(^128) Johns Hopkins Nursing Evidence-Based Practice: Model and Guidelines, Third Edition
asked is, “Can the results be replicated if the test is repeated?” Reliability refers,
in essence, to the repeatability of a test. For example, variation in measurements
may exist when nurses use patient care equipment such as blood pressure cuffs
or glucometers. Two methods used to measure reliability include test-retest reli-
ability and interrater reliability. Test-retest reliability is done by administrating
a measure to the same people on two occasions. If the difference in a person’s
scores is small, reliability is high. A second method to test reliability is interrater
reliability, which involves having two or more observers independently apply an
instrument to the same people.
Measures of Precision
Precision language describes populations and characteristics of populations. An
EBP team should be quite familiar with measures of central tendency (mean,
median, and mode) and variation (standard deviation). The mean denotes the
average value and is used with numerical data. Although a good measure of cen-
tral tendency in normal distributions, the mean is misleading in cases involving
skewed (asymmetric) distributions and extreme scores. The median, the number
that lies at the midpoint of a distribution of values, is less sensitive to extreme
scores and is, therefore, of greater use in skewed distributions. The mode is the
most frequently occurring value and is the only measure of central tendency used
with categorical data. Standard deviation is a measure of variability that denotes
the spread of the distribution and indicates the average variation of values from
the mean.
Another measure of precision is statistical significance, which indicates whether
findings are due to chance. The classic measure of statistical significance, the p-
value, is a probability range from 0 to 1. The smaller the p-value (the closer it is
to 0), the more likely the result is statistically significant and reflects an actual
association or difference between variables/groups. The p-value is influenced by
two factors: the sample size and the magnitude of the difference between groups
(effect size) (Sullivan & Feinn, 2012). For example, if the sample size is large
enough (e.g., 8,000), the results will almost always show a significant p-value,
even if the effect size is small or clinically insignificant. In these instances, it
may not be justifiable or time efficient to translate these results into practice. In