Download presentation

Presentation is loading. Please wait.

Published byPrince Dobbins Modified over 3 years ago

1
Missing Data Analysis

2
Complete Data: n=100 Sample means of X and Y 0.0479 10.1720 Sample variances and covariances of X Y 0.7543 3.1699 3.1699 17.2721 Population mean of Y is 10 Y = a + b X + e with b 0 Grup1: only X (when X =< 0) n1=46 Means -0.7087 S: 0.3187 GRup 2: X and Y when X > 0 n2: 54 Means: 0.6924 12.8277 S: 0.2183 0.9851 0.9851 8.8516 X<0 X>0 X Y (taking an exam if passes X exam) MAR: We want to estimate the mean of Y for the complete sample !! Sample: MCAR, MAR NON-MAR (Rubin, 1984)

3
Ignoring Type of Missingness /TITLE analysis ignoring missing data problem (complete cases) /SPECIFICATIONS CASES=54; VARIABLES=2; ANALYSIS=MOMENT; MATRIX=COVARIANCE; METHOD=ML; GROUPS=1; /EQUATIONS V1 = 3.9*V999 + F1; V2 = *V999 + F2; /VARIANCES F1 = *; F2 = *; /COVARIANCES F1,F2=.3*; /PRINT EFFECT=YES; /MATRIX 0.2183 0.9851 8.8516 /MEANS 0.6924 12.8277 /END

4
... ignoring type of missingness V1 =V1 =.692*V999 + 1.000 F1.064 10.789 V2 =V2 = 12.828*V999 + 1.000 F2.409 31.389 V F --- --- I F1 - F1.218*I I.042 I I 5.148 I I I I F2 - F2 8.852*I I 1.719 I I 5.148 I V F --- --- I F2 - F2.985*I I F1 - F1.234 I I 4.209 I I I V F --- --- I F2 - F2.709*I I F1 - F1 I I I

5
ML estimation (MAR) /TITLE Example of multiple group and missing data /SPECIFICATIONS CASES=54; VARIABLES=2; ANALYSIS=MOMENT; MATRIX=COVARIANCE; METHOD=ML; GROUPS=2; /EQUATIONS V1 = 3.9*V999 + F1; V2 = *V999 + F2; /VARIANCES F1 = *; F2 = *; /COVARIANCES F1,F2=.3*; /MATRIX 0.2183 0.9851 8.8516 /MEANS 0.6924 12.8277 /END /TITLE group 2 with missing y /SPECIFICATIONS CASES=46; VARIABLES=1; ANALYSIS=MOMENT; MATRIX=COVARIANCE; METHOD=ML; /EQUATIONS V1 = 3.9*V999 + F1; /VARIANCES F1 = *; /COVARIANCES /MATRIX 0.3187 /MEANS -0.7087 /CONSTRAINTS (1,V1,V999)=(2,V1,V999); (1,F1,F1)=(2,F1,F1); /LMTEST /END

6
ML estimation (MAR) V1 =V1 =.049*V999 + 1.000 F1.088.560 V2 =V2 = 9.941*V999 + 1.000 F2.488 20.378 V F --- --- I F1 - F1.752*I I.107 I I 7.000 I I I I F2 - F2 19.578*I I 3.237 I I 6.049 I I I V F --- --- I F2 - F2 3.378*I I F1 - F1.543 I I 6.215 I I I correlation.880*I

7
If we had the complete data /TITLE analysis of complete data /SPECIFICATIONS CASES=100; VARIABLES=2; ANALYSIS=MOMENT; MATRIX=COVARIANCE; METHOD=ML; GROUPS=1; /EQUATIONS V1 = 3.9*V999 + F1; V2 = *V999 + F2; /VARIANCES F1 = *; F2 = *; /COVARIANCES F1,F2=.3*; /PRINT EFFECT=YES; /MATRIX 0.7543 3.1699 17.2721 /MEANS 0.0479 10.1720 /END

8
... if we had the complete data V1 =V1 =.048*V999 + 1.000 F1.087.549 V2 =V2 = 10.172*V999 + 1.000 F2.418 24.353 V F --- --- I F1 - F1.754*I I.107 I I 7.036 I I I I F2 - F2 17.272*I I 2.455 I I 7.036 I V F --- --- I F2 - F2 3.170*I I F1 - F1.483 I I 6.566 I correlation.878*I

Similar presentations

OK

Correlation 1. Correlation - degree to which variables are associated or covary. (Changes in the value of one tends to be associated with changes in the.

Correlation 1. Correlation - degree to which variables are associated or covary. (Changes in the value of one tends to be associated with changes in the.

© 2018 SlidePlayer.com Inc.

All rights reserved.

Ads by Google

Ppt on motion for class 9 download Ppt on personality test Ppt on wireless communication Ppt on public speaking skills Ppt on maggi product Ppt on eisenmenger syndrome symptoms Presentations ppt online form Ppt on high level languages in computer Perspective view ppt on mac Ppt on word association test saturday