Alan Mangus Math April 15, 2012
Purpose of the study (the research question): Can the age of adult male humans be used as a reliable predictor of the weight of adult males? Age will be the explanatory variable and weight will be the response variable. Study design: I will be using an observational cross-sectional approach.
Data collection: Since I do not have a population frame, I will be using Systematic Sampling. K = 10 with a random seed of =14, 14+10=24, 24+10=34 and so until a sample size of n=20 was reached. The survey was given to customers of Smith’s Market Place on March 24, 2012 in Salt Lake City, Utah.
Age of Adult MaleWeight (in lbs) of Adult Male
n = 20, ∑x = 70 Mean = 740/20 = 37 Standard Deviation = n = 20, ∑x = 3400 Mean = 3400/20 = 170 Standard Deviation = MinimumQ1MedianQ3MaximumMinimumQ1MedianQ3Maximum Range = = 45Range = 222 – 137 = 85 IQR = 42.5 – 27 = 15.5IQR = – 152 = 37.5 Mode = 27, 31, 39Mode = 154
Y = ax +b a = b = r = y = (x) Critical Value for Correlation Coefficient For n = 20 is r is less than is less than = No Linear Correlation
Distribution of Histogram for age is skewed right. Distribution of histogram for weight is skewed right. Scatter diagram reveals almost no correlation between the two variables. Furthermore the critical value for a sample size of 20 is The correlation coefficient for this survey was 0.137, well below the value needed for a positive correlation.
With such a low correlation between the explanatory variable and the response variable age would not be a very reliable predictor of weight in adult male humans. A more reasonable predictor might be height; which is why it is used in the Body Mass Index formula (BMI).