The key outcome variables in the regression analysis will consist of composite indicators under each of the three outcomes. The groups of services under each outcome (Frequency of Use, Convenience, and Satisfaction) consisted of similar types of service, and held similar patterns in relation to the predictor variables, which implies that they can be combined to form composite indicators in order to include them in the regression models.
The internal consistency of each group is checked using Cronbach’s alpha. The alpha indicates scale reliability, i.e., how closely related the items are as a group, with coefficients of 0.70 or higher considered indicators of good reliability (Upton & Cook, 2008). Table 6.6 summarises the alpha values for the composite outcome indicators, and the services included.
As discussed in Chapter 4, the SHS is somewhat inconsistent in regard to the services included at each time period. In order to make results from comparable, indicators for the two time periods have to consist of the same services. The datasets provide three indicators that are consistent at both time periods: Frequency of Use of Leisure Services, Convenience of Essential Services, and Satisfaction with Leisure Services. In addition to these, there are two further composite indicators for the later period only as this covered a wider range of services and outcomes. These are: Frequency of Use of Necessities and Satisfaction with Public Services.
With Frequency of Use of Leisure Services, the items included in both datasets cover four services (libraries, parks, museums, sports). Reliability coefficients are quite similar at the two time periods (0.59, and 0.60), and are considered acceptable. The 2009-2011 items under Frequency of Use of Necessities yield an alpha of 0.51, which is lower than recommended. This implies that there is more variation in the Frequency of Use of these services, which could be expected considering the somewhat varying types of services included.
However, the grouping with these services is considered important, as they comprise services that most people require access to locally. Therefore all the items are included in the indicator for Frequency of use of Necessities (Table 6.6).
The items measured in Convenience of Essential Services is formed of seven services (post offices, banks, outpatients, small food shops, doctors, chemists, and public transport). Convenience has the highest internal consistency out of the outcomes, with alpha values of 0.83 in 1999-2002 and 0.82 in 2009-2011. While it could be argued that only the ‘very (in)convenient’ responses truly address respondents’ perceptions of convenience, the full scale is retained for the purpose of forming the indicators of Convenience, as the variation in responses will be reflected in the linear indicator.
Finally, Satisfaction was measured for Leisure Services in both samples and additionally for Public Services in the later years only. A consistent indicator for Leisure Services uses the four services included in both surveys: libraries, parks, museums, and sports. The alpha for Leisure Services is slightly lower in the first time period (a=0.59 compared to the later period, a=0.70) and corresponds to the alpha of Frequency of Use, which consists of the same items. Satisfaction with Public Services further shows good internal consistency with an alpha of 0.70.
Table 6.6 The composite indicators for the service outcomes, Cronbach’s alpha.SHS 1999- 2002, 2009-2011.
Outcome indicators Services included N items 1999- 2002
2009- 2011
Frequency of Use of Leisure Services
Libraries, parks, museums, sports 4 0.59 0.60
Frequency of Use of Necessities
Post offices, banks, cash machines, doctors, dentists, grocery/food shops, chemists, outpatients, petrol stations, public transport
8 0.51
Convenience of Essential Services
Post offices, banks, outpatients, small food shops, doctors, chemists, public transport
7 0.83 0.82
Satisfaction with Leisure Services
Libraries, parks, museums, sports 4 0.59 0.70
Satisfaction with Public Services
Health, police, fire, refuse collection, schools, social care, public transport, street cleaning
8 0.70 0.70
To make the composite indicator, the original response categories of the variables are used as detailed in Table 6.7. The composite indicator is formed by taking the average of each respondent’s values ignoring missing responses to any item. In this way, missing responses do not skew the resulting scale. All the indicators in further sections are formed in this way. This provides scales as shown in Table 6.7. Figures 6.13 and 6.14 present the distributions of responses in each scale at both time periods respectively. Frequency of Use of Leisure Services tends to have more responses in the lower end of the scale, indicating lower frequency of use, while responses for Frequency of Use of Necessities are more concentrated around the middle. The scales for Convenience and Satisfaction have higher means compared to Frequency (as shown in Table 6.7), indicating higher average levels, which is reflected in the tendency to higher values in their distributions (Figs. 6.13, 6.14). Satisfaction with Leisure Services stands out with very little variation, with peaks in its distribution at the high end (Figs. 6.13, 6.14).
Table 6.7 Descriptive statistics of the composite outcome indicators.SHS 1999-2002 and 2009- 2011.
1999-2002 Indicator N Mean Std. Dev. Min Max
Frequency of Use of Leisure Services 27,804 3.00 1.20 1 7 Convenience of Essential Services 27,896 3.96 0.83 1 5 Satisfaction with Leisure Services 20,495 4.26 0.74 1 5
2009-2011 Indicator
Frequency of Use of Leisure Services 29,992 2.35 0.84 1 6 Frequency of Use of Necessities 30,044 3.84 0.71 1 7 Convenience of Essential Services 30,029 4.02 0.78 1 5 Satisfaction with Leisure Services 26,118 4.08 0.73 1 5 Satisfaction with Public Services 29,953 4.00 0.61 1 5