Please first define "prevalence of obesity".
i am a newbie and i dont know how to calculate prevalance of obesity in spss using bmi values
Please first define "prevalence of obesity".
obesity is defined as bmi greater than 30 and ,prevalance of obesity means = those cases having bmi greater than 30 * 100 \ total number of cases .... my variable is bmi , and suppose i have 10 cases , out of which 2 are having bmi greater than 30 then my prevalance of obesity will be 20% or 20
Ok, so you need to convert your variables to zeros and ones. Select a column in SPSS. Type your converted bmi values in the rows of that columns. For conversion, simply put 1 instead of bmi values greater than 30, and put 0 instead of the rest. Now in the SPSS toolbars, find and run "Analyze -> Descriptive Statistics -> Frequencies". Now select your column including the zero and one values, and hit OK. It will give you the prevalence of bmi > 30.
In the data view. Just select a column to be used later. For example the first column from the left. Then read your first BMI score. For example, it is =32. Then write down "1" in the first cell of that column. Then read the second BMI. It for example turns out to be = 29.7. Then type 0 in the second cell (I mean the second row in that column). This way, you can fill that column with zeros and ones. Then proceed with calculating the prevalence as mentioned above.
arkabbot (09-06-2012)
ya but i need bmi values as they are coz later i have to corelate them with caloric intake and physical activity (my other variables )..
Keep BMI variable, just create a new variable: > 30 = 1 or </=30 = 0, then run frequencies.
Now you have BMI and new variable Obesity.
arkabbot (09-06-2012)
lol, that is not a problem at all. You don't need to delete them before adding new columns (variables) to your SPSS file. SPSS or all other software allows you to have as many columns as you want. You can put the new values into another column. Or you can put the new values in a new SPSS file, as you are not obliged to necessarily use the same SPSS file your BMI values are stored in. So you can do whatever you like.
thank you , i have two more variables , caloric dietry intake in kcal and caloric consumption due to physical activity in kcal that i want to corelate with my bmi variable , meaning i want to analyze data in such a way so that it proves that caloric dietry intake increase causes greater bmi and increased consumption due to physical activity causes less bmi , how should i proceed ??
Are these variables available for the same individuals, how many individuals do you have? You could possibly run a multivariate linear regression model with BMI (contiuous) as the dependent variable and your two listed variables as independent. You may want to also take a look at the correlation between the two independent variables.
yes these are for the same individuals , i have 400 individuals with bmi , caloric intake , caloric consumption , age , occupation as my variables...can u please explain a little about multivariate regression model..
Then as hlsmith mentioned, you can run a linear regression from "Analyze -> Regression -> Linear".
Then select the variable BMI as your dependent variable, and select the other ones as independent variables. Then hit OK. There are some other ways too but I don't want to make it look confusing.
This way, you can check the effect of each variable when the effect of other variables is controlled for. Its interpretation needs a little search and reading. Read here for any details you need.
The output (or results) for the analyses should provide estimates (beta coefficients) for each of the variables when controlling for the others. You would expect that the caloric dietary to be a positive number (meaning that BMI increases when it increases) and the physical activity to be a negative number (meaning when it increases BMI decreases) - given your referenced assumption.
You should look for the columns with header "beta", and the P values ("sig"). I don't know if you are familiar with correlation coefficients, or not. A correlation coefficient says that when we increase the independent variable by one unit, to which extent the dependent variable changes? The beta is a specific correlation coefficient which shows you the extent of correlation in the absence of the effect of other variables. But I didn't talk about these first, since I think these are information readily available on the net.
Tweet |