00:00What's going on, everybody? Welcome back to another video.
00:02In this series, I show you how easy it is to manipulate and lie with data.
00:12Now, when I toss out the word average, you most likely know what that is.
00:16It's a very common statistical concept.
00:18So in this video, I'm going to show you how you can easily manipulate data using averages
00:21and a way that you can actually make it a little bit more accurate.
00:24All right, so really quickly, you may recognize this off the bat
00:28because I actually made a LinkedIn post with very similar data and kind of explaining this.
00:34And this is what prompted me to create this series in the first place.
00:36But what we're looking at is we're going to take a look at the average.
00:40Now, right here in our data, we have the state.
00:42We have California and we have Georgia.
00:44Vastly different states in terms of average salaries.
00:47Vastly.
00:48California has dramatically higher salaries.
00:51This is just made up data.
00:52So don't take this as real.
00:55Here we have average salaries in the state of California.
00:57The average salary is $200,000.
01:00In the state of Georgia, the average salary is $75,000.
01:04So if we come over here, we're going to say equals, and we're just going to take the average.
01:08We're going to take the average of these two numbers.
01:11And what we can conclude from this is that the average salary for California and for Georgia
01:17is $137,500.
01:20Just as a numbers thing, this number is correct, right?
01:23If you take $200,000 plus $75,000 and divide it by two, you are going to get $137,500.
01:31But this is not correct.
01:33In fact, this isn't just a lie.
01:36This is incorrect data.
01:38What we're working with now is we've already taken the average salary, and now we're taking
01:42the average of an average.
01:44And we're not using the underlying data to conduct and get this average, which is going
01:49to make it completely wrong.
01:51Scroll over just a little bit, and now we're going to be working with what I like to call
01:55raw data.
01:57Now, I like to call it that because that's what it is.
01:59If you don't know what raw data is, it is the actual data before it is manipulated, changed,
02:04or aggregated in any way.
02:06So this is our raw data behind those averages.
02:09And right here, we have California, and we have our salaries, and then we have a lot more
02:13data for Georgia, and here are all the salaries.
02:17So let's come over here, we're going to do the average, just like we did before.
02:22We're going to say the average for California, and look at that, it's that $200,000.
02:27Then we're going to do the average for Georgia.
02:29So we're going to say average, and I did it again, and we're going to take all of Georgia.
02:35So now, you can see we have $200,000 and $75,000.
02:39Those numbers are correct, but let's take a look at the average for everything.
02:44So now, we're going to say average, and we're going to take the average for this entire
02:49salaries column.
02:51And now, the average is actually $100,000, which is very different than $137,500.
02:59In fact, to be more specific, it's $37,500 off.
03:02And that is a big, big difference.
03:05Now, I've seen this in the real world play out.
03:07I've seen this on the news, and I told my wife, and she was like, didn't understand it
03:12at all, and it was kind of hard to explain in that moment.
03:14But if I could have shown her this video, she would have totally gotten it.
03:18But I didn't, and she doesn't believe me, and that's okay.
03:21But this is the actual average.
03:23This is the underlying data behind these averages.
03:27And statistically, if you average an average, most likely, most of the time, you are going
03:32to get an incorrect number.
03:33So if you want to get the correct average, we're going to have to do this right here.
03:37Whereas you take the underlying data, this is our raw data, and you perform an average.
03:42Now, what's interesting is, is if you look throughout all of our salaries here, nobody
03:48makes $100,000.
03:50In fact, most people don't actually come anywhere close to $100,000.
03:53This is $175,000, $225,000.
03:56Then we have $95,000.
03:57That's the closest one.
03:58Now we have $85,000.
03:59No one actually makes $100,000.
04:01And so that's kind of misleading.
04:03Statistically, this is the average, but this may not be what you're trying to convey to
04:08whoever is looking at this.
04:09I think oftentimes when people look at averages, what they really want is something like median.
04:15So let's come in here, and let's take a look at median, and I need to write that right.
04:21But this is going to return the number in the middle of the set of our given or our array
04:26of numbers.
04:27So we're going to do median.
04:28We're going to take all of these numbers.
04:31And what it's going to do is it's going to select the middle point.
04:34So our middle point of our data is actually $82,500.
04:38So if we take our data, we're going to go to data.
04:41We're going to add a filter.
04:43And we're going to do smallest to largest.
04:46Don't worry about these numbers.
04:47I hard-coded those in as L2 and 3 and L4 to 11.
04:51Just ignore that for right now.
04:53But as we go down, we have 10 rows of data.
04:56So it's going to take a look, and it's going to go 1, 2, 3, 4, 5.
05:00And then in between 5 and 6, it's going to say, OK, I need to take the average of these
05:05two numbers, and that's going to be our median.
05:08This number for this data is actually much more representative of what salaries are being
05:13made throughout our raw data, whereas averages can be skewed quite easily.
05:19For example, if I do, this person makes $11 million.
05:22He's the CEO of Starbucks or something, well, we're going to say, OK, the averages for our
05:28raw data is $1.2 million, but nobody actually makes $1.2 million.
05:35But you'll notice that the median stayed the same.
05:37The median didn't change just because of this one outlier right here.
05:41And so oftentimes what I see is a lot of people will use averages to kind of show and prove
05:46something, when if you look at the median, it's a vastly different number.
05:50And so that was a few different things we covered in this lesson on how you can manipulate
05:55and change data, sometimes accidentally, if I'm being honest.
05:57You know, you make these assumptions and you just go for it with these statistical concepts
06:02that you're pretty comfortable with.
06:03You're like, I know what an average is.
06:04You may be misleading or lying about what the underlying data is actually saying.
06:08And so I don't want you to make that mistake because it can be someone embarrassing, especially
06:12to someone who's more senior, who may understand this a lot better.
06:16So don't make this mistake.
06:17Don't be that guy who's averaging averages or using averages where they shouldn't be when
06:22they really, what they need to do is look at the underlying data.
06:25Maybe look at something like median, where it may be more accurate for that use case.
06:28I really do hope that this was helpful.
06:30If you liked this video, be sure to like and subscribe, and I'll see you in the next video.
Comments