Data analysis
Survey data were entered into SPSS (Version 18) and
screened for accuracy and missing values. There was very
little missing data with less than 5% for all variables, except
for approximately 10% for number of shifts. One respondent
with missing data for all items of the PES-NWI was omitted
from the following analyses. Data collected about number of
day, evening, and night shifts were used to create two new
variables that were more informative: the number of total
shifts and the percentage of day shifts per total shifts.
Descriptive statistics were used to summarize sample char-
acteristics and describe the nursing practice environment.
The composite and five subscale scores of the PES-NWI were
calculated according to Lake’s (2002) original instructions.
The scores for all items in each subscale were averaged for
each respondent and the mean score of all subscale scores
was calculated as the composite score of the PES-NWI.
Bivariate associations between sample characteristics and
PES-NWI scores were examined using independent t-tests,
anovas, and Pearson’s correlations. Finally, the associations
between PES-NWI scores and ability to provide quality
nursing care, quality of patient care, and ward morale were
examined using Pearson’s correlations, followed by hierar-
chical multiple regression analyses, which controlled for
nurses’ demographic and work characteristics. In each regres-
sion, variables were entered in three blocks: demographics
(gender, years working as a nurse, education), work charac-
teristics (position, shift type, number of total shifts, percent-
age of day shifts, hours of overtime work, number of patients
in day shifts), and finally the five PES-NWI subscales. For all
analyses, the level of statistical significance was P < 0.05.