Duke Social Support Index (11 Item)
Duke Social Support Index (11 Item)
Duke Social Support Index (11 Item)
Note: DSSI_11 which is not used in derived variables has Index Number FAMF-048
Background
It is generally agreed that social networks and support are important for older people.1 Social
networks consist of family and friends and tend to be larger for older women than for older men.2
As well as instrumental support, these networks provide friendships, confidence and expressive
support that make older women feel supported emotionally.
The DSSI
The DSSI was developed in the United States as a brief, easily administered instrument to
determine an individual’s level of social support.9 Details of wording, response options and scoring
for the scale were not included in the original publication, but were supplied to ALSWH by Koenig
(personal communication, Brendan Goodger, 10 October 1996). There were no instructions
concerning missing items.
The DSSI’s reliability and validity have been confirmed in a sample of community dwelling older
Australian men and women.10 Two sub-scales, social interaction and satisfaction with social
support (subjective support), were identified.
Source items
The form of the DSSI items used in ALSWH survey is shown below.
Social interaction sub-scale
DSSI_1 Other than members of your family how many persons in your local area do you feel
you can depend on or feel very close to?
Code Response
1 None
2 1-2 people
3 More than 2 people
DSSI_2 How many times during the past week did you spend time with someone who does not
live with you, that is, you went to see them or they came to visit you or you went out
together?
DSSI_3 How many times did you talk to someone (friends, relatives or others) on the telephone
in the past week (either they called you, or you called them)?
DSSI_4 About how often did you go to meetings of clubs, religious meetings, or other groups
that you belong to in the past week?
Code Response
1 Hardly ever
2 Some of the time
3 Most of the time
DSSI_11 How satisfied are you with the kinds of relationships you have with your family and
friends?
Code Response
1 Very dissatisfied
2 Somewhat dissatisfied
3 Satisfied
Scale Evaluation
The DSSI was included in the first survey of the Older cohort of the ALSWH.
Item responses
Of the 12 939 older women who returned Survey 1, 12 223 (94.5%) completed all the DSSI items
and another 347 (2.7%) missed one or two items. Percentages who gave the highest scored
response for each of the DSSI items are shown in Table 1. Despite the majority of women being
‘satisfied’ most of the time with their relationships with family and friends, almost 14% answered
‘very dissatisfied’ to the item about satisfaction with “kinds of relationships”. This was contrary to
the pattern of responses for the other ‘satisfaction’ items, where endorsement of the lowest
response was very low (1% to 6%). Many felt they could depend on at least two other people or
spent time with at least two people in the past week, although few women (2% to 3%) went to
religious meetings, clubs or other meetings more than five times a week. Responses to individual
items were missing for between 1% and 3% of respondents.
Internal reliability
The overall sampling adequacy for the two factors was good (Kaiser’s measure of sampling
adequacy = 0.85). Cronbach’s alpha coefficients of 0.80 for the ‘satisfaction’ factor, and 0.58 for
‘interaction’ factor and 0.76 the 10 DSSI items indicated reasonable internal reliability.
Factor Analysis
Exploratory factor analyses using the principal components method with orthogonal (varimax) and
oblique rotations (promax) were performed on the 11 DSSI items. Both rotation methods identified
the same two factors. Based on the varimax rotation, 34% of the variance was explained by the
first factor and a total of 48% of variance was explained by both factors (oblique rotation data not
shown). The last item, ‘How satisfied are you with the kinds of relationships you have with your
family and friends?’ was excluded from the final factor analyses as it did not load on either factor
and had low communality (0.01). The first factor related to satisfaction with social support and
included six items. The second factor consisted of four items relating to social interaction. The
communalities for the first two factors ranged from 0.4 to 0.6 (Table 2), demonstrating that the
shared variance between individual items and all the other items was reasonable.
Derived Variables
Scores
Summed scores were obtained by adding the response scores for the items that loaded together
for each of the factors. Factor scores were calculated as the weighted sum of scores for all items;
weights were the scoring coefficients from the factor analysis with varimax rotation.
Correlations between the factor scores and summed scores were high for ‘satisfaction’ (0.89) and
‘interaction’ (0.98). Hence, summed scores were used in the remaining analyses.
Missing items
Mean imputation was used to replace up to two missing items in the 10-item DSSI; no missing
items were imputed in the two sub-scales.
Construct validity
Summed scores for the sub-scales and the overall DSSI were compared with variables thought to
be related to DSSI. Positive correlations were hypothesised with PCS, MCS, the eight sub-scales
of the SF-36, and life satisfaction, a negative correlation with mean stress, and no association with
body mass index. Differences in mean DSSI scores for different socio-demographic and health
related groups were estimated, taking the over-sampling in rural and remote areas into account,
using the least squares mean option in the general linear model procedure in SAS. The level of
significance was set to 0.005 to reduce the effects of multiple comparisons and the large sample
size.
The construct validity of the 10-item DSSI and the ‘satisfaction’ and ‘interaction’ sub-scales was
supported by significant correlations between the DSSI and its sub-scales, ‘satisfaction’ and
‘interaction’, and stress, life satisfaction, physical and mental health (Table 3, all p-values <0.0001
except for BMI where all p-values were >0.05).
Women with higher social support tended to be Australian born, be widowed or married, have more
education and found it easier to manage on their available income (Table 4; p<0.0001 for all
differences).
In terms of health, they rated their general health more highly, had fewer chronic conditions and
symptoms, took fewer medicines, did not need help with daily tasks and had fewer visits to GPs
(Table 5; p<0.0001 for all differences except ‘need help’, p=0.0009). Social support was positively
correlated with satisfaction with GP (0.21). Women had higher social support if, in the last year,
they had no major illness or injury, there had been no major decline in health, or death, of their
spouse and no major conflict with their children (Table 6; p<0.0001 for all differences except ‘death
of spouse’, p=0.0009).