About this video
The video works through two pairs of scenarios to name the source of variability. Intentionally changing a condition (room temperature) creates induced variability, repeated readings of the same thing show measurement variability, differences between individuals show natural variability, and different random samples from one population show sampling variability.
You'll be able to
- describe induced, measurement, natural and sampling variability.
- identify which type of variability a scenario shows.
- explain why intentionally changing a condition creates induced variability.
Key words
- induced variability
- Variability created when you intentionally change one variable to see if it changes another.
- measurement variability
- Differences in readings caused by imperfections in the measuring instrument or process.
- natural variability
- Differences that exist naturally between individuals, like trees of different sizes.
- sampling variability
- Differences in a statistic between samples, caused by chance in which individuals are picked.
- parameter
- A number describing a whole population rather than a sample.
Transcript
- [Lecturer] So we're told in scenario A, U'ilani wants to test whether room temperature affects reaction time. One session, U'ilani sets the room to a cold temperature. Another session, U'ilani sets the room to a warm temperature. U'ilani compares the average reaction times from each session.
So I guess they're seeing whether if someone's in a hot room or a cold room, maybe they drop something and they see how fast they can react to it. They don't give us a lot of details on the experiment, but that's interesting. Scenario B, a nurse measures the same patient's blood pressure twice using an automatic cuff, but the two readings are slightly different.
Which scenario could show induced variability? Whoops, so pause a video, have a go at that, and then we'll do it together. Here are the four choices. So let's just remind ourselves what induced variability is and actually all the different types of variability.
Induced variability is where you intentionally change some variable to see if it changes some other variable, if you can induce that variability. Now, scenario A is clearly that. U'ilani is changing room temperature and seeing whether that induces a change of a variability in reaction times.
So scenario A is definitely induced variability. We don't know if it actually did induce the change, so that's why it could show induced variability. Maybe when they measure it, they don't see the change, but it could show induced variability. Scenario B, the nurse isn't changing, isn't doing something to induce different blood pressure readings or potentially induce different blood pressure readings either time.
The readings being slightly different, this is just measurement variability that there's some imperfections in the measurement instrument, et cetera, et cetera. Also, someone's blood pressure does change. It's not always exactly the same every nanosecond of every moment.
But I would call this measurement variability. So I would say scenario A, scenario A because U'ilani intentionally changes a condition to create difference in results. Yep, that's what we said. Let's do another example here. So in scenario A, we're told a forester measures the trunk circumferences of mature oak trees in the same forest.
The circumferences vary from tree to tree. In scenario B, a manufacturer has an inventory of identical bolts. Two quality inspectors each randomly select 25 bolts and record the average length of their sample. The averages are different. Which scenario shows sampling variability?
So once again, pause the video, have a go at that. Here are the choices. Well, scenario A, we're not sampling. We're actually, I'm assuming the way it's phrased, measuring the circumstances of all the mature oak trees in the same forest. So, we're actually going for a parameter perhaps here.
And so this is just natural variability. Yes, circumferences will vary from tree to tree. For sampling variability, you're taking different samples from the same population, but just by chance of which sample you happen to pick, you're getting some variability in the statistics that you're calculating based on that sample.
And that looks a little bit closer to here, a manufacturer has an inventory of identical bolts. Although if they were truly identical in every way, then you actually wouldn't see a difference in samples. So maybe they're just using identical and like they're intended to be identical, but the fact that with 25 bolts, you're getting different averages in length of the sample, I would say that is sampling variability, assuming that, yeah, there's just slightly different...
It would be measuring variability if this was truly, if they were truly identical bolt. But since they're asking about sampling variability, I think that's what the intent of this question is. So it's scenario B, because different random samples from the same inventory can have different average lengths.
Yes, and once again, that's only true if they aren't completely identical. There are slight variations in their lengths. Scenario B, because the inspectors intentionally change a condition. No, that would be induced variability.
Video © Khan Academy, used under CC BY-NC-SA. Watch on Khan Academy. Puzzicle is not affiliated with Khan Academy.