SURVEY OF PARTICIPATION IN, AND ATTITUDES TOWARDS, GAMBLING: KEY FINDINGS FROM THE 2004 NOP SURVEY

SURVEY OF PARTICIPATION IN, AND ATTITUDES TOWARDS, GAMBLING: KEY FINDINGS FROM THE 2004 NOP SURVEY By Stephen Creigh-Tyte and John Lepper Technical Paper No 4 April 2004

Professor Stephen Creigh-Tyte, Chief Economist and Head of Research, Department for Culture, Media and Sport. John Lepper, Economic Adviser, Department for Culture, Media and Sport Page 1 of 15

INTRODUCTION 1. DCMS commissioned National Opinion Polls (NOP) to conduct a short survey into gambling participation and some attitudes towards gambling. A quota sample of 2,945 UK adults over 18 years of age were asked four questions between 5 th and 17 th February 2004 as part of the NOP Omnibus Survey. 1 2. Since 1999 participation in most forms of gambling has not increased and may have declined. The exceptions are bingo, betting on events at bookmakers and betting on dog races. On balance, with the exceptions of lotteries and bingo, all forms of gambling are regarded unfavourably. Very few respondents support less strict control of gambling and in the cases of fruit machines and the internet greater strictness of control is strongly supported. These attitudes do not appear vary significantly with sex, income, age or socio-economic group. PARTICIPATION Overall Results 3. The participation in gambling questions were designed to replicate as closely as possible questions asked in the British Gambling Prevalence Survey conducted in 1999 (Sproston et al, 2000). Respondents were asked whether or not they had gambled in the last 12 months or the previous 7 days. The results are shown in Tables 1 and 2. 4. In February 2004, 71% of those interviewed said that they had gambled during the previous 12 months. This is significantly lower than the 73% overall rate of participation in gambling found in 1999. 5. In the past five years, participation in most forms of gambling has fallen. The exceptions are playing bingo and betting on non-race events with bookmakers both of which have risen. There have been particularly noticeable falls in participation in the National Lottery, buying scratchcards, playing fruit machines and football pools and betting with friends and colleagues. 6. A similar pattern was found among those who gambled in the past 7 days. Most forms of gambling showed falling participation although bingo, betting on horse and dog races and betting on events with a bookmaker recorded no change in the past 5 years. 1 The Questions are in Appendix 1. The survey methodology is outlined in Appendix 2. Page 2 of 15

TABLE 1 PARTICIPATION IN FORMS OF GAMBLING IN PAST 12 MONTHS (% of total sample and % change 1999 to 2004) NOP Survey February 2004 National Prevalence Survey September 1999 % change 1999 to 2004 National Lottery 61 65-4* Other lotteries 7 8-1 Scratchcards 16 22-6* Football Pools 5 9-4* Bingo 9 7 +2* Fruit Machines 9 14-5* Private Bets 7 11-4* Horse Races 11 13-2* Dog Races 4 4 0 Events at a bookmaker 4 3 +1* Betting Exchange neg n.a. n.a. Table games at a 2 3-1* casino Internet 1 neg n.a. Other 1 neg n.a. Indicates statistically significant at 5%. neg: negligible. n.a.: not applicable. TABLE 2 PARTICIPATION IN FORMS OF GAMBLING IN PAST 7 DAYS (% of total sample and % change 1999 to 2004) NOP Survey February 2004 National Prevalence Survey September 1999 % change 1999 to 2004 National Lottery 41 47-6* Other lotteries 2 4-2* Scratchcards 5 8-3* Football Pools 3 6-3* Bingo 4 4 0 Fruit Machines 3 6-3* Private Bets 2 4-2* Horse Races 3 3 0 Dog Races 1 1 0 Events at a bookmaker 1 1 0 Betting Exchange neg n.a. n.a. Table games at a casino neg neg n.a. Internet neg neg n.a. Other neg neg n.a. Indicates statistically significant at 5%. neg: negligible. n.a.: not applicable. Page 3 of 15

Regularity 7. Taken together, these data suggest that regularity of gambling (meaning at least weekly gambling) may have fallen for many forms of gambling (see Table 3). Gamblers in the past 7 days as a proportion of all gamblers has fallen for all but two forms of gambling where the ratio can be measured. In the case of horse racing, the regularity has risen and with dog racing it has remained constant. 8. This is not definitive evidence of falling regularity of gambling. The period chosen for the survey may have been an unrepresentative gambling period. Or again, the sample taken for the present study is much smaller than that taken for the 1999 British Gambling Prevalence Survey (which achieved a sample of 6,503) and may have not been sufficiently representative of regular gamblers. Finally, respondents may be unwilling to reveal recent gambling activities. TABLE 3 PROPORTION OF PARTICIPATION IN PAST 12 MONTHS WHICH OCCURRED IN LAST 7 DAYS (%) 2004 1999 Change 1999 to 2004 National Lottery 67.2 72.3-5.1 Other Lottery 28.6 50.0-21.4 Scratchcards 31.3 36.4-5.1 Football Pools 60.0 66.7-6.7 Bingo 44.4 57.1-12.7 Fruit Machines 33.3 42.9-9.6 Private Bets 28.6 36.4-7.8 Horse Races 27.3 23.1 +4.2 Dog Races 25.0 25.0 0 Events with a Bookmaker 25.0 33.3-8.3 ATTITUDES 9. As shown in Table 4, in most cases a majority have neither favourable not unfavourable attitudes to gambling. Only in the cases of fruit machines and internet gambling do more people have an unfavourable attitude than have a neither favourable nor unfavourable attitude. Moreover, more people have a favourable attitude to lotteries than have a neither favourable nor unfavourable attitude. 10. The results from Table 4 can be re-worked as follows. Very unfavourable and unfavourable responses are summed together and compared to the sum of very favourable and favourable responses. The two sums are scaled for don t knows and no opinions and the difference between the two sums taken (see Table 5). Page 4 of 15

11. Lotteries and bingo are regarded favourably on balance and the most negative opinions are reserved for internet gambling, fruit machines, casinos and betting exchanges. However, the betting exchanges result is accompanied by 37% Don t Know / No Opinion and so may be based on lack of information about the actual practice of betting exchanges. TABLE 4 ATTITUDES TOWARDS TYPES OF GAMBLING (% of total row responses; rows may not sum to 100% due to rounding) Very unfavourable Unfavourable Neither favourable nor unfavourable Favourable Very Favourable Don t know/ No opinion Lotteries 6 12 24 42 6 10 Scratchcards 12 24 28 21 2 14 Betting Offices 13 23 30 14 1 18 On-course betting 11 19 28 26 5 17 Bingo 9 15 28 26 5 17 Fruit Machines 18 29 24 12 2 15 Private Bets 12 22 32 12 1 21 Casinos 14 23 30 12 1 24 Internet gambling 22 26 23 4 1 24 Betting Exchanges 13 19 26 4 neg 37 TABLE 5 BALANCE OF OPINION TOWARDS TYPES OF GAMBLING (% scaled for don t knows / no opinions) Very unfavourable + unfavourable A Very favourable + favourable B Balance - A + B Lotteries 20 53 +33 Bingo 17 37 +20 On-course betting 38 27-11 Scratchcards 42 27-15 Betting offices 44 18-26 Private bets 43 16-27 Casinos 46 16-30 Fruit Machines 55 16-39 Betting Exchanges 51 6-45 Internet 63 7-56 Page 5 of 15

REGULATION OF GAMBLING 12. Most respondents feel that the regulation of most modes of gambling is about right. There are also a significant number who believe that a stricter system of control should be put in place. That opinion is particularly clear in the cases of fruit machines and internet gambling. Very few (i.e. well under 5% of the population) favour a regulatory regime that is less strict towards any particular form of gambling (see Table 6). TABLE 6 DEGREE OF REGULATORY CONTROL (% of responses by row; rows may not sum to 100 due to rounding) More strict About Right Less Strict No Opini on Lotteries 16 61 2 21 Scratchcards 24 49 2 26 Betting Offices 17 49 1 34 On-course betting 14 46 1 39 Bingo 8 59 3 30 Fruit machines 35 35 2 29 Private bets 15 36 2 47 Casinos 20 40 2 37 Internet 36 20 1 43 Betting exchanges 17 24 1 58 13. The validity of this part of the survey is open to doubt in view of the relatively large No Opinion response. This presumably reflects a lack of knowledge of gambling regulation in the population at large. 14. By and large, people do not want the control of gambling to be eased even in the forms of gambling (notably lotteries and bingo) which are looked upon with most favour. The satisfaction with present regulation is clearest in the cases of lotteries and bingo and least expressed in the cases of internet gambling and betting exchanges. Table 7 tabulates the results in order of about right responses to current regulation. 15. In 2001, the Home Office Citizenship Survey (Home Office, 2001) found that 55% of those questioned thought that gambling regulation was about right. This appears to be generally consistent with the results reported in Table 7. 16. In addition, in 2001, there was a marked balance of opinion stating that regulation was too loose. 19% stated that regulation of gambling was too loose and only 6% felt it was too tight. 20% did not express an opinion. Page 6 of 15

TABLE 7 SATISFACTION WITH PRESENT REGULATION % Answering About Right Lotteries 61 Bingo 59 Scratchcards 49 Betting offices 49 On-course betting 46 Casinos 40 Private bets 36 Fruit Machines 35 Betting exchanges 24 Internet gambling 20 17. When the 2004 data are reworked a similar balance opinion against looser regulation is found. Indeed, there is a balance of opinion in favour of more strict control for each form of gambling surveyed. This pattern is least evident in the case of bingo and seen most clearly in the cases of internet gambling and fruit machines. The full results are summarised in Table 8. 18. The strength of opinion in favour of stricter control of the internet and fruit machines is contrary to the relatively low participation rates in both forms of gambling and hence may not be based on the experiences of respondents. TABLE 8 BALANCE OF OPINION ON STRICTNESS OF CONTROL (% scaled for no opinions) % More strict Less strict Internet +61 Fruit machines +46 Betting exchanges +38 Scratchcards +30 Casinos +29 Betting offices +27 Private bets +25 On-course betting +23 Lotteries +18 Bingo +7 19. There do not appear to be significant differences between social groups in this pattern except perhaps in the case of socio-economic groups. Even here the differences are marginal. For all gambling modes, except the internet, groups D and E are more in favour of stricter control than other groups. This pattern is particularly pronounced in the case of casinos. The exception in the case of the internet may be due to relatively low access to the internet among people in socio-economic groups D and E. Page 7 of 15

REFERENCES Home Office (2001); Home Office Citizenship Survey: People, Families and Communities, Home Office, London, 2001. Sproston Kerry, Erens Bob and Orford Jim (2000); Gambling Behaviour in Britain: Results from the British Gambling Prevalence Survey, National Centre for Social Research, London, June 2000. Page 8 of 15

APPENDIX 1 QUESTIONS FOR GAMBLING QUESTIONNAIRE Have you spent your own money on any of the following activities in the last 12 months? Tickets in the National Lottery Draw (including Thunderball) Do not include scratchcards Tickets for any other lottery Do not include scratchcards Scratchcards Do not include newspaper scratchcards Football pools or fixed odds coupons Bingo tickets Fruit machines Private bets or playing games with friends and colleagues Do not include betting with a bookmaker Betting on horse races Do not include private bets with friends or colleagues Betting on dog races Do not include private bets with friends or colleagues Betting on other events with a bookmaker like 49s, sports events or election outcomes Do not include private bets with friends or colleagues Betting on a betting exchange Table games (roulette, cards or dice) in a casino Betting on the Internet Another form of gambling for money ( please specify).. Have you spent your own money on any of the following activities in the last 7 days? Tickets in the National Lottery Draw (including Thunderball) Do not include scratchcards Tickets for any other lottery Do not include scratchcards Scratchcards Do not include newspaper scratchcards Football pools or fixed odds coupons Bingo tickets Fruit machines Private bets or playing games with friends and colleagues Do not include betting with a bookmaker Betting on horse races Do not include private bets with friends or colleagues Betting on dog races Do not include private bets with friends or colleagues Betting on other events with a bookmaker like 49s, sports events or Election outcomes Do not include private bets with friends or colleagues Page 9 of 15

Betting on a betting exchange Table games (roulette, cards or dice) in a casino Betting on the Internet Another form of gambling for money ( please specify).. What is your attitude to various forms of gambling? Is it generally very favourable or favourable or neither favourable nor unfavourable or unfavourable or generally very unfavourable or do you have no opinion? Please score your opinion on a scale from 1 for very favourable to 5 for very unfavourable or say if you have no opinion. 1 2 3 4 5 Very Neither favourable Very No favourable Favourable nor unfavourable Unfavourable unfavourable opinion Lotteries Scratchcards Betting at Betting Offices On-course betting Bingo Fruit machines Private bets Casinos Internet gambling Betting exchanges Irrespective of your personal attitude to gambling do you think the control of different forms of gambling should be more or less strict or is the control about right or do you have no opinion? Lotteries Scratchcards Betting at Betting Offices On-course betting Bingo Fruit machines Private bets Casinos Internet gambling Betting exchanges More Strict About Right Less Strict No Opinion Page 10 of 15

APPENDIX 2 STUDY METHODOLOGY NOP RANDOM LOCATION OMNIBUS SAMPLE DESIGN The NOP Random Location Omnibus employs a quota sample of individuals with randomly selected sampling points. The sample design is essentially a 3-stage design, sampling first parliamentary constituencies, then output areas within those selected constituencies and finally respondents within the output areas. The sample is based on 175 sampling points. The selection of Parliamentary Constituencies The first-stage sampling units for the survey are parliamentary constituencies, selected in the following way. The 641* parliamentary constituencies of Great Britain are classified into the Register General's ten Standard Regions. In Scotland, a further classification was by the new Strathclyde Region and the rest of Scotland. In Wales, the South East was classified separately from the rest of Wales. Within each Standard Region, constituencies are classified into four urban/rural types as follows: 1. Metropolitan county Those constituencies which lie completely within the area of the eight Metropolitan Counties of Great Britain. It is appreciated that such areas now technically do not exist but they are still convenient building blocks for sample design. In the case of the North West Standard Region, which contains two Metropolitan Counties, the constituencies of the Greater Manchester MC were classified and listed separately from those of the Merseyside MC. Similarly, for the Yorkshire and Humberside Standard Region, the constituencies of the South Yorkshire MC were listed separately from those of the West Yorkshire MC. In Greater London, constituencies north of the river Thames were listed separately from those south of the river. These were further sub-divided into east and west for each side of the river. * For practical reasons, two constituencies (Orkney and Shetland, and Western Isles) are not included in the sampling frame from which constituencies are selected. 2. Other 100% Urban All urban constituencies, other than Metropolitan County constituencies, in which the population density was greater than 7 persons per hectare. 3. Mixed Urban/Rural Constituencies, consisting of a mixture of urban and rural local authority areas, in which the population was greater than 1.5 and less than 7 persons per hectare. 4. Rural Constituencies, consisting of a mixture of urban and rural local authority areas, in which the population density was less than 1.5 persons per hectare. Within each of the resultant 46 cells, as a final stratification, constituencies are listed in order of the percentage of people resident in households whose head is in socio-economic Groups 1, 2, 3, 4 or 13 (approximates to Social Grades A&B). Page 11 of 15

When all the constituencies have been listed in the above way, the electorate of each constituency is entered on the list and a cumulative total of electors by constituency is formed. The selection is done in the following way. From the file of 639 constituencies, a sample of 175 must be drawn. To draw this sample, the following procedure is undertaken. The total number of cumulative electors (N) on the list is divided by 175 and a random number between 1 and N/175 is selected. This random number identifies an elector, in the cumulative total of electors, and the constituency this elector is in becomes the first selected constituency in the sample. To obtain the other 174 constituencies, the sampling interval N/175 is added on 174 times to the initial random number. This produces 175 cells all containing N/175 electors. Within each cell a random number between 1 and N/175 is selected. This random number identifies an elector, in the cumulative total of electors for that cell, and the constituency this elector is in is selected. This procedure is repeated for all 175 cells. Thus a sample of 175 constituencies is produced. The Selection of Output Areas Within each selected constituency, an output area is selected for each wave of the Omnibus. These output areas are selected at random, but with some stratification control so that the sample of areas drawn is representative of the sample of constituencies and therefore of Great Britain in demographic terms. The variables used for stratification are essentially age, sex, social class, and geodemographic profile (Mosaic classification). Once the areas have been selected, the profile of the aggregated set of areas is checked against the national profile to ensure that is representative. Each area is a small area, containing in average around 150 households. Each output area is therefore homogenous, with the people living within it being fairly similar in social grade terms. Therefore, when quotas are set for interviewing within each output area, the variables we control for are age and sex within working status. No quota is set for social grade, as the selection of areas ensures that the sample is balanced in this respect. This procedure is repeated for each wave of the Omnibus, producing a different sample of areas for each week of fieldwork. The Selection of respondents For each selected output area, a list of all residential addresses is produced. This listing is taken from the Postal Address File, which is a listing of all addresses within Great Britain, and is updated monthly. The interviewer uses this list to identify the households at which they can interview. Overleaf is an example of a typical output area address listing. 12 people are interviewed within each area. Page 12 of 15

SAMPLE POINT WAVE 1 100 LU5 6AR LEIGHTON 81 :83 :85 :91 :97 :101 :110 :112 :114 :116 :117 :121 LU5 6AR LEIGHTON COBBLERS HALL 119 121 LU5 6AS LEIGHTON THE PADDOCKS LU5 6BB MEADOW 1 :2 :3 :4 :5 :6 :7 :8 :9 :10 :11 :12 :14 :15 :16 :17 :18 :19 :20 LU5 6BB MEADOW 21 :22 :23 :24 :26 :28 :30 :32 :34 :36 :7A LU5 6BD FRENCHMANS CLOSE,TODDINGTON,DUNSTABLE,BEDS 1 :2 :3 :4 :5 :6 :7 :8 :9 :10 :11 :12 :14 :15 :16 :17 :18 :20 :21 LU5 6BD FRENCHMANS 22 :23 :24 :25 :26 :27 :29 :31 :33 :35 :37 :39 :41 :43 :45 CLOSE,TODDINGTON,DUNSTABLE,BEDS LU5 6DA RUSSELL 1 :2 :3 :4 :5 :6 :7 :8 :9 :10 :11 :12 :14 :15 :16 :17 :18 :19 :21 LU5 6DA RUSSELL 22 :23 :24 :25 :26 :27 :28 :29 :30 :31 :32 :33 :34 :35 :36 :37 :39 LU5 6DB LAKEFIELD AVENUE,TODDINGTON,DUNSTABLE,BEDS 1 :2 :3 :4 :5 :6 :7 :8 :9 :10 :11 :12 :14 :15 :16 :17 :18 :19 :20 LU5 6DB LAKEFIELD AVENUE,TODDINGTON,DUNSTABLE,BEDS 21 :22 :23 :24 :25 :26 :27 :28 :29 :30 :31 :32 :33 :34 :35 :36 :37 LU5 6DB LAKEFIELD AVENUE,TODDINGTON,DUNSTABLE,BEDS 38 :39 :40 :41 :42 :43 :44 :45 :46 :47 :48 :49 LU5 6EX BRYANT WAY,TODDINGTON,DUNSTABLE,BEDS 2 :3 :4 :5 :6 :7 :8 :9 :10 :11 :12 :14 :15 :16 :17 :18 :19 :20 :21 LU5 6EX BRYANT WAY,TODDINGTON,DUNSTABLE,BEDS 22 :23 :24 :25 :26 :27 :28 :29 :30 :31 :32 :33 :34 :35 :36 :37 LU5 6EX BRYANT WAY,TODDINGTON,DUNSTABLE,BEDS 23A LU5 6EY BRYANT WAY,TODDINGTON,DUNSTABLE,BEDS 38 :39 :40 :41 :42 :43 :44 :45 :46 :47 :48 :49 :50 :51 :52 :53 :54 LU5 6EY BRYANT WAY,TODDINGTON,DUNSTABLE,BEDS 55 :56 :57 :58 :59 :60 :61 :62 :63 :64 :65 LU5 6FE RANDALL DRIVE,TODDINGTON,DUNSTABLE,BEDS 1 :2 :3 :4 :5 :6 :7 :8 :9 :10 :11 :12 :14 :15 :16 :17 :18 :19 :20 LU5 6FE RANDALL DRIVE,TODDINGTON,DUNSTABLE,BEDS 21 :22 :23 :24 :25 :26 :27 :28 :29 :30 :31 :32 :33 :34 :35 :36 :37 LU5 6FE RANDALL DRIVE,TODDINGTON,DUNSTABLE,BEDS 38 :39 :40 :41 :42 :43 :44 :45 :46 :47 :48 :49 :50 :51 :52 :53 :54 LU5 6FE RANDALL 55:56:00 DRIVE,TODDINGTON,DUNSTABLE,BEDS LU5 6JA HOLMFIELD 1 :2 :3 :4 :5 :6 :7 :8 CLOSE,TODDINGTON,DUNSTABLE,BEDS LU5 6JB PEARTREE 1 :2 :3 :4 :5 :6 :7 :8 :9 :10 :11 :12 :14 :15 CLOSE,TODDINGTON,DUNSTABLE,BEDS LU5 6JD FRENCHMANS CLOSE,TODDINGTON,DUNSTABLE,BEDS 47 :49 :51 :53 :55 :57 :59 :61 :63 :65 :67 :69 :71 :73 :75 :77 LU5 6QE RUSSELL 41 :43 :45 :47 :49 :51 :53 :55 :57 :59 :61 :63 :65 :67 :69 :71 :73 LU5 6QE RUSSELL 75 :77 LU5 6QF RUSSELL 38 :40 :42 :44 :46 :48 :50 :52 :54 :56 :58 :60 :62 :64 :66 :68 :70 LU5 6QF RUSSELL 72 :74 :76 :78 :80 :82 :84 :86 :88 :90 :92 :94 :96 LU5 6QP THE HYDE,TODDINGTON,DUNSTABLE,BEDS 1 :2 :3 :4 In addition to the address listing for an output area, the interviewer is also given a quota sheet, which determines what sort of people they must interview. Each interviewer must interview 12 people within an output area, and the quotas are different for each area in order to reflect the demographic profile of that area. Overleaf is an example of a quota sheet. The quotas are set in terms of age and sex within working status. No quota is set for social class, as the selection of output areas ensures that the sample is balanced in this respect. Page 13 of 15

QUOTA SHEET Constituency Constituency number BEDFORDSHIRE MID 2152A47 Sample code 100 SET 15-34 5 ACHIEVED 35-54 4 55+ 3 Men full-time 3 Men not full-time 2 Women working 3 Women not working 4 POST-SURVEY WEIGHTING Given that the sample is controlled by quotas, the final demographic profile should be fairly close to that of the target population. However, the sample will be examined at each Omnibus wave to ensure that the profile is as it should be. The sample will, if necessary, be weighted in order to ensure that it is representative in terms of known population data on age, sex, social class and region, reproduced below. Page 14 of 15

Age/Sex % Class % 15-24 Male 7.7 A 2.4 25-34 Male 8.5 B 15.7 34-44 Male 9.2 C1 28.6 45-54 Male 7.9 C2 21.3 55-59 Male 3.8 D 15.1 60-64 Male 2.9 E 16.9 65-70 Male 3.1 71+ Male 5.2 15-24 Female 7.5 25-34 Female 8.6 35-44 Female 9.4 45-54 Female 8.0 55-59 Female 3.7 60-64 Female 3.1 65-70 Female 3.4 71+ Female 8.0 Standard Region % TV Net Region % North 5.3 London 20.0 Yorkshire & Humberside 8.6 Midlands 16.1 East Midlands 7.3 North West 11.6 East Anglia 3.9 Yorkshire 10.2 GLC 12.8 Central Scotland 5.9 South East exc. GLC 19.5 Wales & West 8.1 South West 8.7 South & South East 9.5 West Midlands 9.1 North East 4.8 North West 10.8 East 7.5 Wales 5.1 South West 2.9 Scotland 8.9 Border 1.2 North Scotland 2.2 Sources:- Standard Region - Registrar Generals Age/Sex - Registrar Generals Social Class - NOP Random Surveys ITV Region - BARB Page 15 of 15