System dynamics models are composed of stocks (representing continuous quantities about a system of interest), flows (controlling the in- and out-flow to the stocks), and variables (that might be constants or also change with the values of the stocks).
I still struggle to grasp when to use a stock and when to use a variable. For example, this model, taken from CLExchange's YT videos shows "Cinderella's happiness" as a stock that can change and the fairy godmother's concern (which rises as Cinderella's happiness decreases) as a variable that influences the flow towards CH. I would have modeled the godmother's concern probably as a stock variable, too. Yet, most papers about SD I find have very few stocks and many more influencing variables. Is there any guideline about this?
Disclaimer: I am mostly familiar with NetLogo's system dynamics modeler, which is quite limited, according to NetLogo Systems Dynamics Modeler for Agents.
Basically, a stock has memory. That is, it knows what value it had in the previous timestep (or tick) and, in the next time step the in-flows are added and the out-flows are subtracted. In contrast, a variable (or auxiliary) is calculated from the current values of all its inputs.
Even if you want to use NetLogo for system dynamics models eventually so that you can connect to agent-based models in some way, you may want to consider starting with specialised system dynamics software to design the SD elements and then redo them in NetLogo once you're happy with the design. I personally like Vensim, which is free for small models (PLE version) and has good tutorials built in.
