Evaluation

[bensoltoff]

Evaluation of final project by Benjamin Soltoff

Topic	Excellent	Satisfactory	Needs Work
Coding style		x
Coding strategy		x
Presentation: graphs	x
Presentation: tables	x
Achievement, creativity	x
Ease of access	x

Remarks:

• Great use of hidden code chunks to combine code and output. Enables reader to view the code, only if they want to
• Rather than printing the table of 130 valid trials, you could use in-line R chunks, like this

`r nrow(filtered)` valid trials are used in the remaining analysis.

• What is "AOI"? Is that "Area of Interest"?
• If you wanted to be really fancy, you could also use in-line R chunks to dynamically produce this paragraph:

Planned comparisons against chance (.5) revealed that when using the Body definition, infants launched predictive fixations systematically to the prior goal (M=0.60, SD=.49, t(88)=2.04, p = .04*), whereas, when using the Hand definition, infants responded at chance levels (M=0.48, SD=.50, t(49)=-0.28, p = .78). Therefore, remaining analyses will be performed on data generated from an initial fixation to the Body AOI.
In case your sample changed in future renderings of the website, these stats would update automatically.

• Pretty basic modeling methods, but because you have experimental data those are more than sufficient
• It reads like a journal article - admittedly a brief one which is nice. My discipline (poli sci) tends to have 20-25 page length articles filled with fluff. This is much easier and quicker to the point. Good starting point for a publishable research paper
• Excellent use of the R Markdown documents to store all your code, output, and analysis. Really captures what I was asking for in the project description
• Your raw data was terribly structured. Even if you aren't going to do analysis in R, is that really how your software generates the data? It seems just atrocious. Is there any way to store it better when it is initially generated? Or export it in a better format?
• Tidying.Rmd probably doesn't need to be an R Markdown file. It could have been stored in an R script, then build a pipeline to run that script and then render the website. See your classmates' runfile.R for an example of how to do this