Hello, this is Dr. Grande. Welcome to my video on using the MCMAR test using SPSS. The Magnimar test is a non-parametric test and it's often used in pre-est post- test research designs. The variables have to be paired or matched data to use a magnimar test and they also have to be dichomous. So in this data set I have here which is fictitious you can see I have 90 participants in this study and the participants are categorized in the pre-EST as either using substances or not. So 70 have are in the substance use category and 20 are not. And then say there's a treatment that takes place that's targeted towards substance use. And for the post test, you want to know again if there's a substance use situation still occurring or if the participant no longer uses substances. And as you can see just by looking at this data without even analyzing it that there are certain participants that moved from substance use to no substance use. And down here a few that went from no substance use to substance use. So again what we're looking for here is there a statistically significant difference between the pre-est and post- test results. So I mentioned one of the assumptions for Magnamar is that the data is paired or matched. Also the frequency the minimum frequency that's preferred for this test for either of the categories in either variable is five. You want five cases minimum. And you can see here there's five cases of at least there's more than that of substance use and no substance use in both of these variables. The MCMAR test will still work if you have less than five for one of the levels of the variables but the result is conservative as compared to a Magnimar test where you have five or more as a minimum frequency. The participants have to be randomly sampled. The groups must be mutually exclusive. And as I mentioned, this test accepts dichomous variables. And both of these are dichomous variables. So there are a couple ways to run the MACMAR test. I'm going to show you both. I'm going to take a look at the one under analyze descriptive statistics first. So it's analyze descriptive statistics then cross tabs and typically the pre-est is in the row and the post test is in the column. I find a little easier to track that way looking at the table, but the results will not be affected if you have those reversed. For statistics, I'm going to check off Magnimar. If you're interested in magnitude and odds ratio, you can also check off risk. Click continue. For cells, I'm going to add the percentages for the row and column. Format is going to remain the same and style I'm just going to leave as default and I'm not going to check off display clustered bar charts or suppressed tables. So the dialogue as it is now I'm going to click okay take a look at the results. So you can see we had uh no missing cases here under the case processing summary. So you can see in the cross tabs for the pre-est substance use the total was 70 which I mentioned and no substance use it was 20 right so there's 20 that were in the no substance use level and 70 that were in the pre-EST looking at the post test though you can see there's only 55 in the substance use and 15 in the no substance use. So most of the participants who are categorized as substance use remain that way but 15 from substance use to no substance use in the post test. And then looking at the no substance use of the pre-EST as I mentioned it's 20 for the post test two went to substance use and 18 remained as no substance use. So if you look at the final count for post test, you have 57 in the substance use category and 33 in the no substance use category. So you can see from the results of the Magnimar test, we have a significant finding. 02 is less than 05. So this would be statistically significant. We would say there's a statistically significant result here because the MACDAR test indicated the significance here indicated as .002 is less than .005. And then as we move down to risk estimate, we can see the odds ratio for pretest is 33. This gives you an idea of the magnitude. As I mentioned, there's another way to run this analysis. So, going back and go back to the main data view. Go to analyze non-parametric tests legacy dialogues and then two related samples. And you can see I'll reset this. This is how it looks by default. So I'm going to move the pre-est to variable one and the post test to variable two. And then I'm going to uncheck will coxin and check magnamar. For options I'm going to leave everything the same and click okay. And you can see the results here. There's not as much output. You have the same numbers here, the 55, 15, the two, and 18 without all the percentages that you had here. So, this calculated the percentages. I checked it off. So, it would do that. You also don't have the totals for row and column here. And under test statistics, you end up with the same value, which is the 02. So, in this case, again, we have a statistically significant difference. we reject the null hypothesis and we would say there is a difference between the pre-est and post- test results. I hope you found this video on using the Magnamar test in SPSS to be helpful. As always, if you have any questions or concerns, feel free to contact me and I'll be happy to assist you.