One Sentence Summary

This table 01 contains a new image quality control that is not based on an MTurk survey, but rather on a no-reference image quality assessment (IQA) algorithm. Our MTurk features increase from 0.015 to 0.021 on a validation set with over 1000 images.

Outline

This markdown contains table 01s for elections with image quality control using only images with above average quality scores. We see our MTurk labels increase from 0.015 to 0.021 on a validation set containing 1003 images. This table also uses the previously updated Election LM with the rolling window vote-share feature. The IAQ algorithm I use here is comparable to SSIM, but doesn’t rely on a direct comparison. To read more about this have a look at these sources:

  1. Table 01 Configuration 4 (using own-party-vote-share-jacknife in the election_lm) with image quality control to only use above average images.

Definitions

Before jumping into regression output, I will put all definitions of important terms here:

  1. Vote-Share : In all tables below, when vote-share is a LHS variable this means the vote-share of a specific candidate election. I.e. how much of the total vote did a specific candidate get in their election.

  2. Own-Party-Vote-Share-Jacknife : This is a feature and computed as the “average vote share of my party in this states house of rep. election over the previous 3 years excluding the current”

Table 1 - Configuration 4

Replacing own_party_win_rate with own_party_vote_share_jacknife in the election lm. This now includes:

  1. Party (limited to Democrat vs. Republican)
  2. Rolling window Vote-share of the candidates own party

Own Party Vote-share-Jacknife is computed as: “average vote share of my party in this states house of rep. election over the previous 3 years excluding the current”

Table 01 - Version 04 - Election Regressions
Fit measured in adjusted R squared and AUC
Model Configuration Election Outcome Vote Share
Adjusted R Squared ROC AUC Adjusted R squared
Single Variable Model
Election LM 0.0715 0.6499 0.0776
Lower 95% C.I. 0.0436 0.6078 0.0483
Upper 95% C.I. 0.1059 0.6921 0.1128
Sex 0.0002 0.5074 0.0050
−0.0015 0.4772 −0.0010
0.0076 0.5376 0.0183
Skine-Tone −0.0091 0.5714 −0.0094
−0.0027 0.5278 −0.0032
0.0399 0.6150 0.0452
MTurk Features 0.0213 0.5888 0.0069
0.0090 0.5448 0.0005
0.0515 0.6328 0.0283
P_hat_cnn 0.1417 0.7187 0.1234
0.1037 0.6793 0.0866
0.1859 0.7581 0.1690
Combined Variable Model
Election LM + P_hat_cnn 0.1932 0.7584 0.1849
0.1531 0.7214 0.1456
0.2369 0.7955 0.2326
Election LM + Sex 0.0738 0.6546 0.0861
0.0464 0.6126 0.0568
0.1101 0.6966 0.1226
Election LM + Sex + P_hat_cnn 0.1986 0.7629 0.1986
0.1583 0.7264 0.1592
0.2462 0.7995 0.2469
Election LM + Sex + Skin-Tone 0.0654 0.6750 0.0759
0.0583 0.6337 0.0673
0.1341 0.7164 0.1462
Election LM + Sex + Skin-Tone + P_hat_cnn 0.1916 0.7721 0.1871
0.1697 0.7361 0.1680
0.2621 0.8081 0.2650
Election LM + Sex + Skin-Tone + MTurk 0.0816 0.6939 0.0775
0.0758 0.6534 0.0746
0.1559 0.7343 0.1537
Election LM + Sex + Skin-Tone + MTurk + P_hat_cnn 0.2001 0.7819 0.1886
0.1842 0.7465 0.1709
0.2749 0.8172 0.2698