1

Nonparametric Assignmentfor PSYC 7431, Spring, 2018

Links to the Data Files

Bradshaw, Michael / Day, Alicia / Fox, Nicole / Gaertner,Lexi
Haley, Erin / Hicks, Kianda / Japczyk Schuler, E. J. / Levy, Rebecca
Long, Kelli / Meier, Emily / Midgette, Emily / Nethercutt, Michael
Reichart, Clay / Richardson, Natalie / Robinson, Demi / Robinson, Samuel
Sharma, Saryu / Tripp, Connor / Vincent, Cameron / Williams, Rachel
Zurlinden, Taylor / ***Vanilla Ice Cream, My Favorite***

You have received a grant to study the effectiveness of Floating Anxiety Reduction Therapy for treatment of gastrointestinal anxiety disorder (GAD). Persons who suffer from GAD experience gastrointestinal symptoms that are accompanied by floating anxiety. From a group of patients with this disorder who have volunteered to participate in your research, you create fifteen matched pairs of patients. You matched the patients on sex, age, race, and their physician’s estimation of the severity of their symptoms. Within each pair of patients you randomly assign one patient to receive the experimental therapy and the other to be put on a three month waiting list. Three months later each participant completes the Gastrointestinal Anxiety Scale (GAS), which is your outcome measure. High scores on the GAS represent frequent and intense occurrence of symptoms of GAD.

You have been provided with the data for this matched-pairs design. In your downloaded data file there is one row of scores for each matched pair. The first score is the GAS score for the untreated patient. The second score is the GAS score for the patient who received the experimental therapy. The third score is a difference score, where positive values indicate that the waitlist patient scored higher on the GAS than did the patient who received the experimental therapy. The delimiter is a blank space.

Your initial task is to use SAS, SPSS, or R to analyze the data with the appropriate Wilcoxon test. The Wilcoxon test is employed because GAS scores are known to be skewed. Employ a .05 criterion of statistical significance. Conduct an exact test, not an approximate (z or t) test. Prepare an APA-style summary report in which you present the following at the top of the Word document:

  • A Word table in which are presented, for each group, the sample size, mean, median, standard deviation, skewness, and kurtosis. Here is a template for that table:

Table 1
Distribution of Gastric Distress Scores in Patients Receiving Floating Anxiety Reduction Therapy or Waitlisted

Group / n / M / Mdn / SD / g1Skewness / g2 Kurtosis
Treated / 15 / 96.3 / 85 / 102.0 / 1.34 / 1.41
Wait List / 15 / 116.7 / 105 / 99.8 / 1.28 / 1.44
  • A statement which identifies who the patients were, what the independent and dependent variables were, what statistical procedure was employed, the computed value of the test statistic, the exact p value, and indication of whether or not the treatment had a statistically significant effect. As always, emphasize the direction of a significant effect. If you are using SAS, the test statistic will be S, the difference between the expected and the obtained sums of ranks. If you are using SPSS, the test statistic will be T, the smaller of the sum of positively signed ranks and the sum of the negatively signed ranks.

At the bottom of your report paste in the output that shows the results of the signed-ranks test. With SAS it will look like this:

The UNIVARIATE Procedure

Variable: DIFF

TestsforLocation:Mu0=0
Test / Statistic / p Value
Student's t / t / 4.352224 / Pr > |t| / 0.0007
Sign / M / 4.5 / Pr >= |M| / 0.0352
Signed Rank / S / 52 / Pr >= |S| / 0.0015

With SPSS it will look like this:

Ranks
N / Mean Rank / Sum of Ranks
V2 - V1 / Negative Ranks / 8a / 9.25 / 74.00
Positive Ranks / 6b / 5.17 / 31.00
Ties / 1c
Total / 15
a. V2 < V1
b. V2 > V1
c. V2 = V1

The smaller of the two sums of Ranks is T. Notice that there is one case with a difference score of zero, which was excluded from the analysis, so the T would be reported this way: T(N = 14) = 31, p = .19. SAS does not tell you if there were any such cases, you need to check on that yourself. See my document on how to handle difference scores of zero.

Test Statisticsa
V2 - V1
Z / -1.351b
Asymp. Sig. (2-tailed) / .177
Exact Sig. (2-tailed) / .187
Exact Sig. (1-tailed) / .093
Point Probability / .004
a. Wilcoxon Signed Ranks Test
b. Based on positive ranks.

Attach your Word document (named “Nnnn_WSRT”, where “Nnnn” is your last name) to email (with subject line “PSYC 7431: WSRT”) and deliver it to Karl by noon on Friday the 23rd of March.

XD_WSRT.docx