![]() Despite its popularity, issues concerning the estimation of power in multilevel logistic regression models are prevalent because of the complexity involved in its calculation (i.e., computer-simulation-based approaches). These issues are further compounded by the fact that the distribution of the predictors can play a role in the power to estimate these effects. We introduce and study a family of robust estimators for the functional logistic regression model whose robustness automatically adapts to the data thereby leading to estimators with high efficiency in clean data and a high degree of resistance towards atypical observations. ![]() To address both matters, we present a sample of cases documenting the influence that predictor distribution have on statistical power as well as a user-friendly, web-based application to conduct power analysis for multilevel logistic regression. MethodĬomputer simulations are implemented to estimate statistical power in multilevel logistic regression with varying numbers of clusters, varying cluster sample sizes, and non-normal and non-symmetrical distributions of the Level 1/2 predictors. Power curves were simulated to see in what ways non-normal/unbalanced distributions of a binary predictor and a continuous predictor affect the detection of population effect sizes for main effects, a cross-level interaction and the variance of the random effects. Skewed continuous predictors and unbalanced binary ones require larger sample sizes at both levels than balanced binary predictors and normally-distributed continuous ones. I already know from the literature that your odds go up! And my piolet study shows that hypertension is the most prevalent of comorbidities in the headache cohort.In the most extreme case of imbalance (10% incidence) and skewness of a chi-square distribution with 1 degree of freedom, even 110 Level 2 units and 100 Level 1 units were not sufficient for all predictors to reach power of 80%, mostly hovering at around 50% with the exception of the skewed, continuous Level 2 predictor. bivariate linear regression, (4) multiple linear regression based on the random predictor model, (5) logistic regression, and (6) Poisson regression. I am only interested in records with a headache diagnosis, but I want to determine what set size I require measuring an increase or decrease in odds given the presence (exposure) to hypertension. Is this what I need to add into the probability textboxes, and what do I add into the X param pi (I think this is the proportion of some distribution)? I assume this is my alternative hypothesis P(headache=1|hypertension=1). The continuous predictors come in two types: normally distributed or skewed (i.e. ![]() The prevalence of a patient having a headache with hypertension if 0.332. GPower z-test: Logistic Regression (continuous predictor) Davey 375 subscribers Subscribe 53 Share Save 17K views 6 years ago Use GPower to compute power for a binary logistic regression. This app will perform computer simulations to estimate power for multilevel logistic regression models allowing for continuous or categorical covariates/predictors and their interaction. I assume this is my null hypothesis P(headache=1|hypertension=1) i.e., hypertension does not impact the outcome of a headache. ![]() To calculate the sample size, the following parameters were established. The prevalence of having a headache, irrespective of hypertension is 0.004 (from ). Sample size calculation was performed using the GPower 3.1 Program for binary logistic regression 54, 55. If you detected an effect more than (e.g.) 80 of the time, you are overpowered - reduce n and start over. Count how often you did detect an effect. I am trying to see if hypertension (my binary independent) makes a difference on having a headache (my binary dependent). Run your analysis Record whether you detect a statistically significant effect Do these steps many times, on the order of 1000 or more times. I need some help understanding what values I need to put into a two-tailed logistic regression power analysis are in G*Power.
0 Comments
Leave a Reply. |