Minutes of the EW working group meeting held on dec. 18 ======================================================= present: A. Venturi (AV), A. Gregorio (AG), A. Lucotte (AL), B. Pietrzyk (BP) B. Bloch (BB) and M.N. Minard (MNM) Topics: / selections / correlations treatment in the fits / / 2-g background for hadrons / notes / ELECTRONS: - L. Thompson reported by E-mail: "for the bhabhas I am still looking and trying to understand the ALISTAR-TOPAZ0 discrepancy for the t-channel subtraction, I managed to run ALISTAR supposedly in cos theta mode (rather than cos theta*) but the results I got were as far from the TOPAZ0 results but in the opposite direction for the full cross-section!. I have started to go into the ALISTAR code to understand exactly how the cuts are applied and am wor -king on that right now. Apart from that I have written some more of the ALEPH note and this is progressing." MUONS: - AV is still working on the pre-scan runs in 1995 which seem to have a wrong bunch identification. - concerning the muons analyses, AV and Imma Riu (IR) met last week and tried to understand the source of the discrepancies between their repsective results (about 1 sigma/peak, 1.7 sigma/P+2). One of the main difference comes from the taus-contamination, which stands at the level of ~0.2% for AV and ~2% for IR. In addition IR did not use yet the tau-MC produced at off-peak energies. AV and IR should meet again next year. to compare their list of events. - AV is also working on the sagitta correction for the tracking and a note is in preparation. TAUS : - AG reports on the status of the taus's note. It will be released in the middle of January, next year. HADRONS: - A discussion arose on the error of 2-gamma: BP reported the uncertainties on 2-g quoted for the different LEP experiments and pointed out their importance on the Z-width uncertainty: D(sig_2g) D(G_z) -------------------------- ALEPH 11 pb 4.7 MeV (4.3 if 0 pb) DELPHI ~5 pb 4.2 MeV (VSAT used for lumi...) L3 3.2 pb 4.3 MeV OPAL 8.9 pb 5.3 MeV (no 95's data) The error quoted here for ALEPH is the combined number of (CALO+TPC) selection for ONE year alone. Presently, the 2-g background is estimated independantly year by year, and the source of uncertainty is purely statistical. It is pointed out that doing this way, the only correlation comes from the choice of the integration range. MNM pointed out that for L3 for instance, which uses calorimetric informations, the acceptance is much reduced with respect to our (they do not select events as close to the beam pipe as we do in Aleph) and so wondered wether the uncertainty concerning these missing events are well estimated. AL will try to re-estimate the 2-gamma background on a fit grouping two different periods of data taking, after checking of triggers condition. - the list of 1993 and 1995 hadrons counts are still not ready. For 1993, the same results as put in AL's note are found. But these numbers were not used at that time because of the problem due to the use of XLUMOK / XLSLUM For 1995, checks still remain to be done with MNM. LUMINOSITY: - BB reports on the checks which have been made using the new BHLUMI version applied to 93 and 94 measurements. She compared present results with those including Z-contribution calculated by Bolek at that time. Some good news fo 1993: () 3 / 4 points show a very good agreement, within 1 sigma_stat (=2,4.10-4) () For P+2 point, a 9,8.10-4 disrepancy is observed. BB will re-run the new BHLUMI for the last point switching OFF the Z contribution in order to see if this can explain such difference. ENERGY: - BB noted that for 1995 the energy per run were not updated in SCANBOOK. She pointed out the importance of knowing these values at a ~10 MeV precision, arguing that a change of 10 MeV may imply (according to the energy) a relative change of 2.10-4 in luminosity. - AL will re-run 1995 selection on 95's energy files (old ones= Warsaw) to produce files for J. Boucrot during the hollidays. AL and BB pointed out that energy files are still not final, and so that central values may change till Moriond. FITS & TREATMENT OF ERRORS: - Following AV, MNM started a discussion on the treatement of the errors in the fitting program. Points to be remembered are the followings (BP and AL): presently in MIZA () Correlation matrices are used for luminosity and energy measurements () For the different selections, the uncertainties are split into multiplicative errors (% to the cross-sections) and constant errors (2-gamma background). The first ones are treated as if they were totally correlated among them while the second one (re-estimated every year in an _almost_ independantly way) is uncorrelated. MNM and AV agree that this way of including uncertainties is not adequat for the different selections. For instance, AV reports that the uncertainty on his off-peak acceptance should be treated as correlated year by year since the same MC is used to compute it... Two decisions have been taken: It is agreed that each analysis should provide correlation matrices year by year and inside each year for systematic uncertainties. It is decided that BP should write a note describing in details the way the uncertainties are presently treated in the fitting programm. As a conclusion, it is decided that all the individual notes should be released in the middle of January, including a part explaining the way to estimate correlations betweeen measurements. Next meeting on 15th January 97 at 14h30. Happy christmas and happy new year ! submitted by AL.