1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100
101
102
103
104
105
106
107
108
109
110
111
112
113
114
115
116
117
118
119
120
121
122
123
124
125
126
127
128
129
130
131
132
133
134
135
136
137
138
139
140
141
142
143
144
145
146
147
148
149
150
151
152
153
154
155
156
157
158
159
160
161
162
163
164
165
166
167
168
169
170
171
172
173
174
175
176
177
178
179
|
def outputfstats(Enum, Eden, dfnum, dfden, f, prob):
Enum = around(Enum,3)
Eden = around(Eden,3)
dfnum = around(Enum,3)
dfden = around(dfden,3)
f = around(f,3)
prob = around(prob,3)
suffix = '' # for *s after the p-value
if prob < 0.001: suffix = ' ***'
elif prob < 0.01: suffix = ' **'
elif prob < 0.05: suffix = ' *'
title = [['EF/ER','DF','Mean Square','F-value','prob','']]
lofl = title+[[Enum, dfnum, around(Enum/float(dfnum),3), f, prob, suffix],
[Eden, dfden, around(Eden/float(dfden),3),'','','']]
_support.printcc(lofl)
return
def paired(x, y):
"""
Interactively determines the type of data in x and y, and then runs the
appropriated statistic for paired group data.
Returns: appropriate statistic name, value, and probability
"""
x,y = map(asarray, (x,y))
samples = ''
while samples not in ['i','r','I','R','c','C']:
print '\nIndependent or related samples, or correlation (i,r,c): ',
samples = raw_input()
if samples in ['i','I','r','R']:
print '\nComparing variances ...',
# USE O'BRIEN'S TEST FOR HOMOGENEITY OF VARIANCE, Maxwell & delaney, p.112
r = obrientransform(x,y)
f,p = f_oneway(_support.colex(r,0),_support.colex(r,1))
if p<0.05:
vartype='unequal, p='+str(around(p,4))
else:
vartype='equal'
print vartype
if samples in ['i','I']:
if vartype[0]=='e':
t,p = ttest_ind(x,y,None,0)
print '\nIndependent samples t-test: ', around(t,4),around(p,4)
else:
if len(x)>20 or len(y)>20:
z,p = ranksums(x,y)
print '\nRank Sums test (NONparametric, n>20): ', around(z,4),around(p,4)
else:
u,p = mannwhitneyu(x,y)
print '\nMann-Whitney U-test (NONparametric, ns<20): ', around(u,4),around(p,4)
else: # RELATED SAMPLES
if vartype[0]=='e':
t,p = ttest_rel(x,y,0)
print '\nRelated samples t-test: ', around(t,4),around(p,4)
else:
t,p = ranksums(x,y)
print '\nWilcoxon T-test (NONparametric): ', around(t,4),around(p,4)
else: # CORRELATION ANALYSIS
corrtype = ''
while corrtype not in ['c','C','r','R','d','D']:
print '\nIs the data Continuous, Ranked, or Dichotomous (c,r,d): ',
corrtype = raw_input()
if corrtype in ['c','C']:
m,b,r,p,see = linregress(x,y)
print '\nLinear regression for continuous variables ...'
lol = [['Slope','Intercept','r','Prob','SEestimate'],
[around(m,4),around(b,4),around(r,4),around(p,4),around(see,4)]]
_support.printcc(lol)
elif corrtype in ['r','R']:
r,p = spearmanr(x,y)
print '\nCorrelation for ranked variables ...'
print "Spearman's r: ",around(r,4),around(p,4)
else: # DICHOTOMOUS
r,p = pointbiserialr(x,y)
print '\nAssuming x contains a dichotomous variable ...'
print 'Point Biserial r: ',around(r,4),around(p,4)
print '\n\n'
return None
def writecc (listoflists, file, writetype='w', extra=2):
"""
Writes a list of lists to a file in columns, customized by the max
size of items within the columns (max size of items in col, +2 characters)
to specified file. File-overwrite is the default.
Usage: writecc (listoflists,file,writetype='w',extra=2)
Returns: None
"""
if not isinstance(listoflists[0], (list, tuple)):
listoflists = [listoflists]
outfile = open(file,writetype)
rowstokill = []
list2print = copy.deepcopy(listoflists)
for i in range(len(listoflists)):
if listoflists[i] == ['\n'] or listoflists[i]=='\n' or listoflists[i]=='dashes':
rowstokill = rowstokill + [i]
rowstokill.reverse()
for row in rowstokill:
del list2print[row]
maxsize = [0]*len(list2print[0])
for col in range(len(list2print[0])):
items = _support.colex(list2print,col)
items = map(_support.makestr,items)
maxsize[col] = max(map(len,items)) + extra
for row in listoflists:
if row == ['\n'] or row == '\n':
outfile.write('\n')
elif row == ['dashes'] or row == 'dashes':
dashes = [0]*len(maxsize)
for j in range(len(maxsize)):
dashes[j] = '-'*(maxsize[j]-2)
outfile.write(_support.lineincustcols(dashes,maxsize))
else:
outfile.write(_support.lineincustcols(row,maxsize))
outfile.write('\n')
outfile.close()
return None
def outputpairedstats(fname,writemode,name1,n1,m1,se1,min1,max1,name2,n2,m2,se2,min2,max2,statname,stat,prob):
"""
Prints or write to a file stats for two groups, using the name, n,
mean, sterr, min and max for each group, as well as the statistic name,
its value, and the associated p-value.
Usage: outputpairedstats(fname,writemode,
name1,n1,mean1,stderr1,min1,max1,
name2,n2,mean2,stderr2,min2,max2,
statname,stat,prob)
Returns: None
"""
suffix = '' # for *s after the p-value
try:
x = prob.shape
prob = prob[0]
except:
pass
if prob < 0.001: suffix = ' ***'
elif prob < 0.01: suffix = ' **'
elif prob < 0.05: suffix = ' *'
title = [['Name','N','Mean','SD','Min','Max']]
lofl = title+[[name1,n1,round(m1,3),round(math.sqrt(se1),3),min1,max1],
[name2,n2,round(m2,3),round(math.sqrt(se2),3),min2,max2]]
if not isinstance(fname, basestring) or len(fname) == 0:
print
print statname
print
_support.printcc(lofl)
print
try:
if stat.shape == ():
stat = stat[0]
if prob.shape == ():
prob = prob[0]
except:
pass
print 'Test statistic = ',round(stat,3),' p = ',round(prob,3),suffix
print
else:
file = open(fname,writemode)
file.write('\n'+statname+'\n\n')
file.close()
writecc(lofl,fname,'a')
file = open(fname,'a')
try:
if stat.shape == ():
stat = stat[0]
if prob.shape == ():
prob = prob[0]
except:
pass
file.write(_support.list2string(['\nTest statistic = ',round(stat,4),' p = ',round(prob,4),suffix,'\n\n']))
file.close()
return None
|
