## Merge files for imputed hormones There are three files with identical content in three formats (CSV, SPSS, R): - fc_days.(csv|sav|rds) - bc_days.(csv|sav|rds) - lh_days.(csv|sav|rds) These files contain as their first column the day in cycle, defined as follows: - fc_day: days since last menstrual onset, from 0 to 39 - bc_day: days until next observed menstrual onset, from -40 to -1 - lh_day: day relative to LH surge (0) from -15 to +15 Then, each file includes imputed steroids (estradiol, progesterone, testosterone, mass-action estimated free estradiol and testosterone) plus a measure of the uncertainty of the imputation. All hormones are in log(pg/ml) units. For example, - `est_free_estradiol_lh`: the estimated mean log(pg/ml) at each cycle day relative to the LH surge. - `est_free_estradiol_lh_se`: the standard error associated with that estimate - `est_free_estradiol_low`, `est_free_estradiol_lh_high`: the 95% credible interval boundaries The variables in the SAV and RDS files contain additional variable labels. Once cycle days have been computed as shown above, the files can simply be merged on the cycle day column. To validate a different assay measure, you could then estimate a correlation between the imputed values and the measured values on the same day. In addition, estimated logged steroid concentrations can be used as a predictor in a regression. Because the steroid estimates are regression-based, the resulting regression coefficients are already adjusted for error.