According to yours truly (see below) there were between 700 and 800 people at the Sunland Park rally where Hillary Clinton spoke for Obama. Rene Romo of the Albuquerque Journal, and Darren Meritz writing for the Las Cruces Sun News, each estimated the same crowd at 2600 people. Meanwhile, Michael Coleman and Colleen Heild of the Albuquerque Journal, estimated the crowd in Mesilla at "about" 3000," while Ashley Meeks of the Sun News refrained from an estimate, simply referring to it as a crowd of "thousands." Finally, the Albuquerque rally was estimated by the Albuquerque Journal at "1000-1500 people."
What gives? Why the discrepancies? How are crowds estimated, anyway?
First, estimating crowds is not easy, unless you have an overhead photograph and have the time to split it into grids, counting bodies in each grid. Most of the time this ain't gonna happen. Second, is a strong tendency for people to exaggerate the number of people in a crowd. For example, I asked a political figure who was at the Mesilla rally for his estimate of the crowd a full hour before Clinton's appearance, and he said there were "at least 3500 people there." At that moment my own count was more like 400-500, with people still trickling in. Third, a ground level you only have an oblique angle with which to estimate the size of the real estate you are trying to cover. This can make it difficult to estimate the density of the crowd over the entire spread of bodies you are trying to count up. What you are trying to do here is to count up, say to 100 people and put an imaginary boundary around these and then try to fill in the rest of the crowd with multiples of this imaginary space.
There are other techniques to estimate a crowd, but these have problems too. You can go by the number of tickets distributed or collected. Problem is, this often puts you at the mercy of those distributing or collecting tickets, and most of the time they have a vested interest in exaggerating the size of the crowd.
This latter approach is what seems to have accounted for the Journal estimates of the crowds in Albuquerque and Mesilla. They apparently asked Republican Party officials how many tickets had been distributed, and found out they distributed about 3000 tickets for the Albuquerque rally and the Mesilla rally and then did rough calculations based on their eyeballing each crowd. The coincidence of the number 2600 in the case of the Sunland Park estimates suggests they both came from a single source rather than an independent eyeballing estimate. It is extremely unlikely that they each calculated that exact amount independently. They might, for example, have asked Luis Avila, field coordinator for Obama in the South Valley, how many people he had prepared for, or asked someone else, say, a secret service agent, for an estimate, and then used that.
I remember a gubernatorial campaign rally I attended in Chihuahua in the mid-1990s. I was with a large contingent of news media people and was offered a ride on the media bus supplied by the PRI, the ruling party of Mexico at that time, struggling to win this governor's election, which was drawing national attention. With a nice vantage point on the speaker's podium, I estimated the crowd at between 5000 and 7000, and, to check out my own possible inadequacies as a crowd-counter, I asked some of the reporters for their take on the crowd size. Most agreed with me, some elevated my estimate up to about 10,000. The PRI media spokesperson heard us talking about this and insisted there were 17,000 people at the rally. She had been told by superiors this was the magic number. Most of the reporters duly wrote down that figure, except for me and another guy, and since I wasn't actually writing about the rally, I didn't think any more about it. It was over 100 degrees on that July afternoon in Chihuahua, and we were happy to drive back to PRI headquarters in the PRI air conditioned bus. The media spokesperson had a nice cold soft drink she handed to each of the reporters getting on the bus, until she got to her holdout reporter, who had told her he would report a crowd size of 10,000. "No coke for you until you give me 15,000," she said jokingly. The next day all but one news source reported a crowd of 17,000. Her candidate still lost the election.
Bottom line: estimating crowds is difficult and there is a strong human tendency to exaggerate the size of a crowd. You should be very skeptical about any estimate of a crowd's size unless you know that the methodology used to calculate was pretty exact. As for me, I will stick to my eyeball estimate of 700-800 persons being in attendance at the Sunland Park rally.