EconDec Test Case

The flat, comma-separated data file econdec.csv contains some prepared trial-level data for the Economic Decision Study.

For clarity, I will provide some explanation of the this flat file and the columns you'll need to manipulate. The structure is hierachical and sequential:

• Each subject completed 12 blocks
• Each block consisted of 6 trials
• Each trial is represented in order by a single row in the dataset, like so:

subjnum	block	trial	values
101	1	1	foo
101	1	2	bar
...	...	...	...
101	1	6	baz
101	2	1	abc
101	2	2	xyz
...	...	...	...
101	12	6	bar
102	1	1	xyz
102	1	2	foo
...	...	...	...
199	12	6	abc

Instructions

You should be able to use Python to read in the trial-level data set and:

1. Perform a couple of basic calculations.
   1. Calculate the difference between each subjects estimation and the value in trueprob.
      1. Note that these two columns represent these percentage values differently.
      2. Put these values in a column called estimation_error.
   2. Exclude outliers (more than 3 standard deviations from the mean) from the values in choicert, esttaskrt, and outcomert.
      1. Label the new columns like so: choicert_exclusive, esttaskrt_exclusive, outcomert_exclusive.
2. Transform the dataset to subject-level with the following values:

subjnum	mean_estimation_error	stdev_estimation_error	mean_choicert	mean_esttaskrt	mean_outcomert
101	xxx	xxx	xxx	xxx	xxx
102	xxx	xxx	xxx	xxx	xxx
103	xxx	xxx	xxx	xxx	xxx
...	...	...	...	...	...