Hello everyone. My name is John MCI and in this short video, we're going to look at one proportion test. We're going to use MiniTab for this, which we will access through the App Store ed dot ac, dot UK system, which gives you access to all the software you would have access to, if you will, using a university machine, such as one of the ones in the library to log on to App Store, ed dot ac dot uk. You just need your university username, which is the same as you would for your email and your password. And then once you have logged in, you will have access to all of the various softwares that you would have access to on a university system. I'm going to use MiniTab 20 for this and click on open Minitab updates every year. So if you are watching on a slightly with a newer version of MiniTab available to you. Don't worry, they tend not to change very much from year to year. One-sample proportion test is a simple test which looks to see if a single sample is significantly different from a defined proportion. Easiest way to think of this is, is your sample different from, for example, a historical proportion that we know of in the literature. In the student research component, we often use the data from the student research component. Tutorials on the statistics App Store edge dot ac dot UK is a, is basically a virtual client which looks at your networked university drives to the drives are available to you whenever you access any university computer. If I go to File and I go to open because I have saved the SRC project, the SRC statistics files. In my OneDrive. I am able to access it from here and open it directly into MiniTab like so. This is not tidy data, it has multiple data tables in it. Columns one to seven, our survey columns 9 and 10 are mere falling data. And columns 12 and 13, flee burden data for cats. For this particular scenario, we're going to look only at column 2, which is the sex of participants on campus. If we scroll down and you'll see that we have 47 participants on campus here. And what we want to know is, is the proportion of males and females on this campus. In this particular sample, the survey, different from the known proportion in the literature. So if we know that historically 30 percent of the people on campus or male is our sample, which is these 47 people in the survey, different significantly from that 30 percent male historical average. There is an intermediary step for running a one proportion, a one-sample proportion test. And that is that we need to tally the data. We can ask Minitab to do this for us by going to Stat Tables, Tally Individual the variables. Only going to tally sex, we have the option to ask or camps presents, cumulative counts and cumulative percentages. I'm only going to ask for the currents in this instance, but you can play with these options if you would like mini tab to tell you percentages, for example. You can also ask for the results to be stored in a new worksheet. In this instance, I'm going to ask for the results to be outputted into the navigator tab. So I'm going to leave that box unchecked. But sometimes it can be useful to store the results in a new worksheet in your Minitab project. I'll click on Okay. And you can see here, if I pull the datasheet time that we have 38 females and nine males in this dataset. Niger on the one proportion test, we go again to start, we go to Basic Statistics and one proportion. We have this drop-down menu here which is highlighted in blue, were asks, do we have one or more samples each column, or do we have summarised data? We just use the tally function to summarise the data. And so we'll select the sunrise data from the drop-down. And now it asks us for the number of events, a number of trials that we have. Event some trials is sort of old biological terminology, thinking about a very standard medical trial, for example, the events are basically the thing that we're interested in. And the trials is the total. So in this case, we're interested in the number of men, which is nine. We can just see there. We know that we have nine mm and the number of trials or the number of things in total we were interested in is 47. The total count we want to perform a hypothesis test. So if I click on perform hypothesis tests, we want to know if the hypothesised proportion. Is 0.3. Remember, I was saying, we know that in the past 30% of campus participants were male. Or we want to know if this particular sample is statistically different from that previously known proportion. If I click on options in C, we have the option to change some of these, we can change the confidence level. 95 percent is very standard, so we'll leave that we can change the alternative hypothesis to a one-tailed direction test. But we just want to know if it is different which two-tailed test. So we'll leave on that. I will use the exact method because we don't have very large sample size. I'll click on, Okay, I'll click on Okay, again. And this is the Output tab on the left-hand side. You can see it tells us, first of all, what it's done is done test and confidence interval for one proportion. It tells us the event proportion is a small p and the exact method was used for this analysis. We get the descriptive statistics, the total n, it's 47. We saw nine events or nine males. And we see the sample P. The sample proportion is 0.19. So 19% of the sample were male as opposed to the 30 percent historically that we know are male. We also get the 95 percent confidence interval for this proportion alongside it, which runs from 0.09 to 0.33. So the 95 percent confidence interval ranges from 9% to 33 percent. That encompasses the 30 percent proportion that we were interested in historically. So looking at this confidence interval, I'm pretty confident that this is not going to be a significant result. That is to say that this 9.190 proportion is not significantly different from 0.3 because 0.3 is within the confidence interval for this proportion. As to say, we can reasonably say that we're confident that the actual value of the sample couldn't be 0.3. And if I scroll down, we can see that the test was or H null was that the proportion is going to equal 0.3. And the alternative hypothesis H1 is the proportional is going to equal 0 and not going to equal 0.3. That symbol means does not equal. And the p-value associated with this, note the difference between p as a proportion and the p-value is 0.13.113 is far above the critical threshold of 0.05, which is the arbitrary number we have decided, is significant. And so we would say this is not a significant result. We can report this in a Word document. Like so. Bring this over here. Might say the proportion of males in the sample was as 0.19. And then we can also produce the 95 percent confidence intervals, which are 0.090.33, which was not significantly different from the historical proportion of 0.3. We should also say any one proportion test. And p is equal to 9.113. Close brackets. Italicised p for reporting standards and get rid of that extra space. And that would be how we would report this test. Plain English alongside the confidence intervals, which help us to see that it is reasonable for this estimate to include 0.3. And the p-value just confirms that we do not think there was a significant difference between these two values. And that is a one proportion test in mini tab.