Path Analysis and PROCESS in SmartPLS: When to Use Them instead of Full PLS-SEM

SmartPLS includes a path analysis module designed to emulate results from Hayes's familiar SPSS mediation and moderation macro, PROCESS. The tools are not identical; understanding the differences prevents inaccurate method reporting.

PLS-SEM & SMARTPLSPrepared by Lê Minh5 min

Start reading

Observed scores and latent variables differ

Scores
State how analysis variables are scored
Path analysis
Relationships among observed variables
Mediation
Define the effect to estimate
Measurement
State limits from measurement error

Conceptual diagram. Using mean scores is not equivalent to estimating a measurement model.

Source: Lê Minh · Conceptual illustration, not software output.

1. What Is Path Analysis in SmartPLS?

Path analysis estimates a system of regression equations involving directly observed variables without constructing latent variables. It originates in Wright's (1921) work decomposing causal relationships into path equations. Unlike a single regression, path analysis accommodates several dependent variables simultaneously, including models combining mediation and moderation, such as moderated mediation.

In SmartPLS, path analysis can use single-item variables or multi-indicator composites. Unlike full PLS-SEM, it does not iteratively estimate indicator weights: SmartPLS assigns equal weights to all indicators within a composite. This resembles McDonald's (1996) path analysis with composite variables, where a simple indicator combination enters the path equations directly instead of using estimated indicator weights.

2. Why SmartPLS Includes a PROCESS-Style Module

PROCESS (Hayes, 2018) is a widely used SPSS macro for mediation, moderation, and moderated mediation through OLS regression and bootstrapping. Because PROCESS operates on individual observed variables, SEM researchers often compute composite scores, such as indicator means, before entering them into PROCESS. This is commonly called tandem analysis.

SmartPLS's path analysis module emulates PROCESS results within SmartPLS, allowing researchers familiar with PROCESS to use path models without manual calculations outside the software.

3. Differences from the Original PROCESS in SPSS

Table 1
Aspect	Original PROCESS (SPSS)	PROCESS-style module in SmartPLS
Input variables	Individual observed variables, such as precomputed average scores.	Single-item variables or multi-indicator composites with automatically assigned equal weights.
Estimation logic	One-step OLS regression using existing variables.	One-step regression-based path analysis, unlike full PLS-SEM's iterative composite score estimation.
Composite score calculation	Users calculate scores before entering them into PROCESS.	SmartPLS aggregates indicators with equal weights, without a separate manual calculation.

SmartPLS documentation also identifies an unresolved issue: no consensus establishes the optimal order of processing when aggregating composites, such as standardizing indicators before summing versus summing raw data before standardization.

4. Implementation in SmartPLS

Build the path model

Construct the model using single-item variables or equally weighted multi-indicator composites according to the mediation, moderation, or moderated mediation hypotheses.

Select the data metric

Choose unstandardized data, mean-centered data (often supporting interaction-term interpretation), or standardized data (subtracting the mean and dividing by the standard deviation).

Specify control variables where applicable

Mark controls if the study needs to account for additional factors beyond the main hypotheses.

Run bootstrapping

Use bootstrapping to test path coefficients, indirect effects, and interaction terms, similarly to the original PROCESS bootstrap approach.

5. When to Use Path Analysis/PROCESS instead of Full PLS-SEM

Sarstedt, Hair, Nitzl, Ringle, and Howard (2020) argue that the common tandem approach, running CFA or SEM for factor scores before entering them into PROCESS, loses information compared with estimating the complete mediation/moderation model directly in PLS-SEM. Following this reasoning, full PLS-SEM with estimated rather than equal weights is often more appropriate for multi-indicator models where measurement accuracy is central.

When SmartPLS path analysis/PROCESS is appropriate

The module suits single-item models, direct comparisons with familiar PROCESS results, or straightforward mediation/moderation models that do not require full PLS-SEM measurement assessment.

6. Example Reporting Wording

Example Methods wording

The mediation/moderation model was estimated using SmartPLS's path analysis module, emulating PROCESS logic (Hayes, 2018). The data metric was [unstandardized/mean-centered/standardized]. Path coefficients and indirect effects were tested with [number of repetitions] bootstrap subsamples. Unlike the original PROCESS macro operating on individual observed variables, multi-indicator composites were aggregated with equal indicator weights.

7. Common Mistakes

Table 2
Weak wording	Why it is inappropriate	Recommended wording
"The PROCESS macro was run in SmartPLS."	The SmartPLS module emulates PROCESS results rather than directly running Hayes's original SPSS macro.	State that SmartPLS's path analysis module, designed to emulate PROCESS results, was used.
"SmartPLS path analysis is identical to PLS-SEM but faster."	The methods differ in estimation logic: path analysis does not iteratively estimate indicator weights as PLS-SEM does.	Explain the choice, such as single-item variables or comparison with PROCESS results.
"The selected data metric does not need reporting."	The choice between unstandardized, mean-centered, and standardized data affects interaction-term interpretation.	State the chosen data metric and rationale, especially with interaction terms.

8. Checklist before Reporting

Have you identified SmartPLS's PROCESS-emulating path analysis module rather than the original PROCESS macro?
Have you stated the selected unstandardized, mean-centered, or standardized metric?
Have you justified path analysis over full PLS-SEM?
Have you reported bootstrap results for path coefficients and indirect effects?
For multi-indicator composites, have you stated that indicator weights are equal?

Conclusion

SmartPLS's path analysis and PROCESS module helps users familiar with Hayes's macro work within SmartPLS without external calculations, but it is not identical to the original tool. The key difference concerns multi-indicator composites: equal weights replace PLS-SEM's iteratively estimated weights.

Choose between path analysis/PROCESS and full PLS-SEM based on the research question and measurement model rather than familiarity alone. When measurement is central, full PLS-SEM usually allows more rigorous assessment of measurement quality.

References

Wright, S. (1921). Correlation and causation. Journal of Agricultural Research, 20, 557-585.
Hayes, A. F. (2018). Introduction to mediation, moderation, and conditional process analysis: A regression-based approach (2nd ed.). Guilford Press.
McDonald, R. P. (1996). Path analysis with composite variables. Multivariate Behavioral Research, 31(2), 239-270. https://doi.org/10.1207/s15327906mbr3102_5
Sarstedt, M., Hair, J. F., Nitzl, C., Ringle, C. M., & Howard, M. C. (2020). Beyond a tandem analysis of SEM and PROCESS: Use of PLS-SEM for mediation analyses! International Journal of Market Research, 62(3), 288-299. https://doi.org/10.1177/1470785320915686

Academic note: This article presents general principles of SmartPLS's path analysis/PROCESS module based on official documentation. Recommendations about when to choose path analysis over full PLS-SEM draw on Sarstedt, Hair, Nitzl, Ringle, and Howard (2020) and require consideration of each model's specific characteristics.