Statistical Testing to Identify Differences Between Groups Using t-tests and ANOVA on Framingham Study Data
Overview
This paper examines statistical methods for testing differences between groups using the Framingham study dataset and SPSS software. The analysis focuses on testing assumptions of normal distribution across four BMI categories (underweight, normal, overweight, and obese) using Kolmogorov-Smirnov and Shapiro-Wilk tests, with Q-Q plots for graphical confirmation of normality.
What this coursework covers
- Introduction to t-tests and ANOVA for group comparison
- Testing normality assumptions across BMI categories using Kolmogorov-Smirnov and Shapiro-Wilk tests
- Analysis of baseline glucose levels stratified by BMI classification with case processing summaries and Q-Q plots
- Discussion of statistical test selection and interpretation of normality results
Preview
Testing to Find Differences Between Groups
Helena Lynch
Capella University
MPH5509
Dr. Cash
November 20, 2020
The t-test and NOVA materials provided helped in examining how to select a statistical test. Therefore, the SPSS and Framingham study data set was used to perform and interpret the statistical tests that provide answers to the following research questions.
Testing for The Assumption of Normal Distribution
The following is a test of nominal distribution of the data.
- BMI (Underweight<18.5)
From the tests for normal distribution, we established that someone was underweight when their BMI was less than 18.5. Moreover, the baseline glucose level for underweight people was 0.26 based on Kolmogorov-Smirnov test. The K-S tests usually compares two samples and determines how far they are from each other. It is a test for normality.
Case Processing Summarya | ||||||
Cases | ||||||
Valid | Missing | Total | ||||
N | Percent | N | Percent | N | Percent | |
Baseline glucose level | 2 | 100.0% | 0 | 0.0% | 2 | 100.0% |
a. Body Max Index = Underweight <18.5 | ||||||
Tests of Normalitya | |||
Kolmogorov-Smirnovb | |||
Statistic | df | Sig. | |
Baseline glucose level | .260 | 2 | . |
a. Body Max Index = Underweight <18.5 | |||
b. Lilliefors Significance Correction | |||
The quantile-quantile (q-q) is usually a graphic representation that is utilized to establish whether a data set is sampled from a population with common distribution. This is confirmed when there is a straight line formed in the plot. In the Q-Q plot show below, we can argue that the data was sourced from the same population and has a normal distribution.
- BMI (Normal 18.5 -25))
In this analysis, we undertook test for people with normal weight. We established that baseline glucose level was 0.