Reproducing OCTO

[andkov]

@annierobi , @ampiccinin

replicating OCTO models with an older data file

[andkov]

After verifying the encoding of two dementia varialbes dementia_entry (positive diagnosis of dementia prior to entry into the study) and dementia_ever (dementia diagnosis at some point in life, but after the entry into the study), the OCTO sample produced the following tally:

from script in .\manipulation\estimation\octo\annirobie\0-prepare-data.R

> ds_wide %>%
+   dplyr::group_by(dementia_entry, dementia_ever) %>%
+   dplyr::count()
Source: local data frame [4 x 3]
Groups: dementia_entry [?]

  dementia_entry dementia_ever     n
           <dbl>         <dbl> <int>
1              0             0   477
2              0             1   126
3              1             1    98
4             NA             1     1

From the total sample, 477 individuals will not be diangosed with dementia at any point in their life. 126 respondents are diagnosed with dementia at some point, but only AFTER they have been enrolled in the study and gave the first wave of measurements. 98 individuals have entered the study already having a diagnosed dementia condition.

[andkov]

Plan for group comparison
Run 1 = 477
Run 2 = 477 + 126

[andkov]

@ampiccinin @annierobi , I have successfully adjusted the automation script to OCTO and have produced the output for three conditions of interests: dem_ever_0 (477 individuals), dem_entry_0 (477+126 =603 individuals) and all (477+126+98 = 701 individuals).

I have been working on processing these results so that I could provide you with three seed reports that print the results in the familiar format. I am getting very close, but i'm not there yet. For now, however, I have zipped the output files and will send them to you so you can inspect individual outputs if you'd like.

The report comparing these groups should follow shortly, maybe today, likely tomorrow.