52
,
Approach
Key Drivers Analysis (KDA) is a multivariate technique that has been used to identify how strongly
attitudes and behaviour towards charities are associated with overall trust and confidence in
charities. It is arguably a more ‘objective’ measure of what drives overall trust and confidence as it
examines a range of responses that people give to a number of questions throughout the survey
rather than relying simply on what people
say
is most important to them when asked directly. A
good example of this working in practice is in the context of staff satisfaction surveys, where
employees often cite pay as most important to them when asked directly, but KDA can reveal that
other factors, such as finding their day-to-day work interesting and varied for example, are in fact
more strongly associated with overall job satisfaction.
KDA performed on the 2005 trust and public confidence survey showed that overall trust and
confidence was primarily explained by five key beliefs in how charities operate, namely: the belief
that charities spend their money wisely and effectively; are well managed; ensure that a
reasonable proportion of donations make it to the end cause; make a positive difference to the
cause they work for; and ensure that fundraisers are ethical and honest.
The 2008 research looked to build on the insights of the 2005 KDA by ‘unpacking’ the five key
beliefs, which have been shown to drive overall trust and confidence, enabling further insight into
the results. A two-tiered approach was used. Question three in the survey asks respondents to
directly state which one of the five qualities is most important to their trust and confidence in
charities overall, showing the reported relative importance of each of the five key beliefs on overall
trust and confidence. For the first (top) tier of the Key Drivers Analysis we conduct regression
analysis to identify the correlation of these five variables with the first question in the survey, in
which respondents state their level of overall trust using a ten point scale. This allows us to identify
which of these attributes at Q3 are strongest drivers of trust at Q1.
The second (lower level) analysis then uses regression analysis to explore the correlations
between these five beliefs and other questions included in the survey to identify the key drivers of
these attributes.
In 2010, a change was made to the questionnaire, replacing the measure
trust in charities to
spend their money wisely and effectively,
with
trust in charities to make independent decisions to
further the cause they work for.
In order to investigate the relative importance of this belief in
comparison to the other four key beliefs a two-tiered KDA approach was again conducted.
In 2012, the two-tiered approach has been maintained, but a slightly different regression technique
has been used called ‘categorical regression’ which provides a more reliable approach to model
categorical variables. This approach is particularly suited to datasets based on large number of
response variables based on Likert scales. In essence, categorical regression enables the
selection of drivers to only those factors that have a strong and unequivocal impact on the
measure of interest, allowing us to focus in on the most relevant drivers for each model.
In 2012, the following questions were included in the second tier analysis (see appended topline
for full details): Q6, Q7, Q10a, Q10b, Q11, Q13 (a and b combined), Q14, Q14g.