Transcript 4. DATA ANALYSISx

DATA ANALYSIS
and
INTERPRETATION
DR. KHAIRUL FARIHAN KASIM
SCHOOL OF BIOPROCESS ENGINEERING
UNIVERSITI MALAYSIA PERLIS
Recaps: Methodology
Methods How did you carry out your work?
• Clear subheadings
• Describe methods in the past tense
• Methods must be described in sufficient detail so
that other researchers can reproduce the experiment
• Describe statistical tests used
• Include setup figures if necessary
Materials and Method Example
3.1 Microbial strains and media
Escherichia coli NovaBlue (Novagen, Inc., Madison, WI) was
used as the host strain for recombinant DNA manipulation. E.
coli was grown in Luria-Bertani medium (10 g/L peptone, 5
g/L yeast extract, and 5 g/L sodium chloride) containing 100
mg/L ampicillin. S. cerevisiae strains were routinely cultivated
at 30ºC in synthetic medium [SD medium; 6.7 g/L yeast
nitrogen base without amino acids (Difco Laboratories,
Detroit, MI), 20 g/L glucose] supplemented with appropriate
amino acids and nucleotides, and in YPD medium (20 g/L
peptone, 10 g/L yeast extract, 20 g/L glucose).
CHAPTER 4: RESULTS AND DISCUSSION
• Presents the findings of the study.
• Presentation should be clear and scholarly done and may come
in the form of tables, figures or charts.
•
Tables and graphs are both ways to organize and arrange data so that it is
more easily understood by the viewer.
•
Tables and graphs are related in the sense that the information used in
tables is frequently also used for the basis of graphs.
•
Analysis refers to the skill of the researcher in describing,
delineating similarities and differences, highlighting the
significant findings or data and ability to extract information or
message out of the presented data.
•
Interpretation is the explanations or suggestions inferred from
the data, their implications but not conclusions.
PRESENTATION OF FINDINGS – HOW??
• VERBAL
•
•
Describes
Explain
• SYMBOLIC
•
•
•
Graphic
Tables/Graphs
Statistical values
Preparing the Tables and
Figure
The rules
•
•
•
•
•
keep format clear and simple.
line up decimal places, note units clearly,
use a large enough typeface
construct a clean orderly arrangement of rows and columns.
Do not construct a table/figure unless repetitive
data must be presented
How to design effective tables
• Two of the columns give standard conditions, not variables and not data
• If temperature is a variable in the experiment, it can have a column
• If all experiments were done at the same temperature, this information should be
noted in Materials and Methods
• The data can easily be presented in text:
“Aeration of the growth medium was essential for the frowth of Streptomycetes
coelicolor. At room temperature (24C), no growth was evident in stationary (unaerated)
cultures, whereas substantial growth (OD, 78 Klett units) occurred in aerated cultures.”
• Has no columns of
identical readings and
looks like a good table
• The independent
variable column (temp)
looks reasonable
enough
• The dependent variable
(growth) has a
suspicious number of
zero
• You should question any
table with large number
of zero or a large
number of 100s when
percentage are used
“ The oak seedlings grew at
temperatures below 20°C and
40°C; no measurable growth
occurred at temperature below
20°C or above 40°C.”
“S. griseus, S. coelicolor, S. everycolor, and S. rainbowensky grew
under aerobic conditions, whereas S. nocolor and S. greenicus
required anaerobic onditions.”
• Not all numerical data must be put in a table
“The difference between the failure rates – 14% (5 of 35) for nocillin and
26% (9 of 34) for potassium penicillin V – was not significant (P=0.21).”
Tips
• In presenting numbers, give only significant figures. Nonsignificant
figures may mislead the reader by creating a false sense of
precision.
• Unessential data, such as laboratory numbers, results of simple
calculations, and columns that show no significant variations,
should be omitted.
• Present the data in the text, or in a table, or in a figure.
• Never present the same data in text, or in a table, or in a figure
• However, selected data can be singled out for discussion in the text.
How to arrange the data?
• The data can be presented either horizontally or vertically
• But “can” does not mean “should”; the data should be organized so
that the like elements read “down”, not across
Read across
Read down
• Its read down, not across.
Well-construct table
• It has headings that are clear
enough to make the meaning of
the data understandable
without reference to the text.
• It has explanatory footnotes, but
they do not repeat excessive
experimental detail.
• Note the distinction. It is proper
to provide enough information
so that the meaning of the data
is apparent without reference to
the text, but it is improper to
provide in the table the
experimental detail that would
be required to repeat the
experiment.
Exponents in table heading
• If possible, avoid using exponents in heading table
• Why – to avoid confusion
• ‘cpm x 103’ and ‘cpm x 10-3 refer to thousands of counts per minute
• If it is not possible to avoid, if may be worthwhile to state in a
footnote (or in figure legend), in words that eliminate the ambiguity
Titles, footnotes, and abbreviations
• Title of a table (or legend of a figure) is like the title of your thesis.
• Should be concise and not divided into two or more clauses or
sentences.
• Unnecessary words should be omitted.
• Give careful thought to the footnotes to the tables.
• If abbreviations must be define, you should list the abbreviations
used in abbreviations list.
How to Prepare Effective
Graphs
When to use graphs
• Graphs are very similar to tables as a means of presenting data in
an organized way
• In fact, the results of many experiment either as tables or as graphs
• How to decide which is preferable? – difficult decision
• A good rules might be:
•
•
•
if the data show pronounced trends, making an interesting picture, use a
graph
If the numbers just sit there, with no exciting trend in evidence, a table
should be satisfactory
Tables are also preferred for presenting exact numbers
• Figure 1 could be replace by one
sentences in the text
“Among the test group of 56
patients who were hospitalized for
an average of 14 days, 6 acquired
infections.”
• Compare Table 9 and Figure 2
• Figure 2 clearly seems superior to Table 9
• In the figure, the synergistic action of the two-drug combination is
immediately apparent
• The reader can quickly grasp the significance of the data
• It also appears from the graph that streptomycin is more effective
than is isoniazid, although its action somewhat slower; this aspect
of the results is not readily apparent from the table
Example of a nice graph
• The lettering was large enough
• It is boxed, rather than twosided (compare with Figure 2),
making it a bit easier to
estimate the values on the
right-hand side of the graph
• The scribe marks point inward
rather than outward
Tips
• Be consistent from graph to graph
•
If you are comparing interventions, keep using the same symbol for the
same intervention
• Do not extend the ordinate or the abscissa beyond what the graph
demands
•
•
•
If your data points range between 0 and 78, your topmost index number
should be 80.
You might feel a tendency to extend the graph to 100, a nice round number
(especially if the data points are percentages)
Your reference numbers should be 0, 20, 40, 60, and 80.
Symbols and legends
• You must define the symbols in the figure legend
• You should use only those symbols that are considered standard
and that are widely available (○, Δ, □, ●)
• Different types of connecting line (solid, dashed) can also be used
• But, do not use different types of connecting line and different
symbols
Other type of graphs
Desirability
Design-Expert® Software
1.00
Desirability
Design Points
1
0
Actual Factor
C: Water volume = 1.00
Prediction
B : Tim e
X1 = A: Temperature
X2 = B: Time
0.50
0.299
0.00
0.449
0.599
0.898
0.748
0.150
-0.50
-1.00
-1.00
-0.50
0.00
A: Temperature
0.50
1.00
How to Prepare Effective
Photographs
• Significant
• Clear
• High-quality
• Crop the features of special interest
What is missing in these three photograph?
DATA ANALYSIS
The purpose
• To answer the research questions and to help determine the trends and
relationships among the variables.
STEPS IN DATA ANALYSIS
• BEFORE DATA COLLECTION
•
•
•
Determine the method of data analysis
Determine how to process the data
Prepare dummy tables
• AFTER DATA COLLECTION
•
•
•
•
•
Process the data
Prepare tables and graphs
Analyze and interpret findings
Consult supervisor/plv/etc
Prepare for editing & presentation
KINDS OF DATA ANALYSIS
• 1. Descriptive analysis
Refers to the description of the data from a particular sample; hence the
conclusion must refer only to the sample.
• In other words, these summarize the data and describe sample
characteristics.
•
• 2. Inferential analysis
•
The use of statistical tests, either to test for significant relationships among
variables or to find statistical support for the hypotheses.
Classification of Descriptive Analysis
• 1. Frequency Distribution
•
A systematic arrangement of numeric values from the lowest to the highest
or vice versa.
• 2.Measure of Central Tendency
•
Average of the set values (mode, median, mean)
• 3. Measure of Variability
•
Statistics that concern the degree to which the scores in a distribution are
different from or similar to each other. (range, standard deviation)
Inferential Analysis
• The use of statistical tests, either to test for significant relationships
among variables or to find statistical support for the hypotheses.
• Inferential Statistics
•
•
ANOVA (significant of differences between means of two or more groups)
T-test
• Hypothesis
•
The outcome of the study perhaps may retain, revise or reject the
hypothesis and this determines the acceptability of hypotheses and the
theory from which it was derived.
INTERPRETATION OF FINDINGS/RESULTS,
IMPLICATIONS AND INFERENCES
• Sufficient data should be used to justify your inferences or
generalization. The implications suggested by the data should be
explained and discussed thoroughly in this portion of your thesis.
• The data analysis involves comparing values on the dependent
measures in statistical cases. In the non statistical approach, these
comparisons usually involve visual inspection of data. Evaluation
depends on projecting from baseline data what findings would be like
in the future if some variables were not experimented.
How to write the discussion
7 key questions to write discussion
1. Does my data agree/disagree with other?
2. Can other people data/hypothesis help to explain my data?
3. Does your data help to explain other people data?
4. What assumption have you made in doing the work and what would
change if you change them?
5. Are there alternative theories as to why the response you observed
occurred?
6. What further knowledge is required for the field (especially for thesis)?
7. What impact does this work have on industry?
Any question?
Thank you