Data:
When you google "General Social Survey Data 2006", you will find many links connecting GSS-2006 data. For example, you can use the following link:
https://www.thearda.com/Archive/Files/Descriptions/GSS2006.asp
Then you can download the dataset, say for example, from the section called "Microsoft Excel File", and save it in your computer. Of course you can rename your data file to, say for example, "gss06".
You can download the codebook for all the variables in the survey as well.
Of course you will only use a few variables (columns) of the data for the final exam. So you can delete as many variables (columns) as you like from the original data file. This will save your computer memory when you analyze the data in R or SAS.
Problem 1:
Please work on this problem using SAS.
Import the data into SAS (referred as gss06), and report the number of subjects and number of variables.
Create a dataset (referred as gss06_sub) including only the following 8 variables, AGEWED, AGEKDBRN, WRKSTAT, AGE, CHILDS, EDUC, SEX, and RACE. Read the codebook for the details of these 8 variables.
Print the first 20 observations of the above data named gss06_sub.
Descriptive analysis of two variables "AGEWED" and "AGEKDBRN". For each variable, describe the number of observations, number of missing values (treat "Don't know" and "No answer" as missing), mean and standard deviation, and five-number summary.
Create two datasets; one dataset named "gss1" excluding those subjects who had missing value on variable "AGEWED", and the other dataset named "gss2" excluding those subjects who had missing value of variable "AGEKDBRN". Export these two datasets as CSV data files to be used in R.
Problem 2:
Please work on this problem using R.
Using R to calculate a number based on your initial. For example, the instructor's name is Yixin Fang, so his initial is "yf". The location of "y" in the letters is 25 and the location of "f" in the letters is 6. Then his number is 256, which is equal to 25 × 10 + 6. Hint: You may want to use built-in R object "letters" and R functions "strsplit" and "which".
Use the number you just calculate as the seed to generate a random number between 0 and 1. If the random number is less than 0.5, then you will use the dataset named "gss1" for the following problems. If the random number is bigger than 0.5, then you will use the dataset named "gss2" for the following problems. Name the selected dataset as "mygss". Hint: Use your seed number to generate random numbers so that the grader can reproduce your results.
In your dataset named "mygss", there are 8 variables, but you will only use 7 variables of them. For example, if your dataset is gss1 originally, then the dependent variable is "AGEWED", and then you can ignore variable "AGEKDBRN" and consider the remaining variables (WRKSTAT, AGE, CHILDS, EDUC, SEX, and RACE) as independent variable. Of course, if your dataset is gss2 originally, then the dependent variable is "AGEKDBRN", and then you can ignore variable "AGEWED". Use R to describe and summarize those 7 variables in your dataset named "mygss".
Randomly divide your dataset "mygss" into two halves. In order to this, use R function "sample" to randomly sample m subjects from n subjects without replacement. Here n is the sample size of your dataset "mygss" and m is the largest integer less than n/2. Name the dataset consisting of these m subjects as "mygss_train" and name the data consisting of the remaining subjects as "mygss_test".
Problem 3:
Please work on this problem using SAS.
Use dataset "gss1" to test if "AGEWED" is marginally associated with "WRKSTAT", "AGE", "CHILDS", "EDUC", "SEX", and "RACE", respectively. If one categorical variable has too many categories, you can decide whether or not to dichotomize it.
Use dataset "gss2" to test if "AGEKDBRN" is marginally associated with "WRKSTAT", "AGE", "CHILDS", "EDUC", "SEX", and "RACE", respectively. If one categorical variable has too many categories, you can decide whether or not to dichotomize it.
Summarize and interpret statistical findings you obtained from the above bivariate tests.
Problem 4:
Please work on this problem using R.
Fit a linear regression model using the dataset named "mygss_train", with "AGEWED" or "AGEKDBRN" as dependent variable, and the other 6 variables as independent variables.
Summarize and interpret statistical findings you obtain from the above regression analysis. Report the statistic called "adjusted R-square".
Identify those independent variables that are significantly associated with the dependent variable under significance level of 0.05.
Fit a linear regression model using the dataset named "mygss", with "AGEWED" or "AGEKDBRN" as dependent variable, and but considering only those independent variables that are identified as significant in the preceding step.
Summarize and interpret statistical findings you obtain from the above regression analysis. Report the statistic called "adjusted R-square".