Voice 023 (mp3cut.net)

The speaker mentions that they couldn't turn something on. They talk about checking the computer for red flags and finding that everything looks good. They mention a modeling window and the need to identify a dependent variable. They also discuss different types of variables and how to switch from continuous to Bernoulli. They talk about different distributions and explain the null model and how to calculate R-squared. They mention a project involving two-level HLM and discuss the goals and variables involved. They provide instructions for creating MDM files and running analysis. By the way, me from now on couldn't turn it on. If you want to hold it up, I'm going to throw it around. It's awesome. Obviously, we're kind of like... Yay. Yay. Yay. Okay. Back to what we were doing before the break. We were just checking the computer. To find out if there are any red flags. So look. The student file looks good. Level 2, variables. I mean, they are all indices. I think the maximum field is 1 to 5. The maximum field is 1 to 5. So nothing beyond the red. So it looks over 5. But we can close this. In case you still need this super statistic. It is actually under checked. You will get the same result. And now, done. Now we have our nbn file. So here is the modeling window. We don't have anything yet. Because we haven't started to build our model. I think today... I'm going to show you the null model. Then you... We may not have time to actually run the null model. Because I want to move to project 2. But I will build my model. You can run it in your leisure time. And see if you can make sense of the result. So basically we do the same thing. But first we need to identify the dependent variable. So if you move your cursor. Let's just use drug abuse. So you will click. I mean, there is only one option. Because we have nothing. So the program assumes that the first thing you want to do is to identify the outcome. So we take that. If you are running a 2-level HLM. Regular 2-level HLM. This is what? This is good. Very good. But... Here we have to do something. Because we are dealing with a dependent variable. But now it is still treated as a continuous. We need to make some changes. If you go with me to basic settings. If you pull down the basic settings menu. What do you see there? The default we have been using is normal. Or continuous. But now we are dealing with 0.1. It is called Bernoulli. Bernoulli. So what we need to do is to switch from continuous to Bernoulli. Click on Bernoulli. And you don't need to do anything else. But here, let me just, by the way, show you the capacity of HLM. Because some of you, if you come from public health, medicine, or whatever. You may need other distributional properties. For example, you can have Poisson distribution. I'm not going to get into detail. But just showing you. Poisson basically is for counting variables. You do not treat counting variables as continuous. Some people do. But seriously speaking, that's wrong. Counting is not continuous. Frequency. No, no, no, not number of events. Frequency. How many times it has happened. Frequency. That's counting. So that's Bernoulli. And then you also have binomial. Many of you probably have heard of that. And then Poisson. Binomial as well. And then see the other panel. You've got multinomial. And then you have ordinal. Multinomial is actually how you run categorical data analysis. You see over there, you will tell how many categories you have. But we're not going to use it. If you choose multinomial, you have to specify how many categories you have. Ordinal. Do you know ordinal scale? I think we do. You have a nominal, ordinal, interval, and ratio. If your dependent variable is nominal, meaning it's just an order of things, you need to use nominal. These are the capacities you may need to use. I think after this course, if you struggle a little, you can learn it by yourself. As I said, categorical data analysis is a lot more complex than Bernoulli. But we are going to use Bernoulli for our purposes. Then you'll see what happens to the equation. Yeah, see? Different now. See the probability statement? The probability formula? This is the null multilevel matrix. This is the null model in a matrix. I'm just going to... I will just go as far as this. If we have time left today, we will come back to this. You can save this, and later we will run this just for curiosity. You check out the results. Next week, we will pay full attention to it. Yeah, just save that. Save this. Maybe I should show it. High school. Sorry, high school. Health behavior. HPLC. Zero. So I know this is the normal one. Yeah. And then you can run. And then you will have the result. But we will also come back to this. Okay. So now let's move on to Project 2. So let's take a look first. Project 2. I have a description of the project. So this is, again, real data coming from my postdoc project. It's a voluntary elementary school study. Basically, we are going to do a two-level HLM. But now, remember, the dependent variable is continuous. Because it is achievement. Achievement is usually considered continuous. So, basically, the goal is to find out... to find out... what student-loved ones who are unresponsible for academic achievement. That's question A. Question B, or Purpose B. Purpose B deals with segments and slopes. So Purpose A actually deals with the intersects. Purpose B deals with the slopes. In terms of the variables, I give it to you. There are four learning outcomes. What happens if you take one, but there are hundreds of them. Here's the scale, mean 50, standard deviation 10. Then there are five student-level variables, SES. Mean 5, standard deviation 1, number of siblings. Then there are three dummy variables. I think gender, native status, and number of parents. Then there are also five variables at the school level. We've got school size, and then there are other school-level variables in that. So our goal is to take one learning outcome of the different variables, and then, given the limitation of the student version, we can only use four variables at each level. So we're going to do four at each level. And then two texts. Here I said enter all your independent variables, but again, I mean, you know, you center your variable or not depends on if zero makes sense. So use that principle to decide if you want to center a certain variable. Like, you know, for dummy variables, usually we do not center them because zero is a valid group. Okay? Ma, create MDM, of course. Then I said, this is the five, and then run a two-level HR model that treats all student-level variables with fixed effects. The way I'm doing the theory, right? With a focus only on the intercept. And then finalize the harmonious model. Meaning, we just keep the second generation. Okay? That answers these five questions. Now, questions one to two, where do you get the information? Sigma squared. Sigma squared and how, from which stage of the model building process? Huh? The null model. Exactly. Okay, remember? Yeah. From the null model, we talk about this thing, level one errors, level two errors, and then the contribution, you know, from each level to a, you know, to the variation. In outcome, then you calculate the IPC over from the null model. Okay? So, three and a four from each model. Of course, the final four models, right? Because we need to intercept the second generation variables at both levels. Okay? And then, do you still recall how to calculate R squared? R squared? What's the general formula? No minus four divided by no. Exactly. Very good. You still remember that. Very good. No minus four divided by no. So, you can apply this formula to level one errors, level two errors, and also, if you calculate the total errors, you can also create a total overall error. Okay? So you apply the same formula to the total errors. That will give you overall. That's what I mean. Okay? Overall. And then, I said, just to find and run another example of a model that treats all student-level variables for random effect. So here, you can do this. Right? Yeah. So, and then still, you get a harmonious model, and then answer these two questions. Basically, these two questions are related to the flow. How do you know I'm talking about flow? How can you be sure I'm talking about flow? What word gives it up? No, not give it up. Give it away. Yeah, relationship. Yeah, relationship. Yeah, that's very good. The slope is relationship, remember? Yeah. Okay. Yeah. So the slope, you know, provides hint. Okay? So here, we are talking about slopes. Okay? You know, relationship provides hint. Good. Okay. So, and then this is the slope-flow interaction again, right? The last question. See if you can find more variables than the slopes. Okay? Yeah. So these are the tasks. Now, let's start from making MDM files, right? So, again, we go back to HRM. Well, no, remember that the HRM allows us to do this, right? Make MDM and then stack packaging plus. Two-level model, yes. And then a cross-sectional person type with moving groups, yes. Then, Browse Level 1. Now, you have to go to the Project 2 folder. Okay. Go to Project 2 folder and take student. Are we good? Good, okay, yeah. Student. For the student. Okay. Yeah. So, Browse Level 1, that's student. The file is called student. And then, choose variables, okay? So, the first one is school ID, and obviously the ID. And then, you see, you have four achievement measures. You've got math, reading, science, writing. And then, you have these demographic variables, sex, and SPS, native status, number of parents, and then number of students. Okay. Okay. And yes, we have missing data. And then again, we want to use pairwise division. So, it is running analysis. Okay. Running analysis. Then, browse for Level 2 file, which is called school. Okay. Choose variable, ID. School ID is the ID. Then, you've got school size, school name SPS, disciplinary climate, academic pressure, and the parental involvement. So, these are all in the two. Okay. Okay. Then, you know, we can give it a name. That's fine. Yeah. Or I suppose, you know, NB. The wrong way. About NBM. Okay. Yeah. You can give it a name. And then we can also save this window into another MDMP file. Okay. So, we can give it the same name, NB. And save. Cool. Everything's ready. And we can make MDMP file. Cool. Oh, well, this message is not material. What it tells you that the thing is some schools just simply did not return the school questionnaire. Okay. So, we don't have school data. But we, but the students in those schools tested. Okay. But the schools did not return the questionnaire. So, we didn't have level two file for those schools. So, in that case, I mean, the student has to be deleted. Okay. Yeah. Will that impact your analysis? Not really. But, if you have centered your variable before making MDMP file, this deletion will ruin your centering. You see, that's why HLM provides you with the function doing centering within HLM. Okay. So, that's coming in. Okay. Yeah. So, here, so here, basically, you don't have to worry about messages like this. It's just, you know, we will need to use the HLM centering function to center this image. Okay. Again, check maximum for red flags on data. There isn't any actually. There isn't any at any level. Okay. It won't make sense if the variable isn't there. So, that's the end of our talk. Okay. We close, and then done. So, basically, this is the modeling phase again. So, let's take math as outcome. Okay. Left click, math. So, again, what do you call this model? The model. Right. Yeah. Very good. So, probably we can save it like HB zero. Okay. We can save it. Any other names? Now we run. Okay. Remind me again what are we going to get from the normal load? See? What do you think of the square? I don't know why but it's out. Number two variable. Okay. So, you will see them again in the last table. Okay. Because that's the final instant of variance. So, if you add up these two numbers, what do you get? The number one variable and the number two variable. If you add up sigma squared, the tau, what do you get? Huh? You add up the variance component what do you get? Total variance. Right? So, here you actually have a variance partition. You partition the total variance into level one, sigma squared, and level two. Okay? Let's see specifically what question we need to answer here. What is the proportion of variance in your learning outcome that is attributable to both tau and sigma squared? Oops. How do you calculate the proportion? How do you calculate the proportion? What is the proportion attributable, the proportion of variance attributable to, to, uh, two? Tau divided by tau plus sigma squared? Yes. Very good. Sigma squared over sigma squared plus tau. Right? Yeah? So, you have actually ninety point three seven seventy three and then ninety point three seven seventy three plus what is tau? Tau is eleven point zero three three nine four. Okay? I will leave the, uh, the rest of my tape to you. What about, uh, level two? What is the proportion of variance attributable to level two? Huh? Tau over sigma squared. Oh, yes. Tau over sigma squared plus tau. Exactly. Do you think you can get it? What is the IPC? This one, right? That's the IPC. See? We answered the first two questions. Right? From here, we need to build, well, we can get extra information, but for the purpose of this project, I mean, that's all. So we answer the first two questions. And then we are going to, uh, um, run a theoretical approach first. Alright? So, level one, let's just use, let's see, we can use sex, uh, dummy variables so we don't have to center it. Regular. FPS? Yes. That's a very difficult variable only for random centering. Right? Uh, we can do native. Native is zero one. Okay? Dummy. So we don't need to center. Let's see, uh, uh, yes, because FPS has artificial grounding. Grounding. What is the first variable of this? What is sex? Sex. And what's sex? Native. Native, yeah. And, uh, maybe we use the last one, sibling. Okay? Yeah. Sibling, I mean, zero just means no sibling. Okay. And native would be the grounded? No. Uncentered. Oh, uncentered. Yeah. You see, from the, uh, the appearance, you'll know. Remember the regular appearance is uncentered. Yeah. What was the last variable? Sibling. Sibling. No, no, no. No. Only FPS is random centered. Okay. Because the other three are either dummy variables or zero. We only have zero. Are we good so far? So then, we need to run because you may need backward division. Right? Yeah. So we can, if you want to save it, you can do so, but, uh, I'm going to run it down. Okay. How long does it take? Okay. So the deletion would start from, uh, the variable with the largest non-cyclical value. Right? So, obviously, it is the sibling. This variable needs to be deleted. Right? Go back. Uh, delete sibling. Okay. If you click sibling again, that's the only option. Deletion is the only option. Okay. So then, we run it again. Run the model down. Ah. Now, it's okay. Okay. So this is our level one model. Okay. Yeah. This is our level one model. Then, uh, uh, level two, of course, um, we model only the random intercept. So again, this model is called random intercept. Okay. Yeah. So level two, that's, that's the, uh, um, oh, uh, that's just useless. Uh, five, um, that's grand mean centric, okay, because zero doesn't make any sense. Right? Two mean SES, grand mean center, uh, disciplinary climate, grand mean center, and, uh, academic pressure, also grand mean center. Because these are independent. Okay. It doesn't mean I have zero. On the scale, because it's from one to five. Okay. Yeah. So they are all grand mean centers, the four variables. So let's run again to see if we need to, uh, again, do backward deletion. Oh, plenty. So first, size needs to be deleted. Okay. Size is the p-value. So we will remove size. Make sure the yellow bar is on beta node. Okay. Run it again. What next? Academic pressure. Right? Academic pressure. Delete. Run it again. Are you all following? Yes. Next one, mean FPS. Right? Make sure you use the one with your box, the standard pair. Okay. So mean FPS also needs to be deleted. Okay. And we run it again. Ta-da! This is our final model. Right? It's a parsimonious because everything is the same, isn't it? Yes. Were we supposed to delete the academic pressure or the... Yes. Academic pressure is the final model. Okay. So we need to delete the academic pressure the second to go. So why is it in the final model? Where? Yeah. In the second? Oh, see that? Yeah, I think you actually deleted it. Oh, I deleted the wrong one. Sorry. Okay. Okay. I know what you mean. The model looks like this, right? Right? Yeah. And academic pressure. Okay. Academic... Yes, yes. That academic pressure. Okay. You know what? Let's do it again. Okay? Yeah, yeah, yeah. Let's do it again. Right? What the heck? But thank you for pointing that out. So, yes. First one, size. Remove size. Okay, run it again. What's next? Academic pressure. Oh, my gosh. Academic pressure. Yes. Run it again. Yeah. Yeah. Yes. Yes. Okay. Remove mean as yet. Only disciplinary climate remains. Let's see. Okay. This is the final. Ta-da. Okay. This is our final model. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay. Okay.