Badge source
- Survey data
├── ETL.py <- Python code for data ETL.
├── ETL_function.py <- Python code with all function for ETL.
├── Factor_Analysis.py <- Python code with all function for factor analysis.
├── DV_FAC.ipynb <- notebook with dependent variable factor analysis.
├── anti_party_FAC.ipynb <- notebook with independent variable factor analysis.
├── Reliability_testing.ipynb <- notebook with reliability testing for all index.
├── media_political_analysis.ipynb <- modeling.
└── raw_data.csv
- Reshape raw data as a dataframe named ml_df
- Filter out dependent variable as a dataframe named IV_df
- Filter out independent variable as a dataframe named IV_df
-
Result of Bartlett’s test
- Chi-square value : 358.638
- p-value : 0.0
- The Bartlett test produces a p-value that is less than 0.05. It means, we reject the null hypothesis or in this case, at least two population variances are different.
-
Result of Kaiser-Meyer-Olkin(KMO)
- KMO value : 0.675
- The KMO test produces a KMO value 0.675 which is great than the standard 0.5
- KMO value : 0.675
-
Communality testing
- Result
- The yellow color indicates that the communality values meet the criteria — greater than 0.5. Eliminated Variable below 0.5 .
- Also according to the Kaiser criteria, the number of factors generated is 2. It means that the 5 columns or well-known variables will be grouped and interpreted into 2 indicators.
- Result
- According to the scree plot we will get the elbow at 2 groups .
- Result
- According to the result above and reference can extract 2 indicators from 6 varaibles
- 極化現象(political_polarization) : anti_1
- 政黨形象(party_image) : anti_3 / anti_4 / anti_5
- According to the result above and reference can extract 2 indicators from 6 varaibles
- Result
-
Result of Bartlett’s test
- Chi-square value : 2616.18
- p-value : 0.0
- The Bartlett test produces a p-value that is less than 0.05. It means, we reject the null hypothesis or in this case, at least two population variances are different.
-
Result of Kaiser-Meyer-Olkin(KMO)
- KMO value : 0.771
- The KMO test produces a KMO value 0.675 which is great than the standard 0.5
- KMO value : 0.771
-
Communality testing
- Result
- The yellow color indicates that the communality values meet the criteria — greater than 0.5. Eliminated Variable below 0.5 .
- Also according to the Kaiser criteria, the number of factors generated is 4. It means that the 5 columns or well-known variables will be grouped and interpreted into 4 indicators.
- Result
- According to the scree plot we will get the elbow at 4 groups .
- Result
- According to the result above and reference can extract 4 indicators from 16 varaibles
- 線上媒體政治參與(online_media_pp) : read_media / like_media / share_media / comment_media -> reference : 劉嘉薇,2019
- 投票參與(voting) : election_mayor / election_18 -> reference : Barnes and Kaase (1979)
- 線下媒體政治參與(offline_media_pp) : read_election_news / read_election_leaflet -> reference : 徐火炎,2001
- 競選工作式政治參與(campaign_worker_pp) : campaign / volunteer -> reference : Mibrath and Goel,1977
- According to the result above and reference can extract 4 indicators from 16 varaibles
- Result
- Method1 : Grouping without FAC but with reference or domain knowledge
anti_party_vars = ['anti_1', 'anti_2', 'anti_3', 'anti_4', 'anti_5']
ml_df['anti_party'] = ml_df[anti_party_vars].mean(axis=1)
- Reliability Testing Result :
- anti_party Cronbach's alpha: 0.7564819668378749
online_pp_vars = ['TV_news_time', 'news_paper_time', 'int_news_time', 'TV_debate', 'read_media', 'like_media', 'share_media', 'comment_media', 'int_discuss']
offline_pp_vars = ['read_election_news', 'read_election_leaflet', 'convince', 'campaign', 'volunteer', 'election_mayor', 'election_18']
ml_df['online_pp'] = ml_df[online_pp_vars].mean(axis=1)
ml_df['offline_pp'] = ml_df[offline_pp_vars].mean(axis=1)
- Reliability Testing Result :
- online_pp Cronbach's alpha: 0.7717699094188335
- offline_pp Cronbach's alpha: 0.809844435651358
- Total Cronbanc's alpha: 0.7908071725350958
- Method2 : Get factors's mean after FAC(因素分析 -> 根據構面取平均)
political_polarization_vars = ['anti_1']
party_image_vars = ['anti_3', 'anti_4', 'anti_5']
ml_df['political_polarization_mean'] = ml_df[political_polarization_vars].mean(axis=1)
ml_df['party_image_mean'] = ml_df[party_image_vars].mean(axis=1)
- Reliability Testing Result :
- political_polarization_mean Cronbach's alpha: 1.0
- party_image_mean Cronbach's alpha: 0.8485172366992396
- Total Cronbanc's alpha: 0.9242586183496198
ml_df['online_media_pp_mean'] = ml_df[online_media_pp_vars].mean(axis=1)
ml_df['voting_mean'] = ml_df[voting_vars].mean(axis=1)
ml_df['offline_media_pp_mean'] = ml_df[offline_media_pp_vars].mean(axis=1)
ml_df['campaign_worker_pp_mean'] = ml_df[campaign_worker_pp_vars].mean(axis=1)
- Reliability Testing Result :
- online_media_pp_mean Cronbach's alpha: 0.8150949007062644
- voting_mean Cronbach's alpha: 0.9821418184416829
- offline_media_pp_mean Cronbach's alpha: 0.9267895495036798
- campaign_worker_pp_mean Cronbach's alpha: 0.9058515655096367
- Total Cronbanc's alpha: 0.9074694585403159
- Method3 : Total score after FAC(因素分析 -> 根據構面取綜合得分)
# Factor analysis with rotation
fa = FactorAnalyzer(n_factors = 2, rotation = 'varimax')
fa.fit(IV_df)
# Create a factor's names
facs = ['Factors' + ' ' + str(i + 1) for i in range(2)]
df_factors = pd.DataFrame(data = fa.fit_transform(IV_df),columns = facs)
df_factors.rename(columns = {'Factors 1': 'political_polarization_score',
'Factors 2': 'party_image_score'}, inplace = True)
ml_df = ml_df.join(df_factors)
- Reliability Testing Result :
- political_polarization_mean Cronbach's alpha: 1.0
- party_image_mean Cronbach's alpha: 0.8485172366992396
- Total Cronbanc's alpha: 0.9242586183496198
# Factor analysis with rotation
fa = FactorAnalyzer(n_factors = 4, rotation = 'varimax')
fa.fit(DV_df)
# Create a factor's names
facs = ['Factors' + ' ' + str(i + 1) for i in range(4)]
df_factors = pd.DataFrame(data = fa.fit_transform(DV_df),columns = facs)
df_factors.rename(columns = {'Factors 1': 'online_media_pp_score',
'Factors 2': 'voting_score',
'Factors 3': 'offline_media_pp_score',
'Factors 4': 'campaign_worker_pp_score'}, inplace = True)
ml_df = ml_df.join(df_factors)
- Reliability Testing Result :
- online_media_pp_score Cronbach's alpha: 0.8311674955098114
- voting_score Cronbach's alpha: 0.9666132108881862
- offline_media_pp_score Cronbach's alpha: 0.8956567852883418
- campaign_worker_pp_score Cronbach's alpha: 0.9058515655096367
- Total Cronbanc's alpha: 0.899822264298994
- Method4 : Grouping and FAC get total score(因素分析 -> 全部算一個綜合得分)
# Formula = (factor1_value * factor1_Proportion + factor2_value * factor2_Proportion) / Cumulative Variance
# Divide by the cumulative variance to get the final scores
ml_df['anti_party_scores'] = (ml_df['political_polarization_score'] * fa.get_factor_variance()[1][0]
+ ml_df['party_image_score'] * fa.get_factor_variance()[1][1]) / fa.get_factor_variance()[1].sum()
- Reliability Testing Result :
- anti_party Cronbach's alpha: 0.7405711981868875
# Formula = (factor1_value * factor1_Proportion + factor2_value * factor2_Proportion) / Cumulative Variance
# online_pp
fa_online = FactorAnalyzer(n_factors = 3, rotation = 'varimax')
fa_online.fit(online_pp_df)
facs = ['Factors' + ' ' + str(i + 1) for i in range(3)]
df_factors_onlnie = pd.DataFrame(data = fa_online.fit_transform(online_pp_df),columns = facs)
ml_df['onlnie_scores'] = (df_factors_onlnie['Factors 1'] * fa_online.get_factor_variance()[1][0] + df_factors_onlnie['Factors 2'] * fa_online.get_factor_variance()[1][1] + df_factors_onlnie['Factors 3'] * fa_online.get_factor_variance()[1][2]) / fa_online.get_factor_variance()[1].sum()
# offline_pp
fa_offline = FactorAnalyzer(n_factors = 2, rotation = 'varimax')
fa_offline.fit(offline_pp_df)
facs = ['Factors' + ' ' + str(i + 1) for i in range(2)]
df_factors_offlnie = pd.DataFrame(data = fa_offline.fit_transform(offline_pp_df),columns = facs)
ml_df['offlnie_scores'] = (df_factors_offlnie['Factors 1'] * fa_offline.get_factor_variance()[1][0] + df_factors_offlnie['Factors 2'] * fa_offline.get_factor_variance()[1][1]) / fa_offline.get_factor_variance()[1].sum()
- Reliability Testing Result :
- onlnie_scores Cronbach's alpha: 0.7865005681513185
- offlnie_scores Cronbach's alpha: 0.8136118677742157
- Total Cronbanc's alpha: 0.800056217962767
online_media_pp_mean = 1.0305
+ 0.1111* C(ethnic, 台灣人[其他])
+ 0.0245* C(ethnic, 台灣人[原住民])
- 0.0189* C(ethnic, 台灣人[大陸各省市人])
- 0.0896* C(ethnic, 台灣人[本省客家人])
- 0.0550* C(ethnic, 台灣人[本省閩南人])
- 0.0433* C(Negative_1, 沒有影響[不知道])
+ 0.1241* C(Negative_1, 沒有影響[可能因此不去投票])
+ 0.0212* C(Negative_1, 沒有影響[轉而支持其他候選人])
+ 0.0017* C(Negative_2, 沒有影響[不知道])
- 0.2322* C(Negative_2, 沒有影響[可能因此不去投票])
+ 0.0998* C(Negative_2, 沒有影響[轉而支持其他候選人])
- 0.0122* C(Negative_3, 沒有影響[不知道])
+ 0.1329* C(Negative_3, 沒有影響[可能因此不去投票])
- 0.0576* C(Negative_3, 沒有影響[轉而支持其他候選人])
+ 0.0613* sex
+ 0.1088* edu
+ 0.0234* income
+ 0.0509* political_knowledge
+ 0.0085* TC_issue
+ 0.0293* political_polarization_mean
- 0.0623* party_image_mean
OLS Regression Results
================================================================================
Dep. Variable: online_media_pp_mean R-squared: 0.076
Model: OLS Adj. R-squared: 0.035
Method: Least Squares F-statistic: 1.874
Date: Thu, 03 Aug 2023 Prob (F-statistic): 0.0111
Time: 15:24:36 Log-Likelihood: -312.35
No. Observations: 503 AIC: 668.7
Df Residuals: 481 BIC: 761.5
Df Model: 21
Covariance Type: nonrobust
===========================================================================================================================
coef std err t P>|t| [0.025 0.975]
---------------------------------------------------------------------------------------------------------------------------
Intercept 1.0305 0.242 4.263 0.000 0.556 1.505
C(ethnic, Treatment(reference="臺灣人"))[T.其他] 0.1111 0.147 0.757 0.450 -0.177 0.400
C(ethnic, Treatment(reference="臺灣人"))[T.原住民] 0.0245 0.166 0.148 0.883 -0.302 0.350
C(ethnic, Treatment(reference="臺灣人"))[T.大陸各省市人] -0.0189 0.155 -0.122 0.903 -0.323 0.285
C(ethnic, Treatment(reference="臺灣人"))[T.本省客家人] -0.0896 0.109 -0.818 0.414 -0.305 0.126
C(ethnic, Treatment(reference="臺灣人"))[T.本省閩南人] -0.0550 0.088 -0.625 0.532 -0.228 0.118
C(Negative_1, Treatment(reference="沒有影響"))[T.不知道] -0.0433 0.089 -0.486 0.627 -0.218 0.132
C(Negative_1, Treatment(reference="沒有影響"))[T.可能因此不去投票] 0.1241 0.098 1.263 0.207 -0.069 0.317
C(Negative_1, Treatment(reference="沒有影響"))[T.轉而支持其他候選人] 0.0212 0.054 0.390 0.697 -0.085 0.128
C(Negative_2, Treatment(reference="沒有影響"))[T.不知道] 0.0017 0.089 0.020 0.984 -0.173 0.176
C(Negative_2, Treatment(reference="沒有影響"))[T.可能因此不去投票] -0.2322 0.134 -1.734 0.084 -0.495 0.031
C(Negative_2, Treatment(reference="沒有影響"))[T.轉而支持其他候選人] 0.0998 0.051 1.960 0.051 -0.000 0.200
C(Negative_3, Treatment(reference="沒有影響"))[T.不知道] -0.0122 0.083 -0.147 0.883 -0.175 0.151
C(Negative_3, Treatment(reference="沒有影響"))[T.可能因此不去投票] 0.1329 0.114 1.162 0.246 -0.092 0.358
C(Negative_3, Treatment(reference="沒有影響"))[T.轉而支持其他候選人] -0.0576 0.053 -1.098 0.273 -0.161 0.046
sex 0.0613 0.046 1.329 0.185 -0.029 0.152
edu 0.1088 0.047 2.295 0.022 0.016 0.202
income 0.0234 0.021 1.132 0.258 -0.017 0.064
political_knowledge 0.0509 0.038 1.355 0.176 -0.023 0.125
TC_issue 0.0085 0.019 0.450 0.653 -0.029 0.046
political_polarization_mean 0.0293 0.025 1.190 0.234 -0.019 0.078
party_image_mean -0.0623 0.022 -2.854 0.004 -0.105 -0.019
==============================================================================
Omnibus: 143.226 Durbin-Watson: 1.967
Prob(Omnibus): 0.000 Jarque-Bera (JB): 422.719
Skew: 1.353 Prob(JB): 1.61e-92
Kurtosis: 6.585 Cond. No. 101.
==============================================================================
Notes:
[1] Standard Errors assume that the covariance matrix of the errors is correctly specified.
voting_mean = 1.5372
+ 0.1619* ethnic('其他')
+ 0.6316* ethnic('原住民')
- 0.2645* ethnic('大陸各省市人')
+ 0.1775* ethnic('本省客家人')
+ 0.0381* ethnic('本省閩南人')
- 0.0573* Negative_1('不知道')
+ 0.1274* Negative_1('可能因此不去投票')
+ 0.3580* Negative_1('轉而支持其他候選人')
- 0.3703* Negative_2('不知道')
- 0.4460* Negative_2('可能因此不去投票')
+ 0.0011* Negative_2('轉而支持其他候選人')
- 0.0986* Negative_3('不知道')
- 0.4686* Negative_3('可能因此不去投票')
- 0.1716* Negative_3('轉而支持其他候選人')
- 0.1267* sex
+ 0.1772* edu
+ 0.1455* income
+ 0.2526* political_knowledge
- 0.0307* TC_issue
+ 0.0695* political_polarization_mean
- 0.1438* party_image_mean;
OLS Regression Results
==============================================================================
Dep. Variable: voting_mean R-squared: 0.107
Model: OLS Adj. R-squared: 0.068
Method: Least Squares F-statistic: 2.741
Date: Thu, 03 Aug 2023 Prob (F-statistic): 6.10e-05
Time: 15:24:36 Log-Likelihood: -801.21
No. Observations: 503 AIC: 1646.
Df Residuals: 481 BIC: 1739.
Df Model: 21
Covariance Type: nonrobust
===========================================================================================================================
coef std err t P>|t| [0.025 0.975]
---------------------------------------------------------------------------------------------------------------------------
Intercept 1.5372 0.639 2.406 0.016 0.282 2.792
C(ethnic, Treatment(reference="臺灣人"))[T.其他] 0.1619 0.388 0.417 0.677 -0.601 0.924
C(ethnic, Treatment(reference="臺灣人"))[T.原住民] 0.6316 0.438 1.441 0.150 -0.230 1.493
C(ethnic, Treatment(reference="臺灣人"))[T.大陸各省市人] -0.2645 0.409 -0.647 0.518 -1.068 0.539
C(ethnic, Treatment(reference="臺灣人"))[T.本省客家人] 0.1775 0.289 0.613 0.540 -0.391 0.746
C(ethnic, Treatment(reference="臺灣人"))[T.本省閩南人] 0.0381 0.233 0.163 0.870 -0.419 0.495
C(Negative_1, Treatment(reference="沒有影響"))[T.不知道] -0.0573 0.236 -0.243 0.808 -0.520 0.405
C(Negative_1, Treatment(reference="沒有影響"))[T.可能因此不去投票] 0.1274 0.260 0.491 0.624 -0.383 0.638
C(Negative_1, Treatment(reference="沒有影響"))[T.轉而支持其他候選人] 0.3580 0.143 2.497 0.013 0.076 0.640
C(Negative_2, Treatment(reference="沒有影響"))[T.不知道] -0.3703 0.235 -1.578 0.115 -0.831 0.091
C(Negative_2, Treatment(reference="沒有影響"))[T.可能因此不去投票] -0.4460 0.354 -1.260 0.208 -1.141 0.249
C(Negative_2, Treatment(reference="沒有影響"))[T.轉而支持其他候選人] 0.0011 0.135 0.008 0.993 -0.263 0.265
C(Negative_3, Treatment(reference="沒有影響"))[T.不知道] -0.0986 0.219 -0.450 0.653 -0.529 0.332
C(Negative_3, Treatment(reference="沒有影響"))[T.可能因此不去投票] -0.4686 0.302 -1.550 0.122 -1.063 0.125
C(Negative_3, Treatment(reference="沒有影響"))[T.轉而支持其他候選人] -0.1716 0.139 -1.236 0.217 -0.444 0.101
sex -0.1267 0.122 -1.040 0.299 -0.366 0.113
edu 0.1772 0.125 1.414 0.158 -0.069 0.423
income 0.1455 0.055 2.662 0.008 0.038 0.253
political_knowledge 0.2526 0.099 2.543 0.011 0.057 0.448
TC_issue -0.0307 0.050 -0.612 0.541 -0.129 0.068
political_polarization_mean 0.0695 0.065 1.067 0.287 -0.058 0.197
party_image_mean -0.1438 0.058 -2.492 0.013 -0.257 -0.030
==============================================================================
Omnibus: 316.969 Durbin-Watson: 2.009
Prob(Omnibus): 0.000 Jarque-Bera (JB): 42.405
Skew: -0.387 Prob(JB): 6.19e-10
Kurtosis: 1.806 Cond. No. 101.
==============================================================================
Notes:
[1] Standard Errors assume that the covariance matrix of the errors is correctly specified.
offline_media_pp_mean = 1.5678
+ 0.3867* C(ethnic, 台灣人[其他])
+ 0.2094* C(ethnic, 台灣人[原住民])
- 0.4852* C(ethnic, 台灣人[大陸各省市人])
- 0.1455* C(ethnic, 台灣人[本省客家人])
- 0.0929* C(ethnic, 台灣人[本省閩南人])
- 0.1142* C(Negative_1, 沒有影響[不知道])
+ 0.1127* C(Negative_1, 沒有影響[可能因此不去投票])
+ 0.2423* C(Negative_1, 沒有影響[轉而支持其他候選人])
- 0.2760* C(Negative_2, 沒有影響[不知道])
- 0.4682* C(Negative_2, 沒有影響[可能因此不去投票])
+ 0.1706* C(Negative_2, 沒有影響[轉而支持其他候選人])
- 0.1731* C(Negative_3, 沒有影響[不知道])
- 0.1712* C(Negative_3, 沒有影響[可能因此不去投票])
- 0.0360* C(Negative_3, 沒有影響[轉而支持其他候選人])
- 0.0826* sex
+ 0.0346* edu
+ 0.0848* income
+ 0.1355* political_knowledge
+ 0.0207* TC_issue
+ 0.1235* political_polarization_mean
- 0.0909* party_image_mean
OLS Regression Results
=================================================================================
Dep. Variable: offline_media_pp_mean R-squared: 0.118
Model: OLS Adj. R-squared: 0.080
Method: Least Squares F-statistic: 3.066
Date: Thu, 03 Aug 2023 Prob (F-statistic): 7.27e-06
Time: 15:24:36 Log-Likelihood: -676.56
No. Observations: 503 AIC: 1397.
Df Residuals: 481 BIC: 1490.
Df Model: 21
Covariance Type: nonrobust
===========================================================================================================================
coef std err t P>|t| [0.025 0.975]
---------------------------------------------------------------------------------------------------------------------------
Intercept 1.5678 0.499 3.144 0.002 0.588 2.548
C(ethnic, Treatment(reference="臺灣人"))[T.其他] 0.3867 0.303 1.277 0.202 -0.208 0.982
C(ethnic, Treatment(reference="臺灣人"))[T.原住民] 0.2094 0.342 0.612 0.541 -0.463 0.882
C(ethnic, Treatment(reference="臺灣人"))[T.大陸各省市人] -0.4852 0.319 -1.521 0.129 -1.112 0.142
C(ethnic, Treatment(reference="臺灣人"))[T.本省客家人] -0.1455 0.226 -0.644 0.520 -0.589 0.298
C(ethnic, Treatment(reference="臺灣人"))[T.本省閩南人] -0.0929 0.182 -0.511 0.610 -0.450 0.264
C(Negative_1, Treatment(reference="沒有影響"))[T.不知道] -0.1142 0.184 -0.622 0.535 -0.475 0.247
C(Negative_1, Treatment(reference="沒有影響"))[T.可能因此不去投票] 0.1127 0.203 0.556 0.578 -0.285 0.511
C(Negative_1, Treatment(reference="沒有影響"))[T.轉而支持其他候選人] 0.2423 0.112 2.165 0.031 0.022 0.462
C(Negative_2, Treatment(reference="沒有影響"))[T.不知道] -0.2760 0.183 -1.507 0.133 -0.636 0.084
C(Negative_2, Treatment(reference="沒有影響"))[T.可能因此不去投票] -0.4682 0.276 -1.695 0.091 -1.011 0.075
C(Negative_2, Treatment(reference="沒有影響"))[T.轉而支持其他候選人] 0.1706 0.105 1.624 0.105 -0.036 0.377
C(Negative_3, Treatment(reference="沒有影響"))[T.不知道] -0.1731 0.171 -1.013 0.312 -0.509 0.163
C(Negative_3, Treatment(reference="沒有影響"))[T.可能因此不去投票] -0.1712 0.236 -0.725 0.469 -0.635 0.293
C(Negative_3, Treatment(reference="沒有影響"))[T.轉而支持其他候選人] -0.0360 0.108 -0.332 0.740 -0.249 0.177
sex -0.0826 0.095 -0.868 0.386 -0.269 0.104
edu 0.0346 0.098 0.353 0.724 -0.158 0.227
income 0.0848 0.043 1.987 0.047 0.001 0.169
political_knowledge 0.1355 0.078 1.747 0.081 -0.017 0.288
TC_issue 0.0207 0.039 0.531 0.596 -0.056 0.098
political_polarization_mean 0.1235 0.051 2.429 0.015 0.024 0.223
party_image_mean -0.0909 0.045 -2.019 0.044 -0.179 -0.002
==============================================================================
Omnibus: 34.965 Durbin-Watson: 1.965
Prob(Omnibus): 0.000 Jarque-Bera (JB): 13.143
Skew: -0.088 Prob(JB): 0.00140
Kurtosis: 2.228 Cond. No. 101.
==============================================================================
Notes:
[1] Standard Errors assume that the covariance matrix of the errors is correctly specified.
campaign_worker_pp_mean = 1.4316
+ 0.2027* C(ethnic, 台灣人[其他])
- 0.1233* C(ethnic, 台灣人[原住民])
- 0.0985* C(ethnic, 台灣人[大陸各省市人])
- 0.0284* C(ethnic, 台灣人[本省客家人])
- 0.0010* C(ethnic, 台灣人[本省閩南人])
+ 0.0542* C(Negative_1, 沒有影響[不知道])
+ 0.0128* C(Negative_1, 沒有影響[可能因此不去投票])
- 0.0341* C(Negative_1, 沒有影響[轉而支持其他候選人])
+ 0.0694* C(Negative_2, 沒有影響[不知道])
- 0.1248* C(Negative_2, 沒有影響[可能因此不去投票])
+ 0.0905* C(Negative_2, 沒有影響[轉而支持其他候選人])
- 0.0097* C(Negative_3, 沒有影響[不知道])
+ 0.2979* C(Negative_3, 沒有影響[可能因此不去投票])
+ 0.0143* C(Negative_3, 沒有影響[轉而支持其他候選人])
+ 0.0426* sex
- 0.0385* edu
+ 0.0065* income
- 0.0500* political_knowledge
+ 0.0165* TC_issue
- 0.0394* political_polarization_mean
+ 0.0110* party_image_mean
OLS Regression Results
===================================================================================
Dep. Variable: campaign_worker_pp_mean R-squared: 0.039
Model: OLS Adj. R-squared: -0.003
Method: Least Squares F-statistic: 0.9308
Date: Thu, 03 Aug 2023 Prob (F-statistic): 0.551
Time: 15:24:37 Log-Likelihood: -352.98
No. Observations: 503 AIC: 750.0
Df Residuals: 481 BIC: 842.8
Df Model: 21
Covariance Type: nonrobust
===========================================================================================================================
coef std err t P>|t| [0.025 0.975]
---------------------------------------------------------------------------------------------------------------------------
Intercept 1.4316 0.262 5.463 0.000 0.917 1.947
C(ethnic, Treatment(reference="臺灣人"))[T.其他] 0.2027 0.159 1.274 0.203 -0.110 0.515
C(ethnic, Treatment(reference="臺灣人"))[T.原住民] -0.1233 0.180 -0.686 0.493 -0.477 0.230
C(ethnic, Treatment(reference="臺灣人"))[T.大陸各省市人] -0.0985 0.168 -0.587 0.557 -0.428 0.231
C(ethnic, Treatment(reference="臺灣人"))[T.本省客家人] -0.0284 0.119 -0.239 0.811 -0.262 0.205
C(ethnic, Treatment(reference="臺灣人"))[T.本省閩南人] -0.0010 0.095 -0.010 0.992 -0.189 0.187
C(Negative_1, Treatment(reference="沒有影響"))[T.不知道] 0.0542 0.097 0.561 0.575 -0.136 0.244
C(Negative_1, Treatment(reference="沒有影響"))[T.可能因此不去投票] 0.0128 0.107 0.120 0.905 -0.197 0.222
C(Negative_1, Treatment(reference="沒有影響"))[T.轉而支持其他候選人] -0.0341 0.059 -0.580 0.562 -0.150 0.081
C(Negative_2, Treatment(reference="沒有影響"))[T.不知道] 0.0694 0.096 0.721 0.471 -0.120 0.259
C(Negative_2, Treatment(reference="沒有影響"))[T.可能因此不去投票] -0.1248 0.145 -0.859 0.391 -0.410 0.161
C(Negative_2, Treatment(reference="沒有影響"))[T.轉而支持其他候選人] 0.0905 0.055 1.640 0.102 -0.018 0.199
C(Negative_3, Treatment(reference="沒有影響"))[T.不知道] -0.0097 0.090 -0.108 0.914 -0.186 0.167
C(Negative_3, Treatment(reference="沒有影響"))[T.可能因此不去投票] 0.2979 0.124 2.402 0.017 0.054 0.542
C(Negative_3, Treatment(reference="沒有影響"))[T.轉而支持其他候選人] 0.0143 0.057 0.251 0.802 -0.098 0.126
sex 0.0426 0.050 0.852 0.395 -0.056 0.141
edu -0.0385 0.051 -0.748 0.455 -0.139 0.063
income 0.0065 0.022 0.290 0.772 -0.038 0.051
political_knowledge -0.0500 0.041 -1.228 0.220 -0.130 0.030
TC_issue 0.0165 0.021 0.803 0.422 -0.024 0.057
political_polarization_mean -0.0394 0.027 -1.476 0.141 -0.092 0.013
party_image_mean 0.0110 0.024 0.464 0.643 -0.036 0.058
==============================================================================
Omnibus: 399.613 Durbin-Watson: 1.889
Prob(Omnibus): 0.000 Jarque-Bera (JB): 5393.751
Skew: 3.545 Prob(JB): 0.00
Kurtosis: 17.391 Cond. No. 101.
==============================================================================