26.11.2012 Views

comparative value priorities of chinese and new zealand

comparative value priorities of chinese and new zealand

comparative value priorities of chinese and new zealand

SHOW MORE
SHOW LESS

You also want an ePaper? Increase the reach of your titles

YUMPU automatically turns print PDFs into web optimized ePapers that Google loves.

The Ecological Fallacy <strong>and</strong> Internal Validity<br />

The essence <strong>of</strong> internal validity is accounting for third-variable influences on<br />

interpretations <strong>of</strong> presumed if A then B relationships in which A represents the<br />

independent variable <strong>and</strong> B the dependent variable. It is here that a source <strong>of</strong> noncomparability<br />

between an individual <strong>and</strong> an aggregate group correlation <strong>of</strong> the same<br />

variables may arise. Concerning shifting from individual to group analyses, “In shifting<br />

from one unit <strong>of</strong> analysis to another, we are very likely to affect the manner in which<br />

outside <strong>and</strong> possibly disturbing influences are operating on the dependent <strong>and</strong><br />

independent variables” (Blalock, 1964: 79). As a result <strong>of</strong> the grouping operation, one<br />

may have controlled for the effects <strong>of</strong> other variables in addition to A, making the<br />

ecological estimate less biased than the individual estimate (Gove <strong>and</strong> Hughes, 1980).<br />

Or one may have introduced various confounding variables, making the ecological-level<br />

correlation more biased (Hanuschek, Jackson <strong>and</strong> Kain, 1974). If a difference occurs<br />

between ecological- <strong>and</strong> individual-level correlations, the problem may be due to a<br />

failure to specify the correct model, <strong>and</strong> not to an inherent logical fallacy in moving<br />

from individual to group correlations. The effects are there, but the reason for finding<br />

differences may be due to a study design failure. Such a failure in cross-cultural studies<br />

is the use <strong>of</strong> country <strong>of</strong> citizenship as a grouping variable; this introduces a nearly<br />

incomprehensible number <strong>of</strong> confounding variables.<br />

The Ecological Fallacy <strong>and</strong> Construct Validity<br />

In this context, construct validity refers to the concept developed by Cook <strong>and</strong> Campbell<br />

(1979), who argue that the term can refer only to constructs at the same level <strong>of</strong><br />

reduction. That is, a construct can only have validity at the individual or group level<br />

unless it can be demonstrated to have validity at both (Sharon Schwartz, 1994). When<br />

discrepancies between individual <strong>and</strong> ecological correlations remain after we believe we<br />

have controlled for all confounding influences, this may be due to (1) further<br />

misspecifications, where other confounding variables are not taken into account, or, (2)<br />

that an aggregate variable may measure a different construct than its namesake at the<br />

individual level (Firebaugh, 1978: 560). The construct referenced on the ecological<br />

level may derive from the context or social environment in which individuals live,<br />

distinct from the attributes <strong>of</strong> those individuals (Blau, 1960; Blau <strong>and</strong> Blau 1982; Davis,<br />

Spaeth, <strong>and</strong> Huson, 1961; Farkas, 1974; Meltzer, 1963). Wealth as an individual<br />

185

Hooray! Your file is uploaded and ready to be published.

Saved successfully!

Ooh no, something went wrong!