![Sample our Behavioral Sciences journals, sign in here to start your access, latest two full volumes FREE to you for 14 days]({"type":"image" , "src" : "/sda/110801/TOC-Subject-Sample-2021-Behavioral-Sciences.jpg"})
Abstract
Primarily from a measurement standpoint, we question some basic beliefs and procedures characterizing the scientific study of human behavior. The relations between observed and unobserved variables are key to an empirical approach to building explanatory theories and we are especially concerned about how the former are used as proxies for the latter. We believe that behavioral science can profitably reconsider the prevailing version of this arrangement because of its vulnerability to limiting idiosyncratic aspects of observed/unobserved variable relations. We describe a general measurement approach that takes into account idiosyncrasies that should be irrelevant to the measurement process but can intrude and may invalidate it in ways that distort and weaken relations among theoretically important variables. To clarify further our major concerns, we briefly describe one version of the measurement approach that fundamentally supports the individual as the primary unit of analysis orientation that we believe should be preeminent in the scientific study of human behavior.
Notes
1 Discovering what attributes of the objects being studied remain invariant under which transformations is the way Keyser (Citation1956) described the objective of scientific inquiry.
2 Accepting the above Kluckhohn and Murray premise, describing the ways “No two people are alike” is more the province of the novelist and the poet than the scientific psychologist. To the extent that is true, it is imperative to discriminate those features that can be studied scientifically from those that cannot.
3 One consequence of this emphasis is the separation it allows between prediction and selection (Nesselroade & Molenaar, Citation2010). The former can have an intraindividual slant by emphasizing predicting future behavior from the individual’s past behavior whereas selection remains an interindividual differences concern.
4 P-technique factor analysis has been further enhanced by the development of dynamic factor analysis (see, e.g., Molenaar & Nesselroade, Citation2012), which also can be applied to the MTS data of one or more individuals.
5 This is not to deny the ultimate necessity of probabilistic statements for behavioral science but to raise the question of which relations are the probabilistic ones. Instead of the mechanisms underlying behavior being the probabilistic ones, we are suggesting that it may be the linkages between manifest and latent variables that should be so viewed.
6 It also may be the case that the occasions of measurement are not synchronous from one individual to another, but this is not a necessary assumption for executing MRSRM designs.
7 Thurstone also mentioned another group of transformations—different sub-populations being studied—but here he was rather pessimistic regarding the likelihood of factorial invariance. For example, he argued that “… the factor composition cannot be expected to be invariant for different age groups, for example, or different groups of subjects, selected by criteria that are related to the factors involved” (p. 360). Meredith (Citation1964a), using Lawley’s selection theorem, systematically explored factorial invariance across subpopulations derived by selection from a common parent population. Meredith (Citation1993) remains today the essential touchstone for the rigorous use of factorial invariance concepts in measurement.
8 Because we are focusing on modeling covariance matrices, we are not dealing with intercepts here. The reader is referred to Meredith (Citation1993) for discussion of the role of intercepts in establishing factorial invariance.
9 The idiographic filter, a proposal for utilizing psychological measurement models to deal with idiosyncrasy challenges some of the bedrock traditions that have driven psychometric developments and practice for decades.
10 We are regarding the factor intercorrelations to be manifestations of a causal mechanism—not the mechanism itself, perhaps, but at least a “projection” of it.
11 Although sets of within-individual variation (P-technique) were used to illustrate higher-order invariance, it should be noted that the arguments apply to sets of between-persons variation (subgroup comparisons) as well (see, e.g., Nesselroade & Estabrook, Citation2008).
12 In a related vein, the handling of DIF in item-response theory still requires pooling across subjects in the respective sub-samples. Such pooling across subjects is a hallmark of analysis of interindividual variation. Molenaar, Huizenga, and Nesselroade (Citation2003) demonstrated by means of a simulation experiment that factor analysis of interindividual variation is insensitive to the existence of large-scale heterogeneity and can provide quite meaningful-appearing factor patterns when the individual factor patterns vary greatly. In Kelderman and Molenaar (Citation2007) this is proven.