![]() ![]() Improvements in the logistic regression module: (1) improved numerical stability (in particular for lognormal distributed covariates) (2) additional validity checks for input parameters (this applies also to the poisson regression module) (3) in sensitivity analyses the handling of cases in which the power does not increase monotonically with effect size is improved: an additional Actual power output field has been added a deviation of this actual power value from the one requested on the input side indicates such cases it is recommended that you check how the power depends on the effect size in the plot window. 31 January 2014 - Release 3.1.8 Mac and Windows ![]() Note, however, that the change affects the results only when N is very small. Negative effect directions, that is, slope|H1 = upper limit. 6 February 2019 - Release 3.1.9.4 Mac and Windowsįixed a bug in t tests: Linear bivariate regression: One group, size of slope. 14 January 2020 - Release 3.1.9.5 Macįixed a bug that caused the “Options” button (which is available for some tests in the main window) to disappear when “Hide distributions & control” was selected. 21 February 2020 - Release 3.1.9.6 Mac and Windowsįixed a bug in z tests: Generic z test: Analysis: Criterion: Compute alpha: The critical z was calculated incorrectly.įixed a bug in t tests: Linear bivariate regression: One group, size of slope: |sy/sx| was sometimes calculated inccorrecty. My approach would be to make a whole lot of simplifying assumptions, and make them on the conservative side (so your power is underestimated, rather than overestimated), and then run a simulation.Changed the behavior of the “X-Y plot for a range of values” which allowed plotting graphs after changing input parameters in the main window without hitting the “Calculate” button which, however, is required to update the “X-Y plot for a range of values” with the new input parameters from the main dialog. SEMs make specifying the model and estimating the power much easier, and you can use a free package like Lavaan (which is part of R).įinally, you can run a simulation, and you can do that in any software you like. Second, you can use a structural equation modeling (SEM) approach. This could be written in any program, but I've never seen it implemented anywhere else. ![]() SPSS has a power option in manova, which is a bit weird and useless (because it's a transformation of the p-value. I know of three approaches to getting the power - first, you can use the SPSS MANOVA function (if you have SPSS). You need to consider that you will be looking at univariate tests and multivariate tests, and for the multivariate tests you need to worry about the sphericity assumption, if that will be violated, if you will use a correction, or if you will use the lower bound estimates. You also need to consider the correlations between the predictors, and you have 8 predictors (5 dummy coded, three continuous) so there are 36 correlations (or covariances, plus 8 variances) that you need to specify. power of manova is affected by the consistency of the effects across predictors and the correlations between the outcome variables - and the correlations between the outcomes isn't something that's often thought about. Specifying a power analysis for a manova (or mancova, same thing really) is hard because there are so many things to think about, and some of these don't get reported in the output. ![]()
0 Comments
Leave a Reply. |