Independent Samples T-Test Homework^ã

    Intra-uterine position (IUP) has been found to affect a number of physical and behavioral variables in rodents. Pups in the 2M IUP (they are sandwiched between two males) apparently are affected by the testosterone being produced by their adjacent brothers. 2M female gerbils show delayed sexual maturation and wean many fewer offspring than do 2F females (2F animals are those that were sandwiched between two sisters). Compared to 2F males, 2M males scent-mark more frequently, have greater anogenital distances, and have larger ventral glands, all correlates of high testosterone. They also mount oestrus females more rapidly and ejaculate after fewer intromissions and with shorter mean latencies than do 2F males.

    Clark, Tucker, & Galef (Animal Behavior, 1992, 43, pp. 215-221) evaluated the reproductive success of 2M and 2F male gerbils. Caesarean delivery was used so IUP could be determined. Compared to 2F males, 2M males were found to sire more young. It was also shown that during oestrus (but not otherwise) female gerbils preferred 2M males over 2F males.

    I have simulated data from a mathematical model based on the results of Clark et al.'s research. I have provided you with simulated data. These data are deposited in my SimData directory on the core server. You can jump directly to your data file by clicking below on the link with your name. It will have a name of the form 'IUP-nn,' where 'nn' is the identification number I use to identify your data files. When you look at the data file you will find that there is one line of data for each gerbil. The first score for each gerbil is the group designation, with 0 coding the 2F condition and 1 coding the 2M condition (you should use SPSS Value Labels to assign these descriptive labels to the numeric codes). The second score for each participant is PUPS (number of pups sired).

    Your assignment is to test the significance of the effect of IUP upon mean number of pups sired. Use the ratio of the two sample variances to decide whether you should report the pooled variances test or the separate variances test. Use nondirectional tests and a .05 criterion for significance. Write an APA-style summary statement. Be sure that your summary statement identifies the subjects, the independent and dependent variables, descriptive statistics for each group (M, SD, and N), the test statistic, the value of the test statistic, the degrees of freedom, the exact significance level, whether or not the difference is significant, the direction of the effect (if significant),  Hedges' g, and  two 95% confidence intervals for the difference in means -- one in raw score units and one in standardized units.  When computing the confidence interval for d, use the pooled variances t and df, even if you used the separate variances t' and df' to get the p value.

    Here is an example (using different variables).  "A Welch-Satterthwaite separate-variances t test was employed to test the significance of the effect of type of psychotherapy on post-treatment wellness scores. Participants who received Rountree therapy scored significantly higher on the wellness measure (M =53.69, SD = 17.02, N = 13) than did those who received Maury therapy (M = 42.69, SD = 1.93, N = 13), t(12.3) = 2.32, p = .039, g = .91.  A 95% confidence interval for the difference in means runs from 1.2 to 20.8 in raw score units and from .09 to 1.71 in standardized units.

    Here is an example for results that fall short of statistical significance.   " Due to the great heterogeneity of variance, a separate variances t test was employed.  Type of therapy did not significantly affect post-treatment wellness (for those in the Rountree program, M = 46.92, SD = 13.38, N = 13; for those in the Maury program, M = 43.23, SD = 3.24, N = 13), t(13.4) = 0.97, p = .35, g = .38."  A 95% confidence interval for the difference in means runs from -11.57 to 4.19 in raw score units and from -.40 to 1.15 in standardized units.

    Deliver your report to me in a Word document named "Yourlastname_Ind_T.doc."   Please do include the underscore character after your last name in the file name.  One of the tables that is produced when using SPSS to conduct an independent sample t test is so wide that it will not fit on a standard page in Word, even in landscape mode.  In earlier releases of SPSS it was possible to copy this table as an image and then resize it to fit, but this is very difficult to do in release 16 of SPSS.  Accordingly, you will solve this problem by telling Word that you are using legal sized paper and printing in landscape mode.  Please read this document for instructions on how to do this.

    At the top of the Word document, following your name, type your summary statement.  After that (at the bottom of the Word document) paste in your SPSS output.

    At the bottom of the Word document paste a screen shot of the NoncT.sav file after you have run the associated syntax file to compute g and a confidence interval for d.  Here is how to do this:

• Enter into NoncT.sav the value of the pooled t (even if you used the separate variances t to test the null hypothesis), the pooled degrees of freedom (n₁ + n₂ - 2), and the values of n₁ and n₂.
• Run the syntax file.
• Delete from the NoncT.sav file the columns "lc2" through "power."
• Resize the window to about the size shown below.
• Hold down the Alt key while you hit the Print Screen Key.
• Place the cursor at the bottom of the Word document.
• Hold down the Ctrl key while you hit the V key.
• Zap, the screen shot is pasted into the Word document, like this:

   Send me a piece of email with the subject line "PSYC 2101:  Independent T Homework."   Attach the word document to that email.  Send it to arrive in my Inbox by no later than 5 PM on Tuesday the 10^th of June, 2008.

If you are using the Internet Explorer, to download the file you should point at your name and then RIGHT CLICK on the mouse. You will get a drop down menu. From that menu select "Save Target As" and then point at the drive/folder where you wish to store the data file.