All work must be handed in as a "hard-copy". No e-mail, e-mail attachments, faxes, etc.
Make it useful to you: although answers are given at the end of the textbook, try to solve the problems on your own; the correct answer is useless if you do not know how to obtain it.
Make it easy to find: label your products with your name, the homework number and date, and page numbers.
Make it easy to read: type all text other than formulae and computations. Spell-check and proofread everything you hand in (note that these are different). Critically evaluate all of your output for correctness, completeness and clarity.
Make it easy to understand: explain your logic. Include computations: it is good preparation for tests, too. Have printouts at the end; state and interpret results referring to them. State conclusions where appropriate--don't leave the reader guessing.
Make it complete: the credit (max 2 points for each set) is given for the effort and for the attempt to tackle all questions, rather than for correct answers. State your findings, and where appropriate provide interpretations.
NOTE: the SPSS data sets to be used for homeworks are found on the n: drive, under n:\kaufman/UST601\. Begin your session by copying the datasets onto your h: drive. You can do this as many times as you need (for example, if you think you have altered your dataset, make a new copy for starting a new homework.)
Spreadsheet: Use Kachigan data. State your findings, and where appropriate provide interpretations.
Enter the data of problem 3.10 in the spreadsheet. Answer questions of 3.10 and 3.11. Same for problem 3.15.
Spreadsheet: create your own data. State your findings, and where appropriate provide interpretations.
Create a sample of at least 25 observations on a variable of your choice (go on the WEB for ideas!) Produce and graph the frequency distribution of the sample, and then compute mode, median, mean, variance, standard deviation.
Produce and graph the frequency distribution of number of years of education (variable EDUC) in the sample.
Examine the ATTEND variable (church attendance).
State your findings, and where appropriate provide interpretations.
A. What is the measurement level? Describe the sample (what measure of central tendency is appropriate?)
A. For the percentage of state population change from 1980-1986 (POPPC) find a. the range, b. the mode, c. the median, d. the mean and e. the standard deviation.
B. For the percentage of the state electorates casting votes in 1988 (TURN88) give a. the range, b. the mode, c. the median, d. the mean and e. the standard deviation and g. the Z score for a turnout of 43%.
A. Compute a confidence interval for the difference between the two means (for 1991 and 1996) of a quantitative variable of your choice. Does it seem that the two means are different? How can you tell?
B. For GSS91, Test at .05 whether respondents were drawn from a population whose mean preference for spending a social evening with friends (SOCFREND) is less than 3.50 on a 7-point scale.
C. For GSS91, selecting only women (SEX = 2) who worked part-time during the previous week (WRKSTAT = 2) test at .01 whether the mean number of hours they worked (HRS1) was more than 20.
A. Some sociologists argue that the amount of education one has is related to the development of more liberal political and economic values. Test this argument by examining the correlation between one's self-rated political conservatism (POLVIWES) and one's years of education (EDUC). Set a = .01 to examine the significance of the correlation coefficient. Use a one-tail test. Interpret your results.
B. Is age linearly related to political outlook? Check the correlation between POLVIEWS and AGE and report the coefficient. Test whether it is significantly different from 0 in the population. Repeat for the GSS96 sample and discuss.
B. One might guess that those who earn less would be more in favor of government intervention to reduce income differences. Using GSS'91 only, test this hypothesis by regressing EQLWLTH on INCOME. Set a = .01 to examine the significance of the regression slope coefficient. Use a one-tail test. Interpret the results. [Recode EQLWLTH as follows: RECODE EQLWLTH(1=7)(2=6) (3=5)(5=3)(6=2)(7=1)]. Perform this test on both data sets and compare your results.
SPSS: Use GSS96.
A. Does the frequency of prayer explain variation in the rate of church attendance? Regress ATTEND on PRAY and report both the R2 and the r, testing H0: r = 0 at a = .01. Repeat for the GSS96 sample and discuss.
Monte Carlo simulation -- sampling from a known population and producing a sampling distribution for a statistic:
Find the MathCAD icon in the MathSoft applications group on the desktop; double-click to open;
Open file n:\kaufman\ust601\mcdlabs\lab4.mcd; or
(
Immediately make a copy of the file on your H: drive by using File, Save as...
A. Which race -- African-American or Caucasian -- is more tolerant of allowing an atheist to speak in public? Crosstabulate SPKATH by RACE after omitting "other" race, and compute the c2. What do you conclude?
(HINT: establish if independent, and if not, check for level among non-independent proportions.)
B. Given the Surgeon General's warnings about tobacco smoking, are people's healths likely to be worse if they smoke? Crosstabulate HEALTH by SMOKE (for both data sets) and calculate chi-square to test the null hypothesis of no health difference between smokers and non-smokers. Set a = .05.
C. Are less-educated people more or less likely than more-educated people to smoke? Find this relationship by crosstabulating SMOKE with EDUC recoded into 3 categories (0-11 years, 12-15 years, and 16 or more years). Compute the c2. What do you conclude? Repeat for GSS96.
D. (50-STATES data) Do states with so-called right-to-work laws (which ban compulsory union membership, or "closed shops") have lower rates of unionized working forces? Dichotomize UNION (at 21.8%) and crosstabulate with RT2WORK, setting a = .05.
Monte Carlo simulation -- sampling from a known population and estimating the regression coefficients to check estimate properties:
Find the MathCAD icon in the MathSoft applications group on the desktop; double-click to open;
Open file n:\kaufman\ust601\lab4.mcd; or
immediately make a copy of the file on your H: drive by using File, Save as...
After reading the text, scroll down to the statement N=20;
Scroll down so you have on your screen both N="20" and the graph of the population regression line and the estimated sample regression line;
Click with your mouse in the area of N=20; hit F9 repeatedly (meaning that you are repeatedly drawing samples of the same size from the know population) and watch the estimated sample line being redrawn on the screen;
When you get comfortable, begin the assignment; for each F9 hit, you get regression coefficient estimates a and b (upper right corner of the graph); copy them into the two prepared arrays below the graph, labeled A and B. For each pair (a, b) enter the value of a in array A, and the value of b in array B.
When you have entered 10 such runs, the average intercept a and average slope b are computed below the arrays.
Record the averages and compare them to the known population values for the intercept (alpha) and slope (beta), which are given at the top of the file.
UST/PAD/PDD 601 
HOMEWORK FORMAT
NOTE: the SPSS data sets to be used for homeworks are found on the n: drive, under n:\kaufman/UST601\. Begin your session by copying the datasets onto your h: drive. You can do this as many times as you need (for example, if you think you have altered your dataset, make a new copy for starting a new homework.)
Assignments
1.5, 1.6, 1.8, 1.11; 2.2, 2.6, 2.11, 2.18, 2.19, 2.20.
(.)