2. SELF ANALYSIS OF THE RESEARCH
This part III of the research methodology focuses on your explanation
of all the work done in your research and your defense of the findings
and approach. This is critical in scientific work as people will critique
your work and you will be called upon to defend.
You will need to re-read your research and critique the method chosen
as well as show how improvements could be made.
4. Problem Statement
Lack of adequate clothing at the
radiography department can affect
radiographers’ health and lead to poor
performance and improper imaging.
7. Qualitative
A questionnaire was
developed and was used to
collect data from
radiographers at three
different clinical sites i.e.
1.Victoria Diagnostic
Imaging Centre
2. San Fernando General
Hospital
3. Couva District Health
Facility .
9. Quantitative:
Instruments were used to measure, Body temp oxygen
level, & heart pulse rate
The radiography room’s temperature, were measured four
(4) times and the average reading was calculated for each
The radiographer’s height were measured
The radiographer’s waist circumference were measured
10.
11.
12.
13. CRITIQUE
• Critique your final submission draft:
• Problem statement and research questions
• Scope and objectives
• Methodology
• Analysis
• Findings
14. CRITIQUE
• Critique your final submission draft:
• Problem statement and research questions
• Scope and objectives
• Methodology
• Analysis
• Findings
15. FINDINGS
• Show how the results were arrived at or problem was solved logically
and deductively.
• Use block diagrams for this
• How do your findings compare with the references in the lit review?
• What can be done to improve your analysis?
• How will your findings and recommendations made improve society?
• What impact will your findings have on cost or on your work
operations?
• Prepare a short defence for your method used in this study.
16.
17. THE PROBLEM
• What have you identified as the symptoms of the problem?
• What relationships have you found to your problem?
• Can you consider these relationships as causes of the problem?
• What was your intuitive reaction to remedy the problem symptoms?
• Is the problem hypothetically a potential system/policy or HR failure?
• What information was required to obtain in order to start the process
of the research?
• List the multiple levels of personal or system failures that may have
led to the problem.
18. RESEARCH QUESTIONS
• How did these questions impact on the research process?
• How was the problem scoped in order to choose these questions?
• How were these used to manage the lit review?
• Give a flow chart of how the process from problem statement (input)
to research questions (process) and lit reviews (output) occurred.
• Indicate a feed back process from your flow chart showing where
information received from the lit review fed back into your research
questions and problem statement, thereby making you adjust them.
19. DATA COLLECTION
• Observational or experimental and errors that may arise in using these
• Variables used
• Data collection process flow chart
• Margin of error
• Secondary or primary data- advantages and disadvantages
• Precision and accuracy
20. DATAANALYSIS
• What are you testing?
• Predictions
• Analyzing to see if your predictions were right
• What may have gone wrong
• Are there other considerations that are more plausible
• We cannot prove things to be true so you may give alternatives based
on what your data are saying
• Pareto diagram of the data
21. METHODOLOGY
• Matching method approach with reference
• What was different/ same
• Errors that may arise and drawbacks of the approach
• Explain the creativity and newness of your method
• Who are the subjects, population
• Exclusion
• Will you be looking for causes, linkages, differences or all
• Instrument used and how you could excite the subjects to want to be part of
a survey
• Expected outcomes- why
22. TECHNOLOGY
• Discuss the effects of technology on your findings and how if
implemented can improve your work.
• Data science and data engineering drive
• Data recording, storage and management
• Instrument change
23. ORGANIZATIONAL CHANGE
• How will the organizational work flow improve with your
recommendations
• Implementation procedures
• Are you violating standards and regulations
• How are you going to evaluate your new process or methods/
recommendations
• What percent improvements are there in cost, scheduling, workflow
speed, customer and worker satisfaction.
• Cultural and work culture barriers involved. Resistance to change.
24. SUMMARY APPROACH
• Define the problem
• Understand the process
• Identify possible causes
• Set a method for solving
• Collect data
• Analyze data
• Identify possible solutions
• Give recommendations for implementation
• Evaluate effects
• Institutionalize the change
25. DATA SCIENCE
• Effects on business world, medicine, science
• How to generate
• What to generate- what is necessary and needed
• Great cash cow – must fill a need
• Will aid man’s progress
• Longevity
• Illness and hospital prevention
• Efficiency increase and scrap reduction
• Entertainment industry
26. RESEARCH CHALLENGE
• Your research shows a process for addressing a problem or
investigating some point of interest that adds to the body of
knowledge. Research is hard work but build character and purpose,
independence in the real world. You must be encouraged to publish
papers from your own work. There must be some area of interest or
shortcoming that you were not satisfied with in your research, or, some
area you were excited about but did not get the chance to expound.
27. PREDICTIONS
• Based on your data series can you plot a cumulative probability
distribution and predict values based on your recommendations?
Example: patients are visiting for diagnosis and you need to determine
how long each person takes for his visit to get some idea if the
hospital’s policy is working. You measure the time taken for each
person and write it down for 26 persons. This is shown in the slide
overleaf.
29. DATAANALYSIS
• You decide to draw a histogram – place the data down in excel- get cut off
intervals for the groupings of data- data- data analysis-histogram- range and
bin- ok – insert chart – bar graph and adjust –format data series to close in
the chart. Add trend line moving average. Look at the distribution you get
and comment.
• You decide to draw a cumulative probability plot – sort data smallest to
largest- i-0.5/n – plot sorted data (x axis) vs (i-0.5/n)+A1 (y axis).
• You decide to get percentile- data array – PERCENTILE.INC - k is the
percentile you want (0.65, say)- the value is given for that percentile and
represents the score at this percentile meaning that 65% of the data fell
below this score. This corresponds to 35.25 based on your data array.
30. HISTOGRAM
Bin Frequency
0 0
7 4
14 7
21 2
28 3
35 1
42 3
49 4
56 1
63 1
More 0
0
1
2
3
4
5
6
7
8
0 7 14 21 28 35 42 49 56 63 More
Frequency
31. PERCENTILE ANALYSIS
A percentile of 0.65 or 65% corresponds to a time of 35.25 minutes. In
other words; 65% of the visitors will take 35.25 minutes or less. Which
means that 35% will take more that 35.25 minutes. This a prediction
based on the observations taken and we are working with normally
distributed data, which they do approximate.
You can now infer that too many visitors are taking longer than some
stipulated time according to policy. Policy says persons should not wait
more than 20 minutes. For 20 minutes the percent will be 49% implying
that 51% of the people will have to wait longer than 20 minutes.
33. INFERENCE
• We see from the cumulative probability chart that any value we are
interested in, we can get a percentile. The 65 percentile , if we go on
the chart we see that we can point off 35 minutes in the same way as
we did in excel for PERCENTILE.INC.
34. NORMAL PROBABILITY PLOT –
PERCENTILE VS DATA
0
0.2
0.4
0.6
0.8
1
1.2
0 20 40 60
CDF PLOT- i-0.5/n vs data
35. NORMAL PROBABILITY PLOT – QI VALUES
OF PERCENTILES VS DATA (QQ PLOT)
-20
-10
0
10
20
30
40
50
60
70
0 10 20 30 40 50 60
NORMAL PROBABILITY PLOT - Z values vs data
36. ANALYSIS OF PLOTS
• Both normal probability plots look the same and give almost a straight
line. One shows the plot against the percentiles (i-0.5/n) and the other
shows the plot against the Q values , which are the z values of the
percentiles.
37. CONCLUSIONS
• In this example we are measuring time, not counting number of
persons. These are quantitative data. For this you divide the data into
groups without leaving any gaps. The bar charts do not have this
joining but stay as separate groups.
• You may want to match the pattern from your data distribution with
theoretical models and also with those obtained from your references
to discuss.
38. FURTHER
• You went further based on your observations. You are trying to
determine whether the speed of the test experienced by the patient has
any bearing on whether he thought the service provided was good. To
do this you use the 65% percentile and see that 17 persons had less
than 35.25 minutes which you consider fast by international standards.
• Those who experienced less than 35 minutes had a proportion of 10/26
thinking it was good quality. Those who experienced more than 35
minutes had a proportion of 9/26 responding that the quality was good.
Can we conclude that those responding better quality is greater among
those who had fast treatment?
39. DEDUCTIONS
• Depending on your solution you may decide that there is no perception
of better quality service simply based on the speed at which the service
is delivered. This information is of importance to those wishing to
undertake a study of the quality of service to the client. You may have
shown no statistical significance/ or may show that there is one.
Whichever way, researchers can use this information to set a
methodology for future work. They may want to use other criteria and
not waste time retesting what you have already done. The point is that
for them to make the judgement, your work must indicate justification
for its findings.