4
and terms into their own activities. In part, this seems to be based on a widespread
acceptance of the premise contained in Lord Kelvin’s credo—that without measurement, knowledge is at best “meagre and unsatisfactory”—with the further implication that, as put by (for example) psychologist James McKeen Cattell, “psychology
cannot attain the certainty and exactness of the physical sciences, unless it rests on
a foundation of experiment and measurement” (1890: p. 373; for more extended
discussions see, e.g., Michell, 1999; Briggs, forthcoming).
In the late nineteenth and early twentieth centuries, scholars working in the field
of psychophysics, such as Gustav Theodor Fechner and Stanley Smith Stevens,
investigated relationships between physical stimuli and their associated human
responses (sensation and perception), on the premise that by establishing quantitative relationships (e.g., through what is now known as Fechner’s law; Fechner,
1860) between known physical quantities such as weight and human sensations, the
latter could be measured as well.
4
Separately, scholars interested in individual differences, such as Francis Ysidro Edgeworth (1888, 1892) and Charles Spearman
(1904, 1907), applied statistical methods and logic originally developed in the context of astronomy (in particular related to the Gaussian distribution) to the study of
human beings, largely in the context of scores on educational and psychological
tests. Starting from the observation that some sets of test items (such as arithmetic
questions) seemed to give more internally consistent results than others, they posited that an individual’s “observed scores” (O) on tests could be decomposed into
“true scores” (T) and “errors” (E), i.e., O = T + E, giving rise to a field that would
later come to be known as psychometrics.
Perhaps unsurprisingly, these early attempts at measuring psychosocial properties were met with skepticism by some members of the broader scientific community. As one important committee concluded, referring in particular to
psychophysics, “to insist on calling these other processes measurement adds nothing to their actual significance but merely debases the coinage of verbal intercourse” (Ferguson et al., 1940: p. 345). Even within the human sciences, many
were (and still are) skeptical of psychosocial measurement, often based on concerns such as that “not everything that can be counted counts, and not everything
that counts can be counted”, as put by the sociologist William Bruce Cameron
(1963: p. 13).
In part, skepticism about psychosocial measurement may have been related to
the very fact that, as previously mentioned, available examples of measurement
pertained exclusively to physical properties, which might have given the impression
that only physical properties are measurable. Interestingly, even many prominent
proponents of psychosocial measurement seem to have accepted this position; for
example, Jum C. Nunnally and Ira H. Bernstein, in the third edition of their influential textbook Psychometric Theory, noted that (1994)
“it is more defensible to make no claims for the objective reality of a construct name such
as ‘anxiety’ and simply use the construct name as a convenient label for a particular set of
4 Work in the psychophysical tradition helped establish a number of relationships between physical
phenomena and sensations (e.g., Fullerton & Cattell, 1892; Thurstone, 1927) that are even today
still foundational for fields such as audiology.
1 Introduction
and terms into their own activities. In part, this seems to be based on a widespread
acceptance of the premise contained in Lord Kelvin’s credo—that without measurement, knowledge is at best “meagre and unsatisfactory”—with the further implication that, as put by (for example) psychologist James McKeen Cattell, “psychology
cannot attain the certainty and exactness of the physical sciences, unless it rests on
a foundation of experiment and measurement” (1890: p. 373; for more extended
discussions see, e.g., Michell, 1999; Briggs, forthcoming).
In the late nineteenth and early twentieth centuries, scholars working in the field
of psychophysics, such as Gustav Theodor Fechner and Stanley Smith Stevens,
investigated relationships between physical stimuli and their associated human
responses (sensation and perception), on the premise that by establishing quantitative relationships (e.g., through what is now known as Fechner’s law; Fechner,
1860) between known physical quantities such as weight and human sensations, the
latter could be measured as well.
4
Separately, scholars interested in individual differences, such as Francis Ysidro Edgeworth (1888, 1892) and Charles Spearman
(1904, 1907), applied statistical methods and logic originally developed in the context of astronomy (in particular related to the Gaussian distribution) to the study of
human beings, largely in the context of scores on educational and psychological
tests. Starting from the observation that some sets of test items (such as arithmetic
questions) seemed to give more internally consistent results than others, they posited that an individual’s “observed scores” (O) on tests could be decomposed into
“true scores” (T) and “errors” (E), i.e., O = T + E, giving rise to a field that would
later come to be known as psychometrics.
Perhaps unsurprisingly, these early attempts at measuring psychosocial properties were met with skepticism by some members of the broader scientific community. As one important committee concluded, referring in particular to
psychophysics, “to insist on calling these other processes measurement adds nothing to their actual significance but merely debases the coinage of verbal intercourse” (Ferguson et al., 1940: p. 345). Even within the human sciences, many
were (and still are) skeptical of psychosocial measurement, often based on concerns such as that “not everything that can be counted counts, and not everything
that counts can be counted”, as put by the sociologist William Bruce Cameron
(1963: p. 13).
In part, skepticism about psychosocial measurement may have been related to
the very fact that, as previously mentioned, available examples of measurement
pertained exclusively to physical properties, which might have given the impression
that only physical properties are measurable. Interestingly, even many prominent
proponents of psychosocial measurement seem to have accepted this position; for
example, Jum C. Nunnally and Ira H. Bernstein, in the third edition of their influential textbook Psychometric Theory, noted that (1994)
“it is more defensible to make no claims for the objective reality of a construct name such
as ‘anxiety’ and simply use the construct name as a convenient label for a particular set of
4 Work in the psychophysical tradition helped establish a number of relationships between physical
phenomena and sensations (e.g., Fullerton & Cattell, 1892; Thurstone, 1927) that are even today
still foundational for fields such as audiology.
1 Introduction