate varies between 77and 41%.

The response rates of the control groups in the ?rst two studies were rather low (42and 41%).However,we have no indication of systematic biased response patterns.The main reason for the non-responses was lack of possibilities for follow-up of those receiving the questionnaire compared with the third study.

Men dominated the case groups (97,88,and 80%),while the control groups included fewer men (75,76,and 50%—Table 3).In general,women are considered to be more safety conscious than men,and if so,the control groups might be biased towards a higher safety level than the case groups.

Table 3

Response rates and characteristics of the respondents in the di?erent studies

Study

Sample size

Responders,number and %of sample

Males/Females among

respondents,number-%

Age under 45/above 45years%

Years of employment in company studied,mode value Safety and emergency activities,Median value

O?-shore oil installation Case group:500

332–66%

322–97%10–3%

62/38

48%

15–19years

Two weeks safety and repetition courses and,four emergency activities Control group:2000

831–42%

623–75%208–25%62/38

30%

0–4years 2

Oil re?nery

Case group:240

128–53%113–88%15–12%58/4228%

10–14years 4Control group:210

85–41%

65–76%20–24%53/47

30%

0–4years 1

Smelting plant

Case group:132

82–62%63–78%19–22%42/5859%

20years +4Control group:39

30–77%

15–50%15–50%

43/57

30%

20years +

2

J.Lund,J.Hovden /Safety Science 41(2003)&–&

9

10J.Lund,J.Hovden/Safety Science41(2003)&–&

The case groups had been employed considerably longer than the control groups in all the companies studied except for the smelting plant(Table3).The exposure to safety activities over longer periods of employment in the same workplace should result in higher safety levels in the case groups than in the control groups.

When comparing participation in safety and emergency activities(Table3),all three control groups demonstrated low participation rates compared with the three case groups,which all had high rates.

When assessing these possible biases between the groups,we concluded that the case groups and the control groups in all three studies represent two distinctly dif-ferent groups with regard to workplaces with,respectively,a high and a low focus on safety and emergency preparedness issues.

The questionnaire and the indexes were designed to obtain a quantitative measure of safety and emergency preparedness behaviours,in order to place the subjects on a scale from low to high degree according to these characteristics.The questions in Tables1and2were anticipated to be typical and relevant for safety and emergency preparedness behaviours in the population.This last anticipation was con?rmed with the low degree of non-responses.For most of the questions,non-responses were lower than5%.

The internal consistency of the indexes was tested by calculating reliability coe?-cients,which varies between0(no consistency)and1(full consistency).The relia-bility coe?cients for the safe behaviour indexes in the three studies were:0.69,0.60, and0.62,and for the emergency preparedness behaviour indexes:0.66,0.61,and 0.57.The indexes seem to have relatively high internal consistency.

4.Results

4.1.Safe behaviour in home and during leisure time among workers

Table1provides responses to the questions regarding safe behaviour for the three studies.The mean value of workers’responses to the di?erent questions on safe behaviour seems rather similar in case and control groups,both between case and control groups,and between the studies.

The similarities and di?erences among the three studies are more visible when the safe behaviour indexes are calculated.The distributions of the safe behaviour index values for the case and control groups in the three studies are compared in Figs.2–4, while Table4presents the means of the safe behaviour indexes.There are small di?erences between the case groups and control groups in the o?shore oil installa-tions studies and the oil re?nery study.In the smelting plant study,although the di?erence is larger it is not signi?cant.There seems to be no signi?cant di?erence in safe behaviour at home and during leisure time between workers from o?shore oil installations,the oil re?nery,or the smelting plant with high focus on safety,com-pared with workers in workplaces with little focus on safety.The H0hypothesis,no transfer of safe behaviour,cannot be rejected.However,in the smelting plant study we see a signi?cant gender e?ect.

4.2.Safe behaviour at home and during leisure time among men and women

In the o?shore oil installation study,women in the case and control groupcom-bined(N=218),showed signi?cantly safer behaviour(P<0.05)with regard to eight questions in Table1(questions1,2,3,5,9,10,11,and14)than did men(N=945). Men had signi?cantly safer behaviour only on question no.8.There were no sig-ni?cant di?erences between the genders for the remaining questions:4,6,7,12,13, and15.For this study the di?erence in safe behaviour between the women in the case and control groupis not tested,p robably due to the low number of women at the o?shore oil installations(N=10).

In the oil re?nery study,women in the case and control groups combined(N=35) had a signi?cantly safer behaviour with regard to three questions in Table1(1,10, and14)than men(N=177).For the rest of the questions:2,3,4,5,6,7,9,13,and 15,there were no signi?cant di?erences between the genders.The female workers at the oil re?nery(N=15)had safer,but not signi?cant behaviour than the women in the control group(N=20).

A way of assessing the in?uence of gender on safe behaviour is to calculate the correlation ratio R2.The values are expressed in percentages from0to100%,indi-cating increasing in?uence with increasing percentage.In both these studies,R2were reported to be in the range between0.01and5.7%,indicating that gender has a low in?uence on safe behaviour.

In the smelting plant study,there were no signi?cant di?erences in any of the questions on safe behaviour between all women(N=33)and all men(N=78),case and control groups combined.However,when case and control groups were treated separately,we found some interesting di?erences.

Table5shows that men in the case grouphad a signi?cantly higher safe behaviour index than men in the control group.Male workers in the smelting plant seem to have safer behaviour at home and during leisure time than male workers from the same village who are in workplaces with little focus on safety.For the men in this study,the H0hypothesis,no transferral of safe behaviour,is rejected.

For the female workers in the smelting plant,the situation was the opposite.They had a lower but not signi?cant safe behaviour index than female workers in work-places with little focus on safety.A closer examination of the data reveals that two

of the female workers in the smelting plant have a safe behaviour index of0,while

the range for the other 16women is from 0.29to 1.0.One woman did not answer three questions and answered ‘‘Never’’on the three remaining of six relevant ques-tions,while the other woman did not answer three questions and answered ‘‘Never’’on the ?ve remaining of eight relevant questions in Table 1.When these two women are excluded from the sample,the safe behaviour index for the remaining 16women was almost the same as for the female workers from the other workplaces.This exclusion of the two women from analysis increases the di?erences between the means of safe behaviour indexes between the case and control groupin Table 4,from 0.06to 0.08.Also,this increased di?erence is close to signi?cance (P =0.064).For the female workers,the H 0-hypothesis could not be rejected.In order to assess if the alternative hypothesis,H 1:transfer of safe behaviour between arenas,can be accepted,we calculated if the number of women in the sample was high enough.We decided on a signi?cance level of 0.05,a power of 33%,and a relevant

Table 5

Mean safe behaviour index with standard deviation at home and during leisure time for men and women in the smelting plant study.In the lower part of table,results of analysis are after exclusion of two women with safe behaviour index of 0(s.:signi?cant,level 5%,n.s.:not signi?cant)Gender

Case groupControl group Di?erence

between means

N Mean Standard deviation N Mean Standard

deviation Men 63

0.660.21150.530.16s.Women 180.530.27150.610.18n.s.Total a 820.630.23300.570.18n.s Women b 160.600.21150.610.18n.s Total c

80

0.65

0.20

30

0.57

0.18

n.s c

a The gender is unknown for one subject.

b

Excluding two women who answered ‘‘never’’or did not answer the relevant questions in Table 1for constructing the safe behaviour index,i.e.safe behaviour index became 0.c

When two women (outliers)are excluded from the analysis,the mean in the case groupincreases from 0.63to 0.65.The di?erence between the means is still not signi?cant however:P =0.064,just above P =0.05.

Table 4

Mean safe behaviour index with standard deviation at home and during leisure time in three studies (n.s.:not signi?cant,level 5%)Study

Case groupControl group Di?erence

between means

N Mean Standard deviation N Mean Standard

deviation O?shore oil installation 332

0.520.238310.510.21n.s.Oil re?nery 1280.600.21850.580.24n.s.Smelting plant

82

0.63

0.23

30

0.57

0.18

n.s.

J.Lund,J.Hovden /Safety Science 41(2003)&–&

13

14J.Lund,J.Hovden/Safety Science41(2003)&–&

di?erence between the means of safe behaviour indexes of0.1.With a standard deviation of0.21(Table5),the standardised di?erence will be0.48(Altman,1991). According to a nomogram in Altman(1991,p.456),the necessary sample size is40 persons,20persons in each group.We had a smaller number of women in our group,leaving us with a less than33%chance(1in3)to accept the alternative hypothesis.

In summary,gender seems to have little in?uence on safe behaviour.For the workers at the o?shore oil installations and the oil re?nery,there seems to be no signi?cant di?erence in safe behaviour at home and during leisure time compared with workers with low focus on safety.However,the tendency towards safer beha-viour at home and during leisure time for the workers in the smelting plant was almost signi?cant.This tendency might be explained by the intervention project in the smelting plant,which aimed to improve the workers’safe behaviour at home and during leisure time.

4.3.Emergency preparedness behaviour in home and during leisure time among workers

The answers on the questions of emergency preparedness behaviour are given in Table2.

The mean value of the answers to the di?erent questions on emergency prepared-ness behaviour di?ers between case and control groups,while there are similarities between the studies(Table2).

The distribution of the emergency preparedness behaviour indexes of the case and control groups in the three studies are compared in Figs.5–7.Table6includes the mean emergency preparedness behaviour indexes.The means between case and control groups di?er signi?cantly in all three studies.There are di?erences in emer-gency preparedness behaviour in home and during leisure time between workers from o?shore oil installations,the oil re?nery,and the smelting plant,all of which have a high focus on safety,compared with workers in workplaces that have a smaller focus on safety.The H0hypothesis,that there is no transferral of emergency preparedness behaviour,is rejected.The safety and emergency preparedness educa-tion and in?uence at the workplace seems to have positive e?ects on the emergency preparedness behaviour at home and during leisure time.

According to Cohen(1965,1977),cited in Ross and Nisbett(1991),the magnitude of an e?ect should be judged relative to the variability of the measure in question. Thus,by Cohen’s criterion,a di?erence between two means that corresponds to a quarter of a standard deviation of the relevant measure would be deemed small,half a standard deviation would be deemed moderate,and a whole standard deviation would be deemed large.Hence,the di?erence between the case and control groups with regards to emergency preparedness behaviour in the o?shore oil installation study is large,the di?erence in the smelting plant study is moderate,and the di?er-ence in the oil re?nery study is small to moderate with regard to emergency pre-paredness behaviour.These di?erences can partly be explained by the fact that emergency training is much more extensive in the o?shore oil and gas industry.

5.Discussion

While the safety programmes conducted in the working environment of the o?-shore oil installations and the oil re?nery seem to have had little e?ect on the safe behaviour in other arenas of the workers’lives,there is an e?ect on the emergency preparedness behaviour.The signi?cant safer behaviour at home and during leisure time found for the male workers at the smelting plant might be explained by the intervention project conducted during the preceding2years.One of the main aims of that project was to improve safety behaviour at home and during leisure time.

The ?rst two studies showed some tendency for women to register safer behaviour at home and during leisure time than men.This was not the case for the female workers in the smelting factory,even when the two outliers who scored 0on the safe behaviour index were excluded from the analysis.The relatively low safe behaviour index at home and during leisure time for the female workers at the smelting plant might be explained by the fact that the majority of these women did o?ce,cleaning,and catering work,where the in?uence of safety measures was likely to be less than among the production workers.

The social and geographic contexts of the three studies di?er.The o?shore oil installation workers lived in locations scattered across the country,as did the

Table 6

Mean emergency preparedness behaviour index with standard deviation at home and during leisure time in three studies (s:signi?cant,level 5%)Study

Case groupControl group Di?erence

between means

N Mean Standard deviation N Mean Standard

deviation O?shore oil installation 332

0.660.208310.480.21s.Oil re?nery 1280.580.20850.510.21s.Smelting plant

82

0.65

0.16

30

0.52

0.23

s.

J.Lund,J.Hovden/Safety Science41(2003)&–&17 control group.The oil re?nery workers lived in the same urban areas as the control group,whereas the smelting plant,dominating life in a relatively remote and isolated place,implies closer relationships and interdependencies between the two groups. Nevertheless,the results of the three studies are quite similar,which indicates some validity,and the possibility of generalising from the?ndings.Some of the more speci?c di?erences in the results between the three studies may also be explained by contextual factors.For example,the safety programme at the plant and the safe community intervention programme in the smelting plant community,and generally the close links between the plant and this small local community,may have had some e?ects on safe behaviour at home and at leisure time for the workers at the plant.

Transfer in general,and transfer of expertise in particular,is in?uenced by a num-ber of factors.The likelihood that transfer will occur at all is determined by the overall similarity between the training tasks and the transfer tasks(Kimball and Holyoak, 2000).The context for the safety training in a factory might di?er substantially from the context in home and leisure arenas.In a factory there might be information campaigns,rules,peer pressure,and a safety culture,which will in?uence and change the safety behaviour of the workers,so that for example,workers start wearing helmets(Fig.1).However,wearing a helmet at work does not necessarily promote wearing of a helmet while bicycling during leisure time.At home and dur-ing leisure time there may be fewer rules,less consciousness of a safety culture,and less peer pressure.The context with bicycling is di?erent to the situation and the context in the factory.The disposition towards safety in general,and towards wearing of a bicycle helmet in particular,does not necessarily change.The safety behaviour is not necessarily consistent across di?erent situations.

Missing transferral of skills was also found in a study among latchkey children 7–9years old(Peterson,1984).When teaching home safety and survival skills to the children,very little cross-situational generalisations,such as treatments for di?erent injuries,was observed.There was little evidence that the training methods utilised a?ected competent or skillful behaviour either positively or negatively in arenas that had not been directly trained.

According to Kimball and Holyoak(2000,p.119):

the degree of transfer is determined jointly by whether encoding conditions permit abstraction of rules that are su?ciently general to cover both tasks(e.g. as a result of experiencing a variety of examples);whether the two tasks share surface and/or structural components;whether similar processing is used in the tasks;and whether prior experiences a?ects(sic)the perception of the tasks.

The transfer of emergency preparedness behaviour from workplaces with high focus on safety to home and leisure arenas might be explained by the similarity of the tasks,or the situation between a?re in a factory and a?re in the worker’s home. It might also be that the knowledge and consciousness of?res and disasters in the workplace and how to handle them and/or escape have the same structural compo-nents as?res in the worker’s house,or in their car.

18J.Lund,J.Hovden/Safety Science41(2003)&–&

Ross and Nisbett(1991)point to di?erent traditions with regard to explaining behavioural consistency.On the one hand,there is the dominating western intellec-tual and moral tradition that supports the habit of dispositional explanations.This tradition emphasises the causal role of the actor,and attributes actions of di?erent kinds to actors of di?erent kinds.On the other hand,the traditions of behaviourism, Lewinan?eld theory,and most sociological frameworks take a situational perspec-tive;that speci?c responses to speci?c situations often might be very stable over time.Ross and Nisbett claim that situational factors exert e?ects on behaviour that are more potent than we generally recognise.

6.Conclusions and implications for prevention

The?ndings in this study indicate that workers in two of the three studies did not transfer safe behaviour from workplaces with high focus on safety to home and leisure arenas.The transferral of safe behaviour among the men in the smelting plant study might be explained by their exposure to the specially designed inter-vention programme at the plant aimed at increasing safe behaviour at home and during leisure time activities.The main aim of that project was to improve the safety behaviour of the workers at home and during leisure time.The intervention programme in the smelting plant community,with the close links between the plant and its surrounding community,probably also had some e?ect on safe behaviour at home and during leisure time for the workers at the plant living in this community.

The safety consciousness acquired by workers at the workplaces with a high focus on work safety without specially designed home and leisure time intervention pro-jects seems not to be comprehensive or deep enough to in?uence safe behaviour in other arenas.The encoding conditions do not permit abstraction of su?ciently general rules that cover both arenas(Kimball and Holyoak,2000).While the safe behaviour was not transferred,the emergency preparedness behaviour was,indicat-ing that these tasks share surface and/or structural components.

We cannot anticipate that information and counselling on safety issues in general will be transferred to speci?c behaviours,and that safety advice in one arena will increase safety consciousness in general.The lack of transferral of safe behaviour as shown in this paper points to the need to design counselling and education of safe behaviour for speci?c situations.The training should be situation-speci?c.For example,the use of safety equipment should be demonstrated in context,the safety advice should be directed to the situation in a way that the subject will recognise, and the desired behaviour should be understood and repeated.

More studies are required in order to developbetter knowledge of transfer mechanisms.A more sophisticated version of the study design used in this article should be developed to control for confounding factors.However,the items inclu-ded in the questionnaire provide a good start for further exploration and testing of the propagation of safety related knowledge,attitudes,and behaviour between risk arenas.

J.Lund,J.Hovden/Safety Science41(2003)&–&19 Acknowledgements

The authors acknowledge with gratitude the contributions from Live Bogen, Eldbj?rg Holmaas,and Richard H?ilund.They supplied the foundation for this paper with their MSc theses on this subject.We also thank Tor Bjerkedal for his contribution during the preparation of this manuscript.The research was sponsored by the Norwegian Research Council.

References

Aar?,L.E.,Rise,J.,1996.Den menneskelige faktor.Kan ulykker forebygges gjennom holdningspa-virkning?(SF-Report5–96).Norwegian Safety Forum,Oslo.

Altman,D.G.,1991.Practical Statistics for Medical Research.Chapman&Hall,London.

Bogen,L.,Holmaas,E.,1994.‘‘Safety First’’-ogsa i hjem og fritid?MSc Thesis,University of Trondheim (NTNU).

Bird,F.E.,Germain,G.L.,1996.Practical Loss Control Leadership,revised ed.Det Norske Veritas, USA.

Cohen,J.,1965.Some statistical issues in psychological research.In:Wolman,B.B.(Ed.),Handbook of Clinical Psychology.McGraw-Hill,New York.cited in Ross and Nisbett(1991).

Cohen,J.,1977.Statistical Power Analysis for the Behavioural Sciences,revised ed.Academic Press,New York.cited in Ross and Nisbett(1991).

Hale,A.R.,2000.Conditions of occurrence of major and minor accidents.Paper at seminar:‘‘Le risque de de faillance et son contro le par les individes et les organisations.’’6–7October2000.Gif sur Yvette. Hale,A.R.,Hovden,J.,1998.Management and culture:the third age of safety.A review of approaches to organizational aspects of safety,health and environment.In:Feyer,A.-M.,Williamson,A.(Eds.), Occupational Injury:Risk,Prevention and Intervention.Taylor and Francis,London.

Heinrich,H.W.,1931.Industrial Accident Prevention.McGraw Hill,New York.

H?ilund,R.,1995.Endring i holdning til sikkerhet.Vil sikkerhetspa virkning i arbeidslivet pa virke risi-koatferd hjemme og pa fritiden?MSc Thesis,University of Trondheim(NTNU).

Kimball,D.R.,Holyoak,K.J.,2000.Transfer and expertise.In:Tulving,E.,Craik,F.I.M.(Eds.),The Oxford Handbook of Memory.Oxford University Press,Oxford,pp.109–122.

Kjelle n,U.,Boe,K.,Hagen,H.L.,1997.Economic e?ects of implementing internal control of health, safety and environment:a retrospective case study of an aluminium plant.Safety Science27,99–114. Lund,J.,2000.Tryggleik-va rt ansvar.Forebygging av ulykker blant ansatte ved smelteverket Bj?lve-fossen ASA og blant befolkningen i A lvik.Report2000:1.National Institute of Public Health,Oslo. Lund,J.,Aar?,L.E.Accident prevention.Presentation of a model placing emphasis on human,structural and cultural factors(submitted for publication).

Lund,J.,Svanstro m,L.,1995.International Journal for Consumer Safety—an issue on‘‘Safe commu-nities’’.International Journal for Consumer Safety2,59–60.

OECD,1994.Improving Road Safety by Attitude Modi?cation.Report prepared by an OECD scienti?c expert https://www.wendangku.net/doc/f27856475.html,anisation for Economic Co-operation and Development,Paris.

Peterson,L.,1984.Teaching home safety and survival skills to latch-key children:a comparison of two manuals and methods.Journal of Applied Behaviour Analysis17,279–293.

Ross,L.,Nisbett,E.N.,1991.The Person and the Situation.Perspectives of Social Psychology.McGraw Hill,New York.

Webber,R.P.,1987.Safety management:the key to injury prevention in Norwegian industry.Paper at ‘‘Sikkerhetsdagene NTH’’,Trondheim.

Wilde,G.J.S.,1993.E?ects of mass media communications on health and safety habits:an overview over issues and evidence.Addiction88,983–996.