1994PASP..106..590T

This paper begins by reviewing the nature and necessity of derived accidental errors, and it describes thier unsatisfactory traditional status in high-dispersion spectroscopy. The paper goes on to review the analysis which yeilds the [Fe/H] averages in subsequent papers of this series. Choices of stars and raw data are discussed first, and teh corrections applied to these data are then considered. These corrections include temperature corrections based on the results in Paper II of this series. The way in which the data are finally assembled is then described. The analysis yields accidental errors as well as means, and these errors are of interest in their own right. For recent data, a compromise decision about assigning the errors yields results which are modest–but usually not decisive–improvements over the errors of older data. The paper concludes with a brief review of some ways to obtain better accidental errors in high-dispersion work.