Time series analyses with psychometric data

8Citations
Citations of this article
9Readers
Mendeley users who have this article in their library.

Abstract

Understanding of interactional dynamics between several processes is one of the most important challenges in psychology and psychosomatic medicine. Researchers exploring behavior or other psychological phenomena mostly deal with ordinal or interval data. Missing values and consequential non-equidistant measurements represent a general problem of longitudinal studies from this field. The majority of process-oriented methodologies was originally designed for equidistant data measured on ratio scales. Therefore, the goal of this article is to clarify the conditions for satisfactory performance of longitudinal methods with data typical in psychological and psychosomatic research. This study examines the performance of the Johansen test, a procedure incorporating a set of sophisticated time series techniques, in reference to data quality utilizing a Monte Carlo method. The main results of the conducted simulation studies are: (1) Time series analyses require samples of at least 70 observations for an accurate estimation and inference. (2) Discrete data and failing equidistance of measurements due to irregular missing values appear unproblematic. (3) Relevant characteristics of stationary processes can be adequately captured using 5- or 7-point ordinal scales. (4) For trending processes, at least 10-point scales are necessary to ensure an acceptable quality of estimation and inference.

Cite

CITATION STYLE

APA

Stadnitski, T. (2020). Time series analyses with psychometric data. PLoS ONE, 15(4). https://doi.org/10.1371/journal.pone.0231785

Register to see more suggestions

Mendeley helps you to discover research relevant for your work.

Already have an account?

Save time finding and organizing research with Mendeley

Sign up for free