soltution
STAT 601 – Assignment #5 (Due Monday, July 2nd )
Review the following:
Lecture Handouts :
10 – Oneway ANOVA – Completely Randomized Block Designs
Narrated Powerpoint Lectures: under heading 11 – parts (a) - (c)
Non-narrated Powerpoint Lectures: under heading 11 – parts (a) – (c)
Please note: part (c) is covered in the Powerpoints for part (a) and (b), i.e. there is no separate Powerpoint for (c).
RESEARCH PAPER QUESTIONS
1 - ANOVA and Multiple Comparisons in Published Research (13 pts.)
The paper “Gestational Diabetes Mellitus Upsets the Proportion of Fatty Acids in Umbilical Arterial but Not Venous Plasma” by Alvino et al. is linked on D2L under the heading “Gestational Diabetes”.
a) Table 1 presents the major results of the study. When comparing the levels of different responses (palmitic acid, stearic acid, total SFAs, etc.) what three populations are being compared using one-way ANOVA? (1 pt.)
b) What multiple comparison procedure did they use for making pairwise comparisons and how did they denote which means were significantly different? (2 pts.)
c) What statistical technique did they use to compare gestational diabetics (GDM) to healthy women (Control)? (1 pt.)
d) For which responses and at which sites were significant differences between gestational diabetics and healthy controls found? (2 pts.)
e) Consider the results for DHA (22:6 n-3) only which are presented below.
When comparing the means across cord blood location for controls or for gestational diabetics one-way ANOVA was used. This is actually an incorrect use of one-way ANOVA, why? (2 pts.)
f) Another potential problem in this study is the experiment-wise error rate. For example, consider the tests used to compare gestational diabetics vs. control on each response (palmitic acid, stearic acid, etc.) at each site (maternal vein, etc.), how many tests like this were done total? How many Type I Errors can they expect to see using an = .05?
Review the following:
Lecture Handouts :
10 – Oneway ANOVA – Completely Randomized Block Designs
Narrated Powerpoint Lectures: under heading 11 – parts (a) - (c)
Non-narrated Powerpoint Lectures: under heading 11 – parts (a) – (c)
Please note: part (c) is covered in the Powerpoints for part (a) and (b), i.e. there is no separate Powerpoint for (c).
RESEARCH PAPER QUESTIONS
1 - ANOVA and Multiple Comparisons in Published Research (13 pts.)
The paper “Gestational Diabetes Mellitus Upsets the Proportion of Fatty Acids in Umbilical Arterial but Not Venous Plasma” by Alvino et al. is linked on D2L under the heading “Gestational Diabetes”.
a) Table 1 presents the major results of the study. When comparing the levels of different responses (palmitic acid, stearic acid, total SFAs, etc.) what three populations are being compared using one-way ANOVA? (1 pt.)
b) What multiple comparison procedure did they use for making pairwise comparisons and how did they denote which means were significantly different? (2 pts.)
c) What statistical technique did they use to compare gestational diabetics (GDM) to healthy women (Control)? (1 pt.)
d) For which responses and at which sites were significant differences between gestational diabetics and healthy controls found? (2 pts.)
e) Consider the results for DHA (22:6 n-3) only which are presented below.
When comparing the means across cord blood location for controls or for gestational diabetics one-way ANOVA was used. This is actually an incorrect use of one-way ANOVA, why? (2 pts.)
f) Another potential problem in this study is the experiment-wise error rate. For example, consider the tests used to compare gestational diabetics vs. control on each response (palmitic acid, stearic acid, etc.) at each site (maternal vein, etc.), how many tests like this were done total? How many Type I Errors can they expect to see using an = .05?