Question:

Using Polit2SetB data set, create a correlation matrix using the following variables: Number of visits to the doctor in the past 12 months (docvisit), body mass index (bmi), Physical Health component subscale (sf12phys), and Mental Health component subscale (sf12ment). Run means and descriptives for each variable, as well as the correlation matrix.

Week 6 â€“ Correlation Exercises

Part I

The aim of this study is to find the influence of maintaining good physical and mental health on the frequency of medical interventions sought. Totally, 997 people were considered for the study, from the dataset Polit2SetB of SPSS. The concept of good health is measured by analyzing the study population in 3 groups, analyzing BMI 29.22(7.37), Physical Health Component Subscale 45.11(10.84), and Mental Health Component Subscale 46.82(10.80). All the 3 groups were analysed for analyzing the number of visits to the doctor in the past year 6.80(12.7)

Null Hypothesis â€“ There is no influence of maintaining good physical and mental health on the frequency of medical interventions sought

Alternate Hypothesis â€“ There is a statistically significant influence of maintaining good physical and mental health on the frequency of medical interventions sought

A correlation matrix was computed, using Pearsonâ€™s bivariate model of correlation, to find the relation between the variables. (Mukaka, 2012)

There are totally 6 original correlations in the matrix, calculated by the formula {N*(N-1)}/2, where N = 4. The diagonal column has a perfect coefficient value of 1, as each variable is correlates perfectly with itself.

The strongest correlation in this matrix is the relation between the Physical Health Component Subscale and the number of doctor visits in the past year

Coefficient of Correlation â€“ -0.316
Strength of Correlation â€“ Moderately strong
Direction of Correlation - Negative
Statistical Significance â€“ Highly significant (0.000)

This implies that the number of doctor visits in the past year were lower with higher scores in the Physical Health Component Subscale. The variable that has the strongest coefficient of correlation with BMI is the Physical Health Component Subscale, with the sample size of 890.

The weakest correlation in the matrix is the relation between the Mental Health Component Subscale and the Body Mass Index.

Coefficient of Correlation â€“ -0.078
Strength of Correlation â€“ Weak
Direction of Correlation - Negative
Statistical Significance â€“ Insignificant (0.022)