Chi-square - Determining Probability Values
Deana Namuth-Covert
Author
01/13/2019
Added
234049
Plays
Description
Video tutorial showing how to read a chi-square distribution table to determine probability values that experimental data supports a tomato breeding hypothesis.
Searchable Transcript
Toggle between list and paragraph view.
- [00:00:00.260]Hi, this is Dr. Deana.
- [00:00:09.300]In this video clip, we'll be taking our chi-squared value that we calculated in the previous
- [00:00:14.840]clip and looking at how we interpret that number.
- [00:00:18.460]Does it support our hypothesis or should we reject it?
- [00:00:22.560]So did we see differences in data which were too great, so we should reject our hypothesis?
- [00:00:29.280]Or can we accept it?
- [00:00:31.240]So in our case that we're looking at with our tomato breeding experiment, we had calculated
- [00:00:36.880]a chi-score value of 1.2335.
- [00:00:40.880]So what do we do with that number?
- [00:00:42.280]We consult a chi-score distribution table, such as one like this.
- [00:00:47.880]So we'll see here on the left, this first column, there's a category called degrees
- [00:00:53.260]of freedom.
- [00:00:55.520]Now what degrees of freedom are, there's a simple formula.
- [00:00:59.060]It's just the number of classes in your experiment minus one.
- [00:01:04.600]Alright, so let's look at our particular example, and we have two different classes.
- [00:01:11.940]We have resistant tomato plants, and we have tomato plants which are susceptible to bacterial
- [00:01:17.780]spot disease.
- [00:01:19.640]So our degrees of freedom is n minus one, so in our case it would be two minus one.
- [00:01:28.840]Which is one.
- [00:01:32.980]So in our case we have one degree of freedom.
- [00:01:35.640]Now let's go back to our chi-square table.
- [00:01:40.400]Here let's write down our numbers which are key.
- [00:01:44.080]So we have one degree of freedom, and our chi-square is 1.2335.
- [00:01:53.780]Okay, so the first thing you do is go to the degrees of freedom
- [00:01:58.620]column, and read down until you find your degrees of freedom for
- [00:02:03.700]your particular experiment.
- [00:02:05.980]So we don't have to go far, we just go to one.
- [00:02:08.100]And then you're gonna read over to the right until you find your chi-square
- [00:02:13.300]value, or where it would fall.
- [00:02:15.100]So I'm looking for 1.2335.
- [00:02:18.880]All right, it's a little bit bigger than .455, but it's less than 1.32.
- [00:02:24.360]So it would fall somewhere here [scribbles between the .455 and 1.32 values].
- [00:02:26.020]And then what you do is you read up
- [00:02:28.400]to this other set of numbers [at the top of the table], and these set of numbers are the probability values.
- [00:02:36.580]Okay, so you may see it as p equals something.
- [00:02:40.440]So it is the probability.
- [00:02:41.760]Okay, so what does that mean?
- [00:02:44.680]What that means is if we were to repeat our experiment,
- [00:02:47.940]what's the probability that we would see a larger chi-squared value?
- [00:02:54.120]And remember, if you have a larger chi-squared, that means your data from
- [00:02:58.180]your experiment deviated even further from what you expected
- [00:03:02.080]than your current experiment, okay so what's the probability that if we were to repeat
- [00:03:06.800]our experiment that we would get even greater deviations and yet be able to still accept
- [00:03:11.800]our hypothesis?
- [00:03:13.680]So in our case we're going to have a probability value that is greater than 0.25 but it's less
- [00:03:26.200]than 0.5 by consulting this chart.
- [00:03:32.780]So somewhere between 25 and 50 percent of the time if we were to repeat the experiment
- [00:03:37.960]we would see data similar to what we had in this particular experiment or even deviating
- [00:03:43.180]greater from what we expected and still we'd be able to accept our hypothesis.
- [00:03:48.920]So what is the cutoff value?
- [00:03:51.440]Well it turns out scientists have determined that this [circles 0.05] is
- [00:03:56.100]the cutoff value.
- [00:03:56.100]It's a critical value in most cases.
- [00:03:59.700]So they're willing to accept a .05 value, meaning that 5% of the time if they were to
- [00:04:05.980]repeat their experiment they would have greater deviations than what they saw and still the
- [00:04:11.500]hypothesis would be correct.
- [00:04:14.120]So sometimes that's a little bit hard to understand, but for our purposes just know .05 is our
- [00:04:20.480]critical value.
- [00:04:22.140]So if you have a chi-squared that falls to the right on this
- [00:04:26.000]table [p-value is less than 0.05], what you're going to do is reject your hypothesis.
- [00:04:31.400]You're going to go back and determine, you know, maybe there's some other kind of genetic
- [00:04:36.440]inheritance pattern that's going on, for one example.
- [00:04:41.200]So any chi-squared values that fall to the left of this .05 probability value [they are greater than .05], then we
- [00:04:48.060]can accept our hypothesis and we know then that our data from our experiment actually
- [00:04:55.900]supports what we were hypothesizing.
- [00:05:00.360]Now it's important that students sometimes think that this proves your hypothesis is
- [00:05:04.120]true, but it doesn't, it just gives you the probability that your data supports it.
- [00:05:09.080]And the only way you can really prove an idea is to do repeated experiments and some real
- [00:05:13.920]thorough investigation.
- [00:05:16.100]So just make sure you know that chi-square doesn't prove anything is true, but it does
- [00:05:21.720]give you a probability value that your data actually supports your hypothesis.
- [00:05:25.800]Alright, so this is how you can solve a chi-square distribution table and being able to interpret
- [00:05:32.940]your chi-squared calculation results and it will help you know what the significance is
- [00:05:38.260]of your data.
- [00:05:39.680]Are any differences that you saw statistically significant so that you should reject your
- [00:05:45.240]hypothesis so they fall to this side of the chart [p-value less than 0.05]?
- [00:05:48.880]Or does your data support your hypothesis?
- [00:05:51.920]So that would be over here on the left side [p-value greater than 0.05].
- [00:05:55.700][music]
- [00:05:56.700][music]
- [00:05:57.700][music]
- [00:05:58.700][music]
The screen size you are trying to search captions on is too small!
You can always jump over to MediaHub and check it out there.
Log in to post comments
Embed
Copy the following code into your page
HTML
<div style="padding-top: 56.25%; overflow: hidden; position:relative; -webkit-box-flex: 1; flex-grow: 1;"> <iframe style="bottom: 0; left: 0; position: absolute; right: 0; top: 0; border: 0; height: 100%; width: 100%;" src="https://mediahub.unl.edu/media/10487?format=iframe&autoplay=0" title="Video Player: Chi-square - Determining Probability Values" allowfullscreen ></iframe> </div>
Comments
0 Comments