Here I look at the data / simulation comparison using different trigger definitions which take into account the fact that JP1 was prescaled in data but not in simulation ...

In the first data / simulation comparison, I saw two major areas of disagreement. The first was an excess of jets seen in the simulation compared to the data in the endcap region which I believe can be attributed to the gains used in generating the simulation. The second area of disagreement was in the jet Pt spectra. Apparently, the JP1 simulation was not prescaled whereas the data was, which, coupled with the way I divided the sample by trigger can lead to discrepencies.

To make a better comparison between data and simulation, I have altered the way I divide the sample by trigger. In my previous study, I divided events into exclusive samples as described in the previous blog page. I now divide the sample into three catagories, jets which fire the L2JetHigh trigger, jets which fire the JP1 trigger, and jets which did not fire either of these triggers. I then scale each sample seperately. This division will lead to double counting when the jets fire both triggers, but it allows for the consistent treatment of the different triggers between data and simulation.

In addition to the new trigger division, I have created a sample where I only look at jets which fall between an eta and detector eta of |0.3|, so that there is no contamination from the endcap.

Figure 1: This figure shows the jet eta distribution for the 'limited eta range' sample.

As in the last blog page, I ran the data and simulation through my dijet code which imposes cuts and selects candidates as described here. Below, I compare data and simulation for the high pt jet and low pt jet components of the dijet event seperately. I also look at the full eta sample and the 'limited eta range' sample.

Figure 2: This figure shows the jet pt data / simulation comparisons for the high pt jet for the full eta range sample. The top pannels show the data (blue) and simulation (red) spectra. The bottom pannels show the ratio of data to simulation. From left to right, the columns show the L2JetHigh events, JP1 events, and untriggered events. The corresponding figrue for the low pt jet can be seen here.

Figure 3: This figure is the same as figure 2 except that all jets in the sample have |eta,detEta| < 0.3. The corresponding figure for the low pt jet can be seen here.

In figure 2, we see better agreement between data and simulation than what was seen in the last blog post. The agreement gets even better when contributions from the endcap are removed as in figure 3. The untriggered sample still shows data / simulation disagreement but this should probably be expected due to the prescale issue.

Figure 4: This figure shows the jet eta data / simulation comparison for the high pt jet. The figure layout is the same as in figure 2. The corresponding plot for the low pt jet can be seen here.

As in the previous blog post, we see a discrepency in the jet eta spectrum in the endcap region. We did not expect the new trigger scheme to fix this issue as it is due to the gains used when generating the simulation.

Figure 5: This figure shows the jet phi data / simulation comparison for the high pt jet, full eta sample. The figure layout is the same as in figure 2. The corresponding plot for the low pt jet can be seen here.

I have also put together pdfs of the above plots along with several other quantities such as jet mass (which has poor data / simu matching), jet RT, track and tower sum pt, and # of tracks and towers (which also has poor data / simu matching). The pdf for the full eta range sample is here and the pdf for the restricted eta range sample is here.