[Frost] Target polarization questions

[Michael Dugger]

Hi,

I'm trying to understand the target polarization and have a couple of 
questions.

Here is what I understand. Please correct me if I am wrong:

>From what I can tell by looking at the "Target Polarization" web page at
http://clasweb.jlab.org/rungroups/g9/wiki/index.php/Target_Polarization
To find the calibration constants for the polarization by cutting the 
data up into 4 types:

1) High B-field after polarization and just before going to low B-field

2) Low B-field right after the switch between low and high B-field

3) Low B-field right before the switch between low and high B-field

4) High B-field just after switch from low B-field and just before 
repolarization

We assume that the polarization can be calculated in the form p = c*A, 
where p is the polarization, c is a calibration constant, and A is the 
area under the peak of a frequency deviation plot. The c calibration 
constant is determined using only high B-field data. For low field data, 
the equation is p = c_{LF1}*A, where the c_{LF1} calibration constant is 
determined by using the c calibration constant and comparing the area A 
between the type 1 and type 2 data (i.e. c_{LF1} = c*A1/A2).

There is a calculation made for c_{LF2}, where c_{LF2} = c*A4/A3. There 
are some measurement where c_{LF2} are deemed as "anomalous", but even 
where the anomalous distinction is not made, there are fairly large and 
systematic differences between c_{LF1} and c_{LF2}. The values of the 
ratio between c_{LF1} and c_{LF2} (for non-anomalous measurements) are 
given as:
1.025
1.036
1.093
1.080
1.037
1.171
1.030
1.074
1.049
1.013

These value of c_{LF1}/c_{LF2} are systematically high and have an average 
value of 1.061.

Question: We are using c_{LF1} and the area A throughout the low B-field 
data to determine the polarization, so I must assume that the High-B field 
type 4 determination of P is not as accurate as type 1. Am I looking at 
this wrong?

Question: Do we use both high B-field types (types 1 and 4) to determine 
c? If so, how is it possible to have a systematic uncertainty in c be only 
1.6% when the c_{LF1}/c_{LF2} ratio is systematically high by 6.1% ?

Thanks for your time.

-Michael