STAS2126 Assignment#9 (Fall 2008)
Submit to Dropbox not by email…Use your first and last name as the file name… Do NOT use # symbol as part of
the file name…Show your work for calculation questions not just final answer or you will lose marks
Testing a Family Intervention Hypothesis: Kangaroo Care
Providing motherinfant body contact during a period of maternal separation was examined for its
effects on parent–infant interactions. Participants were 146 threemonthold premature infants and
their parents.
Half of them received skintoskin contact, or kangaroo care (KC), while the
remainder were in a control group. Interactions between parents and infants were observed
during home visits. Actual results from this study are shown below.
1a
What is the independent variable?
Kangaroo care vs control)
1b
How many dependent variables are listed in the above table?
14
1c
Which dependent variable shows the most significant results? (see F
values)
Father parent intrusiveness (F =15.84)
2.
Calculate Cohen’s d for parent intrusiveness (motherinfant interactions).
Avg SD = (1.0 +0.91)/2 = 0.955
Cohen’s d = (2.532.02)/ 0.955 = .534
3.
If we use a ttest to compare the two groups we would produce a tvalue instead of F value. As indicated on your
Formulas page, t
2
= F.
Therefore, if we know the F value we can calculate the t value and vice versa. Use the F value
given in the table above for “Family cohesiveness” to calculate what the tvalue would be.
t value = √F = √10.33 =3.214
4.
For one of the above tests p=.03. What will be the Type I error rate (or likelihood that the result is due to chance)?
The Type 1 error rate is equal to the p value. Therefore, if the p value is .03, then the Type I error rate is 3%.
5.
Table 2 shows results for many different variables. Having such a large number of statistical tests in one study has a
large impact on
experimentwise
error rate. Briefly explain what this means (consult your course notes)
The more significance tests that are performed, the greater the chance of making a Type I error (believing we
have found a real significant difference when in fact it was just due to chance). The probability that at least one of
these tests will be a Type I error is called experimenterwise error rate (EER).
QUESTIONS BELOW RELATE TO PARENT SENSITIVITY (FOR MOTHER) IN ABOVE TABLE. SHOW WORK.
This preview has intentionally blurred sections. Sign up to view the full version.
View Full Document
This is the end of the preview.
Sign up
to
access the rest of the document.
 Fall '08
 F.Sicoly
 Regression Analysis, Cohen, Sig., family cohesiveness

Click to edit the document details