On the Nature of Measurement

Question 1

Are the necessary and sufficient conditions for measurement?

Answer 1

No, but there are common components that the authors argue for.

Question 2

4 Central Components of Measurement

Answer 2

T (Theory)
I (Instrumentation)
S (Scales & Units)
M (Modeling)

Question 3

Euclidean understanding of measurement.

Answer 3

Defined measure in terms of ratios of magnitudes (lesser vs. greater)

Question 4

Stanley Smith Stevens

Answer 4

Proposed measurement to be assignment of numerals to objects or events according to rules.

Came up with nominal, ordinal, interval, ratio scales

Question 5

Norman Campbell and Ferguson Committee

Answer 5

Campbell proposed that property must be both orderable and additive to be considered measurement

Problem was psychophysics was not additive. Can’t be considered measurement.

Question 6

Classical Test Theory (CTT)

Answer 6

result = true score + error (Spearman)

Question 7

Generalizability Theory (GT)

Answer 7

Includes facets. a “universe score” as true score (?)

Question 8

Item Response Theory (IRT)
- what does it provide that CTT and GT doesn’t?

Answer 8

Provides scales and units.

Question 9

What is Modeling in TISM?

Answer 9

Applying formal models to characterize the uncertainty of measurement results

Question 10

Systematic vs. Random error

Answer 10

Systematic is predictable and constant. Random is unpredictable and unknown.

Question 11

Substantive theory

Answer 11

provides a framework of the nature of a concept (?)

Question 12

Validity according to Samuel Messick (1989)

Answer 12

The degree to which evidence and theory support the interpretations of test scores for the proposed uses of tests.

Question 13

Types of uncertainty (3)

Answer 13

Definitional uncertainty (how property is defined)

Instrumental uncertainty (physical limitations from instrument)

Calibration uncertainty (uncertainty about unit or how scales are defined)