STRIKES AND UNORGANISED CONFLICT: SOME FURTHER CONSIDERATIONS
| Author | P. K. Edwards |
| DOI | http://doi.org/10.1111/j.1467-8543.1979.tb00630.x |
| Published date | 01 March 1979 |
| Date | 01 March 1979 |
STRIKES AND UNORGANISED CONFLICT:
SOME
FURTHER CONSIDERATIONS
P.
K.
EDWARDS*
IN
a recent note, Bean used data from the New Earnings Survey to analyse the
relationship between strike frequency and indices
of
‘unorganised’ conflict in manufac-
turing industry.’This paper reports the resultsof some further analyses
of
these data and
then considers the general significance
of
the findings.
The
1970
New Earnings Survey asked employers how many workers lost pay through
absence and the reason for this absence.
The
results were classified under several
headings, such as holidays, voluntary absence, late arrival at work, and certified and
uncertified sickness. But, because some workers may have worked less than their basic
hours without
loss
of
pay, ‘the survey results do not indicate the full incidence
of
sickness
and other types
of
absence
or
short-time working’.* Bean used the classification
of
the
results by industry to compare, for sixteen manufacturing industries,, the incidence
of
three measures
of
absence with the number
of
strikes per
100,000
employees. The
positive correlations between strike frequency and the measures
of
absence, and among
the absence measures, indicated that ‘organised’ and ‘unorganised’ conflict were addi-
tively related.
In
other words, industries with a high strike rate tended to be high
on
the
other measures
of
‘conflict’.
The three measures used by Bean were voluntary absence, certified sickness, and late
arrival for work. But there are difficulties in using the last two as indices
of
conflict.
Certified sickness covers the relatively
long
periods
of
absence for which a doctor’s
certificate
is
required. It therefore includes a substantial proportion
of
‘genuine’ illness
and cannot be seen as an unambiguous index
of
‘conflictuality’. Shorter, uncertified,
periods
of
absence may provide a more suitable index? The ‘late arrival’ measure
includes early finishing as well as lateness and does not necessarily measure withdrawals
from work which can be seen as conflictual. There is little that can be done about this,
but the problems with certified sickness can be dealt with by considering uncertified
sickness as well. It is also useful to use worker involvement and days lost in strikes, in
addition to the number
of
stoppages, as indices
of
strike activity. Finally, there is no
reason to restrict analysis
to
manufacturing industry.
Rank-order correlations4 for seven indices across twenty-six industry groups are given
in Table
1.
The results generally support Bean’s conclusion that there is an additive
relationship between measures
of
organised and unorganised conflict. Moreover, the
indices
of
unorganised conflict are directly associated with each other. However, not all
the correlations are significant, which suggests that ‘unorganised conflict’ cannot be
treated as a homogeneous phenomenon.
In
particular, the indices of certified and
uncertified sickness are not associated, indicating that they reflect different forces.
Certified sickness is more strongly associated with other indices, which goes against
expectation. This can be attributed to the unsatisfactory nature
of
the uncertified
sickness measure, which is a residual category, and to the way the data are constructed.
Long-term certified sickness is more likely to involve
loss
of
earnings, and therefore to
be included in the figures, than is uncertified sickness, which may involve only brief
absences. Data
on
the former are therefore more statistically valid.
It is necessary
to
go
beyond these interrelationships and to examine possible causal
influences. Table
1
reports the correlation coefficients between the unemployment rate
and average earnings and the seven indices
of
‘conflict’. The unemplc,yment rate is
*
Research Fellow, S.S.R.C.
Industrial
Relations Research
Unit,
University
of
Warwick.
95
To continue reading
Request your trial