The dynamics of employment-related community trust gender and

illbreedinggrouseUrban and Civil

Nov 16, 2013 (3 years and 4 months ago)

68 views


1

EGOS 2006 Bergen
-

Sub
-
theme 11

(due:
1
1 June 2006
)

The dynamics of
employment
-
related community
trust



gender and
experience in the UK Coalfield Communities.

Gráinne Collins

Employment Research Centre,

Trinity College Dublin,


College Green, Dublin 2, E
ire.

c
ollinsg

@

tcd.ie

http://www.tcd.ie/ERC/biogcollins.php


Michael Grimsley

Faculty of Arts, Computing, Engineering & Sciences,

Sheffield Hallam University,

Howard Street, Sheffield, UK, S1 1WB


m.f.grimsley

@

shu.ac.uk

http://aces.shu.ac.uk/cmsmg/


Anthony Meehan
(corresponding author)

Computing Department,

The Open

University,

Walton Hall,

Milton Keynes, UK, MK7 6AA

a.s.meehan

@

open.ac.uk

http://mcs.open.ac.uk/am4469/


Note to (critical) readers: The tables in this draft
paper are in a ‘raw’ state and so are not
always straightforward to interpret. We apologise for this and hope that, in many cases,
close scrutiny of the tables, which are grouped at the end of the paper, is not needed to
engage with the body of the paper.



Introduction


Trust
is

widely

accepted
as
an important factor

in the economic, social and psychological
well
-
being of individuals, communities and societies alike
(
Coleman 1990; Fukuyama
1995;
Giddens 1991
;
Green et al. 2000;
Gilbertson et al. 2005;
Gro
otaert 1998
;

Lin,
2001; Luhmann 2000
;
Putnam,
1993
, 2000, 2001
;
Warren, 2001
).

Recently, it has been
argued that the mechanisms by which these concepts operate are poorly articulated or

2

even fudged (Li et al. 2005).
In this paper
we
seek to
identify

possi
ble ‘mechanisms’ of
trust by probing the way in which
trust
changes in light of peoples’ lived experience

and

specific
ally in their experience of
employment

and worklessness
1
.


To identify mechanisms of trust w
e
consider

the linkages and pathways between
levels of
trust in the community and possible drivers of such trust levels, both within and between
the socio
-
economic boundary of employment status
,

and the socio
-
demographic
boundary of gender.
We look at the dynamics of trust by examining

changing level
s of

expressed

trust
at two time points

a
s people move (or fail to move) across the boundary
between worklessness and employment.
Data
are drawn

from
a
large
-
scale

longitudinal
study

in a sub
-
re
gion of the UK
. H
istorically
a significant proportion of the
male
population
of the area
was

employed in coal mining and heavy industry
;

more
recently
the area is
characterised by
high

unemployment

and very high levels of economic
inactivity
. B
etween
the two
waves of
data collection, employment
in the region
changed

as new types of
work

became available. This new employment was the result of local,
national and European regenerative investment (Gilbertson et al. 2005; Green et al. 2000,
2001, 2005).
The two
waves of the survey
provide community
-
level ‘snap shots’ of

trust
in 2000 and 2004 respectively and allow us to identify both persistent
effects
and volatile
features of trust expression
by people who have work or who are workless
over
the

period of four years.
We are able t
o explore
more fully
the
effects
of
cha
nging

employment status
on

trust

levels

by
examin
ing

the experiences of
a
subset of individuals
of working age for whom we have information
from
both waves

of the survey
.


We
examine the effects on trust of losing, keeping or g
aining

a job, and of enduring

worklessness. Specifically, we look at the
linkages
between trust in employers and other
expressions of trust relations in the community, both in relation to family, friends and
neighbours and in relation to institutions of comm
unity government and
local
employment
.





1

Worklessness includes people who are looking for work as well as people who are economically inactive
for whatever

reason.


3

Recognising the gendered nature of employment,
perhaps
particularly in the study area,
we are alert to the possibility that
employment
-
related trust
may reflect different
gendered experiences (both positive and negative)

of employment and wor
k
lessness
.
Accordingly our analysis considers gender effects from the outset
.



The paper b
egin
s

with
a
summar
y account of

the role of trust in communities
. We

then
examin
e

the evidence for different

arenas of

trust relations (

horizontal


and

vertical

)
a
nd for
some of

the underpinning experiential dimensions of trust
,

from which we
abstract the notion of

empowerment

.
We then
provide
a brief description of the study
area and
the
changes in employment between 2000 and 2004
,
of the
surveys which
provided
the data for our study

and of our statistical analytic techniques
.
In terms of
contributions, we show the effects of two ‘boundary crossings’ in relation to community
trust: between
different
employment categories and between workplace and community.
We p
rovide

reinforcing empirical

evidence that
at least two linked trust arenas

reflecting
vertical and horizontal community relations
may be
distinguished.

We show that
trust in
employers contributes to both these
arenas
of trust

but that the relative contrib
ution

may
change

over time
.

We are able to identify a link between trust
in employers and trust
outside the workplace
between friends, family and local community institutions
.

We
examine
the
nature
of ‘spill
-
over’ effects between the different trust

aren
as

and the
effects of change in employment status
and trust in employers
on these processes.


Gender differences
are
most marked when modelling change in vertical trust and trust in
local employers.
W
e conclude that differences are rooted in different ex
perience
s

as
women and men move between worklessness and employment. In seeking to understand
these differences we relate these results to
research on gender and employment which
suggests that w
omen are more likely to be home
-
centred or have adaptive prefe
rences
and that there is a gendered difference in fear of job loss and gendered affect in
employment generally
(Collins 2005; Collins and Wickham 2004; Hakim 1995)
.




4

W
e conclude
the paper
with a brief
consideration of
how community trust may be
affecte
d
by changes in the predominant forms of employment
available to a community
and

some
possible implications
this
may have

for regenerative employment policy.


Community trust and dimensions of personal experience


Trust can be seen as an 'act of faith' by

one party in respect of the (future) conduct of
another
(Giddens 1991)
.
T
he decision to trust reduces the eff
ort needed
by one party
to
acquire and maintain relevant information on another.
Community relations operate on
the basis of some level of trust which enables people to ‘act’ in the world. If trust is low,
acting in the world requires a considerable ‘act
of faith’ or ‘leap in the dark’. If it is high,
the leap is smaller. The magnitude of the leap may be associated with a proportional
sense of anxiety that may induce a sense of, or even actual, exclusion (Luhmann, 2001).

When trust is high
a
community
may
achieve
objectives that would not be
otherwise
attainable (Coleman, 1990; Fukuyama, 1995)
; c
ommunity trust relations are an
expression of a community’s capacity to achieve a better quality of life than would
otherwise be available if its members acted mere
ly as maximising individuals (
Li et al
2005;
Lin, 2001; Warren, 2001)
.


T
heorists

have been telling us for decades how important it is for business to have the
‘right’ societal environment. This environment includes pro
-
business attitudes, property
rights
, intellectual and human resources but also a level of general trust in the society.
Without this environment, business fails to grow and thrive.
In workplace relations, as in
other social relations, trust facilitates cooperation to the long
-
term benefit

of all parties.
In this context, we note that whilst there is an appreciable literature on
how
trust
is
important for the economic realm
,

less has been said about the effects of changes in the
economic realm and its knock
-
on effects on trust. So trust is

taken to be a prerequisite for
economic success
but
very little work
has attempted

to understand the
two way
linkage
between work
-
related trust and trust in the community.



5

Yet it is also wrong to assume community trust is a unified concept.
Within
commu
nities, i
ndividuals relate
to other individuals, but also to
a wide range of
institutions
. Some institutions are intimate (family, community associations) others
remote (local elected council); some
are
well
-
defined
,
others weakly defined or
amorphous

(fri
endship networks, parents
’ school gate

communities)
.




Different community
relations

entail different
forms of
information

/

knowledge

and risk
and therefore different levels or types of trust

(
Lewicki and Bunker 1996)
.

Braithwaite
and Levi (1998) suggest

that it is possible to differentiate between
two relational arenas
of trust:
vertical and horizontal
.

Vertical trust reflects the quality of relations between
people as ‘
citizens


and institutions of community governance and their associated
services

(edu
cation, health, police, justice, transport, &c).

Horizontal trust reflects
cohesion between
friends, family and

neighbours.

Grimsley
et al.

(2003)
have
demonstrated
empirical evidence for
such a distinction
.
(P
recisely how the workplace fits
into this pict
ure is
something of
an
open
question

that we explore in this paper
.
)




Grimsley et al
.

(2004)
demonstrate
that at least
three experiential factors underpin
expressed levels of trust
:
well
-
informedness

(the

extent to which individuals fee
l

well
informed ab
out the local community),
personal control

(the extent
to which individuals
experience

a sense of
control over their lives), and
influence

(the extent to which people
fe
e
l able to influence community life
)
As indicated above, information or knowledge of
ot
hers is seen as underpinning trust. I
nformation of different forms circulates within
communities and workplaces.
I
nformation
about and
perceptions of competence,
reliability, dependability
,

accountability and responsiveness circulate in the workplace

and
abroad in the community
. All of these contribute to reputation.
The concepts of
influence and control have a number of interpretations in the literature. Inevitably, this
makes it difficult to ensure a consistent interpretation
,

see Skinner
(
1996
)

for a

useful
review). Influence and control are sometimes subsumed in a single dimension
(Bacharach, Lawler 1980).
Other interpretations maintain a disti
n
ction between influence
and control so that influence corresponds to access to, and/or participation in,
decision
processes, e.g. the expectation of having the power to participate in making decisions in

6

order to obtain desirable consequences (Rodin 1990). In the context of employment,
influence is readily interpreted as the ability to contribute
to
the for
mulation

or
determination of employment conditions,
such as

via

trades union and similar
consultative fora. This interpretation appears to be shared by Warren (2001) when
examining political empowerment
,
and
by Hirschman (1970) in considering the nature o
f
‘voice’
.
Seen in this sense
, influence is distinct from a ‘sense of personal control’
. A
sense of control, then,
may be considered as
related to
autonomy, for example, the extent
to which a worker has more or less freedom in relation to choice of their

employer

or
organisation of work vis
-
à
-
vis domestic life
. A sense of control is important for personal
well
-
being and lack of personal control may have deleterious effects on mental and
physical health (Skinner 1996) providing a link between low trust en
vironments and
health.

To facilitate analysis of the
survey
data

(see below)
, we
are led to combine

the three
experiential factors as a single measure of
empowerment

via an analytic
-
synthetic schema
represented graphically in Figure 1.



Figure 1. Experiential dimensions of empowerment.


Well
-
informedness

Personal

co
ntrol

Influence

Informed

choice of action

Informed

participation

Efficacy

(contingency)

Empowerment


7

Pairwise, the three basis dimensions can be seen as giving rise to other compound
experiential dimensions: well
-
informedness and personal control engender a sense of
informed choice of personal
action; well
-
informedness and influence engender a sense of
informed participation (or expression of perceived needs); personal control and influence
induce a sense of efficacy (or contingency, where actions lead to anticipated outcomes).
We consider the e
xperience that combines all of these as a sense of
empowerment
.

The
empowered individual senses that they exert well
-
informed influence, make well
-
informed choices and that their actions are efficacious. Such an individual
, we suggest,

feels well placed to

act in a trusting way
:

.
to

make the leap in the dark which exposes
them to risk in relation to the conduct of others.


Study Area, Data and Methods


The study was undertaken in an area of South Yorkshire

in the

UK.
Compared with the
rest of the UK, the st
udy area has very high levels of economic inactivity in both men and
women (about twice the national average).

This is partly related to
a
high percentage of
long
-
term sickness

and
incapacity. Incapacity in men is conventionally explained in terms
of hist
oric employment in heavy industry,
especially in

coal mining, steel making and
heavy engineering.
Table 1
(after Gilbertson et al. 2005, p35)
presents an
comparative
overview of employment
-
related data for the study area

and the UK,
over
the
approximate
st
udy period
.


During the
relevant four years
the area was the focus of several regenerative initiatives.
T
here was a three point increase

in the percentages
of

both

women and men

of working
age

who
were

economically active (
T
able 1).

In the UK as a whole th
ere was virtually no
change
,
with a one per cent reduction for women and a one per cent increase for men.


For both women and men, the increase in economic activity is largely reflected in
increased rates of full
-
time emplo
yment (three

pe
rcentage points in

each case).
There
was

been a small reduction in the proportion of w
omen and men in part
-
time work.


8

Additionally, more women and men
were

actively looking for work. Against the UK
trend, there
w
as
an

appreciable fall i
n the proportion of people who we
re se
lf
-
employed.


Table 1. Economic (Employment) Status: Working age Study Area and UK


The South Yorkshire Social Capital (SYCS) survey


is a longitudinal study of nine
communities in the
South Yorkshire coalfield
. The survey

involve
d

two linked samples
fro
m the study area. The first survey

(wave 1)
, undertaken in 2000, had 4220 respondents
(Green et al. 2000); the second
(wave 2)
occurred in 2004 had 3771 respondents
(Gilbertson et al. 2005). These surveys include a longitudinal sub
-
sample of 1071
responden
ts

who were common to both surveys
.
The present analyses are mainly
confined to sub
-
samples of respondents who were of working age: 16 to 59 for women
and 16 to 65 for men. There were 2984 and 2431

working age respondents in the
respective surveys.
In

the
longitudinal
sub
-
sample there were 586 people who were still
of working age in 2004.


The survey

datasets were collected to explore,
inter alia
, the social formations and
developments of communities in the South Yorkshire coalfields. For this
,

a
variety

o
f
responses relating to trust w
ere

recorded using ques
tionnaire items

employed in previous
studies
2
. These included questions, with responses measured on a five point rating scale,
on how much the respondent trusted, family, friends, neighbours, local poli
ticians, the
local council and, crucially, local employers
. T
he survey
s

also
recorded the three
experiential items reflect
ing

empowerment

(as considered above): well
-
informedness,
personal control and influence.

The three variables

were

scored on a 4
-
point

and two 5
-
point Likert scales respectively
.



For the entire sample at wave 1, i
nitial principal component analysis (PCA) of the
correlation matrix for the six trust measures demonstrated that around 57 per cent of their
variation could be accounted for b
y two underlying dimensions

(Grimsley et al 2003)
. An



2

For the full list of items and their sources see Green G, Grimsley M, Suokas A et al. 2000. Social capital,
health and economy in South Yorkshire coalfield communities. CRESR, Sheffield Hallam University pp
1
5, 36


37
.


9

oblique rotation of this structure showed one factor, interpreted as vertical trust,
associated with trust in local politicians, the local council and local employers, and a
second factor,
interpreted a
s
horizontal trust, associated with trust in neighbours, family
and friends. These factors had a low, but
statistically
significant, positive correlation (r =
0.18) and overall measures of vertical and horizontal trust were hence obtained by simply
summing

the three trust responses incorporated in the two factors
. E
ach score thus
had

a
minimum value
(lowest trust)
of three and a maximum of fifteen
. A further PCA of the
three positively correlated experiential items
(well
-
informedness, personal control,
inf
luence)
revealed a single component
with approximately equal and high positive
loadings on each item

(Grimsley et al 2004)
. I
t appeared reasonable, therefore, to sum
responses on these three measures to give an overall “empowerment” score
with the
range th
ree to fourteen
(Green e
t al, 2000
). Similar procedures were adopted in the initial
analyses of the sample at wave 2 (Gilbertson et al, 2005).



For this paper, factor analyses of the six trust measures at waves 1 and 2, were repeated
for the sub
-
samples o
f working age respondents. Results were consistent with those
found for the all age samples.

To facilitate multivariate analyses, o
utcome
variables for
individual and overall trust
were
recoded into three discrete ordered categories: Low,
Moderate and High
. M
ultinomial or conditional logistic models (Agresti, 1990) were
then
applied to these three outcomes to explore links with the three experiential factors and
overall “empowerment” both within and across the boundaries
of gender and
employment status
3
.


F
inally, the dynamics of emp
loyment status and trust change in light of
employment
-
related ‘
boundar
y crossings’

were modelled using both multinomial and general linear
methods.

Further
details of methods are provided in the context of the presentation of
re
sults, below.






3

Note that results from
ordinal logistic modelling have not been used as the proportional odds assumption
was invariably violated.


10

Community Trust

Forms
and Trust in Employers


W
hen
considering community
trust
relations

we observed that, following Braith
waite and
Levy (1998) and Grims
l
e
y
et al.

(2003), there appeared to be
(at least) two
different
forms of community tr
ust relations: horizontal trust

between
family, friends and

neighbours,
and vertical
trust expressed in community institutions.
The survey data were
analysed with a view to confirming or rejecting the persistence of this distinction over
time. T
he
re is no
basis to consider
a priori

that these

trust
forms are

independent.
Accordingly, we use
d

PCA and factor analysis

to

identify the underlying components

or
dimensions

of trust.
For the sub
-
samples of working age at wave 1 and wave 2, the first
two components
accounted for 56 per cent and 64 per cent of total variation in the six
trust measures. The obliquely rotated two factor solutions are given in Table 2

(wave 1
analysis was considered above and is incorporated here for purposes of comparison)
.
In
the tabl
e the squared values of the coefficients (in columns) indicate the unique
contribution of each of the six measures to each of the two components or dimensions.

The analysis strongly suggests
the
persiste
nce

of a
distinction between trust component 1,
inte
rpreted as vertical trust
,

and trust component 2, interpreted as horizontal trust.
T
he
obliquely rotated factor structure
s

give

positive correlations between
trust components 1
and 2 (
vertical and horizontal trust relations
)

of 0.18

and 0.2
4
, respectively
-

b
ot
h
significant at the 0.01 level,
which
suggest
s

that vertical trust
may ‘spill

over’ into
horizontal trust and
vice versa
.


Table 2. Components of Community Trust 2000
-
2004: vertical (1) and horizontal (2) trust dimensions in
the two waves (working ag
e only)


Of particular
note

is the change in the relative contribution to trust in employers to
horizontal trust
between 2000 and 2004.

In 2000 trus
t in employers contributed 0.52 to
vertical trust and 0.11

to horizontal trust. In 2004 trus
t in employers

contributed 0.39 to
vertical trust and 0.36

to horizontal trust.
In the sections that follow, we
shall
periodically return to th
is observation

as we
attempt to understand the basis for th
e
(re)emergence of the horizontality of employer trust.



11

Whole
-
surv
ey

Analysis


Levels of Trust and Empowerment

Initial data exploration is via mean profiles for the key boundary defined sub
-
groups of
gen
der and employment status. The
results are shown in Table 3. Average scores over the
two waves, together with relevant
standard errors, have been computed for vertical and
horizontal trust, empowerment and trust in employers.
As these outcomes are not
distributed Normally,
simple, non
-
adjusted,
differences in location between relevant sub
-
groups have been tested using a no
n
-
parametric approach.



Table 3

Trust and empowerment: mean scores by gender & employment status


In both surveys, reported horizontal trust levels were high
, relatively few respondents do
not completely trust friends and familiy
; vertical trust, employer

trust and empowerment
scores were mid
-
range. (It is best to consider these as relative levels rather than absolute
levels as there may well have been framing effects in responses.) Over the four years
there was a small increase in vertical
trust, which ap
peared significant for the employed.
E
mployer trust
showed a small but statistically significant increase for those in work.
Empowerment score rose significantly for women and the employed. There was an

appreciable
decrease in horizontal trust.


People who

we
re employed show
ed

a persistent statistically significant higher level of
trust (all forms) and empowerment than people who are workless
. And the gap between
employed and workless increases, often appreciably
.



In relation to gender, o
n average w
omen s
core consistently higher
on trust and
empowerment
than men.
Between the two surveys, vertical and employer trust levels
converged between women and men, levels increasing for both, but more so for men.
Horizontal trust and empowerment diverged between wome
n and men, the decline being
greater for men.
The fall in horizontal trust was more pronounced for those in
employment than those who were workless, and for men more than women.



12

The
findings
above point to employment and gender effects for trust levels

and
empowerment
. These

sub
-
group comparisons take no account of possible confounding
characteristics such as respondent age, education levels, household tenure or area of
residence. In the following sections, logistic modelling has been used to explore cr
oss
-
sectional relative differences in trust, and possible drivers of trust, between and within the
boundaries of gender and employment status. Such models adju
st for, or take into
account,
possible socio
-
demographic confounding factors.


Trust, Employment
,

Worklessness

and Gender.


We begin by examining expressions of vertical trust
, horizontal trust
and employer trust
at the beginning and end of the study period
using main effects
multinomial
logistic
models
.

Parameters for explanatory variables are expres
sed as odds ratios (together with
confidence intervals). In the multinomial case,
adjusted

odds ratios are given taking one
of the dependent states as base or reference. In the analyses below the base has been
taken as the Low category and the odds ratio i
s given just for the High category.


In Table 4, adjusted odds ratios are first given for employment status and gender for the
entire working age samples in the two years

(model 1a)
. The second part of the table
cons
iders models separately for the

workless

(1b) and employed (1c)

which facilitates
gender comparisons within these two states.


Table 4
.

Relative changes in trust levels between 2000 and 2004 by employment status and gender (main
effects multinomial models).


A
ccording to model 1a,

in the year 2
000
workless respondents were

(on average) 0.85
times as likely as employed respondents to express high vertical trust compared with low
vertical trust; a difference that was not statistically significant.
(
A slightly more
accessible translation might say
that for every 100 employed respondents expressing high
vertical trust

compared with
those expressing
low
, only 85 workless respondents did
likewise.
)

By 2004, workless respondents exhibited an

appreciably reduced likelihood of
expressing either high verti
cal trust or high employer trust.

In

contrast, the
difference in

13

likelihood of
workless respondents

expressing high
as opposed to low
horizontal trust
relative to
employed respondents

has decreased to marginal statistical significance
.

This
is consistent w
ith a secular fall in horizontal trust for both employed and workless but
there being a greater rate of fall for the employed as seen in table 3.



Table 2 indicated that
e
mployer
trust

contributes positively to both
v
ertical trust

and
h
orizontal trust
.
Th
us t
he decline in
v
ertical trust

amongst workless respondents
may be
accounted for in some measure by the relative decline in
e
mployer trust

but there is a
need to understand why trust in employers has fallen amongst the workless (see below)
.

The changes i
n vertical trust serve

to emphasise that the countervailing shift in
h
orizo
ntal
trust

amongst the workless is more appreciable

t
han suggested by the odds ratio

diffe
re
nces of 0.66 and 0.80.
That is, the difference between the employed and workless
in horiz
ontal trust would have been much smaller with
out

the large decrease in
trust in
employers.
This may point to a community level ‘
accommodation
’ to worklessness
status between the workless and their family, friends and neighbours and possibly a
‘community o
f the workless’ effect in which support networks develop amongst the
(enduringly) workless.


In terms of gender, model 1a indicates that, allowing for employment status, men
exhibited lower relative trust levels compared with women for all three outcomes

at both
time points. Relative differences for vertical trust, in contrast to horizontal and employer
trust, were not significant. It is striking that, for horizontal trust,
that is
trust in
neighbours, friends and family, the gendered (relative) gap widen
ed appreciably.


W
e
also
examine
d

whether there are gendered differences in respect of
th
is

picture
within
the employment
category
boundaries (models 1b & 1c)
.
There are two results of
(statistically significant) note. The first is that over the study peri
od, men expressed a
relative
decline in
h
orizontal
trust

compared with

women
,

but that
seem
s

to be
independent of employment status.
In relation to employer trust,
the gap between
workless women and
workless
men widens, but the gap between employed women a
nd
employed
men narrows. For workless people t
he relative decline seems particular
ly

steep

14

suggesting that enduring worklessness
may
ha
ve

a particularly corrosive effect on trust in
employers.


Thus, h
aving adjusted for
a range of demographic factors, the

initial
picture
in which
trust
and empowerment levels are appreciably informed by employment and gender effects is
one that persists.



Trust and Empowerment

In this section

we
use the construct of empowerment, as considered above, to begin to try
explai
n the effects we have observed
. We look to
confirm the persistence of the
relationship between empowerment,
v
ertical trust

and
h
orizontal trust

(suggested by table
2)
and
,

for the first time, we examine the relationship of empowerment to
employer
trust
.


The

empowerment score
was

computed by summing the (Lickert scale) responses in
respect of: well
-
informedness, personal control and influence
and
therefore
score
s range
from 3 to
1
4
.

In the light of the distribution of scores
, and to facilitate logistic mod
elling,

th
ese

were
grouped
into three categories: High (11
-
1
4
), Moderate (9
-
10) and Low (3
-
8)
.

Multinomial logistic modelling was then applied to the three trust outcomes adding this
three category
e
mpowerment explanatory factor to those used in models 1a


1c. The
resultant adjusted odds ratios are given in
Table
5

for
both wave 1 and wave 2

working
age samples (model 2a) and separately for employment status (models 2b, 2c).


The odds ratios
suggest that empowerment is, broadly speaking,
positively
relate
d to
the
likelihood of
the
expression of high relative to low
vertical trust, horizontal trust and
employer trust
.

L
ow empowerment implies a re
duced likelihood of high trust relative to
low trust; high empowerment implies an elevated likelihood of high tru
st relative to low
trust
.


Between 2000 and 2004 the proportion of people expressing elevated levels of
empowerment increased and this accounts, in part at least, for the

steepening gradient


of the relation between empowerment and the three forms of tru
st considered.
Relative to

15

those with high empowerment, people with moderate (and low) empowerment
were

much less likely to express high vertical, horizontal or employer trust in 2004 than in
2000 and this
wa
s especially pronounced for workless compared
wi
th
employed people
(model 2a).
Within employment status sub
-
groups, however, this “steepening” applies to
the workless rather than the employed (models 2b & 2c) and this is particularly the case
with employer trust.


Table 5

Relation of Empowerment to Rela
tive Trust Levels 2000
-
2004 (main effects multinomial model).


This
complex relationship
between empowerment and

trust, particularly

employer trust
,

and the ways it has changed over the four years, leads us to
seek
a deeper

explan
ation

by
explor
ing

the rel
ationship
s

between employment status
and gender
and the basis
dimensions of empowerment
:

the three experiential factors
.



In tables 6 and 7 are given the relevant odds ratios for multinomial logistic models using
the binar
y

forms of the explanatory variab
les: “well informed” (no, yes), “satisfied with
control” (no, yes) and “can influence” (no, yes). The “yes” response is taken as the
reference category in all cases.
The workless and employed are considered in m
odels 3a
and 3b (Table 6)

and males and femal
es are separately analysed in models 4a and 4b
(Table 7).


Table 6
Relation of components of Empowerment to Relative Trust Levels 2000
-
2004 by employment
status (main effects multinomial model).


T
he picture that emerges
in Table 6
is that
,

in
general, th
e workless
are showing
relatively more negative responses across the three
experiential
factors over the two time
points compared with the employed. This is most marked, perhaps, with employer trust
where there was a particularly large relative decline in
the sense of influence.
F
or those in
employment the dominant experiential component of
employer trust

is a sense of well
-
informedness but
this also applies to the workless
.
For workless people, there is evidence
that they come to experience a diminishing s
ense of influence. All t
hese effects, however,

appear to be gendered.


16


Table 7 indicates that, a
cross the study period, for men
the

dominant component
or driver

of

trust, particularly
employer
trust
,

was the sense of well
-
informed
ness
.

Whilst well
-
informe
dness was also important for women,
a
sense of
personal
control was persistent
and
possibly
dominant
in this respect.


Table 7
Relation of components of Empowerment to Relative Trust Levels 2000
-
2004 by gender (main
effects multinomial model).


At the con
clusion of the
survey
analysis
for
the
two
whole sample
s we observe that
worklessness
appears to be
a barrier to increasing vertical and employer trust

but that the
corrosiveness of worklessness on horizontal trust was attenuated, possibly through a
‘horiz
ontal accommodation’ to its effects. E
mpowerment is positively related to levels of
all trust forms but there are significant gender differences in respect to the components of
empowerment that relate to trust, especially trust in employers.



Longitudina
l Subset Analysis

As

indicated
earlier
, the two surveys
(waves 1 and 2 respectively) feature

a set of
common respondents of whom 586 were still of working age in 2004. Of these, 54 had
become workless, 229 had remained workless, 226 had remained in employm
ent, and 77
had become employed

(see
T
able
8
)
.

We choose to order these in this way to suggest a
scale

going from least positive experience to most positive experience (this ordering does
not introduce any assumptions into the models constructed).

The smal
l numbers of people
in the first and last categories lead us to be appropriately
tentative on our findings.


Table
8

Changing
employment between 2000 and 2004 by gender


Wave 1
\
wave 2

Employed

Total (women)

Workless

Total (women)

Employed

226 (158)

54 (38
)

Workless

77 (63)

229 (150)



Figures 2
-
4

illustrate the relationship

between employment status and
employer trust
,
vertical trust

and
horizontal trust

respectively.

Figure 2
shows

a
n

apparently

significant

17

enhancement of
employer t
rust

for those moving

into employment and a small diminution
of
employer trust

for those losing employment
;

allowing for the modest
secular rise
across respondents

whose status did not change, this diminution of trust would be
marginally greater but still not statistically sig
nificant
.



Figure 2 Employment status and Employer Trust


Figure 3 is suggestive (no more) of
enhanced
vertical trust

for those gaining work and
this is consistent with the contribution of
employer trust

to
vertical trust
.

77

226

229

54

77

226

229

54

N =


Workless to


employed


Employed


Workless


Employed to


workless

95% CI

CI

4.0

3.8

3.6

3.4

3.2

3.0

2.8

w1 Trust

employers

w2 Trust

employers


18


Figure 3 Employment status and Vertical Trust


There are no statistically significant effects on horizontal trust discernable in
Figure 4
.


Figure 4 Employment status and Horizontal Trust

77

226

229

54

77

226

229

54

N =


Workless to


employed


Employed


Workless


Employed to


workless

95% CI

CI

10.5

10.0

9.5

9.0

8.5

8.0

7.5

7.0

w1 Vert trust

w2 Vert trust

77

226

229

54

77

226

229

54

N =

Workless to

employed


Employed


Workless


Employed to


workless

95% CI

CI

15.0

14.8

14.6

14.4

14.2

14.0

13.8

13.6

13
.4

w1 Hor trust

w
2 Hor trust


19


No discernab
le gender effect is suggested in respect of vertical trust and horizontal trust
(not shown). However, for employer trust there is a marked gender effect
.
The average
change in employer trust (together with 95% confidence intervals)
for men and women
acros
s the four employment change

categories is given in Figure 5
.


.

workless to
employed
[M:14; F:63]
employed
[M:68; F:158]
workless
[M:79; F:150]
employed to
workless
[M:16; F:38]
Wave 1 to wave 2: Employment status
1.0
0.5
0.0
-0.5
95% CI Change wave 1 to 2 Trust local employers
female
male
Gender

Figure
5.
Employment Status and Employer Trust
change by gender


For women who lose employment ther
e is a marked loss of employer
trust and for those
gaining employment a very marked incre
ase in trust
.

There is, in effect, a linear trend in
employer trust levels for women across the employment status categories.
For men, no
pattern is apparent (
though sub
-
sample numbers are very small
)
.


We explore this
,

perhaps
surprising
,

observation

in t
he following general linear
(regression) models.






20

The dynamics of e
mployment status and employer trust change

To confirm the interpretation of (changing) employment status and
employer trust

by
gender suggested above we produce a general linear model
(GLM) of change in
employer
trust

against changing employment status by gender (Table
8
).
The outcome here is the
difference in recorded trust levels between wave 1 and wave 2.

A positive score indicates
an increase in trust and these trust chan
g
e outcomes

follow approximately Normal
distributions. The estimates sho
w

the average difference between the relevant category
and the base or reference category.
(E
stimates have been adjusted for gender and age of
respondents

but not area of residence because of sma
ll sub
-
sample size. Respondent
education level and household tenure were not statistically significant.
)


Table
8

GLM of
employer trust

against (changed) employment status, by gender
.


For men, no coefficients of the model are statistically significant
ly d
ifferent
from the
reference change category: workless to employed.

By contrast, the coefficients in the
model for women are at least marginally significant or highly significant. Losing or
gaining work is much more significant for women than men in respect

of
employer trust
.
Given the relationship of
employer
trust

to
vertical trust
, the picture presented here is
reproduced when considering change of employment status and
vertical trust

(Table
9
)
;

though
as would be expected
the
relationship is less pronoun
ced
.


Table
9

GLM of VT against (changed) employment status, by gender
.


Consideration of figure

6 suggests
that it is the experience of losing or gaining
employment that has the most appreciable effect on
trust
and so we combine the
categories for people
who are enduringly employed or workless
.

This

give
s

a
“no
cha
nge” group and
is
use
d

as
the reference category
when
further
exploring
the
gendered
effects of employment status change. Table 10 looks at the contribution of the three
empowerment dimensions to

changed

employer trust by gender.


Table 10.
GLM for changed
e
mployer trust related to change in experiential factors

by gender
.



21

Adjusting for base model explanatory variables, it is clear that the only statistically
significant contribution is the cha
nged satisfaction with level of personal control when
experienced by women.


Finally in Table 11, for women only, parameter estimates are given for employer trust
change using employment status change and experiential factor change as explanatory
variables



Table 11
GLM for changed employer trust related to change in experiential factors and employment status
(women only)

.


Thus, even allowing for status change, change in feelings of control is a positive driver of
change in employer trust levels for wome
n. In addition, gaining employment gives a
significant boost to trust in employers
;

whilst “losing” a job appears corrosive of such
trust levels.


Employment status and employer trust change: evidence of community “spillover”
effects

Because of small sub
-
s
ample numbers

it is not realistic to attempt to link explicitly
change in employment status with wider changes in community cohesion levels.
However, at least for women, it appears that gaining or losing a job doe
s positively or
negatively affec
t
s

employer

trust levels. It is therefore feasible
to explore, indirectly, the
pathway from
employment change to community trust change via the intermediary
dimension of
employer trust change.


The results of analyses, again using multinomial logistic models, are gi
ven in Table 12.
The eight outcome variables reflect three category change measures (decrease, no
change, increase) for the longitudinal sample of 586 working age respondents. These
outcomes have been modelled on employer trust change and gender. The odds
ratios are
given for the “decrease” category using
“increase” as reference and have been adjusted
for respondent age, educational attainment and household tenure.


Table 12
Main effects
multinomial

models: Adjusted Odds Ratios*
-

trust change: local employe
rs


22


In terms of trust change outcomes,
employer trust change appears to be consi
s
tently
positively linked with other community trust change measures. In other words, decreases
in employer trust are associated with decreases in other trust measures


and in
creases
with increases. Thus those expressing a decrease in employer trust are, on average, 80
percent more likely to report a decrease in horizontal trust than those who express an
increase in employer trust. Though only two odds ratios are statistically
significant (for
the first four outcomes), all are in the “right” direction to be
consistent with positive
“spill
over”. The most striking result here is
, perhaps,

the significant odds ratio for trust in
neighbours.
Those reporting a decrease in employer tr
ust are, on average, more than
twice as likely as those reporting an increase in employer trust, to record a decrease in
neighbour trust (compared with an increase). Putting this in a more positive way, those
recording an increase in employer trust are, on

average, half as likely as those recording
an employer trust decrease, to report a decrease in neighbour trust.

The second part of the table employer trust change appears to be significantly and
positively associated with empowerment change overall, and
particularly with two
components of empowerment change: satisfaction with control and well
-
informedness.


Summary and Some Points of
Discussion

In
the results presented above we have demonstrated that
crossing the soc
io
-
economic
boundary of employment

vers
us workless and the socio
-
demographic boundary of gender
impacts on levels of trust and its production based on lived
-
experience. We have shown
persistent evidence for
at least two relational arenas of community trust: horizontal and
vertical, and
have see
n
that these are linked s
uch

that an increase in one is associated with
an increase in the other, and a fall in one with a fall in the other. We have shown that trust
in employers
contributes positively to each arena but that the level of the contribution
may change
. We have shown that changes in trust
relations
may be explained, in part, by
change in a sense of empowerment (and its basis experiential dimensions of well
-
informedness, personal control in one’s l
ife, and a sense of influence). In relation to
trust
and empowerment, we have compared trust levels and experience of people who are
workless and people who are employed. We have examined the effects
of changing or

23

maintaining employment status on trust and empowerment and have seen that it is
changed

experience rather
than enduring experience that has the most perceptible effect
.

Having n
ot
ed

from the beginning that there were significant gender effects
in evidence
,
w
e
established that
the impact of change
d employment status

on trust was much more
pro
nounced for women than for men and we related this to

a gender
-
based difference in
respect of

specific components of empowerment.
Specifically, w
e have
demonstrated
that
in relation to trust, and trust in employers particularly, gaining or losing work is
p
redominantly experienced as impacting on
a woma
n

s
sense of personal control in life

but on
a ma
n

s sense of influence

in the community
.

We have shown how cha
nges in one
form of trust ‘spills over’ to effect change in other forms of trust.


Generalisabilit
y


We have highlighted some distinctive socio
-
economic features of the study area,
particularly the significant changes in traditional employment with the close of the coal
-
mining, steel and related heavy industries, the high levels of unemployment and of
worklessness
,

which meant that some of the individual communities within the area were
amongst the most deprived in the EU for a period of time, and the regenerative
investment that produced new but different forms of work. Given this profile, any attempt
to generalise the results reported must be appropriately qualified. This said, there are
many sub
-
regions of the EU, and indeed beyond, where the
profile
of the study areas is
readily recognised.

A comparative study across such regions would be a natural n
ext step
in affirming the validity of the results presented.


The workplace as trust producing environment

One of the observations to emerge was the extent to which trust in employers came to
feature

more prominently
within
horizontal trust over the four y
ears of the study.
The
nature of the SYCS survey instrument does not allow us to ground our understanding of
this issue on the basis of the data available
.
We can consider at least two possible sources
of explanation, which are not mutually exclusive. The
first is to observe that the
diminution of the gap in horizontal trust between workless and employed, possible as a

24

result of a community ‘accommodation’ to worklessness, has eliminated once source of
variance in horizontal trust so that the relative expla
natory contribution of employer trust
has increased. A second source would lie in an
examin
ation of

the
role of the workplace
in the production of trust
in employers

and
both
horizontal
and vertical
trust
.
E
mployment may be seen as a ‘vertical’ relation wi
th an employer
,

characterised by
power, control, regulation (and some measure of negotiation in relation to the application
of these).

Equally,
the workplace may be
experienced as
providing a milieu which
extends the scope for friendship and neighbourly so
lidarity

(Cattell, 2001)

facilitating
production of horizontal trust. An interesting question
therefore
is
how, if at all, the
new
forms of work and the conditions of employment that are available
serve the production
of different forms of trust? It certai
nly seems that research to answer this question must
take account of the evidence presented here that trust production, via the experience of
empowerment, differs for women and men.



Gender

Employment and unemployment are often seen through a masculine p
rism, ‘what is the
male unemployment rate and what types of jobs are men doing?’, for example, the area
where the survey was conducted continues to be ‘defined’ as an
ex
-
coalfield. We do not

underestimate the importance of work for men and the returns thi
s has for families and
communities, but a male
-
focused perspective may obscure the importance of jobs for
women. The above analysis would suggest that if governments want to turn around
failing areas (however failing is defined) through policies designed t
o increase trust in
those areas, then ‘jobs for the girls’ are at very least as just as important as ‘jobs for the
boys’.


We perceive there to be a relative lack of research on women’s attitudes to employment
and what they seek in it.
Previous work on
w
omen
’s preferences to employment

has
often asked on

whether women
are more likely to be home
-
centred or have adaptive
preferences
(Hakim 2000)

-

do women want jobs or to stay at home and mind their
ch
ildren


not of the effects of these choices o
n

community trust. However, there may
have been clues to our findings in a different set of research which looked at the
gendered

25

difference in fear of job loss and gendered affect in employment generally
(C
ollins

2005;
Collins and Wickham 2004
)
.

This
latter
work stresses that it is women
,

rather than men
,

who
particular
ly

fear unemployment,
which

accords with our findings that it is women
whose trust is most affected by the move into or out of employment.
It also accords with
the observation that it is control in one’s life is the dominant empowerment dimension for
women; such control flows self
-
evidently from a personal wage, but the form of work
and the conditions of work
may also contribute to the abilit
y to juggle the demands of
care for children, dependent relatives, a personal social life and other aspects of the non
-
work arena.



Responsiveness
and hysteresis

At a number of points in our analysis we saw that changes in trust levels over time
differed
for one category in relation to another
, and especially in
women

relative to
men.
We would like to explain th
is variation

in response
by reference to trust
and its
production
(
rather than
rely upon

an appeal to some ‘innate’ difference in responsiveness
be
tween women and men
)
. We have pointed to a partial answer by demonstrating that
women
and
men
appear to respond differently to different components of empowerment.
Thus it is not simply the change in, for example, employment status, that matters, but the
w
ay that change is experienced in relation to well
-
informedness, personal control and
influence.

That is, for a man or woman obtaining employment, the character of the work
and the workplace matters too in the production of trust.


A second element of an a
nswer is suggested by the
observ
ation

that
variation in
‘responsiveness’
is suggestive of
the
hysteresis effect in trust
relations. It is generally
more

difficult to repair
a once high but then
broken trust

than it to establish initial trust
.
Thus it may b
e that the recent economic history and (work
-
related) social/domestic
change in the study area have impacted men and women differently in respect of trust
once held but now lost.


Figure 7
seeks to illustrate the scenario
-

i
f
the negative experience in r
espect of employer
trust were more pronounced for
workless
men than
for workless
women

then a more

26

sustained positive going experience would be needed before trust levels began to be
restored. It seems plausible, though no more, to
suggest men in the
study

area
may have
indeed
been at a different point on the trust hysteresis curve than women because of the
effects of the significant and acute loss of traditional forms of employment (under
conditions of
significant
national industrial conflict) in the late
20th Century.

However,
we do not want to in any way underestimate the significance of the change and the
conflict for the communities taken as a whole.


.



Figure 7 Hysteresis of trust


Managing trust
-
led development

In the f
inal point of discussion in this paper we address the issue of trust
-
led regenerative
economic development.
Trust is an ill
-
defined, ethereal, fuzzy concept and, echoing
many others, Li et al. (2005) have highlighted the problem of identifying mechanisms
u
nderpinning trust. Such mechanisms are needed if policies to promote trust are to be
successfully operationalised and managed. In this paper, we have established that
empowerment

partially
explain
s

variation in trust levels
. We suggest that relative to tru
st,
empowerment its basis dimensions of well
-
informedness, personal control and
community influence, are more tangible and more tractable and thus provide
development policy makers, legislators and managers with a more direct means by which
to engineer
the

desired end.

+
ve
experience

Trust

workless
men

workless
women

-
ve
exp
erience


27


In the context trust
,

attracting employment to an area without considering the way in
which the variety,
nature and organisation of
that
work
will differentially impact women
and men may mean that the socio
-
economic outcome is less than opt
im
al. In
demonstrating the ‘spill
over’ of trust in employers into other trust arenas, we have made
clear that
a

good


employer can improve community trust for employees and non
-
employees alike
. And in answer to the question ‘what is a good employer?’ we a
re able to
point to empowerment and its component dimensions
.


Acknowledgement

We

would like to thank colleagues at the Centre for Regional Economic Social Research
at Sheffield Hallam University for enabling us to undertake elements of this work.



Refere
nces

Agresti A.
(
1990
)

Categorical Data Analysis
. New York: John Wiley & Sons


Bacharach. S., Lawler E. (1980)
Power and Politics in Organisations
. San Francisco:
Josey
-
Bass.


Braithwaite, V., Levi, M. (eds) (1998)
Trust and Governance
. New York: Russell

Sage
Foundation, 1998


Cattell V. (2001) Poor people, poor places, and poor health: the mediating role of social
networks and social capital.
Soc Sci Me
;
52
: 1501
-
16.


Coleman, J.
(
1990
)

Foundations of Social Theory
. Harvard University Press, Cambridge,
M
A.


Collins, G. (2005) ‘Seven Reasons Why Mergers Are Gendered. Employment Research
Centre, Dublin: Trinity College Dublin’.
Gender, Work and Organization

12(3): 270
-
289.


Collins, G., Wickham, J. (2004) ‘Inclusion or Exploitation? Irish Women Enter The
L
abour Force’.
Gender, Work and Organization

11(1): 26
-
46.


Fukuyama, F. (1995):
The Social Virtues and the Creation of Prosperity
. London, Free
Press
.


Giddens, A. (1991)
Modernity and Self
-
identity
, Cambridge: Polity Press.


28


Gilbertso
n, J., Green, G., Grimsley, M., Manning, J. (2005).
The Dynamics of Social
Capital, Health and Economy
.
CRESR, Sheffield Hallam University.


Green, G., Grimsley, M., Suokas, A (2000)
Social capital, health and economy in South
Yorkshire coalfield communiti
es.

CRESR, Sheffield Hallam University.


Green, G., Grimsley, M., Stafford, B. (2001)
Capital accounting for neighbourhood
sustainability.
CRESR, Sheffield Hallam University.


Green, G., Grimsley, M., Stafford, B. (2005)
The Dynamics of Neighbourhood
Sus
tainability.

York: Joseph Rowntree Foundation, York Publishing Services Ltd.


Grimsley, M., Meehan, A., Green, G., Stafford, B. (2003) 'Social Capital and Community
Trust and
e
-
Government Services'. In P. Nixon and S. Terzis (eds.):
Trust Management
2003,

First International Conference on Trust Management, Hiraklion, Crete, Greece,
May 2003. LNCS 2692
, Springer
-
Verlag, Berlin Heidelberg, 2003, p165
-
178


Grimsley, M., Meehan, A., Tan, A. (2004a) 'Promoting social inclusion: managing trust
-
based relations be
tween users and providers of public services'. European Academy of
Management, St. Andrews, UK, 5
-
8 May, 2004.

(http://www.st
-
Andrews.ac.uk/~euram04/proceedings/contents.shtml#tracktr also
accessible at http://mcs.open.ac.uk/am4469/ or http://aces.shu.ac.
uk/cmsmg/)


Grimsley M, Meehan A. and Tan A.
(
2004b
)

Designing for Social Inclusion: Computer
Mediation of Trust Relations Between Citizens and Public Service Providers. In Fincher
S, Markopoulos P, Moore D, Ruddle R. (Eds).
People and Computers XVIII
-

D
esign for
Life: Proceedings of HCI
2004

London:.Springer
-
Verlag, 197
-
210



Grootaert, G. (1998).
'Social Capital: The Missing Link'. SCI Working Paper No.3, April
1998, World Bank.


Hakim, C. (1995) Five feminist myths about women's employment.
British Jo
urnal of
Sociology,
46, 3, 429
-
455.


Hakim, C.
(
2000
)

Work
-
Lifestyle Choices in the 21st Century: Preference Theory.

Oxford: Oxford UP'.


Halford, S. and Leonard, P. (2001)
Gender, Power and Organizations
, Basingstoke:
Palgrave.


Hirschman, A. O.
(
1970
)

E
xit, Voice, and Loyalty
, Cambridge, Massachusetts: Harvard
Universi
ty

Press.



29

Lewicki, R. J. and Bunker, B.B. (1996). Developing and maintaining trust in work
relationships. In Kramer. R.M. and Tyler, T.R.
Trust in Organizations: Frontiers of
Theory and Re
search
, Thousand Oaks, CA: Sage 114

139.


Li, Y. J., Pickles, A. and Savage, M. (2005) 'Social capital and social trust in Britain',
European Sociological Review

21(2): 109
-
123.


Lin, N. (2001):
Social Capital: a Theory of Social Stru
cture and Action
. Cambridge:
Cambridge University Press
.


Luhmann, N. (2000) ‘Familiarity, Confidence, Trust: Problems and Alternatives’, in
Gambetta, Diego (ed.)
Trust: Making and Breaking Cooperative Relations
, electronic
edition, Department of Sociology
, University of Oxford, chapter 6, pp. 94
-
107,
<http://www.sociology.ox.ac.uk/papers/luhmann94
-
107.pdf>


Putnam, R. (1993)
Making Democracy Work: Civic Traditions in Modern Italy
, Princeton
University Press, Princeton.


Putnam R.
(
2000
)

Bowling Alone
-

The

Collapse and Revival of American Community.
New York: Simon & Schuster.


Putnam R.
(
2001
)
Social capital: Measurement and consequences. . ISUMA Canadian
Journal of Policy Research; 2 (1) 42


51.


Rodin, J. (1996) Control by any other name: Definitions c
oncepts and processes. In
Rodin, J., Schooler, C., and Schaie, K, (eds.),
Selfdirectedness: Cause and effects
throughout the life course.
Hillsdale, NJ: Erlbaum, 131
-
148.


Skinner E.
(
1996
)

A Guide to Constructs of Control.
Journal of Personality and Socia
l
Psychology
, Vol. 71, 549
-
570.


Warren, M.E. (2001) ‘Social Capital and Corruption’. In: Dario Castiglione:
Social
Capital: Interdisciplinary Perspectives
. EURESCO conference on Social Capital, Exeter,
UK, 15
-
20 September 2001, http://www.ex.ac.uk/shipss/
politics/research/socialcapital/



30

TABLES



Table 1. Economic (Employment) Status: Working age Study Area and UK



Women (%)

Men (%)

Study Area

UK

Study Area

UK


1999

2003

1999

2003

1999

2003

1999

2003

Economically
Active









Full
-
time Work

17

21

36

62*

44

47

61

64
*

Part
-
time Work

27

26

27

5

4

5

Self
-
employed

1

0

4

5

4

2

12

14

Unemployed

5

6

4

3

11

13

6

5

Economically
Inactive

50

47

28

27

36

33

16

17

* UK 2003 figures combine F/T and P/T

Source: Gilbertson et al (2005) p 35; NOMIS: http://w
ww.nomisweb.co.uk/






Table 2. Components of Community Trust 2000
-
2004: vertical (1) and horizontal (2) trust dimensions in
the two waves (working age only)



2000

2004

Trusted Party

Trust
Component 1

Trust
Component 2

Trust
Component 1

Trust
Component

2

Local
Politicians

0.88

-
0.06

0.94

-
0.07

Local Council

0.88

-
0.07

0.94

-
0.03

Employers

0.52

0.11

0.39

0.36

Neighbours

0.15

0.58

0.21

0.65

Friends

-
0.01

0.81

0.00

0.82

Family

-
0.11

0.73

-
0.19

0.75


N = 2984

Correlation(1,2) = 0.18

Variance explaine
d = 56%

N = 2431

Correlation(1,2) = 0.24

Variance explained = 64%



31



Table 3

Trust and empowerment: mean scores by gender & employment status



Arithmetic mean score [standard error]


Wave 1 (2000)

Wave 2 (2004)

Employment
status

Workless

[N = 1553]

Em
ployed

[N = 1431]

M
-
W
*

prob
ability

Workless

[N = 1225]

Employed

[N = 1206]

M
-
W
*

prob
ability








Vertical trust
score [3


ㄵN

8.NT

xM.MTz

8.R4

xM.MTz

<0.01

8.21

[0.08]

8.97

[0.07]

<0.01

Horizontal trust
score [3


ㄵN

NP.8N

xM.M4z

N4.22

xM.MPz

<0.01

13.48

[0.05]

13.82

[0.04]

<0.01

Employer trust

[1




P.M9

xM.M2z

P.PR

xM.MPz

<0.01

3.12

[0.03]

3.57

[0.02]

<0.01

Empowerment
score [3


ㄴN

9.R4

xM.MRz

9.T2

xM.MRz

<0.01

9.68

[0.06]

10.16

[0.06]

<0.01








Gender

Males

[N = 1184]

Females

[N = 1800
]

M
-
W
*

prob
ability

Males

[934]

Females

[1497]

M
-
W
*

prob
ability








Vertical trust
score [3


ㄵN

8.NP

xM.M8z

8.RM

xM.MSz

<0.01

8.45

[0.09]

8.67

[0.07]

0.06

Horizontal trust
score [3


ㄵN

NP.99

xM.M4z

N4.M2

xM.MPz

M.9S

NP.R4

xM.MRz

NP.T2

xM.M4z

0.01

Employer trust

score
[1




P.N4

xM.MPz

P.2T

xM.M2z

<0.01

3.30

[0.03]

3.37

[0.02]

0.09

Empowerment
score [3


ㄴN

9.RT

xM.MSz

9.ST

xM.MRz

M.2S

9.TN

xM.MTz

NM.M4

xM.MRz

<0.01








*Significance probabilities: Mann
-
Whitney nonparametric two sample te
st


32


Table 4
.

Relative changes in trust levels between 2000 and 2004 by employment status and gender (main
effects multinomial models).



Adjusted* Odds Ratios [Lower 95% CI, Upper 95% CI]

(bold indicates statistically significant difference)

Explanator
y
variables

Wave 1 (2000)

Wave 2 (2004)

N in
model

VT
:

High

[
N:
611]

base:
Low

[N:
1110]

HT
:

High
[
N:
1477]

base:
Low
[
N:
715]

ET
:

High
[
N:
981]

base:
Low

[
N:

380]

N in
model

VT
:
High
[
N:
670]

base:
Low
[
N:
900]

HT
:

High
[
N:
943]

base:
Low
[
N:
859]

ET
:

High
[
N:
970]

base:
Low
[
N:
242]

model 1a









Employment
Status









Workless

1553

0.85

[0.68

to

1.06]

0.66

[0.54 to
0.82]

0.64

[0.49

to

0.84]

1219

0.66

[0.53

to

0.83]

0.80

[0.65

to

1.00]

0.34

[0.25

to

0.48]

Employed

1431

1.00

1.00

1.00

1203

1.00

1.00

1
.00










Gender









Male
s

1184

0.86

[0.70 to
1.06]

0.76

[0.62 to
0.92]

0.57

[0.44 to
0.73]

929

0.86

[0.69 to
1.06]

0.62

[0.51 to
0.76]

0.58

[0.43 to
0.78]

Female
s

1800

1.00

1.00

1.00

1493

1.00

1.00

1.00










Separately by

E
mployment statu
s

model 1b

Workless only









Male
s

567

1.02

[0.75 to
1.37]

0.76

[0.58 to
1.00]

0.65

[0.45 to
0.94]

428

0.95

[0.69 to
1.30]

0.63

[0.47 to
0.85]

0.48

[0.32 to
0.72]

Female
s

986

1.00

1.00

1.00

791

1.00

1.00

1.00










model 1c

Employed only









Male
s

617

0.74

[0.55 to
0.99]

0.75

[0.56 to
1.00]

0.53

[0.38 to
0.75]

501

0.78

[0.58 to
1.05]

0.61

[0.46 to
0.82]

0.76

[0.48 to
1.22]

Female
s

814

1.00

1.00

1.00

702

1.00

1.00

1.00










* Adjusted for respondent age, education, housing tenure and
area of residence.

(VT) vertical trust, (HT) horizontal trust and (ET) employer trust.


33

Table 5

Relation of Empowerment to Relative Trust Levels 2000
-
2004 (main effects multinomial model).


Adjusted* Odds Ratios [Lower 95% CI, Upper 95% CI]

(bold indicat
es statistically significant difference)


Wave 1 (2000)

Wave 2 (2004)

N

VT

HT

ET

N

VT

HT

ET

model 2a









Empowerment









Low

(3
-
8)

787

0.19

[0.14 to
0.25]

0.39

[0.30

to

0.49]

0.32

[0.23

to

0.44]

497

0.21

[0.15

to

0.28]

0.27

[0.20 to

0.36]

0
.21

[0.14 to

0.32]

Moderate


(9
-
10)

1087

0.47

[0.38 to

0.60]

0.62

[0.49, to
0.77]

0.59

[0.44 to

0.78]

857

0.27

[0.21 to

0.35]

0.35

[0.28 to

0.45]

0.37

[0.25

to

0.53]

High

(11
-
14)

1110

1.00

1.00

1.00

994

1.00

1.00

1.00










Males

1184

0.89

[0.72 to
1.10]

0.77

[0.63 to
0.94]

0.58

[0.45 to
0.75]

886

0
.95

[0
.75

to
1.19
]

0.69

[0
.55

to
0
.85
]

0.65

[0
.47

to
0
.89
]

Females

1800

1.00

1.00

1.00

1462

1.00

1.00

1.00










Workless

1553

0.88

[0.70 to
1.10]

0.68

[0.55 to
0.84]

0.65

[0.50 to
0.85]

1180

0.72

[0
.57

to
0
.92
]

0
.85

[0
.68

to
1.07
]

0.34

[0
.25

to
0
.48
]

Employed

1431

1.00

1.00

1.00

1168

1.00

1.00

1.00










Separately by

Employment status

model 2b

Workless

only









Empowerment
Low (3


8)

435

0.22

[0.15 to
0.33]

0.37

[0.27 to
0.51]

0.30

[0.1
9 to
0.48]

287

0.16

[0
.10

to
0
.25
]

0.25

[0
.17

to
0
.36
]

0.10

[0
.05

to
0
.18
]

Empowerment

Moderate

(9


10)

565

0.57

[0.41 to
0.78]

0.60

[0.45 to
0.82]

0.64

[0.42 to
0.98]

435

0.28

[0
.20

to
0
.41
]

0.43

[0
.31

to
0
.60
]

0.27

[0
.16

to
0
.46
]

Empowerment

High (11



14)

553

1.00

1.00

1.00

458

1.00

1.00

1.00

Males

567

1.02

[0.75 to
1.38]

0.77

[0.58 to
1.01]

0.64

[0.44 to
0.93]

407

1.14

[0
.81

to
1.59
]

0.71

[0
.52

to
0
.98
]

0.57

[0
.37

to
0
.88
]

Females

986

1.00

1.00

1.00

773

1.00

1.00

1.00










model 2c

Employed

only









Empowerment
Low (3


8)

352

0.16

[0.10 to
0.24]

0.39

[0.27 to
0.57]

0.34

[0.22 to
0.54]

210

0.29

[0
.19

to
0
.45
]

0.33

[0
.22

to
0
.51
]

0.48

[0
.25

to
0
.92
]

Empowerment

Moderate

(9


10)

522

0.38

[0.27 to
0.54]

0.61

[0.43 to
0.87]

0.53

[0.35 to
0.79]

422

0.26

[0
.18

to
0
.37
]

0.29

[0
.21

to
0
.41
]

0.51

[0
.29

to
0
.89
]

Empowerment

High (11


14)

557

1.00

1.00

1.00

536

1.00

1.00

1.00

Males

617

0.80

[0.59 to
1.09]

0.79

[0.59 to
1.06]

0.56

[0.40 to
0.80]

479

0
.81

[0
.59

to
1.11
]

0.66

[0
.49

to
0
.90
]

0
.81

[0
.50

to
1.32
]

Females

814

1.00

1.00

1.00

689

1.00

1.00

1.00

model 2a

* Adjusted for age, education, housing tenure, area of residence.


34



Table 6
Relation of components of Empowerment to Relative Trust Levels 2000
-
2004 by employment
status (main effects

multinomial model).


Adjusted* Odds Ratios [Lower 95% CI, Upper 95% CI]

(bold indicates statistically significant difference)

Empowerment
components

Wave 1 (2000)

Wave 2 (2004)

N

VT

HT

ET

N

VT

HT

ET

Workless

model 3a









Not informed

817

0.39

[
0
.29

to
0
.53
]

0.49

[0
.38

to
0
.64
]

0.48

[0
.33

to
0
.69
]

593

0.28

[0
.20

to
0
.39
]

0.30

[0
.22

to
0
.41
]

0.31

[0
.20

to
0
.49
]

Dissatisfied
with control

472

0.49

[0
.35

to
0
.68
]

0
.80

[0
.61

to
1.05
]

0.34

[0
.23

to
0
.51
]

381

0.57

[0
.39

to
0
.83
]

0.57

[0
.40

to
0
.81
]

0.3
4

[0
.20

to
0
.56
]

No influence

532

0.68

[0
.50

to
0
.94
]

0.59

[0
.45

to
0
.78
]

0
.96

[0
.65

to
1.43
]

491

0
.78

[0
.55

to
1.11
]

0
.84

[0
.61

to
1.16
]

0.53

[0
.33

to
0
.84
]










Males

567

1.00

[0
.74

to
1.36
]

0.75

[0
.57

to
1.00
]

0.63

[0
.43

to
0
.92
]

428

1.12

[0
.80

t
o
1.56
]

0.71

[0
.51

to
0
.97
]

0.58

[0
.38

to
0
.90
]

Females

986

1.00

1.00

1.00

791

1.00

1.00

1.00










Employed
model 3b









Not informed

721

0.35

[0
.25

to
0
.47
]

0.57

[0
.42

to
0
.76
]

0.64

[0
.45

to
0
.90
]

516

0.34

[0
.25

to
0
.47
]

0.37

[0
.27

to
0
.50
]

0.
45

[0
.27

to
0
.75
]

Dissatisfied
with control

378

0.56

[0
.39

to
0
.81
]

0
.73

[0
.53

to
1.01
]

0
.68

[0
.46

to
1.01
]

273

0.56

[0
.38

to
0
.83
]

0.63

[0
.43

to
0
.91
]

0
.72

[0
.40

to
1.30
]

No influence

452

0.49

[0
.35

to
0
.69
]

0
.75

[0
.55

to
1.02
]

0.68

[0
.47

to
0
.99
]

390

0
.98

[0
.70

to
1.38
]

0
.85

[0
.61

to
1.18
]

1.05

[0
.62

to
1.80
]










Males

617

0
.78

[0
.57

to
1.05
]

0
.77

[0
.58

to
1
.04
]

0.55

[0
.39

to
0
.78
]

501

0
.88

[0
.65

to
1.20
]

0.67

[0
.50

to
0
.90
]

0
.82

[0
.51

to
1.33
]

Females

814

1.00

1.00

1.00

702

1.00

1.00

1.00










*Adjusted for age, education, tenure, area of residence
.


35


Table 7
Relation of components of Empowerment to Relative Trust Levels 2000
-
2004 by gender (main
effects multinomial model).


Adjusted* Odds Ratios [Lower 95% CI, Upper 95% CI]

(bold indicat
es statistically significant difference)

Empowerment
components

Wave 1 (2000)

Wave 2 (2004)

N

VT

HT

ET

N

VT

HT

ET

Males
model 4a









Not informed

591

0.38

[0
.27

to
0
.53
]

0.44

[0
.32

to
0
.59
]

0.62

[0
.43

to
0
.89
]

451

0.33

[0
.23

to
0
.48
]

0.31

[0
.22

t
o
0
.44
]

0.21

[0
.12

to
0
.37
]

Dissatisfied
with control

360

0.47

[0
.32

to
0
.70
]

0
.93

[0
.67

to
1.30
]

0.64

[0
.42

to
0
.97
]

300

0.60

[0
.40

to
0
.90
]

0.58

[0
.40

to
0
.86
]

0
.89

[0
.51

to
1.55
]

No influence

396

0.38

[0
.26

to
0
.57
]

0.56

[0
.41

to
0
.77
]

0.64

[0
.43

to
0
.96
]

366

0
.79

[0
.54

to
1.17
]

0
.80

[0
.55

to
1.17
]

0
.60

[0
.35

to
1.03
]










Workless

567

0
.94

[0
.66

to
1.36
]

0.62

[0
.44

to
0
.87
]

0
.72

[0
.48

to
1.07
]

428

0.67

[0
.46

to
0
.98
]

0
.89

[0
.62

to
1.29
]

0.22

[0
.13

to
0
.38
]

Employed

617

1.00

1.00

1.00

501

1.00

1
.00

1.00










Females

model 4b









Not informed

947

0.36

[0
.27

to
0
.48
]

0.59

[0
.46

to
0
.76
]

0.50

[0
.35

to
0
.72
]

658

0.30

[0
.22

to
0
.40
]

0.34

[0
.26

to
0
.44
]

0.57

[0
.38

to
0
.88
]

Dissatisfied
with control

490

0.55

[0
.39

to
0
.76
]

0.69

[0
.53

to
0
.91
]

0.40

[0
.28

to
0
.58
]

354

0.50

[0
.35

to
0
.72
]

0.63

[0
.45

to
0
.88
]

0.30

[0
.19

to
0
.48
]

No influence

588

0
.76

[0
.56

to
1.03
]

0.72

[0
.56

to
0
.94
]

0
.88

[0
.61

to
1.28
]

515

0
.95

[0
.70

to
1.31
]

0
.82

[0
.61

to
1.10
]

0
.81

[0
.52

to
1.26
]










Workless

986

0
.77

[0
.57

to
1.03
]

0.69

[0
.52

to
0
.90
]

0.59

[0
.41

to
0
.86
]

791

0.70

[0
.52

to
0
.95
]

0
.84

[0
.63

to
1.12
]

0.44

[0
.28

to
0
.69
]

Employed

814

1.00

1.00

1.00

702

1.00

1.00

1.00










*Adjusted for age, education, tenure, area of residence
.


36

Table 8 GLM of empl
oyer trust against (changed) employment status, by gender
.


Working age at
wave 1(2000) &
wave 2 (2004)

[N = 586]


Employer Trust Change


Men [N]

Women [N]


Adjusted*
estimate

significance

Adjusted*
estimate

significance

Employed to
workless

0.456 [
16]

0.216

-
0.880


[38]

<0.001

Workless

0.204 [79]

0.504

-
0.542

[150]

0.001

Employed

-
0.186 [68]

0.531

-
0.320

[158]

0.051

Workless to
employed (base)

0 [14]

NA

0 [63]

NA

*Adjusted for respondent age. Sub
-
sample sizes are given in square bra
ckets.


Table 9 GLM of VT against (changed) employment status, by gender
.


Working age at
wave 1(2000) &
wave 2 (2004)

[N = 586]


Vertical Trust Change


Men [N]

Women [N]


Adjusted*
estimate

significance

Adjusted*
estimate

significance

Employed to
wor
kless

0.687 [16]

0.541

-
1.224

[38]

0.029

Workless

1.275 [79]

0.172

-
1.029

[150]

0.012

Employed

-
0.039 [68]

0.966

-
0.643 [158]

0.112

Workless to
employed (base)

0 [14]

NA

0 [63]

NA

*Adjusted for respondent age. Sub
-
sample sizes are given
in square brackets.



Table 10. GLM for changed employer trust related to change in experiential factors by gender.


Explanatory variables

(working age)

Outcome: change Employer trust scale


(wave 2


wav攠NF

䵥n

tomen

Adjus瑥tG
敳瑩m慴a

卩pnif楣i
n捥

Adjus瑥tG
敳瑩m慴a

卩pnif楣in捥






Chang攺e汥le氠lf
“informedness”

M.M8N

M.222

M.M44

M.PR4

Chang攺es慴asf慣瑩tn w楴h 汥ve氠
of 捯n瑲ol

M.M2N

M.TMM

0.105

0.025

Change: degree of influence

-
0.033

0.581

-
0.020

0.638

*Adjusted for base model explana
tory variables: age, (gender). Note that tenure, NVQ level: both
consistently not significant.


37

Table 11 GLM for changed employer trust related to change in experiential factors and employment status
(women only)

.

Working age at
wave 1 & wave 2

[N = 406]
women

Frequency

Employment (FT, PT, self): change from wave 1 to wave 2,
experiential change &
employer
trust change



Adjusted* estimate

Significance

Employed to
workless

38

-
0.467

0.014

Workless to
employed

61

0.461

0.003

No change (base)

307

0

NA


Change: level of “informedness”

M.M4M

M.P9S

Chang攺es慴asf慣瑩tn w楴h 汥ve氠lf
捯n瑲ol

0.118

0.011

Change: degree of influence

-
0.013

0.748


Table 12
Main effects
multinomial

models: Adjusted Odds Ratios*
-

trust change: local employers


Working age

N

=
586



Adjusted Odds ratios
with
lower & upper 95% CI

Trust change: 2000 & 2004 [N]

Variables

Categories

N in
model

Local
politicians

Decrease [165]

base:

Increase [199]

Local council
Decrease [146]

base:

Increase [212]

Neighbours

Decrease [112]

base:

I
ncrease [160]

Horizontal

Decrease [125]

base:

Increase [174]

Trust in local
employers:

2000 & 2004

Decreased

129

1.43

[0
.80

to
2.56
]

1.54

[0
.84

to
2.81
]

2.07

[
1.02

to
4.23
]

1.81

[0
.91

to
3.57
]

Unchanged

276

1.41

[0
.85

to
2.31
]

2.23

[
1.33

to
3.75
]

1.47

[
0
.83

to
2.60
]

1.12

[0
.66

to
1.92
]

Increased

181

1.00

1.00

1.00

1.00








Gender

Male
s

177

0
.87

[0
.54

to
1.40
]

0
.70

[0
.42

to
1.17
]

0
.93

[0
.51

to
1.68
]

1.08

[0
.62

to
1.88
]

Female
s

409

1.00

1.00

1.00

1.00











Empowerment factors change: 2000 &

2004 [N]




I
nformed

Decrease [129]

base:

Increase [181]

Control

Decrease [113]

base:

Increase [228]

Influence

Decrease [168]

base:

Increase [176]

Overall

Decrease [155]

base: large

increase [127]

Trust in local
employers:

2000 & 2004

Decreased

129

2.00

[
1.11

to
3.63
]

2.51

[
1.28

to
4.91
]

1.08

[0
.60

to
1.93
]

2.76

[
1.42

to
5.35
]

Unchanged

276

0
.89

[0
.53

to
1.50
]

2.35

[
1.31

to
4.23
]

1.12

[0
.68

to
1.84
]

2.30

[
1.31

to
4.05
]

Increased

181

1.00

1.00

1.00

1.00








Gender

Male
s

177

0
.84

[0
.49

to
1.42
]

1.
81

[
1.08

to
3.03
]

1.00

[0
.61

to
1.65
]

1.00

[
0.58

to
1.73
]

Female
s

409

1.00

1.00

1.00

1.00








*Adjusted for age, education, tenure
.