Is there any fixed rulee or fixed steps for developing a measurement instrument?
Outside an organization that may have defined its own rules, there is no formal rule to follow. Though the 'best practice' approach for all sorts of developments applies as well:
The above is a rough generalization of the so-called 'V-Model'. Other names for this model and other models exist, but the approach is virtually the same - applying to pure software designs as well as to hardware, mechanical engineering . . . you name it.
What's hard to show: expecially during the design phase (steps 1-4): do intensive critical reviews of the outcome of each stage - as well as vs. the specifications of the prior step. The word 'critical' cannot be over-emphasized: this is the only way to unveil hindrances, bad design, contradicting requirements etc.
Following these simple steps helps to avoid unnecessary 'iterations'. Nonetheless be prepared for unforeseen 'surprises' during all stages. Especially regarding the term 'V-model' you will find a lot of materials "in the net".
What's not covered by the model itself, but should happen prior stage 1: assess what you want to develop, why you want to do that (really nothing available ? Need more accuracy? ... and so on ...). This is denominated the "make or buy" decision that may or may not stop a development.
Hope this helps
So far I know there are churchill's (1979), DeVellis (2003), and Vazquez et al.'s (2002) multi-phased protocol for developing a Multivariate measurement scale in social science area.
Is there anyone who can add more please.
I think it depends a lot on what construct you want to measure, what sort of instrument you want to use, and what the final purpose of the measure is. Development of a psychological questionnaire involves different steps than a knowledge test, than an equation to estimate caloric expenditure, than a poll of voting preferences, etc.
- Badges
- Science topic
- Similar topics
- Psychometrics
- Measurement