조사연구 권 호 연구논문 응답자의성격특성과성향점수매칭기법을활용한모바일조사향상방안연구 * 2)3)4) a) b) c)
조사연구 주제어 반사실적매칭 성격특성 모바일조사 성향조정가중법 This study aims to reduce estimation biases from mobile survey by employing propensity score matching based on respondents' personality. Firstly, we selected 2009 and 2011 Korean General Social Surveys(KGSS) as reference data and then sampled mobile survey respondents by one-to-one matching to those of 2011 KGSS based on sociodemographic dimensions. Secondly, we chose subjective political orientation as an outcome variable for estimation whose mean values are statistically identical between the two years of KGSS but more conservative than that of mobile respondents. Thirdly, we performed propensity score matching by 10 items of personality in a consistent way with propensity score weighting and produced a comparable but better performance than that of weighting. Our conclusions are threefold: First, personality traits are stable enough to be a basis of propensity score. Second, combining pre-survey matching by sociodemographic traits and after-survey matching by the small number of items can provide a cost-effective way of bias correction by coping with impatient mobile respondents. Third, propensity score matching provides us with more intuitive tools than weighting to improve the validity of propensity score methods. Key words : propensity score matching, big five personality, mobile survey, propensity score weighting
응답자의성격특성과성향점수매칭기법을활용한모바일조사향상방안연구 Ⅰ. 문제제기
조사연구
응답자의성격특성과성향점수매칭기법을활용한모바일조사향상방안연구 Ⅱ. 기존연구검토 1. 모바일조사의가능성과문제점
조사연구
응답자의성격특성과성향점수매칭기법을활용한모바일조사향상방안연구 2. 반사실적매칭과성향점수를활용한보정방법
조사연구 3. 응답자의성격특성을활용한성향점수매칭
응답자의성격특성과성향점수매칭기법을활용한모바일조사향상방안연구
조사연구 Ⅲ. 분석계획 1. 연구자료
응답자의성격특성과성향점수매칭기법을활용한모바일조사향상방안연구 2. 변수 1) 성격특성
조사연구 < 표 1> 분석에사용된 TIPI 의설문문항과성격특성 설문문항.. (extraverted, enthusiastic).. (critical, quarrelsome).. (dependable, self-disciplined).. (anxious, easily upset).. (open to new experiences, complex).. (reserved, quiet).. (sympathetic, warm).. (disorganized, careless).. (calm, emotionally stable).. (conventional, uncreative) 성격특성 (extraversion) (-) (agreeableness) (conscientiousness) (-) (emotional stability) (openness to experience) (-) (extraversion) (agreeableness) (-) (conscientiousness) (emotional stability) (-) (openness to experience) 2) 보정확인을위한변수들
응답자의성격특성과성향점수매칭기법을활용한모바일조사향상방안연구 3. 분석방법
조사연구 Ⅳ. 분석결과 1. 응답자의특성
응답자의성격특성과성향점수매칭기법을활용한모바일조사향상방안연구 표 응답자의사회인구학적특성 KGSS 2009 KGSS 2011 모바일조사 빈도비율빈도비율빈도비율 단위 명 계 769 48.09 689 44.89 689 44.89 2,147 830 51.91 846 55.11 846 55.11 2,522 1,599 100.00 100.00 100.00 4,669 10 64 4.02 51 3.32 51 3.32 166 20 236 14.81 200 13.03 200 13.03 636 30 405 25.42 336 21.89 336 21.89 1,077 40 393 24.67 366 23.84 366 23.84 1,125 50 495 31.07 582 37.92 582 37.92 1,659 1,593 100.00 100.00 100.00 4,663 328 20.51 285 18.57 285 18.57 898 355 22.2 342 22.28 342 22.28 1,039 91 5.69 81 5.28 81 5.28 253 44 2.75 51 3.32 51 3.32 146 61 3.81 33 2.15 33 2.15 127 46 2.88 55 3.58 55 3.58 156 49 3.06 49 3.19 49 3.19 147 112 7.00 122 7.95 122 7.95 356 121 7.57 113 7.36 113 7.36 347 38 2.38 36 2.35 36 2.35 110 83 5.19 93 6.06 93 6.06 269 75 4.69 87 5.67 87 5.67 249 65 4.07 63 4.10 63 4.10 191 50 3.13 54 3.52 54 3.52 158 63 3.94 55 3.58 55 3.58 173 18 1.13 16 1.04 16 1.04 50 1,599 100.00 100.00 100.00 4,669
조사연구
응답자의성격특성과성향점수매칭기법을활용한모바일조사향상방안연구 < 표 3> 연속형범주들의기술통계량 변수명, 1,597 KGSS 2009 KGSS 2011 모바일조사 2012 평균 ( 표준편차 ) 4.446 (1.499) 1,534 평균 ( 표준편차 ) 4.499 (1.508) 평균 ( 표준편차 ) 4.235 (1.277), 1,597 3.589 (1.597) 1,533 3.606 (1.534) 3.909 (1.415), 1,596 4.940 (1.350) 1,533 4.945 (1.317) 4.784 (1.308), 1,596 4.062 (1.601) 1,534 3.997 (1.603) 4.365 (1.488), 1,597 4.359 (1.538) 1,534 4.364 (1.547) 4.389 (1.388), 1,595 3.954 (1.659) 1,534 3.897 (1.636) 3.973 (1.530), 1,597 5.105 (1.276) 5.023 (1.266) 5.061 (1.326), 1,595 3.535 (1.709) 1,533 3.482 (1.736) 3.669 (1.697), 1,596 4.292 (1.514) 4.378 (1.499) 4.083 (1.431), 1,598 3.773 (1.539) 1,534 3.815 (1.601) 3.786 (1.457) 1,534 1,548 1,575 2.993 (0.984) 1.904 (0.642) 1.971 (0.596) 1,427 1,498 1,501 2.990 (0.996) 2.068 (0.666) 2.128 (0.596) 2.873 (0.910) 2.326 (0.540) 2.448 (0.523) 1,561 2.540 (0.929) 1,504 2.705 (0.925) 2.556 (0.863) 1,569 3.734 (1.017) 1,486 3.505 (1.061) 3.816 (0.995)
조사연구 < 표 4> 순서형로짓을통한 4 개태도변수에대한자료간차이비교 독립변수 군대지도층신뢰금융지도층신뢰북한에대한인식국정운영평가 계수값 표준오차 계수값표준오차 계수값표준오차 계수값표준오차 ( :KGSS 2011) KGSS 2009-0.527*** 0.074-0.567 *** 0.077-0.385 *** 0.067 0.402 *** 0.066 0.772*** 0.072 1.066 *** 0.074-0.356 *** 0.067 0.534 *** 0.066 1-1.654 *** 0.060-2.086 *** 0.066-2.152 *** 0.061-3.439 *** 0.104 2 1.212 *** 0.056 1.160 *** 0.056-0.538 *** 0.050-1.620 *** 0.057 3 - - 1.416 *** 0.055-0.056 0.049 4 - - - 1.429 *** 0.053 4,562 4,594 4,578 4,569 pseudo 0.038 0.060 0.003 0.006
응답자의성격특성과성향점수매칭기법을활용한모바일조사향상방안연구 2. 자료간의특성비교
조사연구 < 표 5> 성격특성비교를위한로짓회귀분석결과 독립변수,,,,,,,,,, KGSS 2011 (KGSS 2009 와비교 ) 모바일조사 (KGSS 와비교 ) 계수값표준오차계수값표준오차 0.028 0.029-0.157 *** 0.027 0.007 0.025 0.138 *** 0.023-0.012 0.030-0.042 0.026-0.012 0.024 0.090 *** 0.022 0.013 0.026 0.006 0.024-0.023 0.026-0.008 0.023-0.066 * 0.030 0.046 0.026-0.017 0.022 0.015 0.019 0.048 0.028-0.064 ** 0.025 0.027 0.025-0.016 0.023 0.034 0.288-0.667 ** 0.256 3,112 4,647 pseudo 0.003 0.023
응답자의성격특성과성향점수매칭기법을활용한모바일조사향상방안연구 < 표 6> 정치적성향비교를위한회귀분석결과 독립변수 계수값표준오차 (: KGSS 2009) KGSS 2011-0.003 0.035-0.120 *** 0.035 2.993 *** 0.025 4,496 0.003 3. 성향점수가중방법
조사연구 < 표 7> 주관적정치성향편의 ( 괄호안은추정계수의표준오차 ; N=4,476) 모형 1 모형 2 모형 3 모형 4 - 검정다중회귀 a 성향점수가중성향점수매칭 b (: KGSS) *** -0.116 (0.030) *** -0.097 (0.030) ** -0.085 (0.031) * -0.077 (0.034) vs. 1 16.4 26.7 33.6 % vs. 2 12.4 20.6 vs. 3 9.4 다중회귀모형은성격특성 개문항을통제한결과 매칭법은 표 의결과중인근이웃매칭결과를제시 4. 매칭기법
응답자의성격특성과성향점수매칭기법을활용한모바일조사향상방안연구 Propensity scores.2.4.6.8 1 mobile KGSS 그림 조사자료간성향점수상자그림 비교
조사연구 < 표 8> 매칭기법적용전후정치적성향비교 정치적성향 KGSS 모바일조사차이표준오차 2.873 0.116 *** 0.030 1:1 2.911 0.079 * 0.040 2.989 2.912 0.077 * 0.034 2.901 0.088 ** 0.031 2.902 0.087 ** 0.031
응답자의성격특성과성향점수매칭기법을활용한모바일조사향상방안연구 < 표 9> 매칭방법에따른성격특성문항별자료균형화정도 구분 1:1 변수 KGSS 평균 모바일조사 편의 (%) 편의 감소량 (%) -test >, 4.480 4.235 17.6-5.45 0.000, 3.632 3.909-18.6 - -5.8 0.000, 4.950 4.784 12.6-3.98 0.000, 4.035 4.366-21.4 - -6.71 0.000, 4.398 4.389 0.6-0.18 0.855, 3.927 3.973-2.9 - -0.9 0.369, 5.074 5.061 1.0-0.33 0.738, 3.496 3.669-10.1 - -3.2 0.001, 4.337 4.083 17.2-5.43 0.000, 3.778 3.786-0.6 - -0.18 0.857, 4.480 4.444 2.6 85.4 0.98 0.327, 3.632 3.626 0.4 97.8 0.15 0.877, 4.950 4.964-1.1 91.6-0.41 0.685, 4.035 4.018 1.1 95.0 0.41 0.684, 4.398 4.416-1.2-110.6-0.47 0.637, 3.927 3.902 1.6 44.6 0.61 0.544, 5.074 4.953 9.4-796.6 3.54 0.000, 3.496 3.529-1.9 80.7-0.75 0.454, 4.337 4.324 0.8 95.2 0.32 0.751, 3.778 3.713 4.2-639.3 1.63 0.102, 4.480 4.462 1.3 92.8 0.48 0.631, 3.632 3.649-1.1 93.9-0.43 0.666, 4.950 4.959-0.7 94.6-0.26 0.794, 4.035 4.012 1.4 93.3 0.54 0.586, 4.398 4.383 1.0-66.1 0.37 0.710, 3.927 3.926 0.1 96.6 0.04 0.970, 5.074 5.032 3.3-212.9 1.25 0.213, 3.496 3.474 1.3 87.1 0.5 0.617, 4.337 4.339-0.2 99.1-0.06 0.951,. 3.778 3.718 3.9-582.2 1.49 0.135 계속
조사연구 계속 구분 변수 KGSS 평균 모바일조사 편의 (%) 편의 감소량 (%) -test >, 4.480 4.422 4.2 76.3 1.61 0.108, 3.632 3.689-3.9 79.2-1.47 0.141, 4.950 4.930 1.5 87.7 0.6 0.550, 4.035 4.066-2 90.6-0.76 0.446, 4.398 4.383 1.0-71.1 0.38 0.701, 3.927 3.937-0.6 79.5-0.22 0.822, 5.074 5.043 2.4-133.7 0.93 0.351, 3.496 3.492 0.2 97.8 0.08 0.932, 4.337 4.311 1.8 89.7 0.68 0.495, 3.778 3.750 1.8-217.7 0.7 0.485, 4.480 4.425 3.9 77.8 1.5 0.133, 3.632 3.684-3.5 81.1-1.33 0.182, 4.950 4.933 1.3 89.3 0.52 0.602, 4.035 4.059-1.6 92.5-0.61 0.544, 4.398 4.383 1.0-69.0 0.38 0.705, 3.927 3.935-0.5 84.0-0.18 0.861, 5.074 5.042 2.5-138.2 0.95 0.342, 3.496 3.489 0.4 96.1 0.15 0.880, 4.337 4.315 1.5 91.4 0.57 0.569, 3.778 3.750 1.9-224 0.71 0.476
응답자의성격특성과성향점수매칭기법을활용한모바일조사향상방안연구
조사연구 < 표 10> 매칭방법별성격특성편의 (bias) 의보정정도 성격특성편의 ( =10) 소속자료예측모형 평균 표준편차 최대값 pseudo LR > 10.250 8.348 21.384 0.022 128.25 0.000 1:1 2.431 2.668 9.363 0.002 19.52 0.034 1.423 1.239 3.911 0.001 5.50 0.855 1.942 1.281 4.165 0.001 7.99 0.630 1.802 1.182 3.904 0.001 6.98 0.727 Ⅴ. 토의
응답자의성격특성과성향점수매칭기법을활용한모바일조사향상방안연구
조사연구
응답자의성격특성과성향점수매칭기법을활용한모바일조사향상방안연구 참고문헌. 2011. KGSS 2010... 2011.. 12(2): 51-76.. 2010.
조사연구. 17(1): 55-66.. 2003.. 56: 143-166.. 2011.. 20: 1-16.. 2012. : 2011. 13(1): 135-158.. 2009.. 10(3): 21-36.. 2003. :. 4(1): 1-29.. 2011. :. 12(2): 145-157.. 2012. () (). 13(3): 1-31.. 2010.. 11(2): 1-28.. 2001. 16. 2(1): 1-36. Bosnjak, M. 2011. Self-Administered Mobile Surveys. Presented at the Mobile Research Conference 2011, London, Market Research in the Mobile World. April 18 th. Gerber, A. S., G.A. Huber, D. Doherty, C. M. Dowling, and S.E. Ha. 2010. Personality and Political Attitudes: Relationships across Issue Domains and Political Contexts. American Political Science Review 104(1): 111-133. Gosling, S. D., P. J. Rentfrow, and W. B. Swann Jr. 2003. A Very Brief Measure of the Big Five Personality Domains. Journal of Research in Personality 37: 504 528. Guo, S. and M. W. Fraser. 2010. Propensity Score Analysis - Statistical Methods and Applications. California: Sage Publication, Inc.
응답자의성격특성과성향점수매칭기법을활용한모바일조사향상방안연구 Kim, H. S. 2011. Consequences of Parental Divorce for Child Development. American Sociological Review 76: 487-511. Macer, T. and S. Wilson. 2011. Globalpark Annual Market Research Software Survey 2010. London: Meaning. Malhotra, N. and A. J. Krosnick. 2007. The Effect of Survey Mode and Sampling on Inferences about Political Attitudes and Behavior: Comparing the 2000 and 2004 ANES to Internet Surveys with Nonprobability Samples. Political Analysis 15(3): 286-323. Morgan, S. and C. Winship. 2007. Counterfactuals and Causal Inference: Methods and Principles for Social Research. New york: Cambridge University Press. Robert, B. W. and W. F. Delvecchio. 2000. The Rank-order Consistency of Personality Traits From Childhood to Old Age: A Quantitative Review of Longitudinal Studies. Psychological Bulletin 126(1): 3-25. Taylor, H., J. Bremer, C. Overmeyer, J. W. Siegel, and G. Terhanian. 2001. Using Internet Polling to Forecast the 2000 Elections. Marketing Research 13(1): 26-30. Traugott, M. W. 2001. Trends: Assessing Poll Performance in the 2000 Campaign. The Public Opinion Quarterly 65(3): 389-419. Triandis, H. C. and E. M. Suh. 2002. Cultural Influences on Personality. Annual Review of Psychology 53: 133-160. Shapiro, H. 2011. Mobile Versus Online Research. http://www.resolvemr.com/mobile_vs_online_research.pdf Wright, M., J. Citrin, and J. Wand. 2012. Alternative Measures of American National Identity: Implications for the Civic-Ethnic Distinction." Political Psychology 33(4): 469-482. < 2014/1/1, 2014/3/31, 2014/4/15>