used may be 1=male and 2=female or for an item which uses a Likert scale, 1=strongly
agree, 2=agree, 3-uncertain, 4=disagree and 5=strongly disagree. All missing data should
be indicated through codes. For example, if a respondent did not indicate his/her sex, the
code 9 may be used. Items with missing data are then excluded from subsequent analysis
of data.
Statistical analysis
Once the data has been coded into a format that can be entered on a spreadsheet or a
statistical analysis package, the appropriate procedure may be used to process the data
into a format that can be analysed (frequencies, tables, diagrams, etc.).
The tables and diagrams most relevant in addressing the needs of the study will eventually
app
ear in the student’s/researcher’s dissertation. Tables and diagrams can be used in an
exploratory analysis
of data to identify trends, show proportions
and the distribution of
values and to visually compare the relationship among variables.
With descriptive statistics, the researcher can describe (and compare) variables numerically
with means and standard deviations, while with inferential statistics,
the researcher can
reach conclusions about how the data collected relates to the original research objectives
and hypothesis and how these results might be generalised to the research population.
Guidelines for the successful application of statistics in the analysis of data
1. Statistics must be appropriate to integrate into the study
and not use as a device to
create an impression of scientific analysis and objectivity.
2. The researcher must develop a good working knowledge of the principles underlying the
statistical procedures in the study to analyse data to enable appropriate use and meaningful
interpretation of the results.
3. The researcher who has access to statistical software
packages such as SPSS for
Windows should first work through examples of coding and analysing research instruments
before proceeding with an analysis of the research data.
4. Before the researcher embarks on any kind of statistical analysis, they should determine
the characteristics of the data in terms of its type (i.e. nominal, ordinal, interval or ratio) and
its distribution (normal or skewed) since these inherent properties will determine what kind
of statistical test is possible to use.
Dostları ilə paylaş: