Best statistical test for longitudinal study?

Dear Colleagues

I am wondering if any one could help suggest the most appropriate statistics for my query.

I am looking at the time taken for patients to complete an electronic questionnaire and would like to assess whether the time taken is impacted by age group, treatment modality and employment status. Patients have completed the questionnaires over 10 different time points.

I know the time points will impact completion time (patients become familiar with questionnaire over time). So have accounted for this as a variable.

I would like to adjust for the time points but the focus of analysis would be on the dependent variables listed above.

I have considered multivariable analysis or a linear mixed model, but am now getting confused for which is the best to proceed with.

Grateful for your advice!

Dear colleagues,

I am new to R, I have been using Amos for conducting CFA but came to learn that when dealing with Likert scale, Lavaan in R is a better tool to do such analysis.

I started using lavaan to conduct CFA and report its fit indices.

I have four factors after conducting EFA, 4 to 5 items under each factor.

I did the following:

cfamodel <- "

communication =~ Com1+Com2+Com3+Com4+Com5

networking=~ Net1+Net2+Net3+Net4+Net5

problemsolving =~ ProblemS1+ProblemS2+ProblemS3+ProblemS4

tech =~ Tech1+Tech2+Tech3+Tech4+Tech5"

fit <- cfa(cfamodel, data = rdata, ordered = c("Com1", "Com2", "Com3", "Com4", "Com5", "Net1", "Net2", "Net3", "Net4", "Net5", "ProblemS1", "ProblemS2" ,"ProblemS3" , "ProblemS4", "Tech1", "Tech2", "Tech3" ,"Tech4", "Tech5"))

However, I get this warning:

Warning message:

In lav_model_vcov(lavmodel = lavmodel, lavsamplestats = lavsamplestats, :

lavaan WARNING:

The variance-covariance matrix of the estimated parameters (vcov)

does not appear to be positive definite! The smallest eigenvalue

(= -2.260671e-18) is smaller than zero. This may be a symptom that

the model is not identified.

I tried finding similar issues online for potential answers, but I failed to find so.

I found one place in the internet talking about the same warning as for second order models, my model is not a second order model. in fact is only CFA after EFA, as I don't have dependent variables to go for SEM structure model.

I would like to know if this warning means I can not continue?

if there is a similar thread that I could not find, kindly share it here.

The other doubt I have is; can I draw a neat CFA model through lavaan (attached the output of my try)?

Thank you very much in advance.

Mugaahed

If I have two correlations: one between variable 1 and variable 2; second between variable 1 and variable 3. What's the best way to analyze the data so that I am able to make the assumption that the correlation between variable 1 and variable 2 is stronger than the one between variable 1 and variable 3. Can I make that assumption just based on the fact that one correlation coefficient is higher than the other, or do I have to make further tests? If so, what's the best way to do that in SPSS.

Thank you very much for helping me with this question.

I recently included GEE models in the statistics and calculated it with the wald chi square test.

Does anyone know how to correctly report the findings considering APA-Guidelines?

e.g we would report the findings of a rANOVa the following:

"No main effect of group factors F(1,92)=.52, p > .05"

How do you report these findings? Please find an output of the model attached. Thank you!

In a mediation analysis, the direct effect between DV and IV is non-significant. However, the indirect effect becomes significant when the mediating variable comes into play. I often read in sources that for mediation analysis there must be a relationship between DV and IV. According to my logic, if a variable makes a normally non-significant relationship significant, it should be possible to talk about a significant mediating effect there. How would you interpret this? And would such a situation pose a problem? Thank you in advance for your answers along with your source suggestion on the subject.

Dear researchers,

As I want to run multilevel models, I face some errors that I think originate from Grand Mean Centering or Group Mean Centering procedures. However, these two methods are somehow vague and confusing.

If my data-set has the same scale, for example, the seven-point Likert scale, Is it necessary to accomplish grand or group mean centring?

Would you please tell me when I should apply grand mean centering, and when group mean centering?

Are there any resources that teach how to accomplish these two methods in multilevel analysis?

Hello,

I am currently writing my dissertation and have a few questions regarding reporting my results. I am not very confident in statistics, so please bear with me if this doesn't make any sense.

I am using a two-sample Wilcoxon test for my statistical analysis which I am doing through R. I am struggling with how to report the results as am only given a W and p-value. However, I have seen that most people report their results using an additional U and Z value. Is it ok to report using only the W and p value?

Additionally, I have read conflicting advice on whether to report the mean or median values! Could someone please help? My data doesn't follow a similar distribution pattern, so I am unsure if reporting medians would be viable. However, I have been told that I shouldn't report means and standard deviations on non-parametric tests. I am a little confused about what the best thing is to do.

Please can someone help!

In a simple mediation analysis I have a significant indirect effect, but no direct or total effect. I know some studies report this indirect effect as evidence of mediation, but from the examples I found it appears this exists for cases of suppression. From Kenny's website suppression exists if c' is the opposite sign to ab (and cites MacKinnon et al., 2007). However, in my analyses the c' (the direct effect) is the same sign to ab. When c' is positive ab is positive, and in another analysis when c' is negative ab is negative. Can I conclude mediation from this? If so, does anyone have examples of papers that do this? Thanks in advance!

What is the best test to use to compare mean scores between two groups of unequal size? One group is n=4 and the other is n=68. The n=4 group doesn't have enough subjects to really test for normality so I'm not sure if a t-test for independent means will work. I'm thinking probably a Mann-Whitney U test. Any suggestions? Is it even possible to compare the means between the two groups with such a difference in size?