-
Notifications
You must be signed in to change notification settings - Fork 0
/
replication-ts.log
325 lines (263 loc) · 13.2 KB
/
replication-ts.log
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
180
181
182
183
184
185
186
187
188
189
190
191
192
193
194
195
196
197
198
199
200
201
202
203
204
205
206
207
208
209
210
211
212
213
214
215
216
217
218
219
220
221
222
223
224
225
226
227
228
229
230
231
232
233
234
235
236
237
238
239
240
241
242
243
244
245
246
247
248
249
250
251
252
253
254
255
256
257
258
259
260
261
262
263
264
265
266
267
268
269
270
271
272
273
274
275
276
277
278
279
280
281
282
283
284
285
286
287
288
289
290
291
292
293
294
295
296
297
298
299
300
301
302
303
304
305
306
307
308
309
310
311
312
313
314
315
316
317
318
319
320
321
322
323
324
325
-----------------------------------------------------------------------------------
name: <unnamed>
log: /Users/pablobarbera/Dropbox/research/twitter-trolls/replication/replic
> ation-ts.log
log type: text
opened on: 19 Oct 2016, 14:43:29
. clear all
.
. ** opening and appending data
. use "germany-time-series.dta"
(Written by R. )
. append using "uk-time-series.dta", force
(note: variable party was str17, now str35 to accommodate using data's values)
(note: variable name was str26, now str28 to accommodate using data's values)
. append using "spain-time-series.dta", force
(note: variable name was str28, now str69 to accommodate using data's values)
. append using "greece-time-series.dta", force
(note: variable party was str35, now str60 to accommodate using data's values)
.
. ** strings to factors
. encode(country), gen(code)
. encode(electability), gen(elec)
. encode(twitter), gen(id)
. gen logfoll = log(followers_count+1)
(328 missing values generated)
. gen voteshare = votenl/100
. table country
----------------------
country | Freq.
----------+-----------
Germany | 468
Greece | 396
Spain | 900
UK | 1,216
----------------------
.
. ** computing weights
. by twitter, sort: egen sumtweets = sum(ntweets)
. sum ntweets
Variable | Obs Mean Std. Dev. Min Max
-------------+--------------------------------------------------------
ntweets | 2980 36.07584 76.91307 0 1160
. gen weight = 1/(sumtweets / r(mean))
(404 missing values generated)
.
. ** dropping missing values
. drop if engaging_tweets == .
(727 observations deleted)
. drop if ntweets < 2
(183 observations deleted)
.
. ** keeping only first 3 weeks
. drop if week == 4
(418 observations deleted)
.
. ** setting TSCS structure of data
. xtset id week
panel variable: id (unbalanced)
time variable: week, 1 to 3, but with gaps
delta: 1 unit
.
.
. ** Table 3 ***
.
. xtreg d.impolite_mentions l.engaging_tweets [pweight=weight], fe cluster(id)
Fixed-effects (within) regression Number of obs = 907
Group variable: id Number of groups = 505
R-sq: within = 0.1001 Obs per group: min = 1
between = 0.0010 avg = 1.8
overall = 0.0000 max = 2
F(1,504) = 4.12
corr(u_i, Xb) = -0.7911 Prob > F = 0.0428
(Std. Err. adjusted for 505 clusters in id)
---------------------------------------------------------------------------------
D. | Robust
impolite_ment~s | Coef. Std. Err. t P>|t| [95% Conf. Interval]
----------------+----------------------------------------------------------------
engaging_tweets |
L1. | .2784413 .1371277 2.03 0.043 .009029 .5478537
|
_cons | -.1160898 .0560135 -2.07 0.039 -.2261385 -.0060412
----------------+----------------------------------------------------------------
sigma_u | .05977672
sigma_e | .08280675
rho | .34258703 (fraction of variance due to u_i)
---------------------------------------------------------------------------------
. estimates store m1, title("All")
.
. margins, dyex(*)
Average marginal effects Number of obs = 907
Model VCE : Robust
Expression : Linear prediction, predict()
dy/ex w.r.t. : L.engaging_tweets
---------------------------------------------------------------------------------
| Delta-method
| dy/ex Std. Err. z P>|z| [95% Conf. Interval]
----------------+----------------------------------------------------------------
engaging_tweets |
L1. | .1137368 .0560135 2.03 0.042 .0039524 .2235212
---------------------------------------------------------------------------------
. sum engaging_tweets
Variable | Obs Mean Std. Dev. Min Max
-------------+--------------------------------------------------------
engaging_t~s | 1652 .4124117 .1892352 .0045098 .9833333
. di .1892352 * .2784413
.0526909
. sum impolite_mentions
Variable | Obs Mean Std. Dev. Min Max
-------------+--------------------------------------------------------
impolite_m~s | 1519 .0625076 .0720724 0 .805
. di (.1892352 * .2784413)/.0720724
.73108284
.
. xtreg d.impolite_mentions l.engaging_tweets [pweight=weight] if country=="UK", fe
> cluster(id)
Fixed-effects (within) regression Number of obs = 339
Group variable: id Number of groups = 212
R-sq: within = 0.0261 Obs per group: min = 1
between = 0.0137 avg = 1.6
overall = 0.0138 max = 2
F(1,211) = 2.29
corr(u_i, Xb) = -0.1612 Prob > F = 0.1317
(Std. Err. adjusted for 212 clusters in id)
---------------------------------------------------------------------------------
D. | Robust
impolite_ment~s | Coef. Std. Err. t P>|t| [95% Conf. Interval]
----------------+----------------------------------------------------------------
engaging_tweets |
L1. | .0749597 .0495367 1.51 0.132 -.0226906 .17261
|
_cons | -.0499184 .0263272 -1.90 0.059 -.1018164 .0019796
----------------+----------------------------------------------------------------
sigma_u | .0337002
sigma_e | .0450672
rho | .35863314 (fraction of variance due to u_i)
---------------------------------------------------------------------------------
. estimates store m2, title("UK")
.
. xtreg d.impolite_mentions l.engaging_tweets [pweight=weight] if country=="Spain",
> fe cluster(id)
Fixed-effects (within) regression Number of obs = 370
Group variable: id Number of groups = 187
R-sq: within = 0.1300 Obs per group: min = 1
between = 0.0153 avg = 2.0
overall = 0.0022 max = 2
F(1,186) = 0.97
corr(u_i, Xb) = -0.8320 Prob > F = 0.3265
(Std. Err. adjusted for 187 clusters in id)
---------------------------------------------------------------------------------
D. | Robust
impolite_ment~s | Coef. Std. Err. t P>|t| [95% Conf. Interval]
----------------+----------------------------------------------------------------
engaging_tweets |
L1. | .3524678 .3582612 0.98 0.326 -.35431 1.059246
|
_cons | -.1511933 .1524074 -0.99 0.322 -.4518627 .1494761
----------------+----------------------------------------------------------------
sigma_u | .04425627
sigma_e | .08474519
rho | .21428223 (fraction of variance due to u_i)
---------------------------------------------------------------------------------
. estimates store m3, title("Spain")
.
. xtreg d.impolite_mentions l.engaging_tweets [pweight=weight] if country=="Germany
> ", fe cluster(id)
Fixed-effects (within) regression Number of obs = 123
Group variable: id Number of groups = 64
R-sq: within = 0.2257 Obs per group: min = 1
between = 0.0510 avg = 1.9
overall = 0.0003 max = 2
F(1,63) = 3.86
corr(u_i, Xb) = -0.8970 Prob > F = 0.0539
(Std. Err. adjusted for 64 clusters in id)
---------------------------------------------------------------------------------
D. | Robust
impolite_ment~s | Coef. Std. Err. t P>|t| [95% Conf. Interval]
----------------+----------------------------------------------------------------
engaging_tweets |
L1. | .4277292 .2177745 1.96 0.054 -.0074584 .8629168
|
_cons | -.0870226 .0479011 -1.82 0.074 -.1827452 .0087
----------------+----------------------------------------------------------------
sigma_u | .07903904
sigma_e | .11079814
rho | .33725834 (fraction of variance due to u_i)
---------------------------------------------------------------------------------
. estimates store m4, title("Germany")
.
. xtreg d.impolite_mentions l.engaging_tweets [pweight=weight] if country=="Greece"
> , fe cluster(id)
Fixed-effects (within) regression Number of obs = 75
Group variable: id Number of groups = 42
R-sq: within = 0.0580 Obs per group: min = 1
between = 0.0007 avg = 1.8
overall = 0.0013 max = 2
F(1,41) = 1.94
corr(u_i, Xb) = -0.5505 Prob > F = 0.1713
(Std. Err. adjusted for 42 clusters in id)
---------------------------------------------------------------------------------
D. | Robust
impolite_ment~s | Coef. Std. Err. t P>|t| [95% Conf. Interval]
----------------+----------------------------------------------------------------
engaging_tweets |
L1. | .4109312 .2950947 1.39 0.171 -.1850246 1.006887
|
_cons | -.057414 .0488508 -1.18 0.247 -.1560703 .0412422
----------------+----------------------------------------------------------------
sigma_u | .09688706
sigma_e | .11565276
rho | .41239035 (fraction of variance due to u_i)
---------------------------------------------------------------------------------
. estimates store m5, title("Greece")
.
.
. ** Generating table
. estout m1 m2 m3 m4 m5 /*displays models 1-5*/ ///
> using "tabl3.tex", replace style(tex) ///
> title(OLS regressions of impolite tweets received on engaging tweets sent
> , with candidate fixed effects.) ///
> prehead( \def\one{\footnotesize{$\ast$}} ///
> \def\two{\footnotesize{$\ast\ast$}} ///
> \def\three{\footnotesize{$\ast\ast$$\ast$}} ///
> \begin{table}[h!]\caption{@title} ///
> \label{tab:table4} ///
> \centering\begin{threeparttable}\begin{tabular}{l
> *{@M}{r@{}l}}\hline\hline ///
> ) ///
> collabels(none) /*hides labels of columns*/ ///
> posthead(\hline\\) ///
> cells(b(star fmt(%9.2f)) se(par)) /*displays coefs with 1 decimal*/ ///
> starlevels(\one 0.10 \two 0.05 \three 0.01) ///
> /*reported levels of significance can changed*/ ///
> varlabels(_cons Constant L.engaging_tweets "\% Engaging tweets sent (lagg
> ed)") /*modify label of constant*/ ///
> prefoot(\\\hline) ///
> stats(N_clust N r2, fmt(%9.0g %9.0g %9.2f ) labels("N (candidates)" "N (o
> bservations)" "$R^2$")) ///
> nolegend /*hides significance symbols legend (I'll do it manually)*/ ///
> label /*make use of variable labels*/ ///
> stardetach /*display stars in separated column*/ ///
> wrap varwidth(30) ///
> postfoot( ///
> \hline\hline\end{tabular} \begin{tablenotes} ///
> \item \footnotesize{Dependent variable: Change in
> proportion of engaging tweets sent, by week. ///
> Robust standard errors in parentheses. //
> /
> Signif.: \one 10\% \two 5\% \three 1\%.} ///
> \end{tablenotes}\end{threeparttable}\end{table} /
> //
> )
(note: file tabl3.tex not found)
(output written to tabl3.tex)
.
.
.
. log close
name: <unnamed>
log: /Users/pablobarbera/Dropbox/research/twitter-trolls/replication/replic
> ation-ts.log
log type: text
closed on: 19 Oct 2016, 14:43:30
-----------------------------------------------------------------------------------