How to Make Recommendations for Educational Practice from Correlational Data Using Structural Equation Models


METADATA ONLY
Loading...

Date

2023-06

Publication Type

Other Journal Item

ETH Bibliography

yes

Citations

Altmetric
METADATA ONLY

Data

Rights / License

Abstract

In this commentary, we outline a five-phase process by which recommendations for educational practice can be distilled from correlational data using structural equation modeling (SEM). First, meta-theoretical beliefs associated with latent variables-that mental attributes cause behavior and can therefore be measured indirectly by observing multiple indicators of that behavior-must be adopted and made explicit. Next, an SEM must be formulated with relevant pathways and covariates that exhaustively represent our theoretical knowledge and assumptions about the structure of the psychological phenomena being studied. Third, model-data-fit indices and estimated parameters associated with the SEM should be carefully interpreted. Fourth, the model should be replicated across educational contexts, and any necessary changes should be incorporated into the relevant psychological theory. Fifth, the results of multiple studies can then be interpreted together with other sources of evidence as a basis for communicating our current theoretical understanding and caveats to practitioners. We also point out that educational recommendations should likely never be entirely prescriptive, and instead lie on a continuum of specificity based on the strength of the evidence.

Permanent link

Publication status

published

Editor

Book title

Volume

35 (2)

Pages / Article No.

48

Publisher

Springer

Event

Edition / version

Methods

Software

Geographic location

Date collected

Date created

Subject

Structural equation modeling; Recommendations for practice; Research design; Causation; Correlational Data

Organisational unit

Notes

Commentary

Funding

Related publications and datasets