Over the past week, border security has been the biggest sticking point in Senate discussions about immigration reform.

Many Republicans, like Sen. John Cornyn (R-Tex.) have argued that the United States shouldn't provide a pathway to citizenship for the 11.5 million unauthorized immigrants currently in the country until the U.S. border with Mexico is secure.

But what does it actually mean for that border to be secure? Who decides? And is a completely secure border even possible?

1) At the moment, the U.S. government considers border security in the Southwest about 84 percent "effective."

It was only last year that the nine Border Patrol sectors along the Southwest border finally standardized their metrics for how effectively they were securing the border. The results are laid out in a 2012 report from the Government Accountability Office:

Each sector measures three things: "Apprehensions" are the portion of border-crossers who are caught and detained. "Turn backs" are people who had either climbed a fence or crossed the border but were sent back to Mexico. "Got aways" are, well, border-crossers who made it through without getting caught.

These numbers are thought to be reasonably accurate, says Rebecca Tallent, the director of immigration policy for the Bipartisan Policy Center. The Border Patrol will use cameras and other evidence to get a rough count for "turn backs" and "got aways." Workers at the Yuma sector, for instance, will comb the desert every morning and count footprints in the sand to see who slipped through in the night. That said, they're still rough estimates. It's very possible that other border-crossers are slipping by undetected.

In 2011, the GAO reports, border security along the Southwest was thought to be about 84 percent effective. That is, 16 percent of attempts to cross the border were successful — which amounted to about 85,000 people getting through. Another 61.3 percent resulted in apprehensions. And 22.7 percent of attempts were turned back.

Now, effectiveness varied by sector — the Yuma sector, which is very well-staffed, had an effectiveness rate of 93.7 percent. By contrast, border security in the Rio Grande Valley sector was just 70.8 percent effective. Here's a map of the sectors, via the BPC's Matt Graham:

And keep in mind that these are just the government's numbers. Other studies, like this recent reported from the Council on Foreign Relations, have suggested that the apprehension rate might be even lower than the GAO thinks — and that the Border Patrol is missing a large number of border-crossers. Still, when it comes to legislation, the official "effectiveness rate" is what counts...

2) Many Republicans originally wanted to postpone a pathway to citizenship until after that effectiveness rate was 90 percent — or even higher.

According to the New York Times, Sens. Bob Corker (R-Tenn.) and John Hoeven (R-N.D.) originally wanted an amendment to the immigration bill that "would have required a 90 percent effectiveness rate in apprehending or turning back illegal crossers, using a combination of conventional border infrastructure, like fencing and observation towers, and high-tech elements including heat-sensing cameras and drones."

Whether that's possible is an open question. As noted above, border security with the Southwest was considered 84 percent effective in 2011. The number of successful crossings dropped dramatically, to just 85,000 that year. But it's hard to know how much of that reduced flow is due to better border security and how much due to the weak U.S. economy.

Sen. John Cornyn, meanwhile, wanted to go even further than Corker and Hoeven. He wanted the Southwest border to achieve a 90 percent apprehension rate before any of the provisions offering a pathway to citizenship for current unauthorized immigrants kicked in. So for his measure, turn backs don't count.

That's a much stricter standard. Only about 61.3 percent of border-crossers are actually apprehended. Anther 22.7 percent are turned back. Tallent says there's a good reason for that — it's often expensive to detain and process border-crossers, particularly depending on where they've crossed. Cornyn's proposal would have required a big uptick in apprehensions.

Many senators seemed to agree that this was unworkable. On Thursday, the Senate voted  54-43 to put Cornyn's amendment aside for now.

3) In the end, Corker-Hoeven went with much more money and staff for border security rather than a hard metric.

The latest Corker-Hoeven amendment would scrap the 90 percent effectiveness requirement, and just make it a "goal." Instead, the senators have come up with an amendment to the bill that would spend $30 billion to double the number of federal border agents, complete 700 miles of fencing, and expand radar and aerial drone surveillance along the border.

"The deal is expected to secure at least a dozen more Republican 'yes' votes for the measure and could help ensure its passage by the sizable margin that proponents have said they need to make it viable in the House," my colleague Ed O'Keefe reports.

It's not clear what all this extra staff and technology will do for border security. There are also good questions about whether the Department of Homeland Security can even train that many guards, or whether there's actually room for 700 miles of fence — Corker called the provisions "almost overkill." But for Republicans looking to convince their base that they're taking border security seriously, overkill might be exactly what they need.

4) Even if Congress does bolster border security, that won't, by itself, eliminate illegal immigration.

Case in point: The Congressional Budget Office expects that the annual flow of new illegal immigrants would drop by just 25 percent if the Senate immigration bill became law. Now, that’s just a rough estimate — other experts I asked thought the real number could be much higher or lower in any given year. But the CBO's reasoning is worth following.

On the one hand, the CBO notes, beefed-up security measures would likely tamp down border crossings. But there’d also be more immigrants who overstay their temporary visas under newly established programs, such as the W-visas for low-skill workers. Many of these immigrants might just stay in the country once their visas lapse and work off the books.

Granted, policymakers have devised all sorts of strategies to crack down on illegal hiring, such as a E-Verify system for employers. But many of these policies don't work perfectly, and there's political resistance to more sweeping measures, such as creating a tamper-proof biometric card. My colleague Dylan Matthews recently wrote a nice rundown of the pros and cons of strategies to prevent companies from hiring illegally.