I would suggest reading the following article by Paul Spector on Physical Symptoms Inventory.  here he makes a very clear explanation on causal vs effector indicator scales.

Spector, P. E., & Jex, S. M. (1998). Development of four self-report measures of job stressors and strain: Interpersonal Conflict at Work Scale, Organizational Constraints Scale, Quantitative Workload Inventory, and Physical Symptoms Inventory. Journal of occupational health psychology, 3(4), 356.

Thank you. I read the piece you suggested. I understand what the difference is between a formative and reflective latent variable. What I'm more interested in is if "health" should be considered formative or reflective.

I spoke with someone about this issue yesterday and his thought was it could be either.  I have both clinical and self-report measure of both physical and functional health. The suggestion was that it makes more sense to treat the clinical measures as formative, and the self-report as reflective.

The clinical measure could be considered "measured without error" as they readings from instruments. For example, for clinical physical health my measures are blood pressure (categorized into normal, prehypertensive, hypertension 1, and hypertension 2), BMI (categorized into underweight, normal, overweight, and obese), and waist circumference (categorized according to risk for heart disease). For the clinical functional health I have measures of flexibility, strength, and fatigue for different body areas. These measures could be viewed as "forming" physical and functional health. I am thinking that I would not have to model the measurement error given that these items can be considered "objective" and measured without error. Does my logic follow here?

I am separately modeling self-reported physical and functional health. These latent variables could be considered reflective because the indicators are based on perceptions of these facets of health. Again, does my logic make sense?

I would appreciate feedback on my reasoning stated above. Does it sound reasonable? Are there problems I might encounter in the review process when I submit this work for publication?

Hi Gretchen,

Health is a possible formative construct especially if the component parts of "Health" are construed as being dimensions of the Health construct. This means that, if without a dimension(s) Health will not be fully defined, then, a formative construct makes sense. 

In summary, whether you specify it as a formative construct will depend on:

I have included below two very informative papers that address your concern. 

Thank you,

Madison

Great question! I noticed you're saying physical health. My first response was tied to how people perceive their health and how it is tied to culture, expectations, and reference groups. Similarly, when I see people have one impediment, even if severe, they don't always say poor health. 

Thank you, Madison. I will add that to my growing library of reference articles (e.g., Diamontopoulos (2006), and Edwards & Bagozzi (2000)).

Charles, the way I am defining physical health is that it represents health conditions, whereas functional health is how those conditions interfere with daily functioning. I hope that helps.

I´m agree with Madinson. So, I think you must think in a proposal more simple, more specific.

Gretchen, I think we get into trouble when we try to turn "physical [or mental] health" into an outcome measure. There's no agreed-on measure for that, if we even agreed on what "that" is.

I'd try for more precise constructs like physical function, perceived health status, medical history and distinct biometric values like BP and BMI.