Reply to thread

Message

<blockquote data-quote="Cadence" data-source="post: 8467599" data-attributes="member: 6701124">Measurement of latent variables - the ones not directly observable in the way height supposedly is -&nbsp; has a long history (the field now called psychometrics - as opposed to psychometry which would be much more lucrative).&nbsp; Markus and Boorsboom have a nice book on discussing how to conceptualize the validity of such measurements (although the later chapters are especially dense, and the early ones wouldn't hurt by having either some college level statistical and/or psychological training).&nbsp; <a href="https://www.amazon.com/Frontiers-Test-Validity-Theory-Multivariate/dp/1841692204" target="_blank">Frontiers of Test Validity Theory: Measurement, Causation, and Meaning (Multivariate Applications Series): Markus, Keith A., Borsboom, Denny: 9781841692203: Amazon.com: Books</a>&quot;IQ&quot; as an example of a latent variable certainly has been abused and much of the early work is closer to phrenology.&nbsp; That's a separate issue from whether there are underlying things the general public would label as &quot;intelligence&quot;.&nbsp; Which is separate from if it is worth the consequences of pursuing.As far as average height, that's only easy because we gloss over the technicalities.&nbsp; People for example shrink over the course of the day, so when you measure them makes a difference.&nbsp; As far as weight, one fun data set in an intro stat book is the daily measurements of &quot;the kilogram&quot;.&nbsp; The statistical measurement error models for these physical measurements have some relationships to the psychometric ones.</blockquote>

[QUOTE="Cadence, post: 8467599, member: 6701124"] Measurement of latent variables - the ones not directly observable in the way height supposedly is - has a long history (the field now called psychometrics - as opposed to psychometry which would be much more lucrative). Markus and Boorsboom have a nice book on discussing how to conceptualize the validity of such measurements (although the later chapters are especially dense, and the early ones wouldn't hurt by having either some college level statistical and/or psychological training). [URL='https://www.amazon.com/Frontiers-Test-Validity-Theory-Multivariate/dp/1841692204']Frontiers of Test Validity Theory: Measurement, Causation, and Meaning (Multivariate Applications Series): Markus, Keith A., Borsboom, Denny: 9781841692203: Amazon.com: Books[/URL] "IQ" as an example of a latent variable certainly has been abused and much of the early work is closer to phrenology. That's a separate issue from whether there are underlying things the general public would label as "intelligence". Which is separate from if it is worth the consequences of pursuing. As far as average height, that's only easy because we gloss over the technicalities. People for example shrink over the course of the day, so when you measure them makes a difference. As far as weight, one fun data set in an intro stat book is the daily measurements of "the kilogram". The statistical measurement error models for these physical measurements have some relationships to the psychometric ones. [/QUOTE]

Verification