How to Use a Monte Carlo Study to Decide on Sample Size and Determine Power Export

@article{muthen-power, author = {Muthen, B. O. and Muthen, L. K.}, citeulike-article-id = {2151504}, citeulike-linkout-0 = {http://www.statmodel.com/discussion/messages/22/449.html?1183045053}, comment = {"I'm curious to learn how the values of -2, -1.5, -1, and zero map onto the proportions of missingness that you report on page 605. In other words, how does one work backwards from an expected amount of missing data per wave of measurement to obtain the proper regression weights to use in the MODEL MISSING section of the syntax?" "The values are logits. To turn them into probabilities, use the formula p = 1 / (1 + exp (-logit)) for y1 the logit of -2 results in the probability of 0.12. For y2-y4, the logit is based on the covariate also. The logit for y2 is logit = -1.5 + bx; For x=1, logit = -1.5 + 1*1 = -.5 The probability for a logit of -.5 is .38." http://www.statmodel.com/discussion/messages/22/449.html?1183045053 ---=note-separator=--- -2 =0.119202922, -1.5 =0.182425524, -1 =0.268941421, 0 =0.5, 0.5 =0.622459331, }, journal = {Structural Equation Modeling}, keywords = {carlo, equation, modelling, monte, mplus, power, sample, size, structural}, number = {4}, pages = {599--620}, posted-at = {2007-12-20 12:34:45}, priority = {0}, title = {How to Use a Monte Carlo Study to Decide on Sample Size and Determine Power}, url = {http://www.statmodel.com/discussion/messages/22/449.html?1183045053}, volume = {9}, year = {2002} }

TY - JOUR ID - muthen-power L3 - citeulike-article-id:2151504 IS - 4 JF - Structural Equation Modeling EP - 620 TI - How to Use a Monte Carlo Study to Decide on Sample Size and Determine Power VL - 9 SP - 599 KW - carlo KW - equation KW - modelling KW - monte KW - mplus KW - power KW - sample KW - size KW - structural N1 - "I'm curious to learn how the values of -2, -1.5, -1, and zero map onto the proportions of missingness that you report on page 605. In other words, how does one work backwards from an expected amount of missing data per wave of measurement to obtain the proper regression weights to use in the MODEL MISSING section of the syntax?" "The values are logits. To turn them into probabilities, use the formula p = 1 / (1 + exp (-logit)) for y1 the logit of -2 results in the probability of 0.12. For y2-y4, the logit is based on the covariate also. The logit for y2 is logit = -1.5 + bx; For x=1, logit = -1.5 + 1*1 = -.5 The probability for a logit of -.5 is .38." http://www.statmodel.com/discussion/messages/22/449.html?1183045053 N1 - -2 =0.119202922, -1.5 =0.182425524, -1 =0.268941421, 0 =0.5, 0.5 =0.622459331, AU - Muthen, BO AU - Muthen, LK PY - 2002/// UR - http://www.statmodel.com/discussion/messages/22/449.html?1183045053 ER -