Ask Power - Recent questions and answers in Use of WebPower

Answered: clarification of input for moderated mediation model 7

Sat, 20 Sep 2025 01:08:00 +0000

Thank you for the feedback.

1. the coefficients are unstandardized ones.

2. the value needs to decide by the user based on the research questions.

Answered: WebPower removed from CRAN?

Wed, 20 Jul 2022 15:19:33 +0000

This was because of the update of package WebPower depends on. We will update the package in the next couple of days.

At the same time, you can download and install it from the archive of CRAN here: https://cran.r-project.org/src/contrib/Archive/WebPower/WebPower_0.7.tar.gz

Thanks.

Answered: Power Curve for multigroup mediation models, non normal data and a continious moderator

Sun, 23 Jan 2022 15:03:49 +0000

Yes, you can generate a power curve for multiple group analysis, but not automatically through the wp.mc.sem.power.curve function. You can calculate the power for each sample size and then plot the curve. Below is the code used in Zhang (2013) for multiple group power analysis.

library(bmem)
ex3model<-"
y ~ start(c(.283, .283))*x + c(c1,c2)*x + start(c(.36, .14))*m +c(b1,b2)*m
m ~ start(c(.721, .721))*x + c(a1,a2)*x
m =~ c(1,1)*m1 + start(c(.8,.8))*m2 + start(c(.8,.8))*m3
x ~~ start(c(.25, .25))*x
y ~~ start(c(.81, .95))*y
m ~~ start(c(.87, .87))*m
m1 ~~ start(c(.36, .36))*m1
m2 ~~ start(c(.36, .36))*m2
m3 ~~ start(c(.36, .36))*m3
"
indirect<-'
med1 := a1*b1
med2 := a2*b2
diffmed := a1*b1 - a2*b2
'
bootstrap<-power.boot(ex3model, indirect, nobs=c(400,200), nrep=2000, nboot=1000, parallel='multicore')
summary(bootstrap)

No. Currently, only continuous data are supported. If x is only used as a predictor, you might ignore it.
We have implemented a few calculators (from the R package pwr2ppl) based on the moderated mediation models discussed by Hayes. However, they are not fully tested yet. See here https://webpower.psychstat.org/wiki/models/index#mediation_analysis

Answered: A question on the syntax for power analysis using the "wp.mc.sem.basic" function

Mon, 03 Jan 2022 20:27:47 +0000

You can try adding std.lv = T in the function so that

power.2f <- wp.mc.sem.basic(model = model.2f, indirect = NULL, nobs = 100, nrep = 5000, std.lv=T)

Answered: Power analysis for data-model fit using a Monte-Carlo approach

Mon, 03 Jan 2022 19:26:16 +0000

We may add this in the future but don't have a specific timeline for it. Thanks.

Answered: how to perform sensitivity power analysis for independent samples t-test using our effect size?

Mon, 03 Jan 2022 19:25:02 +0000

The question needs clarification but you can calculate your effect size based on t-test for power analysis.

Answered: how to determinate ng and nm in wp.rmanova

Mon, 03 Jan 2022 19:23:00 +0000

Currently, wp.rmanova does not support the analysis with two within-subject factors. It can conduct analysis for one within-subject factor or one within-subject factor and one between-subject factor.

Answered: Power calculation for model with ordinal outcome variable

Fri, 01 May 2020 16:19:33 +0000

wp.regression is for continuous outcomes. WebPower supports logistic regression with a binary outcome and one predictor with different distributions but does not support regression with a categorical outcome.

Answered: Does wp.regression() alpha parameter specify a one-tailed or two-tailed test?

Fri, 01 May 2020 16:03:42 +0000

It only used two-sided test ~ An F-test is used.

Answered: Effect size in RM ANOVA: f or d?

Fri, 24 Apr 2020 00:22:55 +0000

You are right that the effect sizes are defined differently for d and f in t-test and repeated-measures ANOVA. See Section 8.3 in the manual for the definition of f: https://webpower.psychstat.org/wiki/_media/grant/webpower_manual_book.pdf

Answered: post hoc power analysis

Wed, 19 Feb 2020 01:37:28 +0000

Using a?0.39 and b?0.39 should work. Please include a link to your analysis in the error output if something went wrong.

In general, post-hoc power analysis isn't recommended since it does not add anything new to your sample.

Answered: Please send us the URL to your analysis day=20200207&base=67d262748ad87d3963c776f660248cad so that we can fix it.

Fri, 07 Feb 2020 18:43:21 +0000

Thank you for the question. This is a bug in our program and should have been fixed.

Answered: Sample size for a 3-way moderation and moderated mediation

Wed, 25 Sep 2019 19:17:29 +0000

It might be too late to input on this. But you can simulate X, Z, W, Y first and then calculate the interaction terms.

Answered: Effect Size for 3-arm cluster randomized trial

Wed, 25 Sep 2019 19:15:57 +0000

You can find the formular in https://webpower.psychstat.org/wiki/_media/grant/webpower_manual_book.pdf.

Representing interaction terms in montecarlo SEM power

Tue, 05 Mar 2019 16:37:41 +0000

Hello,

I've been rethinking my previous question and I think, I would be able to solve that one if I knew how to represent interaction terms between exogenous variables with lavaan based syntax for a sem montecarlo power estimation.

Is there anyway I can do it?

For example when dealing with moderated mediation?

Answered: multilevel modeling power analysis

Wed, 21 Nov 2018 06:58:43 +0000

Great question. Really interesting

Answered: accounting for clustered data

Tue, 31 Jul 2018 12:42:41 +0000

Currently, WebPower does not suppose this.

Answered: Where can I download the WebPower package for R?

Sat, 21 Apr 2018 17:43:12 +0000

The R package WebPower can now be installed within R.

https://cran.r-project.org/package=WebPower

Answered: Could you please guide me in calculating skewness and kurtosis (Mardia's multivariate coefficient): load data, etc

Sun, 27 Aug 2017 00:28:36 +0000

See if this helps:

http://nd.psychstat.org/blog/estimate_sample_skewness_and_kurtosis_in_popular_statistical_software

Answered: But why you equalize S with the standard error (S.E.)?

Mon, 14 Mar 2016 11:55:04 +0000

I assume they were sample standard deviations instead of s. e. s. Otherwise, the variances seemed to be a little bit too different. But you can simply change Tue formulas to reflect they are s. e. s.

Answered: Can I estimate the detectable differences with Web Power?

Fri, 11 Mar 2016 14:56:28 +0000

Let me provide a little more details here.

First, use the two-sample t-test power analysis, one can get the effect size is 0.967 with a power of 0.8 and the sample sizes \(n_1=16\) and \(n_2 = 20\).

Suppose the the common standard deviation of the two population can be estimated using the pooled standard deviation such that

\[ s_p = \sqrt{ \frac{(n_1-1)s_1^2+(n_2-1)s_2^2}{n_1+n_2-2} } = 7.32\]

assuming \(s_1=8.38\) and \(s_2=6.35\). Then the detected difference should be 0.967*7.32 = 7.07.

Answered: but in that case, where do I put the mean values of my samples?

Thu, 10 Mar 2016 14:54:34 +0000

There is no need to input mean value and standard deviation. The purpose is to estimate the effect size which is a function of mean and standard deviation. You would need to input information on sample size and power. The output is the effect size. Note that the sample effect size is equal to the ratio of mean or mean different divided by the standard devaition, you can calculate the mean or mean difference to be the effect size multiplied by the standard deviation. But usually, effect size itself is enough.

Answered: How do I calculate sample size using a power

Tue, 23 Sep 2014 16:00:27 +0000

I am sorry we haven't compiled a complete manual to use our calculators.

To calculate the sample size using any of the calculators, you can input the desired power in the "Power" field and then remove the sample size in the "Sample size" field (leave it blank). Then the output will display the needed sample size according to the power you specified.