Sei sulla pagina 1di 2

Simple Formatting stuff (5 points)

Title page contains useful Title, authors, class, and date.


Length is appropriate and pages are numbered
Client identified and affiliation given.
Typos and grammar stuff.
Title page is good, and the length is spot on! Many groups struggled to keep their
report under 7 pages. The clients are identified, although it would have been nice
for you to state that they are graduate students in the psychology department.
There are few if any typos or major grammar issues.
Introduces scientific problem and questions of interest stated (5 Points)
This report does a great job identifying the problem at hand and the questions of
interest.
Describes data both in text and figures/tables. (30 points)
No raw code included
Tables and figures numbered and labelled in a practical manner
Data is described in an effective way
Tables and figures are well-formatted, visually appealing, and useful
There is no raw code shown in the report. The tables and figures are well-formatted
and are very useful in the report. Some of the figures are a little confusing though.
One example is Figure 3 and Figure 5; the data seem so full of zero valued
observations that its hard to see much of anything in the histograms or boxplots of
the data. The overwhelming presence of zeroes is a large obstacle here and I hope it
can be overcome quickly next semester. The numbering and labeling of the figures
Is good here too.
Exploratory analysis (30 points)
Useful variable names
Correct analysis
General utility of EDA
Context given and expanded upon
The use of achronyms in this report is helpful, but how you introduce them is a little
off. Generally, Id recommend giving achronyms like the University of Georgia
(UGA) immediately after stating the full name, as written above. There appear to be
some problems in the analysis itself which are never addressed. The first issue
deals with zero-valued voxels. Based on Figure 1, it seems that many voxels are not
even in the brain and are therefore not really useful in your analysis. Since you
never mentioned removing empty voxels, I assume that youve included them in

your analysis, which seems like an error. Also, its a bit of a stretch to talk about
taking the median of two observations. Generally, a median is used to help avoid
problems related to outlying values. When you take the median of two observations
youre really just taking the mean, and presenting this value as a median, while
technically true, implies a level of resistance to outliers which is not actually
present.
Even though Ive outlined what I feel are two pretty meaningful shortcomings
in this EDA, I think that this sections is quite good and Im confident that this group
understands their data better than many other groups. I look forward to reading
this report again after these have been fixed.
One other question that Im confused about: you mention specifically that
you plan omn using non-parametric tests to do voxel-wise comparisons, but then
you talk about visualizing these differences using something called a parametric
map. Isnt it a bit of a misnomer to call that plot a parametric map? Or is there
perhaps a deeper problem with SPMs that actually uses the parametric methods
that you wish to avoid in this analysis?

Writing quality (30 points)


Appropriate language and tone
Written for the appropriate audience
Good perspective (what will be hard, what has already been done)
Future Plans
First, Id like to thank you greatly for your research into fMRI and your inclusion of
sources. Im giving you bonus points in this section because this was a great item
to include and really improved the overall quality of the report. Additionally, I feel
like I have a pretty clear vision of the analysis that you plan to use. One thing which
was a little confusing is that although you mention voxel-wise, cluster-wise, and set
analysis, I dont see any description of your plan for the cluster level analysis. You
should make sure to verify that plan before you spend too much time on this
analysis. The language and tone in this report are appropriate for the intended
audience and perspective is provided and explained well throughout the report!

Potrebbero piacerti anche