PSYC 220 (Research Methods in Psychology), Fall, 2004     Steven J. Gilbert, SUNY-Oneonta


V ASSIGNMENTS FROM PAVKOV & PIERCE, READY, SET, GO!

SPSS Assignments (Revised 9/05/04)

> Each SPSS assignment should contain two parts:

1.                              Part 1: SPSS Output sheets.

                                 Part 2: Answers to questions (see below).

      > Part 1:  The work you do on SPSS will result in output sheets.  These should be emailed directly to me at gilbersj@oneonta.edu.  Be sure to include your Name, Assignment #, Dataset (1-5), and Data Form (A-E) in the body of the email.  You also should make a paper copy for your own use in answering the questions in Part 2.

3.   > Part 2:  The questions for each SPSS assignment are given below.  The identical questions for each assignment are appear on a computer survey instrument.  You access this survey by clicking on the appropriate hyperlink on the question sheet.  Enter your answers to questions on the survey, and then press the SUBMIT button.  This will send an email containing your answers, directly to me.  You will receive an email containing feedback (comments and grades) from me, at a later time.

 

Grading of SPSS Assignments

>Each assignment is worth 7.5 points.

>7.5 points are awarded for each assignment containing no errors, that is turned in on time, along with the accompanying SPSS output sheets.

>An assignment containing any errors is returned to you for correction.  Correct the errors, and then email a correction sheet to me containing all identifying information (Name, Assignment #, Dataset & Data Form) along with the answers to questions that had contained errors (and only these answers; be sure to number these answers,  to avoid any confusion).  This process will continue until there are no errors.  1.5 points are deducted for each required resubmission.  Thus, you receive 6.0 points (80%) if 1 resubmission  is required, 4.5 points (60%) if 2 resubmissions are required, etc. 

>1 point is subtracted from the final score of each SPSS assignment turned in late.

 

ASSIGNMENT 1

>>> Read Assignment 1 in the Pavkov & Pierce text.

>>> Attached to your copy of the Course Outline is DATASET #1. This contains the data you will be working with for Assignment 1. (Note: The actual numbers in your copy of each DATASET are unique to you, so be sure to work only with your copies of the DATASETS as you complete your SPSS assignments.)  Input the data from DATASET #1 into the DATAVIEW WINDOW of SPSS.  Be sure to SAVE your work onto a floppy disc, for future use.

 

DATASET #1.  20 students in a section of PSYC 100 were given their second test on either WHITE or BLUE paper. The test contained 50 multiple-choice questions, and the SCORE a student received was the number of correct answers.

 

>>> Using the DEFINE VARIABLE WINDOW, give each variable a VARIABLE NAME (e.g., COLOR), and a VARIABLE LABEL (e.g., COLOR OF TEST).
>>> Give the VALUES of the variable SEX the LABELS "1 = male" and "2 = female".
>>> Give the VALUES of the variable COLOR the LABELS "1 = white" and "2 = blue."
>>>  Be sure to SAVE the file on a floppy disc, as well as on the hard drive or "P" drive of your computer.
>>> Send the contents of the DATAVIEW WINDOW to gilbersj as an email attachment.


          Answer the following Questions using the online Survey Form for Assignment 1
Which variables that you inputted into SPSS would be considered...
(A1-1) DEPENDENT VARIABLES;
(A1-2) MANIPULATED INDEPENDENT VARIABLES;
(A1-3) NON-MANIPULATED (i.e. measured), SUBJECT-CLASSIFICATION VARIABLES?

 

 

ASSIGNMENT 2

>>> Read Assignment 2 in the Pavkov & Pierce text.
>>> Load  DATASET #1 into the DATAVIEW window.
>>> Using the FREQUENCIES procedure, generate descriptive statistics for the SCORE subjects received on their test.

>>> Save the OUTPUT FILE containing the descriptive statistics on a floppy disc, as well as on the hard drive or "P" drive of your computer.
>>> Send the contents of the OUTPUT FILE containing the descriptive statistics to gilbersj as an email attachment.

 

          Answer the following Questions using the online Survey Form for Assignment 2
(A2-1)  What is the Mean score?
(A2-2)  What is the Median score?
(A2-3)  What is Standard Deviation of the distribution of the scores?
(A2-4)  Given the Standard Deviation, what is the range within which we would expect approximately 2/3rds of the test scores of the PSYC 100 class to fall?
(A2-5)  Given the Standard Deviation, what is the range within which we would expect approximately 95% of the test scores of the PSYC 100 class to fall?
(2-6)  Read Box 1 (below).  Then report the skewness statistic.  Given this statistic, would you conclude that the distribution of PSYC 100 test scores is NORMAL, SKEWED RIGHT, or SKEWED LEFT?

 

Box 1. 
A skewness statistic of 0 denotes a distribution that is perfectly symmetrical around the mean. If  the skewness statistic is OUTSIDE the range of -2.0 to +2.0 (i.e., lower than -2.0 or higher than +2.0), then we can be confident that the distribution is skewed--that it is NOT symmetrical around the mean.  Specifically, a skewness statistic lower than -2.0 denotes a distribution that is negatively skewed; it has a longer tail on the negative (left) side of the mean, reflecting more extreme low scores than high scores.  A skewness statistic higher than +2.0 denotes a distribution that is positively skewed; it has a longer tail on the positive (right) side of the mean, reflecting more extreme high scores than low scores.

 

What if  the skewness statistic is not exactly zero, but neither is low enough to exceed the -2.0 figure or high enough to exceed +2.0 figure necessary to confidently establish that a distribution is skewed (not symmetrical)?  In such cases, we consider the distribution to be symmetrical, and treat it as such.


 

ASSIGNMENT 3

>>> Read Assignment 3 in the Pavkov & Pierce text.
>>> Load  DATASET #1 into the DATAVIEW window.
>>> Using the CHARTS option of the FREQUENCIES procedure, produce a HISTOGRAM for the SCORE subjects received on the test.

>>> Send the OUTPUT FILE containing the HISTOGRAM to gilbersj as an email attachment.

Following the examples in Box 2 below, write a concise description of the SCORE variable from DATASET #1, using the information developed in Assignments 2 & 3.
 

Box 2.
The mean number of movies that our participants reported seeing each month was 6.23 (N = 37, SD = 3.12).  The skew statistic (-1.37) suggests that the distribution can be treated as normal in shape, with approximately 2/3rds of the participants reporting seeing between 3.11 and 9.35 movies each month.

 

The mean number of CDs found in the dorm rooms of our sample of students living in Hays Hall was 11.35 (N = 48, SD = 17.21).  The distribution of scores was positively skewed (+4.53), with 45 students owning between 0 and 12 CDs, and 3 students owning more than 80.

 


 

ASSIGNMENT 4

>>> Read Alison's story in Box 3.
 

Box 3.
Alison was interested in determining whether the color of the paper a test was printed on, affected students' performance on the test.  She had a hunch that students will do better on a blue paper test than on a white paper test. To test her hunch, Alison gave the 20 students in a particular PSYC 100 section a 50 multiple choice question psychology test either on blue paper or white paper.  Which color test a student received was determined by a flip of a coin.  Later, Alison calculated the mean test score earned by the blue paper test group, and the mean for the white paper test group, and compared the two means.

       

     Answer Questions (A4-1) - (A4-16) using the online Survey Form for Assignment #4
(A4-1) What is the DEPENDENT VARIABLE in Alison's experiment?
(A4-2) What is the INDEPENDENT VARIABLE in Alison's experiment?
(A4-3)  How many levels of the INDEPENDENT VARIABLE are there? 
(A4-4) In some experiments, each participant is exposed to only one of the levels of the INDEPENDENT VARIABLE -- these are called BETWEEN- or INDEPENDENT GROUPS experiments.  In other experiments, each participant is exposed to all levels of the INDEPENDENT VARIABLE -- these are called WITHIN SUBJECT or REPEATED MEASURES experiments.  Which kind of experiment is Alison's?
(A4-5) Is Alison's INDEPENDENT VARIABLE a MANIPULATED or SUBJECT variable? 
(A4-6) State a NULL HYPOTHESIS concerning the effect of the independent variable on the dependent variable in Alison's experiment.
(A4-7) State Alison's RESEARCH HYPOTHESIS.
(A4-8) Is Alison's RESEARCH HYPOTHESIS DIRECTIONAL or NONDIRECTIONAL? 

 

>>> Read Assignment 4 in the Pavkov & Pierce text.
>>> Load  DATASET #1 into the DATAVIEW window.
>>> Make sure you understand the definition in Box 4.

 

Box 4.
The INDEPENDENT SAMPLES t-test is used when a researcher has divided subjects into TWO GROUPS, and wishes to compare the means of these two groups on a dependent variable that has been measured using an interval or ratio scale of measurement .

 

>>> Perform an INDEPENDENT SAMPLES t-test appropriate to test Alison's hypothesis.

>>> Send the contents of the OUTPUT FILE containing the results of the t-test to gilbersj as an email attachment.

(A4-9) Report the dependent variable means and standard deviations for each of the two groups.
(A4-10) Speculate as to whether the means appear to be significantly different from each other (i.e., more different than you might simply expect by chance).
(A4-11) According to LEVENE'S TEST FOR EQUALITY OF VARIANCES, are you entitled to assume that the variances of the two groups are equal?  Report the information that enables you to make this decision.
(A4-12) Report the results of the appropriate t-test for comparing the two group means, including the t value, the df, and the significance level obtained [e.g., t(35) = 1.23, p < .23].
(A4-13) Based upon the results of the t-test, should Alison ACCEPT or NOT-ACCEPT the research hypothesis?
(A4-14) Let's restate (4-13).  Do the results of Alison's experiment SUPPORT, or NOT SUPPORT her hypothesis?
(A4-16) Make up a phony t and p that would compel the opposite conclusion from the one you reached in questions (A4-13) and (A4-14).


 

ASSIGNMENT 5

>>> Read Greg's story in Box 5.
 

Box 5.
Greg noticed that it seemed easier to read and understand printed material when the letters were in a bold font rather than the standard font.  To test out this hunch, Greg gave 20 PSYC 100 students their 100-question final exam in a special format.  The even numbered questions appeared in the standard font, but the odd numbered questions were in a bold font.  Greg calculated a separate mean score for the 50 standard-font (even numbered) questions and the 50 bold-font (odd numbered) questions, and compared these two means.

 

       Answer Questions (A5-1) - (A5-13) using the online Survey Form for Assignment #5
(A5-1) What is the INDEPENDENT VARIABLE in Greg's experiment?
(A5-2) How many levels of the INDEPENDENT VARIABLE are there?
(A5-3) What is the DEPENDENT VARIABLE in Greg's experiment?
(A5-4) In some experiments, each participant is exposed to only one of the levels of the INDEPENDENT VARIABLE -- these are called BETWEEN GROUPS or INDEPENDENT GROUPS experiments.  In other experiments, each participant is exposed to all levels of the INDEPENDENT VARIABLE -- these are called WITHIN SUBJECT or REPEATED MEASURES experiments.  Which kind of experiment is Greg's?
(A5-5) State a NULL HYPOTHESIS concerning the effect of the independent variable on the dependent variable in Greg's experiment.
(A5-6) State Greg's RESEARCH HYPOTHESIS.
(A5-7) Is Greg's RESEARCH HYPOTHESIS DIRECTIONAL or NON-DIRECTIONAL? 

 

>>> Read Assignment 5 in the Pavkov & Pierce text..
>>> Input the data from DATASET #2 into the DATAVIEW WINDOW of SPSS.  Be sure to SAVE your work onto a floppy disc, for future use.

 

>>> Using the DEFINE VARIABLE WINDOW, give each variable a VARIABLE NAME (e.g., STANDARD), and a VARIABLE LABEL (e.g., STANDARD FONT SCORE).
>>> Give the VALUES of the variable SEX the LABELS "1 = male" and "2 = female".
>>> Send the contents of the DATAVIEW WINDOW containing DATASET #2 to gillbersj as an email attachment.

>>> Make sure you understand the definition in Box 6.


Box 6.
A PAIRED t-test compares the means of two measures taken on the same people. These can be the same measure given twice (e.g., the same 50-item psychology test given first as a pretest, and later as a posttest), or two different measures that use the same or comparable scales (e.g. the 50 even and the 50 odd numbered questions on a 100 question psychology test).

 

>>> Do a PAIRED t-test to test Greg's research hypothesis.

>>> Send the contents of the OUTPUT FILE containing the results of the t-test to gilbersj as an email attachment.

(A5-8) Report the MEANS and STANDARD DEVIATIONS for the two sets of scores that you are comparing with your t-test.
(A5-9) Speculate as to whether the means appear to be significantly different from each other (i.e., more different than you might simply expect by chance).
(A5-10) Report the results of the t-test, including the t value, the df, and the significance level obtained [e.g., t(41) = 1.32, p < .18].   Ah, but check out Box 7 before you do!

 

Box 7.
Note:  The significance level (p) reported by SPSS is appropriate for tests of non-directional ("two-tail") hypotheses (e.g., men differ from women on a measure of happiness).  Because you are testing a directional ("one-tail") hypothesis (e.g., men have higher happiness scores than women), you must divide the reported significance level by 2.  Thus, if the (two-tail p reported by SPSS is .064 (which would NOT be statistically significant), the (one-tail) p that actually applies to the test of your hypothesis is .032 (which IS statistically significant).  p < .032 is what you should report.

 

(A5-11) Based upon the results of the t-test, should Greg ACCEPT or NOT-ACCEPT the research hypothesis?
(A5-12) Let's restate (5-11).  Do the results of Greg's experiment SUPPORT, or NOT SUPPORT his hypothesis?
(A5-13) Make up a phony t and p that would compel the opposite conclusion from the one you reached in questions (5-11) and (5-12).



 

ASSIGNMENT 6

>>> Review Alison's story in Box 3.
>>> Read Jenny's story in Box 8.
 

Box 8.
Jenny was intrigued by Alison's experiment on the effect of blue vs white paper on how students did on tests.  She agreed with Alison's hunch that students will do better on a blue paper test than on a white paper test and wanted to replicate that aspect of Alison's experiment.  But Jenny went farther.  She reasoned that students should door POORER on a test written on RED paper than one written on the standard white paper.  To test her hunch, Jenny gave a different group of 20 students in a PSYC 100 section a 50 multiple choice question psychology test either on blue, white, or red paper.  Which color test a student received was determined by a roll of dice (1 or 2 = blue; 3 or 4 = white; 5 or 6 = red).  Later, Jenny calculated the mean test score earned by the blue, white and red paper test groups, and compared the three means.

   

     Answer Questions (A6-1) - (A6-9) using the online Survey Form for Assignment #6
(A6-1) What is the DEPENDENT VARIABLE in Jenny's experiment?
(A6-2) What is the INDEPENDENT VARIABLE in Jenny's experiment?
(A6-3)  How many levels of the INDEPENDENT VARIABLE are there? 
(A6-4) In some experiments, each participant is exposed to only one of the levels of the INDEPENDENT VARIABLE -- these are called BETWEEN GROUP or INDEPENDENT GROUPS experiments.  In other experiments, each participant is exposed to all levels of the INDEPENDENT VARIABLE -- these are called WITHIN SUBJECT or REPEATED MEASURES experiments.  Which kind of experiment is Jenny's?
(A6-5) Is Jenny's INDEPENDENT VARIABLE a MANIPULATED or SUBJECT variable? 
(A6-6) State a NULL HYPOTHESIS concerning the effect of the independent variable on the dependent variable in Jenny's experiment.
(A6-7) If we consider the WHITE group to be a CONTROL group, to which the BLUE and RED groups are compared, then  Jenny actually has TWO research hypotheses.  State Jenny's two RESEARCH HYPOTHESES.
(A6-8) Are Jenny's RESEARCH HYPOTHESES DIRECTIONAL or NON-DIRECTIONAL? 

 

>>> Read Assignment 6 in the Pavkov & Pierce text.
>>> Input the data from DATASET #3 into the DATAVIEW WINDOW of SPSS.  Be sure to SAVE your work onto a floppy disc, for future use.

>>> Using the DEFINE VARIABLE WINDOW, give each variable a VARIABLE NAME (e.g., COLOR), and a VARIABLE LABEL (e.g., COLOR OF TEST).
>>> Give the VALUES of the variable SEX the LABELS "1 = male" and "2 = female".
>>> Give the VALUES of the variable COLOR the LABELS "1 = white", "2 = blue," and "3 = red."
>>> Send the contents of the DATAVIEW WINDOW containing DATASET #3 to gillbersj as an email attachment.

 

>>> Read Box # 9..


Box 9.
An INDEPENDENT SAMPLES (BETWEEN GROUP) ANOVA compares the means of MORE THAN TWO groups of subjects on a dependent variable that has been measured using an interval or ratio scale of measurement. 

 

Sometimes, RESEARCH hypotheses and STATISTICAL hypotheses differ.  Suppose I gather PSYC, SOC, and POLI-SCI majors and measure how funny they are.  My RESEARCH hypotheses might be: (a) PSYC majors are funnier then SOC majors; and (b) SOC majors are funnier than POLI-SCI majors.  These are two DIRECTIONAL, RESEARCH hypotheses.

 

Statistically, the first step in testing these two hypotheses is to do an INDEPENDENT SAMPLES (BETWEEN GROUP) ANOVA, directly comparing the funniness means of the three groups.  The primary STATISTICAL HYPOTHESIS in an ANOVA always is NON-DIRECTIONAL; it is the hypothesis that the means differ from each other more than would be expected by chance.  If the STATISTICAL ANOVA hypothesis is NOT SUPPORTED -- if the three means DON'T differ more than would be expected by chance -- then we conclude that neither of our RESEARCH HYPOTHESES are SUPPORTED (i.e., PSYC majors are NOT funnier than SOC majors, and SOC majors are NOT funnier than POLI-SCI majors).

 

But, what if the STATISTICAL ANOVA hypothesis that the three groups differ in funniness IS SUPPORTED (at p < .05)?  Then, we instruct SPSS to perform a SHEFFE POST-HOC test.  This is like doing three INDEPENDENT SAMPLES t-tests (comparing the funniness means of PSYC vs SOC majors, SOC vs POLI-SCI majors, and PSYC vs POLI-SCI majors).  The first two comparisons are direct tests of our two RESEARCH HYPOTHESES.

 

 

>>> Perform an Analysis of Variance (ANOVA) on the data from Jenny's experiment.  Be sure to instruct SPSS to include descriptive statistics (Mean & Standard Deviation of the dependent variable measure for each group), and the Sheffe Post-Hoc Comparison test.

>>> Send the contents of the OUTPUT FILE containing the results of the ANOVA and the Post-Hoc Comparison test to gilbersj as an email attachment.

(A6-9) Write a short paragraph describing the results of the ANOVA and Sheffe tests, and interpreting the meaning of these results. Include the following information:
(a) a statement of the Means and Standard Deviations of the three groups;
(b) a speculation as to whether the means appear to be more different from each other than would be expected by chance;
(c) a statement of the result of the ANOVA in terms of the F, df-between, df-within, and significance level
        (e.g., F(2, 197) = 18.716, p < .001);
(d) a statement of the decision the ANOVA compels concerning whether or not you can ACCEPT the STATISTICAL HYPOTHESIS that the three group means differ more than chance expectation; and

(e) what conclusions the Sheffe tests suggest concerning whether or not there are significant differences between each pair of means (i.e., whether the results support your specific, directional RESEARCH HYPOTHESES.


Note: Like the Independent Samples t-test, the SPSS ANOVA procedure produces a test of HOMOGENEITY OF VARIANCES, to determine whether or not the variance of the scores of each group are roughly equivalent. If they are not (i.e., if the significance of the homogeneity of variances test is p < .05) then the results of the ANOVA might be inaccurate or misleading. The text does not instruct you as to what to do in this case!  We will discuss it in class.

 

 

 

ASSIGNMENT 7

>>> Review Greg's story in Box 5.
>>> Read Ken's story in Box 10.
 

Box 10.
Ken could see how it would be easier to read and understand material in a bold font than material in a standard font.  Ken noticed that in some of his textbooks, certain blocks of material (like a quotation) are written in italics, and he always found that hard to read.  Ken speculated that just as students would do better on test questions written in a bold font (than in a standard font), they would do poorer on questions written in italics (than in a standard font).  To test these predictions, Ken gave 21 students in a section of ANTH 100 their 150 question final exam in a special format.  Questions 1,4,7,10....148 appeared in a standard font, questions 2,5,8,11....149 were in a bold font, and questions 3,6,9,12....150 were in an italics font.  Ken calculated the students' mean scores on the standard, bold, and italic questions, and then compared the means.

      

      Answer Questions (A7-1) - (A7-11) on the online Survey Form for Assignment #7
(A7-1) What is the INDEPENDENT VARIABLE in Ken's experiment?
(A7-2) How many levels of the INDEPENDENT VARIABLE are there?
(A7-3) What is the DEPENDENT VARIABLE in Ken's experiment?
(A7-4) In some experiments, each participant is exposed to only one of the levels of the INDEPENDENT VARIABLE -- these are called BETWEEN or INDEPENDENT GROUPS experiments.  In other experiments, each participant is exposed to all levels of the INDEPENDENT VARIABLE -- these are called WITHIN SUBJECT or REPEATED MEASURES experiments.  Which kind of experiment is Ken's?
(A7-5) State a NULL HYPOTHESIS concerning the effect of the independent variable on the dependent variable in Ken's experiment.
(A7-6) If we consider the STANDARD FONT questions to be the CONTROL treatment, to which the BOLD FONT and the ITALIC FONT treatments are compared, then Ken actually has TWO research hypotheses.  State Ken's two RESEARCH HYPOTHESES.
(A7-7) Are Ken's RESEARCH HYPOTHESES DIRECTIONAL or NON-DIRECTIONAL? 

 

>>> Read Assignment 7 in the Pavkov & Pierce text.  Skip the subsections titled "Testing Between-Subjects Effects" and "Testing Sphericity" and "The Univariate Test" under the heading "INTERPRETING THE OUTPUT."
>>> Input the data from DATASET #4 into the DATAVIEW WINDOW of SPSS.  Be sure to SAVE your work onto a floppy disc, for future use.

>>> Using the DEFINE VARIABLE WINDOW, give each variable a VARIABLE NAME (e.g., STANDARD), and a VARIABLE LABEL (e.g., STANDARD FONT SCORE).
>>> Give the VALUES of the variable SEX the LABELS "1 = male" and "2 = female".
>>> Send the contents of the DATAVIEW WINDOW containing DATASET #4 to gillbersj as an email attachment.

>>> Make sure you understand the definition in Box 11.


Box 11.
A RELATED SAMPLES ONE-WAY ANOVA in many ways resembles the PAIRED SAMPLE t-test (Assignment #5). It compares the means of MORE than two measures taken on the same people. As with the PAIRED SAMPLE t-test, it only makes sense to compare the means of measures that are taken on the same scale.

 

Note:  The following discussion will resemble one that you read in Box 9, for Assignment 6.  There the discussion related to a BETWEEN (or INDEPENDENT) GROUPS design.  Here it relates to a REPEATED MEASURES (or WITHIN -SUBJECT) design. 

 

Sometimes RESEARCH hypotheses and STATISTICAL hypotheses differ.  Suppose I gather 100 students, have them listen to 10 minute samples of CLASSICAL MUSIC, 50's ROCK, and contemporary HIP-HOP music, and measure how much they like each on a 1 (hate) to 10 (love) scale.  My RESEARCH hypotheses might be: (a) students will like HIP-HOP better than ROCK; and (b) students will like CLASSICAL less than ROCK.  These are two DIRECTIONAL research hypotheses.

 

Statistically, the first step in testing these two hypotheses is to do a ONE-WAY REPEATED MEASURES ANOVA, directly comparing the mean ratings given to the three kinds of music.  The primary STATISTICAL HYPOTHESIS in an ANOVA always is NON-DIRECTIONAL; it is the hypothesis that the means differ from each other more than would be expected by chance.

 

If the STATISTICAL ANOVA hypothesis is NOT SUPPORTED -- if the three means DON'T differ more than would be expected by chance -- then we conclude that neither of our RESEARCH HYPOTHESES are SUPPORTED (i.e., students do NOT like HIP-HOP better than ROCK, and students do NOT like CLASSICAL less than ROCK).

 

But, what if the STATISTICAL ANOVA hypothesis that the three mean music ratings differ, IS SUPPORTED (at p < .05)?  Then, you must do what you learned in Assignment 5!  Instruct SPSS to do a RELATED SAMPLES (WITHIN-SUBJECTS) t-test, comparing the means for HIP-HOP and ROCK (testing the first RESEARCH HYPOTHESIS), and a second t-test comparing  the means for CLASSICAL and ROCK (testing the second RESEARCH HYPOTHESIS).  These two t-tests directly test your two RESEARCH  HYPOTHESES.

 

>>> Do a RELATED SAMPLES ONE-WAY ANOVA on these three variables.

Note: You will want the output to include the mean and standard deviation for each variable.  In order to obtain these statistics, you must click on the OPTIONS button on the lower right hand corner of the GLM dialog box, and then click on DESCRIPTIVES.

(A7-8) Write a brief statement of the relevant statistics produced by the RELATED SAMPLES ONE-WAY ANOVA procedure (i.e., the means, standard deviations, F, dfs, and p).
(A7-9) Do the results enable you to accept the STATISTICAL ANOVA HYPOTHESIS that the means for the test questions written in the three font styles differ more than would be expected by chance?
(A7-10) Do the results warrant performing POST-HOC t-tests comparing the means of each pair of scores?
(A7-11) If the answer to (7-10) is "yes," then do the appropriate POST-HOC PAIRED t-tests (see Assignment #5).  Report the results of these tests as you did in Assignment 5, Q (A5-10). Write an explicit statement as to whether the results of the POST-HOC tests SUPPORT or FAIL to SUPPORT the two RESEARCH HYPOTHESES of Ken's experiment, that you stated in (A7-6).

>>> Send the contents of the OUTPUT FILE containing the results of the ANOVA, and thePOST-HOC t-tests (if necessary) to gilbersj as an email attachment.

 


 

ASSIGNMENT 8

>>> Read Rachel's story in Box 12.
 

Box 12.
In her Psychology of Personality class, Rachel learned about the "Big Five" personality factors, and was interested in how three of them -- Extroversion, Agreeableness, and Conscientiousness -- related to each other, and to how many friends a student was.  Rachel gave the three "Big Five" subscales that measure Extroversion, Agreeableness, and Conscientiousness, to 22 women living in Hays Hall.  In addition, she asked the women to make a list of people on campus whom they consider to be a friend.  Rachel counted the number of  names on each subject's list and used that number as a measure of popularity.  Rachel expected each personality scale to show a positive relationship with her measure of popularity. 

 

       Answer Questions (A8-1) - (A8-5) using the online Survey Form for Assignment #8
(A8-1) Name two variables that Rachel expected to have a positive relationship with each other.
(A8-2) State a NULL HYPOTHESIS and a DIRECTIONAL (one-tail) ALTERNATE (RESEARCH) HYPOTHESIS concerning the expected relationship between these two variables.
(A8-3) If the Null Hypothesis were true, then what value CORRELATION COEFFICIENT would you expect to obtain for these two variables?

 

>>> Read Assignment 8 in the Pavkov & Pierce text.
>>> Input the data from DATASET #5 into the DATAVIEW WINDOW of SPSS.  Be sure to SAVE your work onto a floppy disc, for future use.

>>> Using the DEFINE VARIABLE WINDOW, give each variable a VARIABLE NAME (e.g., AGREE), and a VARIABLE LABEL (e.g., AGREEABLENESS FACTOR).
>>> Produce a SCATTERPLOT that graphically shows the actual relationship between the variables you chose in (8-1). 

(A8-4)  Does the shape of the scatter suggest that a significant correlation has been obtained?  Why (or why not)?

>>> Perform the CORRELATION procedure, making an intercorrelation table showing the relationships among all four variables that Rachel measured.

>>> Send the contents of the DATAVIEW WINDOW containing DATASET #5, the OUTPUT WINDOW containing the SCATTERPLOT and INTERCORRELATION TABLE to gillbersj as an email attachment.

(A8-5) Describe the results of the correlation procedure for the two variables you chose in (8-1), including the relevant statistics (r, df, sig.). Do the results support Rachel's  hypothesis (i.e., are the results statistically significant)?

 

Assignment 9.

>>> Read Assignment 9 in the Pavkov & Pierce text.
>>> Review Rachel's story in Box 12, and the intercorrelation table you produced for Assignment 8.
>>> Read Dan's story in Box 13.
 

Box 13.
Rachel showed Dan the intercorrelation table she had made.  Dan was intrigued, and wondered how well you could predict a young woman's popularity, from knowledge of her levels of Extroversion, Agreeableness, and Conscientiousness.  To find out, Dan obtained Rachel's data, and performed a Multiple Regression Analysis, with POPULAR as the DEPENDENT (Y) VARIABLE, and Extroversion (X1), Agreeableness (X2), and Conscientiousness (X3) as three INDEPENDENT, or PREDICTOR variables..

>>> Load  DATASET #5 into the DATAVIEW window.
>>> Do the MULTIPLE REGRESSION that Dan did.

>>> Send the contents of the OUTPUT FILE containing MULTIPLE REGRESSION to gilbersj as an email attachment.

 

       Answer Questions (A9-1) - (A9-5) using the online Survey Form for Assignment #9
(9-1) What percentage of the overall variance in subjects' POPULARITY scores is explained by (related to) the three predictor variables ?
(9-2) What percentage of the overall variance in subjects' POPULARITY scores is NOT explained by (related to) the three predictor variables ?

>>> Report and interpret the statistics (F, dfs, p) that show whether the percentage described in (9-1) represents a statistically significant level of prediction, i.e., whether the R SQUARE is greater than that which we would expect to occur by chance.

>>> Read Box 14.
 

Box 14.
The Multiple Regression also tells us whether each predictor (independent) variable significantly contributes to the prediction of the outcome (dependent) variable after the contribution of the other predictor variable(s) have been factored out--in effect telling us whether each predictor accounts for a UNIQUE portion of the variance in the outcome variable -- a portion NOT explained by any other predictor.  

(9-3) Report the relevant statistics (t, p), and determine whether they support the contention that Extroversion predicts a SIGNIFICANT, UNIQUE portion of the variance in POPULARITY.
(9-4) Report the relevant statistics (t, p), and determine whether they support the contention that Agreeableness predicts a SIGNIFICANT, UNIQUE portion of the variance in POPULARITY.
(9-5) Report the relevant statistics (t, p), and determine whether they support the contention that Conscientiousness predicts a SIGNIFICANT, UNIQUE portion of the variance in POPULARITY.


 

Assignment 10.

>>> Load  DATASET #1 into the DATAVIEW window.
>>> Review Alison's story in Box 3, Assignment 4.
>>> Read Box 15.
 

Box 15.
The INDEPENDENT VARIABLE of Alison's experiment was COLOR of PAPER (white vs blue).  Ideally, the selection of subjects should have produced the following breakdown.

                      COLOR OF PAPER

 

WHITE

BLUE

MEN

n = 5

n = 5

WOMEN

n = 5

n = 5

                                                            N(white) =10        N(blue) = 10   

Alison was concerned that the proportion of MEN to WOMEN in the WHITE group might be SIGNIFICANTLY different from the proportion of MEN to WOMEN in the BLUE group.  If this were true, then any differences between the test scores of subjects in the WHITE vs the BLUE group MIGHT be due, at least in part, to differences between how well MEN and WOMEN do on tests.  This would seriously compromise the INTERNAL VALIDITY of Alison's experiment.

>>> Read Assignment 10 in the Pavkov & Pierce text.

 

      Answer Questions (A10-1) - (A10-4) using the online Survey Form for Assignment #10

(10-1) To determine whether the proportion of MEN to WOMEN differed in the two groups, Alison did a CHI SQUARE test of INDEPENDENCE, with SEX and COLOR as the two variables.  State the NULL HYPOTHESIS and the ALTERNATE (Research) HYPOTHESIS of Alison's CHI SQUARE test.

Make a CROSSTABULATION TABLE and perform Alison's CHI SQUARE test.

>>> Send the contents of the OUTPUT FILE containing the CROSSTABULATION TABLE and CHI SQUARE test to gilbersj as an email attachment.

(10-2)  Describe the essential information contained in the CROSSTABULATION table.  What conclusion do the data appear to suggest, concerning whether or not the proportion of MALES to FEMALES differ in the WHITE vs BLUE groups?
(10-3) Report and interpret the relevant CHI SQUARE TEST statistics (CHI-SQ, df, p).
(10-4) Has the internal validity of Alison's WHITE vs BLUE paper experiment been compromised by the SEX VARIABLE?  What tells you that?