I have two datasets with different sample sizes. For example, I have 5 plant species richness levels (different sample sizes), all of which were corresponding with α diversity. However, as the limitation of minimum numbers of samples (>6) constructing the networks, I only used 3 plant species richness levels to construct species networks. So, can I construc a Structural Equation Model based on two datasets?
From what I can understand, whenever we run the model fit test(we utilise Amos graphics) and if the model fit is okay, then definitely you can continue the analysis.
You can check this paper for further information. Regards.
https://ecologicalprocesses.springeropen.com/articles/10.1186/s13717-016-0063-3
Xiangdeng Du It's not entirely clear to me what you are asking. Are you wondering whether it is OK to combine two separate samples (data sets) into a single data set and analyze the combined data set with a single structural equation model (SEM)? That would be OK if the two samples (data sets) are randomly drawn from the same population. If, however, the samples represent different populations, multigroup SEM analysis would be more appropriate. In a multigroup SEM analysis, separate samples/data sets are analyzed simultaneously (without combining the data across groups), and parameter estimates are statistically compared across groups. Multigroup analysis allows you to find out whether different groups are causally homogeneous (i.e., whether and which parameters differ significantly across groups). If none of the parameters differ across groups, the groups/data sets can be combined into a single group/data set and analyzed with a single SEM.
- Badges
- Science topic
- Similar topics
- Mathematics
- Statistics
- Sampling Design
- Sample Size