Transcript Reporting Results of an Analysis

1
PS1006 Lecture 5
Results sections - General format
Sam Cromie
2
This week:
• No Seminars
• No labs
• Just the exam
3
This week’s exam:
• Survey research
• T-tests – independent groups and
repeated measures
• Non-parametric equivalents of t-tests
– Mann Whitney U
– Wilcoxon signed rank
4
The bow-tie shape of a research
report:
5
6
Introduction
Method
Why we did
the study
What
we did
General to
Specific
Results
What we
found
Discussion
What it
means
Specific to
General
7
Introduction
General background > Specific background > Aims and Hypotheses
8
Method
Subjects
Design
Equipment
Procedure
9
Results
Descriptive
Statistics
Inferential
Statistics
10
Discussion
Explain study results > Fit to other research > Implications
11
Results in reports - sequence
• Step 1 - quote descriptive statistics
– Graph or table – need to be discriminatory
• Salient trends identified including those contrary to
hypothesis
• to establish patterns in your data - not just for the sake of it
– Clarify the direction and importance of any potential
effect
• A significant t result
– does not tell us whether values increased or decreased
– Does not tell us if the difference is large or important – just if it
is significant
– Include measure of variability
12
Step 2 – Inferential statistics
• state the purpose of the analysis
– E.g., “a repeated measures t-test was carried
out to determine whether there was any
difference between the pre and post-therapy
anxiety levels”
– helps reader follow the logic of the analysis
– helps you formulate the analysis in your own
head
13
Step 2 – Inferential statistics
• Should always report
– Symbol or name for the statistic used (r, t, T, U)
• Note statistical symbols should always be italicised
– Degrees of freedom or n
– Value of test statistic; t =
– Whether the probability is greater or less than the
alpha value; p < or > .05
– Whether the test was one-tailed or two-tailed
• A few of the more complicated analyses that you will
encounter next year do not have tails e.g., ANOVA.
14
Step 3 – Do not interpret the results
– This is the job of the discussion
– Helps the reader to clearly distinguish what
you found from the interpretation you give to it
15
Step 4 – Summarise the results at
the start of your Discussion
– Without reference to results of statistical tests
e.g., group A was found to be larger than
group B
– Indicate whether your hypothesis was
supported or rejected
16
Language advice
• You can assume the reader has knowledge
of inferential statistics:
– Do not have to mention the null hypothesis
– Critical alpha will be assumed to be 0.05. Only
mention if you use a more or less conservative
value.
– You MUST refer to any difference which you
investigate as being either statistically
significant or non-significant