Science method

Structural Equation Modeling - Science method

This group is intended for researchers interested in various applications of structural equations
Questions related to Structural Equation Modeling
  • asked a question related to Structural Equation Modeling
Question
3 answers
While doing SEM, even though all other model fit indices are acceptable, PClose value stays 0. What does this indicate? Is it due to non-normality of sample???
  • asked a question related to Structural Equation Modeling
Question
2 answers
I am attempting to conduct a genomic SEM with a GxE term and a term for the effect of G on E. The model would also include 11 other covariates. The study was a case-control design.
I used the R package gwsem to conduct a GxE analysis without the effect of G on E. However, I can not use it for the current analysis since the package does not permit a term to be a covariate and a dependent variable.
I have tried to create a custom model by writing mxPath statements from OpenMx, but I ran into two problems. One, I am not sure how to specify the interaction term. Two, I get an error message when I run mxRun that the E term is not specified although it is included in two path statements: once as from=E and once to=E.
How would I specify such a model in gwsem or OpenMx? Or is there another R package or program to create this type of model?
I would greatly appreciate any help in specifying my model.
Relevant answer
Answer
Hello Qamar Ul Islam,
Thank you for the references. I will definitely check them out.
Paul
  • asked a question related to Structural Equation Modeling
Question
10 answers
Is it possible that the particle size computed from atomic force microscope image analysis be much lower than that computed from the scanning electron microscope image results?
Relevant answer
Answer
Jamal,
There is 78x78 nm area for the AFM image, and 2890 x 2890 nm for the SEM image. It looks like you measure with AFM nanoparticles on the surface of a micro particle, observed by SEM.
  • asked a question related to Structural Equation Modeling
Question
5 answers
I have extended the UTAUT2 model and I have already developed my model hypotheses. I read about various methods/techniques used such as (Factor Analysis, Partial Least Square (PLS), Structure Equation Model (SEM), Regression Analysis) and tools such as (SPSS, Smart-PLS, Mplus, R, PLS-graph, AMOS). But I am not sure which one to use.
Please I would like some advice about recommended techniques and tools. I am really looking to find out what you consider to be the most efficient method and your rationale as cost and time are limited factors.
Many thanks
Relevant answer
Answer
smart pls is do better than spss based my experiences
  • asked a question related to Structural Equation Modeling
Question
10 answers
1) In the first situation which I am facing, both indirect (a*b) and direct (c') effects are insignificant, while their sum, i.e. total effect [(a*b)+c'] is significant.
2) In the second scenario, the indirect effect is insignificant, while direct effect is significant, so there is no mediation, but yet, the total effect is significant.
I would appreciate if someone can share interpretation of these scenarios. What role does total effect play in such a scenarios? What is the best way to report such situations in a research paper?
P.S. I am using confidence interval method to assess significance (whether 0 falls in between LL and UL or not). Attached are the screenshots of both the scenarios.
Relevant answer
Answer
Thanks Michael R. Frone this was really helpful!
  • asked a question related to Structural Equation Modeling
Question
3 answers
Hi dear scholars! I have a 64 sample size for applying SEM. I know this sample size is samll. I have 7 latent variables. I am developing covariance-based SEM analysis using AMOS. my data is not normal that's why I want to apply Unweighted Least Square (ULS). when I select the ULS option in AMOS then I got no result. For ULS analysis does not run. I don't obtain Red arrow for obtaining results. I don't understand where the problem is. when I select the Maximum likelihood option, I obtain results, but my data is not normal and I want to apply ULS. I am using 64 countries' data yearly. I want to get international trade data from UNCTAD Statistics Database. but I did not understand the data from there. so I took data from the world bank. dear scholar, please guide me I am really worried. I will myself do work just guide me and tell my mistake. why ULS is not running and Maximumlikehod provides covariance greater than 1. I also don't want to run Maximumlikehod because my data is not normal.
Relevant answer
Answer
You may use smart PLS software. If sample size below 100 this software more effective than AMOS.
  • asked a question related to Structural Equation Modeling
Question
5 answers
I am planning to investigate the socio economic factors: age, income, gender, labor affecting for the adoption of farming practice in rural farmer community. Structural equation model is using to build a relationship among the factors. In what way do I need to prepare the questionnaire for that.
Relevant answer
Answer
Structural Equation Modeling (SEM) is a multivariate statistical analysis procedure that examines the structural relationships between variables of interest. In other words, SEM yields an informative representation that explains the relationships between latent and manifest variables, as well as between latent variables. Manifest variables are observed/measured variables capturing or quantifying a construct of interest. Latent variables, on the other hand, are unobservable variables representing constructs of interest. While developing your survey, you should make sure you have adequate latent and manifest variables in line with your study’s objectives. You could go through the article by Davvetas et al. (2020) for an elaboration on ten basic questions germane to SEM. Besides, you might consider the book by Saris and Gallhofer (2014) for insights into questionnaire development and validation.
Davvetas, V., Diamantopoulos, A., Zaefarian, G., & Sichtmann, C. (2020). Ten basic questions about structural equations modeling you should know the answers to – But perhaps you don’t. Industrial Marketing Management, 90, 252–263. https://doi.org/10.1016/j.indmarman.2020.07.016
Saris, W. E., & Gallhofer, I. N. (Eds.). (2014). Design, evaluation, and analysis of questionnaires for survey research. John Wiley & Sons. https://doi.org/10.1002/9781118634646
Good luck,
  • asked a question related to Structural Equation Modeling
Question
4 answers
Are multigroup analyzes via structural equation model an alternative to investigate the influence of background variables, considering the model of the Theory of Planned Behavior?
If so, would this also be a way to deal with confounding factors?
Relevant answer
Answer
Thanks for your response, Profa. Mengye Yu !
  • asked a question related to Structural Equation Modeling
Question
7 answers
Dear researchers,
I am looking for a brief and comprehensive resource that illustrates running various types of Structural Equation Modeling (SEM) in Mplus software. I am so grateful if you share any information about this topic.
Relevant answer
Answer
sound advice, for sure. I have been working with Mplus for a while and I concur with your assessment. I merely suggested it because I was unsure as to what the actual question was.
Best
Marcel
  • asked a question related to Structural Equation Modeling
Question
5 answers
I performed a CFA and got model fit. however, while performing path analysis by taking composite variable to test the hypothesis, CFI, GFI, Chi-square of SEM path model are above threshold limit but RMSEA is above 0.9. Is it ok to report only CFI, GFI and Chi-square? If report RMSEA too, Is model still called fit?
Relevant answer
Answer
Rare cases you may find this, I believe you are running AMOS with more than 500 sample data set.
RMSEA must be less than 0.8 according to the references
  • asked a question related to Structural Equation Modeling
Question
6 answers
Hello! I'm currently testing out a SEM model using R for the first time and I was wondering whether I might have some help interpreting my RMSEA output. I received an RMSEA value .044 (CI .041, .046) and a p-value of 1. Why would the p-value be 1?
Relevant answer
  • asked a question related to Structural Equation Modeling
Question
3 answers
We have a Ga2O3 sample to perform cathodoluminescence with. The CL system uses an SEM machine for the incident beam. In SEM its traditional to plate the sample in Au/Pd, or something similar, prior to observation. Should the sample be plated in Au/Pd for cathodoluminescence as well? Will this greatly affect the results if it is plated versus not?
Relevant answer
Answer
Dear Jacob Blevins , according to Gatan, Delmic and Horiba, specimen preparation for cathodoluminescence (CL) should follow the same principles that SEM´s samples, so basically you only need to coat your sample with a thin conductive film when it is non conductive, with conductive samples, including semiconductors, no conductive coat is necessary. Carbon or metals can be used for coating.
The publications I could read about CL over Ga2O3 did not use any sort of conductive coating, so I imagine the charge dissipation would be good enough through the carbon tape.
May be the following papers would be useful to your own work:
Hope this helps. Good luck with your experiments.
  • asked a question related to Structural Equation Modeling
Question
2 answers
I performed a dry ball-on-disk wear test on In-718 plate, but in my wear profile, I am getting a "hump" in the middle of the wear track. By looking at the SEM images, it appears that this extra material that forms the "hump" is not uneroded material, but is just In-718 that has adhered to the wear track.
Does anyone know of any methods to remove this adhered material on the track? It could be mechanical or chemical, but my objective is to remove this material to get the proper profile of the wear track.
Relevant answer
Answer
I am not sure if you should remove it. It could be wear debris adhered as well. It is known that wear debris can get adhered forming a glazed surface during wear tests. I think you can rather think about why it is forming the bump and then write accordingly in your manuscript.
  • asked a question related to Structural Equation Modeling
Question
3 answers
I am currently working on earthquake risk perception. I have information of perceived probability of occurrence of an earthquake, perceived damage to property, perceived damage to life, perceived level of fear of earthquake (on an ordinal scale). All these four risk perception variables, loads into single latent variable.Let's name the factor score value from factor analysis as a variable Overall Risk Perception. I have information on sociodemographic factors, no of earthquake experienced, time gap from last earthquake experienced etc which I can hypothesize that influence these four risk perception variables. I found from Multiple logistic regression that risk perception parameters have significant relation with four risk perception parameters and overall risk perception.
I would like to hypothesis perceived damage to life and perceived damage to property are related with perceived fear (it would be a two way arrow) and develop a SEM. I do not have any other latent variable. As I have only one latent variable will it be possible to develop Structural Equation Modelling.
Relevant answer
Answer
Marcel Grieger and Sebastian Gyamfi Thank you for suggestions. I am clear now
  • asked a question related to Structural Equation Modeling
Question
6 answers
Calling SEM experts and those with SEM in bacteriology. We ran an air dried layer of LB cultured bacteria over the SEM stub. This is like some canalicules being witnessed. can this be bacteria. ON LB agar, we see swarming effect of this bacteria. Please help?
Relevant answer
Answer
It's crystals from salt in your buffer. You need to fix your specimen and thoroughly wash it with water to get rid of any salts.
  • asked a question related to Structural Equation Modeling
Question
1 answer
I am trying to create an SEM for a microcosm experiment. However, some of the stressor effects or trophic interaction effects by the end of the experiment have disappeared but are evident when we look at the temporal data. My question is whether I can create an SEM that includes selective temporal data with the final data, or would this not make sense?
  • asked a question related to Structural Equation Modeling
Question
2 answers
Respected members,
I want to know how to report such result in research where second order relationship comes into existence.
For eg
Variable A, B and C ; all are individual and separate constructs (First order) as per previous studies.
But in my survey, post checking validity parameters in CFA using AVE, MSV , ASV etc values , it is found that Variable A and Variable B are making second order constructs, then how to justify this second order relationship with theory if Previously no such relationship is established.
Relevant answer
Answer
Hello Antriksha,
First, remember that sample to sample variation can often lead to apparently different implied or observed structures. Ideally, you'd like to see affirmation of your more complex structure in new/different data.
You haven't indicated how much better, if at all, the more complex second-order factor model is compared to what we would presume to be just correlated first order factors. If allowing for factor correlation is sufficient, then why stray from parsimony (and, apparently, previous work)?
If you can make a compelling case that: (a) not only is the second order model superior in terms of reproducing observed relationships among the manifest variables; and (b) may be justified by a sensible modification to the underlying theory, then maybe you have good grounds to argue for an amended theoretical framework. Otherwise, you'll have a hard time selling the idea.
Good luck with your work.
  • asked a question related to Structural Equation Modeling
Question
6 answers
I am using Structural Equation Modeling (SEM) to determine the relationship between job demands and job strain. Five job demands are measured using 3 items, and job strain is measured using 4 items. A competing measurement model strategy utilizing a Confirmatory Factor Analysis (CFA) approach revealed that a model where job demands is estimated by a first-order five factor model comprised out of the five measured job demands and where job strain is estimated as a one-factor model fits the data best.
The next step in my analysis was to add directionality, resulting in the structural model. Estimating the standardized regression weights of the five job demands showed that two job demands, work overload and emotional demands, have a beta higher than |1|. I already checked for multicollinearity using the VIF score and this revealed that the highest VIF score, a score of 2.1, was assigned to emotional demands. This does not clearly indicate multicollinearity.
The emotional demands variable has a significant correlation of .42** with job strain, work overload has a significant correlation of .20** with job strain and emotional demands and work overload have a correlation .56**. Interestingly is that the beta of work overload equals -1.44, which is negative, whereas the its correlation is positive. Further, the beta of emotional demands is 2.11. When all variables are included, the R^2 in job strain equals 0.73.
When removing the work overload variable, the beta of emotional demands decreases to .58 and the R^2 decreases to 0.39. Likewise, when removing the emotional demands, the beta of work overload increases to -.08 and the R^2 decreases to 0.28. Looking at this effect, it seems to me that work overload is a suppressor variable in my model. However, I am not sure if this is the case, nor if it is correct for my standardized regression weights to be larger than |1|.
Does anyone know what to do with this issue?
If you require any additional information or data, please let me know.
Thank you in advance!
Relevant answer
Answer
It looks like a typical suppressor situations due to highly correlated predictor variables as indicated by the standardized path coefficients > |1| and the opposite signs of the path coefficients compared to the zero-order correlations. Since these path coefficients are partial regression coefficients (not: zero-order correlations), they can definitely be > |1|. This does not necessarily point to an improper solution (unless you get a latent variance or residual variance estimate that is negative or a latent R^2 > 1).
The two predictors emotional demand and work overload are probably highly correlated with each other, and so one suppresses variance in the other when predicting job strain. This is not necessarily problematic. I would check out the general literature on suppression effects in path models, e.g.,
  • asked a question related to Structural Equation Modeling
Question
3 answers
I am working on a machine learning based SEM image super resolution. To train my model, I got image pairs with two different resolutions (2048*2048 and 4096*4096).
For each image pair, both images were supposed to show the exact same region on a specimen.
However the image pairs turned out to be slightly misaligned. It is easiest to see on the edges of the attached images.
Is there a tool that is able to cut out the region that both images of one image pair have in common?
Relevant answer
Answer
Downsample the 4096*4096 to 2048*2048, open this image with original 2048*2048 as a stack. Align these two slices using mutual info or SSD algorithm (I used Dragonfly, but it can be done with a plugin in ImageJ). Check the shift (7 pixels in X and 5 in Y), crop it from right bottom corner of original 2048*2048 image. Crop 14 and 10 pixels from original 4096*4096 picture.
  • asked a question related to Structural Equation Modeling
Question
5 answers
Hello everyone,
I am conducting Ph.D. in Finance in which I will use panel data methodology and data collection method will be secondary (Quantitative data). In one of the objective of thesis, I will use simultaneous equation model and in another, I will use structural equation model (SEM). But I do not know any expert who used to conduct workshops on these topics in case of secondary data. Can anyone tell me about the name of some famous experts, I should follow. Please share your knowledge and experience. Thanks a lot.
Relevant answer
Answer
  • asked a question related to Structural Equation Modeling
Question
6 answers
Dear all
I want characterization of green synthesis of silver nanoparticles by SEM,TEM and XRD .
the problem is that my sample is liquid form and all above analysis required sample in power form or thin film
so kindly let me know how to make thin film for SEM ,TEM and XRD
Relevant answer
Answer
XRD part was explained by Dirk Luetzenkirchen-Hecht . As for TEM and SEM you can use standard drop cast method. Take a TEM grid with carbon film. Put a drop of liquid with suspended particles on it. From its side carefully bloat it with a filter paper to remove excess of liquid. Let it air dry for an hour. You can use similarly prepared grids for both TEM and SEM. For SEM just put a double coated carbon sticky tape on a stub and place a grid on it. Environmental SEM can not help you because is will show only the surface layer of a liquid. And of course, no grinding, no ovens.
  • asked a question related to Structural Equation Modeling
Question
9 answers
I need to investigate bacteria morphology
Relevant answer
Answer
You can use HMDS (Hexamethyldisilazane) instead of CPD. It is a liquid with very low surface tension and it can be bought for electron microscopy suppliers. Replace last change of alcohol with HMDS; after 5 minutes replace with fresh HMDS; after 5 minuts pippet out most of HMDS, leaving only enough to barely cover your specimen; let it dry out on air. Cheap and simple replacement of CPD. In most cases works as well or even better.
  • asked a question related to Structural Equation Modeling
Question
4 answers
I have four variables i.e. selfie posting, selfie viewing, group posting and group viewing, respondents were asked to rate frequently and infrequently similarly another variable self-esteem ( 10 statements), need for popularity (12 Statements) and life satisfaction(5 statements) which were asked on 7 points Likert scale (strongly disagree to strongly agree). My question is that how to deployed SEM using AMOS 20.0
The same has been employed in the mentioned paper. bt how?? Let anyone know me.. Thanks well in advance..... Wang, R., Yang, F., & Haigh, M. M. (2017). Let me take a selfie: Exploring the psychological effects of posting and viewing selfies and groupies on social media. Telematics and Informatics, 34(4), 274-283
Relevant answer
Answer
I just done my paper on same concept and luckily got inspired from same mentioned paper. I would be glad to collaborate with you. I am good at smart pls and analysis section.
  • asked a question related to Structural Equation Modeling
Question
3 answers
Dear all,
I trust this email finds you well.
I am currently testing a theorical model using the PLS-SEM and Smartpls. The aim is to assess the influence of certain soft skills (empathy) on client relationship outcomes (e.g customer loyalty).
I would like to control to assess the effect of certain (control) variable on some of my dependent variables.
The issue that I am encountering is the following.
I have 5 control variables:
· Gender (w/ 2 groups: male female)
· Salary (w/ 5 groups: below 20k, 21-30k, 31-50k, 51-70k, +70k)
· Education (w/ 5 groups: no diploma, high school diploma, bachelor, master, PhD)
· Visit frequency (w/ 5 groups: once a year, once every 6 months, ect...)
If I understood things correctly to yield significant result and interpret findings correctly, I will need to use dummy variables. Which means that, based on the above, there will be a total of 17 control variables.
Based on your experience, is there a way to simplify the above? I was thinking to reduce the number of groups for each variable. For instance, instead of having 5 salary categories, reduce the number of categories down to 2 such as:
- Salary: above and below the national average salary.
- Education: those who have at least a master, those who do not.
Etc…
What do you think of the above?
Of course, if you have any information / resources / material / that may allow me to address the above issue, that will be appreciated.
I thank you once more for your assistance and wish you a nice day.
Best regards,
Relevant answer
BITT
  • asked a question related to Structural Equation Modeling
Question
4 answers
Do we need to rescale items that have different categories for a latent variable in SEM? For instance, I have three items measured on a 1-5 scale, one measured on 1-4 scale, and one measured on a 1-6 scale. I have put them all together to create a latent variable and running an SEM analysis. I had the impression that categories of items do not matter to form a latent variable in SEM. But, a reviewer has asked to justify the method with citation of relevant reference, and I am needing a reference for this approach.
Relevant answer
Answer
Ramesh Paudyal, The indicators of a latent variable do not have to be measured using the same scale. If you need an example illustraing this, see page 94 in James B. Grace's book titled "Structural Equation Modeling and Natural Systems". Here, Grace measures Body Size using body mass, wing lenght, and beak length.
  • asked a question related to Structural Equation Modeling
Question
3 answers
Is there any one can help me to answer these comments:
1/More information need to be added in SEM image based on shape and figure prints
2/Please comment on the SEM result. What insight can be obtained from the micrographs? is the applied magnification suitable???
FIGURE: Scanning electron micrographs showing structural changes of pretreated OPWP. (a) untreated OPWP (A and S). (b) Acid and steam explosion pretreatment of OPWP (A1 and S1).
DISCUSSION:
The micrographs showing the surface morphology of untreated and pretreated Orange peel waste (OPWP) are depicted in Figure 3. From Figure 3 (a), the surface of untreated OPWP A and S is irregular, rough and uneven. Also, the particle size and shape are different. Whereas Figure 3 (b), it has more irregular, rougher and more porous surface than the untreated OPWP. In addition, it has a swollen structure. (Borah et al. 2016) described that high pressure and temperature pretreatment causes lignin to fluidize and coalesce, resulting in a globular shape. Thus, the dilute acid sulfuric removes the hemicellulose and causes holes on the biomass surface as seen in Fig. 3 (b). (Saha et al. 2016; Yi et al. 2013) demonstrate that the dilute acid has major effect on cellulosic fibers by compromising the integrity of the cell wall of orange peel with release of sugar molecules.
Relevant answer
Answer
I believe so.
  • asked a question related to Structural Equation Modeling
Question
4 answers
I have fit a random intercept factor model (Maydeu-Olivares & Coffman, 2006) in Mplus by the following syntax. One substantial factor where all item loadings are freely estimated and one random intercept factor where all item loadings are constrained to be 1. The random intercept factor variance is estimated.
Syntax:
ANALYSIS: ESTIMATOR = MLR;
ROTATION = GEOMIN(ORTHOGONAL);
MODEL: G BY a-a20 (*1); # G represents substantial factor
RI BY [email protected]; # RI represents random intercept/method factor
But I got output as following. It looks like that the algorithm automatically takes G as the method factor because the factor loadings from this factor are much smaller than RI and so does its variance (if I delete [email protected] to let the variance be estimated too). What I want the algorithm does is to take G as the substantial factor and estimate its factor loadings and take RI as the method factor and only estimate its variance. Does anybody know how to specify this in Mplus? Thanks a lot!
MODEL RESULTS
Two-Tailed
Estimate S.E. Est./S.E. P-Value
G BY
A1 0.445 0.118 3.771 0.000
A2 -0.085 0.137 -0.624 0.533
A3 0.148 0.111 1.334 0.182
A4 0.106 0.118 0.898 0.369
A5 0.184 0.162 1.137 0.255
A6 0.239 0.095 2.523 0.012
A7 0.444 0.125 3.556 0.000
A8 0.193 0.093 2.077 0.038
A9 -0.008 0.084 -0.092 0.926
A10 0.395 0.133 2.972 0.003
A11 0.094 0.166 0.563 0.574
A12 0.258 0.137 1.879 0.060
A13 0.318 0.117 2.717 0.007
A14 0.712 0.091 7.819 0.000
A15 0.316 0.104 3.027 0.002
A16 0.504 0.129 3.911 0.000
A17 0.290 0.104 2.786 0.005
A18 0.957 0.086 11.064 0.000
A19 0.485 0.164 2.951 0.003
A20 0.484 0.122 3.974 0.000
RI BY
A1 1.000 0.000 999.000 999.000
A2 1.000 0.000 999.000 999.000
A3 1.000 0.000 999.000 999.000
A4 1.000 0.000 999.000 999.000
A5 1.000 0.000 999.000 999.000
A6 1.000 0.000 999.000 999.000
A7 1.000 0.000 999.000 999.000
A8 1.000 0.000 999.000 999.000
A9 1.000 0.000 999.000 999.000
A10 1.000 0.000 999.000 999.000
A11 1.000 0.000 999.000 999.000
A12 1.000 0.000 999.000 999.000
A13 1.000 0.000 999.000 999.000
A14 1.000 0.000 999.000 999.000
A15 1.000 0.000 999.000 999.000
A16 1.000 0.000 999.000 999.000
A17 1.000 0.000 999.000 999.000
A18 1.000 0.000 999.000 999.000
A19 1.000 0.000 999.000 999.000
A20 1.000 0.000 999.000 999.000
RI WITH
G 0.000 0.000 999.000 999.000
Intercepts
A1 3.251 0.096 33.728 0.000
A2 3.246 0.098 32.988 0.000
A3 2.071 0.079 26.213 0.000
A4 1.976 0.083 23.804 0.000
A5 2.915 0.107 27.167 0.000
A6 1.773 0.069 25.742 0.000
A7 2.668 0.091 29.478 0.000
A8 2.431 0.077 31.708 0.000
A9 1.498 0.066 22.761 0.000
A10 2.545 0.092 27.682 0.000
A11 2.910 0.101 28.903 0.000
A12 2.720 0.098 27.665 0.000
A13 1.891 0.079 23.824 0.000
A14 2.346 0.083 28.217 0.000
A15 1.739 0.070 24.819 0.000
A16 3.057 0.091 33.654 0.000
A17 2.066 0.072 28.682 0.000
A18 2.445 0.089 27.498 0.000
A19 3.289 0.113 29.187 0.000
A20 2.654 0.098 27.014 0.000
Variances
G 1.000 0.000 999.000 999.000
RI 0.492 0.073 6.751 0.000
Reference: Maydeu-Olivares, A., & Coffman, D. L. (2006). Random intercept item factor analysis. Psychological methods, 11(4), 344.
Relevant answer
Answer
Christian Geiser Thanks a lot for further suggestions!
I have tried exploratory factor analyses but, unfortunately, the resulting factors are perfectly correlated (above .80). That's why I prefer a one-factor solution...
  • asked a question related to Structural Equation Modeling
Question
3 answers
The Probability Level (P-value?) of Structural Equation Model (SEM) is ,560. Some sources claim that a high Probability Level makes The Default Model insignificant. In contrast, some other sources claim that a low Probability Level indicates a poor fit. Which one is correct? How should the model be interpreted?
The sample size is 444. The Model Fit Indices' screen shoot in the attachment.
Relevant answer
Answer
A p value of (p =.56) indicates that the model is not fit. You might want to delete certain items with low factor loading (below .40) and check for large modification indices and high standardized residual covariance to get acceptable model fit values.
  • asked a question related to Structural Equation Modeling
Question
7 answers
When we are evaluating the remineralization in the interface layer between the sealant and enamel using scanning electron microscopy (SEM)could we assess in the same time the mineral content in a quantitative way of the interface enamel?
Relevant answer
Answer
I do not understand what type of remineralization you expect to find, but crystal structure of healthy enamel is pretty dense and do not allow any significant infiltration of a resin. Of course, dental bore makes rather rough surface of an enamel, and applied resin will fill all the crevices, so there will be no flat, even interface. Additional difficulty is that resolution of standard EDS procedure for an enamel is about 2 microns (some paper erroneously state that infiltration is the same 2 micron, their authors just do not know what "resolution" means). In your case mapping is rather useless, even if pretty. I would advise to use line scan, as was already suggested by Ahmed Samir Bakry .
  • asked a question related to Structural Equation Modeling
Question
7 answers
Hello, I have a few samples of SS316, polished and etched. They are mounted in bakelite (which in non-conductive). I would like to take a few SEM images to see the microstructure. I was wondering if I deposit a carbon layer, is it going to cover the surface of the samples, so the microstructure will not be shown? Any thought on how to make it conductive.
As of Ni paste, I need to add carbon coating to avoid charging of bakelite.
Relevant answer
Answer
Avoid any coating, especially with metals (Au, Pt, and such). It is possible you'll need to use EDS and coating will give artefacts. Use conductive paste/glue (carbon or silver), you can find it in any SEM lab. Make a conductive pass (thick line) from you specimen to the bottom of Bakelite disk, where it will come in contact with SEM stage. You specimen will be sufficiently grounded.
  • asked a question related to Structural Equation Modeling
Question
4 answers
I am going to fit structural equation modeling despite the violation of normality assumption,Am I correct to do SEM?
Relevant answer
Answer
I had the same problem of multivariate nonnormal data. You may use bootstrapping method. I'm using AMOS and it provides bootstrapping in View>Analysis properties>Bootstrap. This method helps to take random samples from your data and creates a population of which you specified the total number (such as 500, 1000, 5000). Hence, you may report estimates created by bootstrapping method.
  • asked a question related to Structural Equation Modeling
Question
6 answers
Hi everyone,
I tested an SEM model with 2 IV, 4 mediators and 1 DV on a sample of 1000 participants (see attached figure). Could you please help me to find an estimation for a good sample size using power analysis for this multiple-mediator model.
Best,
Robin
Relevant answer
Answer
You could refer to Jak et al. (2020) and Wang and Rhemtulla (2021) for recent insights. Here are the full citations.
Jak, S., Jorgensen, T. D., Verdam, M. G. E., Oort, F. J., & Elffers, L. (2020). Analytical power calculations for structural equation modeling: A tutorial and Shiny app. Behavior Research Methods. https://doi.org/10.3758/s13428-020-01479-0
Wang, Y. A., & Rhemtulla, M. (2021). Power analysis for parameter estimation in structural equation modeling: A discussion and tutorial. Advances in Methods and Practices in Psychological Science, 4(1), 251524592091825. https://doi.org/10.1177/2515245920918253
Good luck,
  • asked a question related to Structural Equation Modeling
Question
13 answers
Hi, I am running path analysis with latent variables. My model fit indices are good, however some of the factor loadings are negative. Also some of the standardized estimate are more than 1 like chemical on N2O is more than 1 (1.60) and topographical on N2O is -1.03.
Is it alright to have negative loadings in the attached path diagram. How can i correct this diagram?
Thanks
Relevant answer
Answer
Hello Waqar,
beyond what Christian already said, I have some problems understanding the latent variable. It seems that in your model, these are rather categories but not underlying common causes.
If a (measurement) model is substantially misspecified, you'll get bogus effect estimates.
HTH
--Holger
  • asked a question related to Structural Equation Modeling
Question
1 answer
Hi,
I have a dyadic dataset with distinguishable dyads (husbands and wives). I want to test a model with a predictor (X(f) X(m)), a mediator which is a similarity/difference in a trait, and dependent variable (Y(f) Y(m)). I would normally use SEM, but I'm not sure how to treat the difference score. Thank you for your help!
Relevant answer
Answer
A very interesting approach
  • asked a question related to Structural Equation Modeling
Question
5 answers
Hi, I want to ask about the total effect and indirect effect. Is it possible that one of my variables is insignificant and negative indirect effect and total effect? For instance, Brand Image (independent variable) has insignificant and negative effect (indirect and total) towards customer satisfaction (mediating variable) and Brand Loyalty (dependent variable).. But the rest of my independent variables are significant and have positive effect.. Is there any problem? Thank you..
Relevant answer
Answer
Imran Anwar Okay then. Thank you for explaining it to me. It really helps me.
  • asked a question related to Structural Equation Modeling
Question
3 answers
Hi!
I ran a meditational model using SEM. I have an independent variable (X), a mediator (M), and two outcomes (Y1 and Y2).
Using bootstrap I found a mediation effect of M between X and both Y1-Y2 (i.e., full mediation, the direct paths between X and Y1-Y2 become n.s. when in the model is introduced the mediator).
Using SEM, I want to deepen this result highlighting for which outcomes M is the best mediator.
Is it enough to check the magnitude of the two indirect effects?
Alternatively, I thought to constrain the coefficient of the paths M->Y1 and M->Y2 to be equal and then check the models' fit.
If the model fit of the constrained model would result worsen compared to the unconstrained model could help me to sustain that the two paths are likely different in favor of the strongest ones? Thus, if the coefficient of M->Y1 is stronger than M->Y2, then I would sustain that for Y1 (vs. Y2), M is the stronger mediator.
Thanks
Relevant answer
Answer
Very interesting question. I am interested in knowing about it
  • asked a question related to Structural Equation Modeling
Question
4 answers
Is it right to face socioeconomic status as a latent variable? May you give an example?
Relevant answer
Answer
Hello Matheus,
in 2007, there was a whole special issue in Psychological Methods on formative variables that centered around the example of socio-economic status as a latent formative variable. It started with the very critical view by Howland and colleagues:
Howell, R. D., Breivik, E., & Wilcox, J. B. (2007). Reconsidering formative measurement. Psychological Methods, 12(2), 205-218.
The debate continued later with the same folks in the Journal of Business Research:
Wilcox, J. B., Howell, R. D., & Breivik, E. (2008). Questions about formative measurement. Journal of Business Research, 61, 1219-1228.
Best,
--Holger
  • asked a question related to Structural Equation Modeling
Question
7 answers
Hello!
I have a few images of fungal biofilms acquired through SEM but they all have different magnifications (i.e. 160x, 140x, 110x and 100x). Does anyone know a way to change the magnification of images so I can have them all at the same magnification?
Thanks in advance!
Relevant answer
Answer
Hello,
You can use Fiji (https://fiji.sc/); menu Image/Scale (Ctrl+E).
But changing magnification for images means rescaling, changing size of your images.
For example, you have one image at 1K X (1000 pixels by 1000 pixels) and one image at 2K X (1000 pixels by 1000 pixels).
If you want the 2K X image at the same magnification than the 1K X image, The final image will have the size of 500 pixels by 500 pixels). You lose 3/4 of the information because you only have 1/4 of 1000 pixels by 1000 pixels image.
If you want the 1K X image at the same magnification than the 2K X image, The final image will have the size of 2000 pixels by 2000 pixels). But you won't get more information neither the same sharpness than an original 2K X image.
The best is to put a scale bar on original images (with Fiji; menu Analyze/Set scale - may need manual image scaling if fiji can't find the µm or nm per pixel information coming from the microscope in the image metadata).
or take all pictures at the same magnification.
  • asked a question related to Structural Equation Modeling
Question
3 answers
I'm writing this post because there is limited help on how to use Mplus as a mac user (Catalina). This was a HUGE headache to figure out so I thought I would share...
Checking system preferences
Find the Mplus folder in your Applications folder. Double click on each of the following, and try to open all of them (diagrammer, mplus, mplus editor) one-by-one. If they don’t open, click on each of them one-by-one and add them under the system preference/security & privacy/GENERAL tab. Make sure this is done for the diagrammer, mplus, and mplus editor.
Then, click on the PRIVACY tab; and allow diagrammer, mplus, mplus editor to have full disk access, and access to files and folders.
Then click on startMplus, followed by: diagrammer, mplus, mplus editor (in any order)
Syntax and Data Files:
-You can use either .txt or .dat files EVEN on a mac!
-Save syntax and data files in same place (e.g., desktop)
.dat
-save .dat data file directly from your source
.txt
-copy and paste your numerical data into textedit, and save as plain text.
-format, ‘make plain text’, convert this text into plain text? Yes. If when you clicked format, it said ‘make rich text’, the file is already in the format you want it to be, so leave it alone.
-enter FILENAME.txt, include the .txt even if you have the box checked “if no file extension is provided, use “.txt”
Running files
Open Mplus Editor, click the folder and open the syntax file (FILE.inp).
Depending on whether you are using .dat or .txt enter this!
E.g.,
DATA:
FILE IS 'FILE.dat' ;
OR
DATA:
FILE IS 'FILE.txt' ;
Press run.
Save the output.
Good luck!
Relevant answer
Answer
I cannot thank you enough. I have been tearing my hair out for two days trying to figure this out.
  • asked a question related to Structural Equation Modeling
Question
3 answers
I am working on SEM, and I am now at that stage where I got to think on what are the functions of SEM? where does it fit in?
Relevant answer
Answer
The standard error of measurement is a concept of classical test theory (CTT). It is the standard deviation of the measurement error variable as defined in CTT. As such, it is directly related to (un)reliability as explained by Imran Anwar. The SEM can be used to construct a confidence interval (CI) around individual observed test scores to express uncertainty associated with individual test scores. The lower the reliability of test scores, the larger the SEM and the wider the CI around individual test scores.
A related concept is the standard error of the estimate (SEE) in linear regression analysis (the standard deviation of the residual variable in regression). The SEE is related to prediction error (1 - R^2) whereas the SEM is related to unreliability (1 - Reliability) in measurement.
  • asked a question related to Structural Equation Modeling
Question
19 answers
I expect the suggestion about how to calculate the particle size of hydroxyapatite extracted from biowastes.
Relevant answer
Answer
The Scherrer equation calculates crystallite size, not particle size. It also is complicated by the fact that the XRD peak broadening is caused by the instrument and strain, as well as crystallite size.
  • asked a question related to Structural Equation Modeling
Question
3 answers
1. data were collected by Non probability sampling Techniques.
2. Data is non- normal.
with references.
Thanks in Advance.
Relevant answer
Answer
Yes you can but the statistical tests will be meaningless since they require a probability sample. See any statisics book to find out why this is so. David Booth
  • asked a question related to Structural Equation Modeling
Question
3 answers
I am working on a Fractal analysis of some shale sample SEM images using FracLac. However, it doesn't include the Succolarity calculation. I hope someone can help me with this issue. I appreciate that.
Relevant answer
Answer
Dear Bo Li ,
The general problem is the problem of topological properties of fractals, in particular, connection properties. The study of concepts related to percolation is just a way to explore such properties. I have not studied this subject for a while. I have just reviewed old papers and I have found two relevant reeferences:
Robins et al, "Computing connectedness: disconnectedness and discreteness", Physica D 139 (2000) 276–300.
Luo and Liu, "On the classification of fractal squares", arXiv:1404.6619
The last two authors are your compatriots and I think that they belong to a school of mathematicians that are doing very good work in this regard.
I hope it is useful. Best wishes.
  • asked a question related to Structural Equation Modeling
Question
10 answers
I have chitosan-TPP nanoparticles, prepared by probe sonication. Basically, I add 4 mL of TPP solution (1.76 mg/ml) to 10 mL chitosan solution (3mg/ml). Then I probe sonicate it (5 min; 30-10 on/off). Finally I separate the NPs by centrifugation (15000 rpm, 20 min, room temp).
I get quite good DLS results (app 250 nm size; 0.1 PDI). The problem is, when I try to get SEM image, I get micrographs as attached (either with the ddH20 resuspended pellet or with NP suspension obtained right after the probe sonication).
I air dry the NP (1:100 dilution) directly on SEM stub. What is that background? It is also observed on the NP as well? Did anyone observe such a micrograph? What should I do? Really appreciate your help. Thank you in advance
Relevant answer
Answer
Could you try to view the sample uncoated? The image quality will be lower but it might show you if you are looking at gold (alternatively see if someone could coat with platinum or another metal that gives a finer grain size). I see you have an FEI FEGSEM. if it is an ESEM you could try the 5kV cone and large field detector. You could also try the GSED detectors. If it is not an ESEM you could try 2kV or lower with the ETD. It will be difficult and you may not be able to to resolve the features. Good luck.
  • asked a question related to Structural Equation Modeling
Question
4 answers
I am conducting confirmatory factor analysis using lavaan in R with several both first- and second-order factors in the model (20 latent constructs and 97 items in total; n = 511; no missing values). After running the CFA, the model fit is not sufficient, but some low factor loadings, a couple Heywood cases and several indications from the modification index give some pointers for model improvement. Deleting some items from the model and re-structuring two factors based on an EFA eliminates Heywood cases and results in acceptable model fit (Chi-square=4117, DF=2646, CFI=.94, RMSEA=.038, SRMR=.052), however, the CFA of the adjusted model now returns the following error message: "covariance matrix of latent variables is not positive definite". There aren't any negative values in the covariance matrix though and also not in the correlation matrix.
I'm fairly new to this kind of statistical analysis and have no clue what causes this error and how to fix the issue. I'd appreciate any help. Please find attached the original and adjusted CFA measurement model as well as the covariance and correlation matrices of the adjusted model causing the NPD error.
Thanks a lot for your input!
Relevant answer
Answer
Sometimes a message like this occurs when several factors are highly (but not necessarily perfectly or > |1|) correlated. High correlations among factors can cause linear dependencies (e.g., one factor being perfectly predictable from a set of other factors) without there being negative variance estimates or correlation estimates > |1|.
  • asked a question related to Structural Equation Modeling
Question
3 answers
Is it possible to conduct SEM before collecting the actual data based on previous researches??
Relevant answer
Answer
If the results of the previous research are available in sufficient detail that you can produce a covariance matrix (or at least, a correlation matrix) that includes all the variables you want in your model, then yes, you can do SEM on that matrix, without (new, raw) data. Otherwise, I believe the answer is no, you can't.
  • asked a question related to Structural Equation Modeling
Question
4 answers
This picture was taken during my recent research, on lignite deposits during the Pliocene, in France. ("Silica flowers"?)
Have you seen this structure before? personal SEM photos. Thanks a lot for your help.
Relevant answer
Answer
Thank you for your participation !
  • asked a question related to Structural Equation Modeling
Question
5 answers
Attached with the question is a SEM image of an endo-parisitic mite of lepidotera.
Relevant answer
Answer
Dear Joyce,
That's a lovely SEM. I think this looks like a species of Suidasia, probably the "flour mite" or "scaly grain mite" S. pontifica. I'm not sure what they're doing living internally in a moth, but they do seem to get into all sorts of places. There's even a couple records of them from human ears. If you're working on stored product moths, then I'd be a bit more confident of my suggestion of Suidasia.
You can find some SEMs of this mite and other diagnostic and biological information as it's a well-studied species. See for example:
and
Kind regards,
Owen.
  • asked a question related to Structural Equation Modeling
Question
2 answers
Observation of bacteria under a SEM or MET
Relevant answer
Answer
I did not understand your answer can you explain to me, please?!!
  • asked a question related to Structural Equation Modeling
Question
6 answers
Dear all
I have used AMOS for confirming my structural model based on parallel mediation. However, since my trial version of SPSS has expired I would like to know , if I can use SMART-PLS or some other software to calculate VIF and effect size values and include it in my final analysis.
Relevant answer
Answer
Hello Oshin,
I fully agree with Imran that it does not make sense to switch analytic methods mid-stream, unless your research question(s), data, or both have materially changed.
May I suggest you have a look at the freely available R library, lavaan? It is quite capable, and the R system can handle just about any kind of statistical analysis you might wish to try.
The Jamovi package is built on R (as is JASP), and you may find that data entry, manipulation, and so forth is a bit easier than using straight R. All are free.
Good luck with your work.
  • asked a question related to Structural Equation Modeling
Question
22 answers
this samples under SEM , could you please help me to explain the features ?
@geology
Relevant answer
Answer
I am not certainly the right person to give you an answer on your question. What it cough my attention was your way you start your message: "Hello gents", I do know at least 3 women that can give you a Master class on this, so please be more inclusive next time.
  • asked a question related to Structural Equation Modeling
Question
3 answers
Dear colleagues,
It would be very helpful if someone of you could tell me, how to test configural measurement invariance in an intervention study with two groups and two measurment points.
I was wondering, if I have to compare the SEM of the first measurement with the second measurement (both groups together) or if I have to compare the SEM of the first measurment with two SEM of the second measurement (one SEM for each intervention group).
Thank you in advance!
Best regards,
Robin
Relevant answer
Answer
Robin Junker, Brown's book on confirmatory factor analysis has a specific section (p. 221) entitled "Longitudinal Measurement Invariance" with a detailed example that I recommend you to have a look at.
  • asked a question related to Structural Equation Modeling
Question
3 answers
I performed E-beam lithography(EBL) with multiple doses of the same pattern to determine which dose the pattern is optimally produced.
Now the dvelopment of resist is done, and I want to see how the pattern was formed.
I think verification using SEM is a good way to do it, but I'm hesitant for several reasons.
1. Is CD-SEM typically used to image EBL result?
I heared CD-SEM is good method for imaging results of EBL, because it uses lower voltage.
Is CD-SEM good method for imaging EBL result?
Or is there other method commonly used for imaging EBL result?
2. CD-SEM can be replaced by typical FESEM with lowered voltage?
CD-SEM is usually compatible with 6 or 8 inch wafer, but my sample is piece(1 inch * 1 inch).
Therefore, I can't use CD-SEM.
CD-SEM can be replaced by typical FESEM with lowered voltage?
3. SEM imaging can damage EBL results, although it uses lowered voltage?
I know that E-beam resist is weak to electron beam energy.
CD-SEM, or FESEM with lowered voltage still damage E-beam resist?
If so, how to lower damaging from imaging?
I want to use the EBL results in the next process(lift-off, etching…) after imaging. Is it possible?
4. How to avoid charge-up in non-destructive method?
Since the EBL results are dielectric, so charge-up is inevitable in the SEM.
I know sputtering metal is good method, but I want to use EBL result in the next process(lift-off, etching…) after imaging.
In that case, how to avoid charge-up in non-destructive method?
I heared conductive paste, conductive tape, or electrification dissipating material(Espacer, AquaSave) can be alternatives.
Can I use them for SEM imaging of EBL results?
Thank you in advance.
Relevant answer
Answer
You always can do your imaging on samples and then once you find the sample with the right patterns, replicate its fabrication steps, the results should be reproducible and then you can apply the next fabrication steps.
(4)- Conductive tape or stickers are attached to the SEM stub, the substrate is sticked to it to hold it firmly and to remove charge, additionally some add conductive ink all around the edge of the substrate, to ease the charge dissipation of the top layer/s.
In my experience the best method to remove charge is usign a local flow of N2 gas, injected directly on the are you are taking the image. This N2 flow is injected through a thin plastic needle that is hold very close to the sample.
If your structures are micron sized you could use some alternative methods to SEM, like optical profilometry. With nano-patterns you could use AFM.
Hope it helps.
  • asked a question related to Structural Equation Modeling
Question
15 answers
It would be great to listen from you as this latent variable is an important one for my research !
Relevant answer
Answer
Could it be a RESULT to be discussed in your research that this important scale does show the reliability expected from the litterature?
  • asked a question related to Structural Equation Modeling
Question
5 answers
I googled to conduct common method bias (CMB) tests in R but there is no information as such. Maximum information (including videos) is related to conducting CMB tests (like Harman's single factor) using SPSS or using AMOS (like common latent factor technique). All the constructs in my model are latent constructs and I am using covariance-based SEM techniques to analyze my model using Lavaan in R.
Now I want to conduct common method bias tests using R. Any help in this regard (e.g. sample code etc.) is welcome.
Relevant answer
Answer
Hi Harmanjit,
neither the Harman's one factor test nor the additional-general-factor model work. The last thing I heard was focused on the marker method. Again, highly disputed at least, there were some recent papers defending the approach:
Williams, L. J., & McGonagle, A. K. (2016). Four research designs and a comprehensive analysis strategy for investigating common method variance with self-report measures using latent variables. Journal of Business and Psychology, 31(3), 339-359. doi:10.1007/s10869-015-9422-9
Williams, L. J., Hartman, N., & Cavazotte, F. (2010). Method variance and marker variables: A review and comprehensive CFA marker technique. Organizational Research Methods, 13(3), 477-514. doi:10.1177/1094428110366036
Williams, L. J., Hartman, N., & Cavazotte, F. (2010). Method variance and marker variables: A review and comprehensive CFA marker technique. Organizational Research Methods, 13(3), 477-514. doi:10.1177/1094428110366036
HTH
Holger
P.S. And I say this as a reviewer: NEVER use the Podsakoff paper to defend Harman. The paper just reports this as part of its review but clearly recommends not to use it.
  • asked a question related to Structural Equation Modeling
Question
4 answers
I tried synthesizing silver nanoparticles using plant extract. 1mM has not changed any colour. i then tried using 10mM and colour changed from green brown. i was monitoring the biosynthesis using UV-vis but then my peak remains at 328nm even after 72hours. i have no idea on how to explain the peak with the image i received from SEM. anyone to help on this please
Relevant answer
Answer
I second the suggestion from Sai Prasad Nayak
Also, please dilute the sample, drop cast on clean silicon wafer, dry well and then record SEM. If your plant extract used for synthesis is colored, then it will mask the absorbance of your Ag nanoparticles even if Ag nanoparticles are formed. My strong suggestion is dilute the plant extract like 100-1000 times and then use 1mM or 10mM Ag+. It must work. Also, perform EDAX and look for elemental analysis. Best is to confirm via, XRD or TEM.
Good luck!
  • asked a question related to Structural Equation Modeling
Question
1 answer
Hey there,
I am currently working on a SEM with an N=274. The textoutput shows, that my model is unidentified, but I don't know why. Hopefully someone can point out the issue. Thank you in advance :)
Best regards,
Morris
Relevant answer
Answer
@Morris AMOS runs the model with only two manifest variables under a latent variable if regression weight is put on both the paths from respective latent variable to its manifest variables. In your case, you have two latent variables with only two manifest variables viz. V (converging with V1, and V2) and NA (converging with NA1 and NA2) but for NA, regression weight has been put on only one path from NA to NA1 and no regression weight have been put on both the paths from V to V1 and V2. Put regression weight on both the paths from V and NA to their respective observed variables. Hopefully it will run.
  • asked a question related to Structural Equation Modeling
Question
6 answers
Hello all
I have to get an SEM image for my perovskite LED device cross section but I am not able to get a clear image. It is very blurry and I am not able to get any information from there. The machine I am using is Tescan FIB FERA. Can anyone help me with this?
Relevant answer
Answer
You need to look for a microscope JEOL
  • asked a question related to Structural Equation Modeling
Question
3 answers
Estimation technique for count level dependent variable
SEM analysis
Relevant answer
Answer
Hi,
as i said; I don't think that ML is appropriate here and an estimator reflecting the nature of the variables (even if they are only indicators) should have been used.
In addition, I would not trust the model as the latent variables do not seem reasonable to me.
Best,
--Holger
  • asked a question related to Structural Equation Modeling
Question
4 answers
Hello all,
I am analyzing a model using SEM, but my result seems hard to interpret the effect between non-formal learning experience(X), volition(Y), and expectancy(M) in the context of beginning farming.
1. indirect effect is positively significant (coef=0.172, p<.001)(The effect of X->M(coef=0.155, p<.001) and M->Y(coef=1.106, p<.001) is both positively significant), while direct effect(X->Y) is negatively significant(coef=-0.116, p<.001).
Q. I am confused about whether it can happen and how I could interpret the result if is possible. (See the path diagram).
2. I am concerned about the suppressor effect(X and M) because there is a low correlation between X and Y(r=.2045*) while a high correlation between X and M(r=.3497*). (correlation between M &Y is 0.6685). This positive correlation(X and Y) turns out to be negative when I analyze SEM. (When I check the partial correlation between X and Y controlling M, it turns out to appear -0.1066. )
Q. Can I say this is a suppressor effect? If it is, can someone help me with how I can deal with this suppressor effect? Should I describe all these situations in my paper? (or hopefully are there any other options to solve this problem?)
Note 1: All variables are continuous.
Note 2: observations are enough(n=407)
I would highly appreciate it if someone can help me!!! Thank you very much.
Relevant answer
Answer
I don't think it's because multicollinearity for the value of VIF is both1.14 when I do regression. I think it is kind of suppression effect according to MacKinnon et al, 2000. Do you have an idea how I can deal with this problem? Any idea will be appreciated. Thanks
  • asked a question related to Structural Equation Modeling
Question
3 answers
Hi,
I'm writing a short paper using SEM (Structural Equation Modeling), and I just faced a problem while drawing a research model.
I wanted to use a variable which can show a 'paradox' effect; variables similar to 'distortion', but seems there isn't really such variables. I've been googling a lot of papers, but couldn't find any :(
If you know any, could you please share?
Even a short word will be a big help for me.
Thank you very much :D
Relevant answer
Answer
This makes no sense. If (Un)satisfaction has any relationship with intention then identifying an unrelated variable will not change anything. I recommend googling for "path tracing" and you will learn three simple rules how causal structures are related to empirical correltions. This delivers the knowledge to play with options yourself.
Good luck
--Holger
  • asked a question related to Structural Equation Modeling
Question
10 answers
Theory of Reasoned Action
Theory of Planned Behavior
Marketing research
Business research
Relevant answer
Answer
  • asked a question related to Structural Equation Modeling
Question
8 answers
I am working with Tungsten oxide nanoparticles and getting a mix of different morphologies, I wish to understand and predict the cause of the structures I am getting, the rods, the plates, and the valleys in between. I would also wish to learn about different morphologies in detail. Please suggest a book that has a good command of this.
Thank you
Relevant answer
  • asked a question related to Structural Equation Modeling
Question
1 answer
I am characterising a volcanic deposit in which I am currently imaging at SEM the fine ash (>4 phi). For each sample I do a 500um-wide elementary cartography. I know that by superposing filter you can know which component are present but for the proportion I am worried since here I would have a ratio of surfaces and not of weight (As I have for bigger fractions between -7.0 and -0.5 phi).
Should I multiply the surface area of each component by it's density (using densities of the various components) assuming EVERY component is 1 um thick (which it is the worring assumption for me) ? This last assumption allow me to transform a surface to a volume and then to end up with a mass and not a mass.m-1.
Thank you in advance !
Relevant answer
Answer
yes ,you can
  • asked a question related to Structural Equation Modeling
Question
3 answers
Dear everyone
I will process a biological material (polychaetes) to perform SEM. I am in doubt if after dehydration in alcohol the specimens can be stored in absolute alcohol for later drying at a critical point , and how long can I store the material for?
Best regard
Relevant answer
Answer
Hi. I hope the following article could help you:
  • asked a question related to Structural Equation Modeling
Question
4 answers
I'm trying to make a SEM model where my final endogenous variable is of unordered category (the variable is having three category which are not in any hierearchy/order). I want to know if it is possible to do this in AMOS, the AMOS documentation talks only aboit ordered categorical variable.
I have tried using Tool > Data Recode > Choosing the Variable and selecting Ordered Categories and clicking on Details > Then changing all three categories into the upper column of "Unordered Categories". This is not working and showing the error "Index was out of range"
Relevant answer
Answer
Smartpls handle both ordered and unordered categories. It can also be used for complex model. Thus, is far better than AMOS.
  • asked a question related to Structural Equation Modeling
Question
2 answers
We are running an SEM model for the scale development of a construct. We first ran a confirmatory factor analysis model. Since the data was non-normal, we used the ADF method of estimation. The CFI and the GFI are way below the required cutoff (around 0.72) but the RMSEA is doing very well, i.e. below 0.05. The SRMR is also 0.07. What does this imply? Which model fit statistics should be reported for ADF estimation in SEM? Is it acceptable in this case to report on RMSEA and RMR?
  • asked a question related to Structural Equation Modeling
Question
7 answers
Hi Everyone,
We are evaluating a proposal for SEM. The instrument is offered with a standard Tungsten filament or LaB6 filaments. I know LaB6 is better in terms of image quality, spot size, and filament life as compared to Tungsten filament (of course price is high). What are the other points we should consider such as high vacuum chamber, maintenance cost, operational cost, etc? Any help from an experienced user will be much appreciated.
Relevant answer
Answer
Dear Dr. Muhammad Zahid ,
I suggest to have a look at the following, interesting document:
- Choosing the Right Microscope – Part 1
Choosing the Right Microscope: a Step-by-Step Guide – #1 Tungsten, LaB6, or Field Emission By Tss Microscopy (2019)
My best regards, Pierluigi Traverso.
  • asked a question related to Structural Equation Modeling
Question
10 answers
I have reached to higher-order CFA --- 4 latent factors and one general factor. The question here is:
Is it possible to build SEM from CFA including one general factor?
I have attached a picture to clarify Higher-order CFA.
Relevant answer
Answer
Hello Osama,
as I said, technically, you can embed the second-order factor in any structure you propose. However, you said a strange thing:
"in other words, to consider the common or general factor (intercultural competence ) as the dependent variable and its dimensions as an independent variable in SEM"
The dimensions of a second-order factor (SOF) are *consequences* of the
SOF--not independent variables. Or do you mean that there are one or several variables (not yet in your path diagram) on which the primary factors (e.g., "procedural skills") are supposed to have an effect? It is hard to belief how this variable (let's call it Y) would be dependent on the primary factors and a cause of the second-order factor....In such a case, Y would be a mediator by which the primary factor(s) cause the SOF which runs against the main premise of the whole SOF structure (and would imply a loop which cannot be estimated without either longitudinal data or instrumental variables).
Hence, such a model would not be reasonable and technically not possible.
But perhaps I have misunderstood. We could spare some time if you draw exactly the model that you have in mind. A great tool is dagitty.net in which you can draw a path diagram and "publish it" (what means that you simply get a link that you can include here.
Holger
  • asked a question related to Structural Equation Modeling
Question
5 answers
We look to compare effects of several operational conditions on the texture of the deposit.
Relevant answer
Answer
You can measure the porosity of coated according to ASTM E2109. The percentage area of pores on coating can be measured using ImageJ analyzer software. Measurements based on the circularity range of pores between 0.3 to 1. The circularity of pores was measured from the formula, 4 ×pi × (cross-section area of the pores shape/perimeter of the pore shape ^2). The value of 1.0 circularity indicates that the shape is a perfect circle. As detailed in the following reference;
  • asked a question related to Structural Equation Modeling
Question
4 answers
Hey there,
how can I calculate the significance level (p-value) for total effects in an SEM?
I used SMSS AMOS to calculate a structural equation model (SEM). I applied the Maximum Likelihood Estimator. I get total, indirect and direct effects. However, I do not get significance levels for the total effects. Is there a way to do so? May I be able to derive the significance somehow from the direct and indirect effects?
Thank you so much for you answers!
Best
Benedict Heblich
Relevant answer
Answer
Hello again Benedict,
Bootstrapping is quite useful regardless of sample size, as it still allows for valid and robust estimation of parameters even when all the usual assumptions can't be made regarding the data (normality, homogeneity). If you're fully content with the conformance of data to the assumptions, then you might wish to consult the AMOS user's manual (as AMOS frequently defaults to not reporting certain elements, standardized path coefficients being one such example: http://www.csun.edu/itr/downloads/docs/IBM_SPSS_Amos_User_GuideV24.pdf).
Good luck finishing your project.
  • asked a question related to Structural Equation Modeling
Question
3 answers
Hi, I am working on my project for which I need assistance on this question: “Is it possible to test five continuous moderators on four independent variables and two dependent variables using SEM?” If yes, what technique or tool is appropriate? AMOS SEM, CFA, PATH ANALYSIS, MODERATED MULTIPLE REGRESSION OR HIERARCHICAL MODERATED MULTIPLE REGRESSION. I need a generous assistance to complete my project
Relevant answer
Answer
Hi Ghana,
I hope you are doing well! Without seeing a path diagram it is hard for me to see what this model would actually look like, but I will do my best to answer the question. Although this analysis is technically possible in AMOS or Lavann in R (AMOS would be better) using the path analysis, I worry about the complexity of the interpretation. The output would contain dozens of moderated values across potentially eight relationships. So, I am wondering if there is a more parsimonious way to analyze these data.
I would recommend being a little more narrow with your hypotheses by picking out the most theoretically sound and empirically supported moderated relationships. Then, I would consider running a simpler model or conducting multiple moderation analyses (potentially in the PROCESS macro). If cutting the model down or making it into multiple models is not an option, then yes I believe it is possible in AMOS, but running the model and interpreting the output might be quite complicated.
Hope this helps!
Best,
Bryant
  • asked a question related to Structural Equation Modeling
Question
5 answers
Hi everyone,
What I want to do is to observe the cross section of a polymer/silver sample inkjet printed on thin kapton film under SEM. What I did to prepare the sample is :
Glue the sample on a thicker and more rigid substrate (this is to avoid distortion) and put the sample in a arcylic mould and wrap it with epoxy. After curing the epoxy I polish it mechanically to remove the front side of the epoxy and half of the sample , so that cross-section is revealed.
But what I observe in SEM is of un-satisfactory quality, I can find material in very un-realistic location. I feel like it is like polishing move some of the material elsewhere. I wonder does anyone has met similar problem?
Relevant answer
Answer
Hello,
What is the thickness of kapton and silver?
You can try to put your sample in liquid nitrogen, and to break it, you will have a net cross section.
Regards,
Emile HAYE
  • asked a question related to Structural Equation Modeling
Question
3 answers
Dear all,
with the "group" argument in the "sem" function in lavaan (R) I have created two separate models for men and women for my research. Accordingly to prior recommendations I have first tested for configural, metric, scalar and strict invariances, and the model worked well. Now I'm wondering if there is a way to compare beta coefficients between those two models, as I would like to see whether there are any significant differences between the same paths on models for different sexes. As such I would like to ask for your help!
Best regards and thanks!
RT
Relevant answer
Answer
Hi Radoslav,
impose equality constraints on the unstandardized coefficients and test the difference with the chisquares of the unrestricted and restricted model (the same logic as in any other form of invariance test).
HTH
Holger
  • asked a question related to Structural Equation Modeling
Question
3 answers
I'm analyzing mediating using secondary data for 3 years. For now, I've done the regression analysis using Baron and Kenny method. My question is, can I also test the mediating effect using SEM? Since so far as I concern I can't find any related articles that use secondary data in testing mediating using SEM.
Relevant answer
Answer
6 pasos para crear una buena estrategia SEM
  1. 1: Marca tus objetivos. ...
  2. 2: Busca las keywords ganadoras. ...
  3. 3: Organiza tus anuncios por grupos y segmenta. ...
  4. 4: Aprovecha los diferentes tipos de concordancias. ...
  5. 5: Establece y revisa tu landing page. ...
  6. 6: Usa el remarketing. ...
  7. 7: Haz un seguimiento continuo
  • asked a question related to Structural Equation Modeling
Question
5 answers
In a CFA, we have very good fit indices. For example, the CFI = 1. This, however, is not a just-identified model because degrees of freedom is not 0. When I see such super excellent fit indices in a not-just-identified model, I can't help but be suspicious.
Would such a well fitting CFA be appropriate to use in further analyses? Or should this particular CFA be thrown out?
Relevant answer
Answer
Hello Michael
1) I would not do the split of the sample as you loose power and efficiency
2) The model is very small and has not many testable implications.
3) Despite the non-significant chisquare I would be suspicious about the strong differences among the factor loadings. Does this make sense when looking onto the question wordings?
I would recommend to enlarge the model with external validation variables. This has two purposes: You add further testable implications that the model has to fulfill if it is correct and you get estimates (the links to the validation variables) that further support or falsify your assumption about the meaning of the latent variable.
See the first study published in this paper that I described here (with a link).
All the best,
Holger
  • asked a question related to Structural Equation Modeling
Question
3 answers
I am university student trying to understand the interpretation of SEM. Specifically, I'd like to understand how the study of Seo & Park (2018) applied SEM in their study. The link of the study is this:
This is the part I'm having trouble understanding:
χ^2 = 576.887, df = 219, CMIN/DF = 2.634, p < 0.001, GFI = 0.855, AGFI = 0.817, RMR = 0.085, CFI = 0.918, TLI = 0.906, and RMSEA = 0.074
What is χ^2, df, CMIN/DF, p, GFI, AGFI, RMR, CFI, TLI, and RMSEA?
I'd appreciate any help.
Relevant answer
Answer
Hello Ram,
Timo Mandler 's response is on target. I would just add that measures of model-data fit address the general question: How well do the estimated model indices/parameters allow one to reproduce the observed relationships (e.g., correlations or covariances) among the measured variables in the data set? Closer is better!
David Kenny's web page is also a quick overview of most of the fit indices used in SEM applications: http://davidakenny.net/cm/fit.htm
Good luck with your work.
  • asked a question related to Structural Equation Modeling
Question
3 answers
I have used SEM before, but the results are not clear, and I have used an optical microscope ( Nikon Eclipse Lv100POL), but the light couldn't pass through. I am interested in impurities, voids, and pullouts.
Relevant answer
Answer
Amal Mlhem If what you want to look into (impurities etc.) is on the surface that we can see now, I think SEM should be OK. If your SEM images are not clear, I am thinking if it's due to the low conductivity of your polymer-matrix composites. You can coat a thin layer of Pt or C on it, it's the common method for the low-conductivity samples. You mentioned "the light couldn't pass through", so, I am wondering the defects you want to observe exist inside of your sample, not on the surface? If you want to look into the interior defects, SEM can't work, I think maybe you need to grind to remove your surface.
  • asked a question related to Structural Equation Modeling
Question
7 answers
Can we apply SEM to our data to discover a path model based on exploratory factor analysis or should we develop a proposed path for our exogenous and endogenous variables and use path analysis to test the "goodness" of the path model?
Relevant answer
Answer
Hi,
you just need the correlation matrix of the variables and TETRAD will print a single path diagram that prints all equivalent models matching this particular correlation matrix. These models will have the same "skeleton" (generic structure) but differ (marked by an innovative style of representing the lines between two variables) in their causal directions. Most--unfortunately--will be marked as totally ambiguous--that is--x may affect y, y may affect x or both result from an unobserved confounder. Others may be partially "oriented" and others may result in a fixed direction, meaning that in all models of the class the link between the two respective variables could only be in a certain direction.
And no, this is not based on simply permuting the variables but--as mentioned, on the identifying equivalence class by means of graphical principles (i.e., d-separation).
With regard to the progression, you creating a strawman in a very (sorry) despicable manner. I would of course not recommend to submit this output as this will probably result in an immediate rejection (unfortunately, as knowing the partial ambiguity of certain partitions of a model is of high interest for the field instead of comprehensively accepting or rejecting the whole model). Some parts may be completely ambiguous (thus, demanding future research) while others may have solid evidence.
My ideal progression would be to start with a confirmatory approach WHEN your theory is so solid that you can come up with a predefined model. In most cases of applying path models or SEM, this is not the case. Why not starting with a pre-study using TETRAD and then--based on the result--enlarge the model or extending it in such a way that the ambiguous parts become less ambiguous (e.g. by incorporating instruments or by planning an intervention). Likewise, a failure in a confirmatory test of a predefined model could be succeeded by a TETRAD analysis to find possible errors. Science is always the circle between deduction and abduction. Model testing vs. exploration is no difference.
Best,
--Holger
  • asked a question related to Structural Equation Modeling
Question
8 answers
I am looking for a script to run moderation analysis in R (which I am unable understand how to add to the following script for SEM wherein I want to test moderation of F upon relationship between D & E, i.e. E~D+F+D*F). The background of the model is that it is based on cross-sectional data and all are latent constructs:
#Load package
library(lavaan)
#Model Development
Model=
'
A =~ A1+A2+A3
B =~ B1+B2+B3+B4
C =~ C1+C2+C3+C4
D =~ D1+D2+D3
E =~ E1+E2+E3
F =~ F1+F2+F3+F4
D ~ A+B+C
E ~ D
'
#Fitting SEM Model
fit1<- sem(Model,data=mydata)
fit1
summary(fit1, standardized=T)
modindices(fit1, sort. = T)
inspect(fit1, what = "std")
Relevant answer
Answer
The key is that you first have to understand what the residual centering approach does. This is in detail explained in the paper. Then connect this rationale with the attached R code. Actually the code is explanatory and I assume you are confused by the variable / object names. In this example "wfcdata" was the dataframe and the w's and na's were variable names (standing for work-family conflict and negative affect). Based on that, try to understand the pieces.
1. The first section multiplies all combinations of w and na items (resulting in product term items
2. Then conduct singular regressions and save the residuals
3. Then you have your residual indicators of the latent product term with which you can specfify your model.
The tricky part is to specify the estimation of appropriate error terms but again this should be doable with the knowledge from the paper.
Good luck