2016-09-01-Prep

[andkov]

To Do

• [ ] investigate the misClassification logic
• [x] Develop a stable script for Single model exploration (https://github.com/IALSA/ialsa-2016-amsterdam/commit/3c6d4821add4d2873e64a1a1a9a57cf2c4e963a1)
• [x] Correct model specification ( #27) (https://github.com/IALSA/ialsa-2016-amsterdam/commit/665340b3cce0668c6efe3e2d7c54a6957bf54f83)
• [ ] Outsoure the object with model specification to the external object (idea : have model specification + utility functions + starter)
• [ ] Edit the Stage 2, namely add another step between 2 and 3 that would prepare data for modeling (centering, subsetting)
• [ ] develop a cascading estimation flow, in which the transition estimates from the less complex model are used as the starting values of the following, more complex model

  Questions

• [x] what is the substantive and implementational logic of fixing parameters with fixedpars option in msm() call?

[andkov]

Discussion questions

• what options in msm() call can be standardized across the studies? which ones will need tweaking during the estimation? if we can create such a list, I can create a wrapper function that would standardize (at least partially) the estimation call.