Psych 45: N-back stats



In [41]:

    
%matplotlib inline

import pandas as pd
import seaborn as sns
sns.set(style='ticks', context='poster', font_scale=1.5)

Import data file



In [52]:

    
data = pd.read_csv('http://web.stanford.edu/class/psych45/demos/Psych45-nback_stats.csv')

# filter RT column so just a number (ms and comma not included)
data.avg_rt = data.avg_rt.str.strip(' ms').str.replace(',', '').astype(float)

task_list = ['2-back', '3-back', '4-back']



In [53]:

    
data.head()









    Out[53]:






  
    
      
      task
      percent_correct
      avg_rt
      combined
      when
      date_downloaded
    
  
  
    
      0
       3-back
       40.00
        962.18
       2,405.45
       16 hours ago
       12:15 PM, Sunday
    
    
      1
       3-back
       46.67
        880.35
       1,886.46
       16 hours ago
       12:15 PM, Sunday
    
    
      2
       3-back
       86.67
        894.30
       1,031.88
       16 hours ago
       12:15 PM, Sunday
    
    
      3
       3-back
       60.00
       1069.83
       1,783.04
       16 hours ago
       12:15 PM, Sunday
    
    
      4
       3-back
       26.67
       1258.98
       4,721.18
       16 hours ago
       12:15 PM, Sunday

Analyses

Number of data points:



In [59]:

    
data.groupby(['task']).count().when









    Out[59]:





task
2-back    25
3-back    25
4-back     4
Name: when, dtype: int64

Mean accuracy and response time:



In [63]:

    
data.groupby(['task']).mean().reset_index()









    Out[63]:






  
    
      
      task
      percent_correct
      avg_rt
    
  
  
    
      0
       2-back
       77.622
       889.6416
    
    
      1
       3-back
       50.866
       925.0708
    
    
      2
       4-back
       54.285
       781.2725

How does task affect percent of trials correct?



In [54]:

    
g = sns.factorplot(x='task', y='percent_correct', 
                   x_order=task_list,
                   data=data, 
                   palette='Set1', ci=68, 
                   aspect=1.5)
g.set_ylabels('% correct')
sns.despine(trim=True)

How does task affect average response time?



In [56]:

    
g = sns.factorplot(x='task', y='avg_rt', 
                   x_order=task_list,
                   data=data, 
                   palette='Set1', ci=68, 
                   aspect=1.5)
g.set_ylabels('avg RT')
sns.despine(trim=True)



In [ ]:

	task	percent_correct	avg_rt	combined	when	date_downloaded
0	3-back	40.00	962.18	2,405.45	16 hours ago	12:15 PM, Sunday
1	3-back	46.67	880.35	1,886.46	16 hours ago	12:15 PM, Sunday
2	3-back	86.67	894.30	1,031.88	16 hours ago	12:15 PM, Sunday
3	3-back	60.00	1069.83	1,783.04	16 hours ago	12:15 PM, Sunday
4	3-back	26.67	1258.98	4,721.18	16 hours ago	12:15 PM, Sunday

	task	percent_correct	avg_rt
0	2-back	77.622	889.6416
1	3-back	50.866	925.0708
2	4-back	54.285	781.2725