Comment on fitting MA time series by structural equation models
| Authors | |
|---|---|
| Publication date | 1999 |
| Journal | Psychometrika |
| Volume | Issue number | 64 | 1 |
| Pages (from-to) | 91-94 |
| Organisations |
|
| Abstract | In a recent paper by van Buuren (1997) it is concluded that parameter estimates in pure moving-average (MA) models, obtained by software for fitting structural equation models (SEMs), are biased and inefficient. In this comment it is shown that this negative finding may be due to a particular feature of van Buuren's simulation experiment. A modified procedure for fitting MA models by means of SEM software is proposed, and some of its implications are discussed. |
| Document type | Comment/Letter to the editor |
| Language | English |
| Published at | https://doi.org/10.1007/BF02294322 |
| Permalink to this page | |