SmartPLS - Science topic

SmartPLS and Amos for structural equation modelling. I‘m looking for mentorship in that area of research.

I and some of my friends are working on a project that require AI and to solve the issue, we decided to combine two theories. In order to successfully infuse AI, we had to go through the dimensions of both theories, to form their respective constructs then combine the formed constructs to form our final IV which is third-order construct. We've successfully gone through those stages. However, we got stuck at the interpretation and the discussion of the findings.

The major concern is, should we start the discussion from the firs-order? Even though that is not the main focus of the identified issue.

Or, we should start from the second-order constructs?

Thank you all in advance.

Hello members!

My query concerns moderation analysis in PLS-SEM for SmartPLS 4 software users...

Currently, I am writing a research gap on a certain topic. I have found that three latent variables, X, Y, and Z, have been modeled differently in existing studies. Some studies have modeled the effect of X on Y, others Z on X, others Z on Y, and others Z moderated the effect of X on Y. My novel idea is to combine these variables in a single model (as shown in the attachment).

My question is, can I draw this model as it is in SmartPLS4 and proceed with validation and estimation, or are there steps that I have to follow? My main worry is the moderation part and mediation, as the X now takes the mediator position in this model.

When introducing new constructs or variables into a model and examining their effects in accordance with specific established theory, especially in a context where the theory has not been tested before. Is this theory development or theory testing and confirmation? Given that I need to justify my use of PLS, which is generally recommended for models where the emphasis may be more on theory development.

I am attaching a screenshot of the model which is done on SMARTPLS 4 using Secondary data . Is it technically correct

Is it the difference between value for d_ULS and value for d_G ?

or

Is it the difference between value for d_ULS and value for d_G of Saturated Model and d_ULS and value for d_G of Estimated Model?

To my knowledge, the total effect in mediation reflects the overall impact of X on Y, including the magnitude of the mediator (M) effects. A mediator is assumed to account for part or all of this impact. In mediation analysis, statistical software typically calculates the total effect as: Total effect = Direct effect + Indirect effect.

When all the effects are positive (i.e., the direct effect of X on Y (c’), the effect of X on M (a), and the effect of M on Y (b)), the interpretation of the total effect is straightforward. However, when the effects have mixed or negative signs, interpreting the total effect can become confusing.

For instance, consider the following model: X: Chronic Stress, M: Sleep Quality, Y: Depression Symptoms. Theoretically, all paths (a, b, c’) are expected to be negative. In this case, the indirect effect (a*b) should be positive. Now, assume the indirect effect is 0.150, and the direct effect is -0.150. The total effect would then be zero. This implies the overall impact of chronic stress on depression symptoms is null, which seems illogical given the theoretical assumptions.

Let’s take another example with mixed signs: X: Social Support, M: Self-Esteem, Y: Anxiety. Here, the paths for a and c’ are theoretically positive, while b is negative. The indirect effect (a*b) should also be negative. If the indirect effect is -0.150 and the direct effect is 0.150, the total effect would again be zero, suggesting no overall impact of social support on anxiety.

This leads to several key questions:

1. Does a negative indirect effect indicate a reduction in the impact of X on Y, or does it merely represent the direction of the association (e.g., social support first improves self-esteem, which in turn reduces anxiety)? If the second case holds, should we consider the absolute value of the indirect effect when calculating the total effect? After all, regardless of the sign, the mediator still helps to explain the mechanism by which X affects Y.

2. If the indirect effect reflects a reduction or increase (based on the coefficient sign) in the impact of X on Y, and this change is explained by the mediator, then the indirect effect should be added to the direct effect regardless of its sign to accurately represent the overall impact of both X and M.

3. My main question is: Should I use the absolute values of all coefficients when calculating the total effect?

Good day, everyone. I am analyzing a moderated mediation model, and I need to examine the conditional indirect effects at various levels of the moderator.

The PROCESS in SmartPLS 4 reports the conditional indirect effect, but only at + 1 and 0 values.

My moderator in SmartPLS, I set it as a Ordinal Scale, because I use Likert Scale for my survey.

Can I or where I can get the indirect effect at -1 values?

Hello network,

Is there anyone who could help me with my empirical research - especially with the statistical elaboration - on the topic of entrepreneurial intention and business succession in German SMEs, or who has experience with the preparation of Structural Equation Modeling?

Please feel free to send me a private message.

Thank you and best regards

Julia

Hello guys, I Have a question, I have structural equation model with multiple mediators, 2 mediator variables are dummy and 2 are continuous. I also heard that there is different manners of running moderation analysis in different software.

In terms of dificulty, which software is less dificult to run SEM with multiple mediator variables? AMOS ou SmartPLS? And Why?

Thank you in advance.

Kind regards?

Factor loading is more than 0.92 also ave and cr are more than 0.95? Is this acceptable? If it's not i tried to delete items but the more i delete the more factor loading is higher! how can i treat it?

I have data from population based observation (not questionnaires but yearly observation from secondary database) and I already have a common model for each populations (6 groups, each has the same latent variables, observed variables, and the structural models are also the same shape). As the study is some kind of longitudinal basis (not independent to each other), am I still be able too use MGA (Multi Group Analysis)? My result does not pass the MICOM procedure, is doing a MICOM procedure an obligatory prior MGA in my specific case?

I will be very grateful for this kind act.

My email: hanan.eid@esri.usc.edu.eg

my email : souarwalid36@gmail.com

In my testing of statistical significance by bootstrapping, p_value shows that there are no specific indirect effects, but confidence intervals bias corrected 2.5% - 97.5% does not contain the value 0, this indirect effect has significance (followed source I read).

So there was a conflict and I wondered to use which result. Am I missing something?

I need a sample paper that uses second-order SEM reflective models using SmartPLS.

VIF test should be used for the independent variables only!!!. I get this comment from the reviewer ... what does that means?

My research simply contains one independent variable with two sub-variables and one dependent variable, and I only counted the VIF for the main independent variable ?!!

Since all my measurement scales (for my indicators) are the same, I would like to report unstandardized path coefficients, since they are easier to conceptualize than standard deviations. However, everything I read says to use standardized results for MGA. Why can't I use unstandardized?

Is ex ante power analysis the same as a priori power analysis or is it something different in the domain of SEM and multiple regression analysis? If it is different, then what are the recommended methods or procedures? Any citations for it?

Thank you for precious time and help!

Why do some indicators (NFI, chi-square, D_g) not appear in smartpls results?

When I analyze my research data, the previous tests are all ok, only the NFI test and D_g appear as N/A, and the Chi-square appears as infinite. I suspect this is because the model is composed of second-order constructs.

I would like to know why they appear like this and how I interpret them.

If you can help me, I'd appreciate it!

Suppose, I have indicator x1 which is part of a reflective construct A, can I also use x1 as an indicator for another reflective construct B.

Is it possible to conduct a Meta-SEM (meta-analysis of structural equation models) using the SmartPLS 4 software?

I have a sample composed of two categories, say males and females, when I test a certain path it gives insignificant results in both categories, surprisingly, when I test the same path for the whole sample it gives significant result!!! This doesn't appear when I use Amos (in Amos the path is insignificant for the whole sample too). If testing structural model is not yet reliable in SmartPls cb-sem, can I use it for CFA only and then test hypotheses using Amos? Given that the two softwares give very close results regarding CFA, but SmartPls is just more user-friendly

i have an adopted scale.do i need to still run EFA and CFA. or there is an alternate approach.

Please cite the use of SmartPLS: Ringle, C. M., Wende, S., and Becker, J.-M. 2022. "SmartPLS 4." Oststeinbek: SmartPLS GmbH, http://www.smartpls.com.

please if ANY ONE CAN GUIDE REGARDING USE OF THIS RESOURCE. PLZ SHARE

Dear community, I am working on my thesis data and my data analysis shows that there is no relationship between IV and DV, only the indirect relationship exist in presence of 2 mediators (serial mediation). Whereas the literature supports the positive relationship between the IV and DV.

The r square value for the same is very low 0.052.

*all the measurement model criteria i.e, loadings, Cronbach alpha, ave, htmt, vif value meet the criteria.

* all other relationships are significant (iv-mediator-dv)

Now my question is what could be the possible reason for insignificant relationship between IV and DV.

How should I proceed with this.

Why the value of r square for main DV is so low , whereas for mediators its 0.45 and 0.52.

*If I proceed with the same, would it create any problem at time of my PhD Defence.

If iI use EFA on SPSS to explore the factors and later I intend to check the reliability and validity of these explored factors using SmartPLS rather than CFA on AMOS, will it be valid approach?

These questions explore the distinctions between SmartPLS and AMOS in terms of their statistical methods, modeling capabilities, treatment of data, and user experience. It is essential to understand these differences to make informed decisions about which software tool is best suited for specific research projects and objectives. Researchers can use these questions as a starting point to delve deeper into the comparative analysis of SmartPLS and AMOS for their individual research needs.

To handle second order factor, I used the latent variable scores calculated by smartpls. After attaching that data file with my path, i tried to create groups on the basis of a categorical variable. But it simply doesn't work. It doesn't give an error either. Has anyone faced the similar problem ?

These questions explore the distinctions between SmartPLS and AMOS in terms of their statistical methods, modeling capabilities, treatment of data, and user experience. It is essential to understand these differences to make informed decisions about which software tool is best suited for specific research projects and objectives. Researchers can use these questions as a starting point to delve deeper into the comparative analysis of SmartPLS and AMOS for their individual research needs.

is it necessary to run CFA, if instrument is adopted.(role of EFA and CFA before structural model

is this approach different/ alternate/ substitute

Please cite the use of SmartPLS: Ringle, C. M., Wende, S., and Becker, J.-M. 2022. "SmartPLS 4." Oststeinbek: SmartPLS GmbH, http://www.smartpls.com.

can anyone explain that are these two different approaches. in what ways it can be understood.

Attention Scholars,

I, an assistant professor and researcher in Finance and FinTech, extend a warm invitation for collaborative research projects. My expertise lies in conducting quantitative analysis using SmartPLS, a powerful tool for modeling and analyzing complex data relationships.

If you are working on groundbreaking research in Finance, FinTech, or related Administrative and Financial Sciences and wish to explore the realms of quantitative analysis, I welcome the opportunity to collaborate as a co-author. Together, we can produce impactful research that contributes to our fields.

Please check my Google Scholars and Scopus

Mousa Ajouz (Ph.D)

Palestine Ahliya University

When we check our data set for univariate (e.g., SPSS) and Multivariate Skewness and Kurtosis as measures of non-normality as defined by Mardia (via Mardia' s web of Power) in order to establish if our data set is normally distributed or data is non-normal and on the basis of this we decide to proceed with parametric or non-parametric test such as SmartPLS SEM. My question is that our data is a multivariate type of data, so do still we need to test and report a Univariate normality test performed in SPPS, or we should only report Marida' Multivariate Skewness and Kurtosis results. Particularly, when our aim is to use later on SmartPLS, SEM or AMOS for further data analysis.

Reply to this question from researchers and scholars will be highly appreciated.

Thank you

As a reviewer, I saw that a manuscript effect size (f square) was found to be greater than 1. The study was conducted in SmartPLS. For this reason, I first asked SmartPLS, but did not get any answers. I was able to find "one" answer on SmartPLS's forum that it might answer a similar question and the answer explained "it has a big effect" (https://forum.smartpls.com/viewtopic.php?t=1902#:~:text=There%20is%20no%20such%20thing,effect%20on%20your%20endogenous%20variable.), but the sources it gives are not satisfactory and the answers are not also satisfactory. I asked google if this is possible, but did not get a very satisfactory answer. And I said "try ChatGPT". Here is the answer;

This answer is also not a very satisfying and technical explanation. Can anyone explain this to me?

Other than the independent and dependent variables, there is a moderator and a mediator.

Hello All,

I have one or two articles about PLS-based multi group analyses, but would like to read more on this topic. I saw some on the SmartPLS academy literature, but those are already a few years old. Can anybody recommend others?

Thanks.

Dear colleagues,

I am having trouble inputting moderating variables in newly introduced CB-SEM model in SmartPLS. CB-SEM does not provide a moderating effect function (as in PLS-SEM version). Can someone please direct me on how to perform moderation analysis and access moderating slope diagram for CB-SEM models? Thank you.

The data I collected for my research yielded a non-normal distribution.

I aim to test a hypothetical model using SEM, and AMOS is said to be better for confirmatory research. However, I don't want an inflated model (since the data are not normally distributed).

Accordingly, I have the following questions:

1. Is SmartPLS a good fit for conducting SEM and path analyses, and is that more accurate than Amos for the data that are not normally distributed?

2. Moreover, is it better to use VB-SEM?

I asked the second question because VB-SEM is said to be more flexible regarding non-normality.

I sincerely thank the researchers who will answer these questions.

I am currently working with a study titled "Knowledge, Competency, Adoptability and Sustainability of Artificial Intelligence (AI) Technology Among Physician Entrepreneurs in GCC Countries". I am in dilemma that which SEM model AMOS/SmartPLS will suit for the study. The total responses received so far is 220 out of the population size of 400. The main objective is to find out knowledge , competency, adoptability and sustainability of Artificial intelligence among Physician entrepreneurs.

I will be much thankful to you, if you can extend your kind feedback on this.

Regards

Dr.Sharfras Navas

I am looking for one but I could not get an info. I only found SmartPLS for such a tool, but I am looking for a free alternative.

SmartPLS is an excellent program for structural equational models, however on several occasions it does not save the files well due to compatibility issues. I would like to know if you know of any alternative.

How can you assess the model fit in SmartPLS when you have a dataset with a large number of variables and indicators, and some of these variables may be highly correlated with each other?

What are the best practices for assessing the model fit in SmartPLS when dealing with data that may violate the normality assumption, such as non-normal distributions, outliers, or skewness?

In SmartPLS, how can you evaluate the model fit in the presence of multicollinearity issues, such as high correlations among the predictors or factors, and what are the potential consequences of ignoring these issues for the validity and reliability of your results?

In SmartPLS, what are the most appropriate criteria to evaluate the goodness of fit for a model with both reflective and formative constructs, and how can you address issues related to the reliability and validity of the measures used?

I want to inquire about various methods to prove the contingency perspective in social sciences.

21 items 👉🏼 4 sub-constructs 👉🏼 1 construct

33 items 👉🏼 4 sub-constructs 👉🏼 1 construct

so there are three levels:

lower order 👉🏼 medium order 👉🏼 higher order

The results are significant. But tell me, can we squeeze these 21 and 33 items into above 2 constructs, respectively?

The constructs are metric and ordered .

I need help, pls.

I run the consistent PLS-SEM algorithm on SmartPLS to check for Model fit. However, results showed d_G as N/A, Chi-square infinity, and NFI as N/A.

I have read from past discussions that I should check the inner VIF, fused two multicollinear constructs, but still, the issues on these three parameters are not addressed.

Hello,

I am conducting a survey using a 5 Likert scale. When I tried to analyze the primary results, (none to all) and (strongly disagree to strongly agree) scales. By the way, I have only one question in the survey that uses none to all. It shows a meaningless (negative) correlation between them. I think the issue is (none to all) scale indicates none at 1. However, another scale(strongly disagree to strongly agree) indicates a negative at 1. Maybe that is the reason why I see a negative correlation. Is there any solution to analyze or interpret it? I use SmartPLS for structural equation modeling. Thank you!

I am currently working with a study titled "Knowledge, Competency, Adoptability and Sustainability of Artificial Intelligence (AI) Technology Among Physician Entrepreneurs in GCC Countries". I am in dilemma that which SEM model AMOS/SmartPLS will suit for the study. The total responses received so far is 220 out of the population size of 400. The main objective is to find out knowledge , competency, adoptability and sustainability of Artificial intelligence among Physician entrepreneurs.

I will be much thankful to you, if you can extend your kind feedback on this.

Regards

Dr.Sharfras Navas

I'm using SmartPLS to investigate a model. My question is if the moderator variable (Z) can be also used as predictor of variables (X₁, X₂) which their affects on another variable (Y) is moderated by Z? Or will it cause certain problems?

I'm adding a figure to be clearer.

what is the difference between reflective and formative measurement models?

in smartPLS

I have a SEM study where some variables were collected with a 7-points scale, while others with a 5-point. Is there any literature that I can look at? or any opinion? Thanks.

I have only ever used similarly scaled items in my PLS models. Logically, I can see that would be possible to rescale items with more categories to smaller scales, but with some loss of information - say to convert items in a 5 point Likert having Strongly Disagree, Disagree, Neutral, Agree, Strongly Agree to a 3 point Likert scale of Disagree, Neutral, Agree. I am less convinced that one could go the other direction without adding information that isn't there, or that one could standardize dichotomous variables alongside likert scale variables - but I am not a mathematical statistician. Hence, my question is about mixing indicators in the model...

Suppose we want to create a latent reflective construct using 4 items from a recently administered survey, as follows:

Item 1 - Y or N answers

Item 2 - 3 point likert answers

Item 3 - 5 point likert answers

Item 4 - Continuous variable that can be binned into several categories

Is this advisable or is it ill advised? Please include your rationale.

In SmartPLS 4's Construct reliability and validity output have Composite reliability (rho_a) and Composite reliability (rho_c).

What is the difference between Composite reliability (rho_a) and (rho_c)?

Which one should use for my study?

I am currently assessing a conceptual framework (Model 3 of the Hayes' Process) involving a moderated moderation mechanism. I am using SmartPLS for other analysis. However, Process Macro using SPSS does share support.

I am specifically trying to figure out if it is possible using a way (in SmartPLS) to perform the moderated-moderation analysis?

I was using smartPL3, and the Q square effect size was calculated manually. Currently I'm using the smartPLS4, do you know if it can be computed automatically? because I tried to compute it manually using the following equation, but I received weird results.

q square = (q square included - q square excluded)/(1-q square included).

So far, I knew that we can modify a construct with either formative or reflective indicators in SmartPLS. Is there anyone doing a MIMIC Model in SmartPLS (PLS-SEM)?

Are there any references/published articles that used PLS-sem modeling where the effect size (f-square) was higher than 1?

Hi,

SmartPLS is known as non-parametric software. However, can we use it in a data set that is considered normal according to the kurtosis and skewness values?

In addition, even if the fornell-larcker test is performed, should the correlation values be given from spss before testing the effect hypotheses?

Thanks,

When I want to open my project file in smartPLS 3.0, nothing appears on it, even on the indicator's pannel. In the indicator's pannel there is a message that'' Your data file setup is corrupt. Please open the data file and revise the setup''. I have also changed the file format to csv. I appreciate your kind help in advance!

Hi,

I am using SmartPLS to compare two competing established models and using BIC/CVPAT for that purpose. However, what changes in the models is the causal direction, so the key dependent variable is different in each model (Model 1: Y1; Model 2: Y4; please see the attached representations).

I read in Hair et al. (2022, p. 205), that "Some research situations call for the establishing and comparing of alternative theoretical models (...) Such alternative models all focus on the same endogenous variables but differ in their structure (...)". Does this mean that I can't use model selection criteria in this case? If so, how can I identify which model best approximates the data generation?

Thank you in advance.

Hi, I am using SmartPLS for moderation analysis.

I have three moderators: gender, age, and experience. Two gender groups (male and female), four age groups (less than 30 years old, 30-39 years old, 40-49 years old, and more than 50 years old), and four experience groups (no experience, less than 2 years, 2-4 years, and more than 5 years).

Given the fact that age and experience have more than two groups, which approach in PLS should I use for moderation analysis? Interaction effect? or Multigroup Analysis (MGA)? And how?

If you happen to have some readings, like articles, theses, or books for sharing, I would be very grateful for that! Thank you very much!!

Simpson’s paradox is a statistical phenomenon where the relationship between two variables changes if the population is divided into subcategories. In the following animation, we can see how the linear relationship between two variables is inversed, if we take into account a third categorical variable. Simpson's paradox highlights the fact that analysts should be diligent to avoid mistakes.

How to identify this phenomenon in SmartPLS?

can anyone tell me the difference among Discriminant Validity: specially HTMT - Matrix, HTMT - list and HTMT - Bar chart?

i got them in SmartPLS 4

I am trying to run q square using blindfolding option but start calculation tab is inactive and omission distance tab of.red color.. one specific data set..

For other data set and models it is working? What could be the reason?

Using AMOS in CFA measurement model, can I use control for latent variable? if so what is the appropriate method?

Thanks

I need to analyse the impact of Values perception on Purchase Intention between 2 countries.

Results of MLR is different from PLS, why?

Please enlighten

If we use SmartPLS to analyse the structural equation modeling (SEM) then what could be the appropriate sample size? Is there any minimum and maximum sample size is required to analyse the PLS-SEM?

Hello Fellow SmartPLS Researchers,

I conducted a multi-sample analysis in SmartPLS to assess the significance of structural path differences between business and non-business majors for a hypothesized model. The results only revealed one statistically significant between-group difference, and that was somewhat moot as the paths for both groups were significant and in the predicted direction. However, among the paths for which no significant between-group differences were measured, two were significant for non-business majors only, and one was significant for business majors only. (I should note that all three of these paths were significant and in the predicted direction for the full sample). Does anyone have a suggestion as to how to interpret and/or write-up the results under these circumstances? It would seem important to note that there are deviations from the full sample results for each group even though the differences are not statistically significant between groups. Any assistance with addressing this issue would be greatly appreciated.

Ken

SmartPLS: What should I do if discriminant validity HTMT>0.90? Is any alternative available for SMART PLS?

One of the revisions that the panel is for me to place a mediation table to finally show that the said study has a mediation. I already placed a table (the one from SmartPLS reference) in the Research Methodology chapter prior to the recommendation of the panel. I used SmartPLS as the statistical tool. How to textually present it?

Hello everyone,

I have difficulties with my analysis of the UTAUT2 model. I have been working intesively with the model and the different analysis options.

However, I have not changed the model structure.

Still, I do not understand how to proceed in my analysis:

1. do I have to analyze the use behavior directly as a reflective construct?

2. do I have to perform two analyses with and without use behavior?

3. what can I do if my values for discriminant validity are above 0.97?