QI in organizational work groups ICA ‘04
12
assessing measuring organizational characteristics related to QI were in the first section of the
instrument, those measuring communication, commitment, and identification came next, followed by
items measuring work unit structure and outcome variables. Job stress, communication networks,
response to CQI and demographics were measured in the final part of the questionnaire.
Measurement
Items measuring each of the constructs in the study were incorporated into the questionnaire. When
possible, multiple-item indices were utilized for each construct, but due to restrictions on the length of
the instrument, some constructs were measured with a single item. As a check that the indices tapped
distinct, independent constructs, principal component analyses (equamax rotation) were run for the
structural and process variables. Although a few items had to be dropped or combined into new
indices, for the most part the constructs were independent and items loaded on the indices they were
intended for. Judgments were made based on reliability estimates (alphas) and previous research using
this set of variables (cite withheld for blind review) and are explained below. Final items for the
questionnaire are available from the first author on request.
The eight structural variables initially loaded onto four separate components that explained
70.9% of the variance in the items. However, four variables that loaded onto two components proved
to be both analytically and theoretically distinct – alpha coefficients were .33 and .14 respectively and
these had been independent items in prior research. As such, a second principal component analyses
(equamax rotation) were run whereby six factors were extracted, explaining 90.4% of the variance.
The final indices were as follows. Task variability was measured by two items developed by Van de
Ven and Ferry (1980) that assessed the degree to which individuals performed the same tasks and
faced the same problems day to day (
α
= .69). Task interdependence was assessed by a single item
that measured the extent to which members work on their own tasks independently (Van de Ven and
Ferry, 1980). Rate of change in unit processes was measured with a single item, “The way work is
done in my work unit is reorganized often.” External communication was assessed by a single item
Convention