Information Technology Reference
In-Depth Information
quantitative measurement
. Qualitative measurement is what we obtain
from using a nominal scale of measurement. Researchers sometimes call
qualitative variables by other names:
Categorical Variables.
Nonmetric Variables.
Dichotomous Variables (when there are only two values or cate-
gories).
Grouped Variables.
Classification Variables.
It is useful for our purposes to think of quantitative measurement
in a somewhat restrictive manner. Although the ordinal scale certainly
presumes an underlying quantitative dimension, we would generally pro-
pose thinking in terms of those scales for which it is meaningful and
informative to compute a mean. With the ability to compute a mean and
all that this ability implies, the gateway is open to performing a whole
range of statistical procedures such as the ANOVA. Summative response,
interval, and ratio scales meet this standard. Researchers sometimes call
quantitative variables by other names, such as
Continuous Variables.
Metric Variables.
Ungrouped Variables.
2.2 CENTRAL TENDENCY AND VARIABILITY
In the ANOVA designs that we will cover in this topic, the dependent
variables will be measured on a quantitative scale of measurement. We are
therefore able to generate summary statistics that quickly convey a great
deal of information about a set of scores. We will focus on two classes of
summary measures: central tendency and variability.
Measures of
central tendency
provide an index (or single-value sum-
mary) of the most typical score in a set or distribution of scores. We use
a measure of central tendency, the mean, to provide a convenient way to
describe the impact of our independent variable on the dependent mea-
sure. Thus, we might hypothesize that the mean of one group (e.g., the
experimental group) will be greater on the dependent variable than the
mean of another group (e.g., the control group).
Variability
addresses the issue of how scores within a group or treat-
ment condition vary or deviate from one another. Knowledge of variability
(as we will see later in this chapter) helps us gauge whether the manipula-
tion of our independent variable is really producing differences between
or among the means of our treatments, or if these observed differences are
simply due to random or chance fluctuation. The difference between the
means of the experimental and control groups is evaluated with respect
to the variability within each of the two groups.
