Longitudinal Structural Equation Modeling by Newsom Jason T

Author:Newsom, Jason T. [Jason T. Newsom]
Language: eng
Format: epub
ISBN: 9781317975335
Publisher: Taylor & Francis (CAM)

Unequally Spaced Observations

Another deviation from the simple, ideal time metric is when observations are not equally spaced. In practice, this is likely to be true in many data sets, in which interviews, for instance, are not conducted precisely one year apart. Small deviations from time structure of this nature are unlikely to create serious biases, and data are usually treated as time structured.11 Provided that data are time structured (i.e., consistent across cases) and there is true unequal spacing of observations, there also is no special problem with estimating a latent growth curve. Loadings may be altered to appropriately reflect the difference in spacing. For example, observations from a survey conducted annually for the first three waves yet biennially for three subsequent waves could be assigned time codes 0, 1, 2, 4, 6, and 8. The slope factor mean would then represent change in yt for each yearly increase, with the annual rate of increase estimated over the last three biennial waves for the shorter interval. In other circumstances, data may be considerably less orderly than this and the length of interval may vary for different cases. Depending on the degree of variation across cases, a combination of missing data estimation and appropriate coding of the slope loadings may be sufficient to handle the problem. If there is considerable variation across cases in the regularity of the intervals, special estimation of these individually varying time points can be used, a method described in greater detail shortly.

Download

Copyright Disclaimer:
This site does not store any files on its server. We only index and link to content provided by other sites. Please contact the content providers to delete copyright contents if any and email us, we'll remove relevant links or contents immediately.