Hence, descriptive statistics enable a concise portrayal of
the information regarding
measurements, for example rate, frequencies, means and standard deviations. Inferential
statistics/measurements go further. While descriptive st
atistics describe a sample’s
attributes based on the information gathered from respondents, inferential measurements
are utilised to obtain information about the population from which the sample was drawn
based on the data outlined in the descriptive measurements (Denzin & Lincoln, 2011)
Mind map 6.1 list the different types of quantitative data and their characteristics.
Mind map 6.1: Types of quantitative data
In analysing quantitative data, it is important to understand the inherent nature of the data
to collect, because this will determine the type of analysis that is appropriate and legitimate
in the context of the study and the data collection
Four categories of quantitative data:
Nominal measures are descriptive measures that serve only to indicate the alternative
states of the variable. For example, in measure of gender, a respondent may be male or
female or
in measure of religion, a respondent may be Christian, Jewish, Hindu, Moslem,
78
Buddhist or belong to some other unspecified faith. The respondent is either male or
female, either a Christian or not a Christian.
Nominal measures in this manner have the particular qualities of thoroughness and shared
restrictiveness. They represent the most reduced conceivable dimension of measurement.
In endeavouring to evaluate such information, try to convert it and express the findings/
discoveries as frequencies or percentages. Although it is usual practice to appoint numerical
codes to these attributes to facilitate computer analysis, these numbers cannot be used for
mathematical purposes.
Ordinal
measures are rank estimates, usually reflect choices made by the subject or
categories predetermined by the researcher. Ordinal measures
can be logically rank-
ordered and the different attributes represent relatively more or less of the specific variable.
As with nominal data, there is little scope for treating such data mathematically.
Frequencies and percentages are calculated, but not the arithmetic mean and statistical
analysis which are largely precluded.
Interval measures refer to those variables of which the attributes are not only rank-ordered,
but they are separate by a uniform distance between them. An example would be the IQ
scale or temperature scales. Interval measures allow some degree of mathematical and
statistical treatment. Thus, the arithmetic mean IQ of a group of respondents with a range
of IQs can be calculated and correlated and regression analyses can be carried out.
Ratio measures are based on an absolute scale, which has a fixed zero point. This means
that the scale readings are exactly proportional to the variables being measured.
Ratio
measures represent the highest possible level of precision and are amenable to all forms
of statistical analysis. Saunders et al. (2003:328) stated that quantitative data can also be
divided into two distinct categories:
• Categorical data cannot be measured numerically but can be classified into sets
(categories) according to specified criteria (e.g. gender,
religion, profession,
qualification) or placed in rank order (e.g. level
of experience, consumer preference,
etc.). Nominal and ordinal data fall into this group.
• Quantifiable data is data of which the values can be measured numerically. The more
precise the measurement the greater the range of statistical techniques that can be
used to analyse the data. Interval and ratio measures fall into this group. Generally it
is better in quantitatively-oriented studies to collect data that enables highest possible
level of statistical analysis.
Dostları ilə paylaş: