Table 1 Variables included and expected associations

Socio-demographics

Marital status and living situation

We expect people with a partner or those living with others to feel safer and expected an association with Safety [25]. We also expected a positive association with Social participation as co-residents provide a source of social contact and stimulation.

Well-being

Global 7-point self-rated QoL

Since all ASCOT attributes are aspects of QoL, positive associations were expected for each attribute with global self-rated QoL.

GHQ-12 [15–17]

GHQ-12 was used as an indicator of psychological well-being. Research has demonstrated the relationship between psychological and emotional well-being and QoL, so we expected positive associations with all attributes [36, 37].

Health and disability

EQ-5D [13, 14]

There is a close relationship between health and QoL, so we expected positive associations with all attributes [36–39].

ADLs and IADLs

ADLs and IADLs are frequently used as measures of 'need' in social care research since they capture how well the individual functions in their daily life unaided. They do not capture the compensatory activity of social care. However, since these measures capture restrictions in a person's ability to manage independently, we would expect positive associations with Control. We also expected a positive relationship between the personal care ADLs and Personal cleanliness, and the food-related ADLs and IADLs and Food and drink, but with the relationship restricted to those reporting they could manage on their own being more likely to choose the top level.

Control and autonomy

Control and autonomy subscale of CASP-12 [18, 19]

We anticipated a positive association between this measure and Control. We also expected to find differences in subscale scores between the top and second response option for each attribute, reflecting the sense of choice conveyed by the wording of the top level.

Nature of locality and environment

5-point interviewer-rated cleanliness and tidiness of respondent's home

We expected interviewer judgements of cleanliness and tidiness of the home to be positively associated with Accommodation.

4-point self-rated design of home [20]

A poorly-designed home will make it more difficult for optimal care to be provided in the home [40, 41], so we anticipated positive associations with Control, Personal cleanliness, Accommodation and Safety attributes.

Type of area (London, city/large town, suburb of city/large town, small town, rural/village)

We anticipated that area would be associated with Safety, as people living in more urban areas are likely to feel less safe due to fears about crime [25, 42, 43]. We also expected urban areas to be associated with better outcomes in the Occupation and Social participation attributes due to better transport links and amenities.

4-point self-rated accessibility of local area

We expected poor accessibility of the local area to be a barrier to achievement of outcomes in the Control, Occupation and Social participation attributes.

Social contact and support

Frequency of meeting up with friends and relatives, frequency of speaking to neighbours, frequency of speaking to friends and relatives

UCLA 3-item loneliness scale [21].

Measures of contact with people outside of the home were expected to be strongly positively associated with Social participation. We also expected a positive association with Occupation because social activities are a way of occupying one's time. In addition, a positive association with Safety was hypothesised since supportive networks are likely to enhance a person's sense of safety [23, 25].

A negative association with Social participation was expected as loneliness has consistently been shown to be related to social contact [44, 45]. A negative association with Safety is likely due to people who feel lonely being more likely to feel vulnerable [25]. Negative associations with all the other attributes, except Dignity, are likely to be observed as people who are lonely are likely to lack support to achieve good outcomes and loneliness is closely associated with depression [46], which has been shown to be related to poor QoL [37].

Participation

Involvement in organised formal and informal groups and activities in last 12 months, unpaid volunteering in last 12 months

We anticipated a positive association between taking part in groups and volunteering and Occupation, in particular. A positive association with Social participation was also expected as organised groups are a source of social contact.

Service quality

Items capturing the quality of care delivery by care workers (see same care workers, come at suitable times, do the things you want done, arrive on time, in a rush, spend less time than supposed to, informed about changes in your care, global rating of way treated by care workers) [27]

A positive association was expected between these items capturing aspects associated with the quality of the delivery of care by care workers and Dignity since Dignity is included in the measure to capture the effect of the way care is delivered on a person's sense of self-worth. In particular, we expected the global rating of the way the person felt they were treated by the care worker to be associated with Dignity.