The Relationship Between Multiple Measures of Asthma Diseases' Status Over Time

Assessment of the Relationship between Multiple Measures of Asthma Diseases' Status over Time Using Bivariate and Multivariate Procedures
Name
Institution
Assessment of the Relationship between Multiple Measures of Asthma Diseases' Status over Time Using Bivariate and Multivariate Procedures
Introduction
This study was conducted upon a randomized sample of 119 disadvantaged inner-city children aged between 5-12 years. All of the children were suffering from moderate severe asthma. The participants live in/around San Francisco and San Jose- California. They are all characterized by low income families.
The study is to assess the relationship between multiple measures of asthma diseases' status over time. The time frame takes a range of 52 weeks measured at baseline, 32 weeks and 52 weeks.
The measures taken in the assessment are; Diary Data, Pulmonary Function Tests, CHSA Spirometry (child health survey for asthma) data, Parent-Reported Disease Symptoms and Parent-Reported Health Care Utilization (Sharek, et al., 2002). Since there is no standard measure for asthma outcomes, we have to analyse (statistically) in order to select the measures with the highest correlation.
Bivariate Analyses
There are three types of bivariate analysis:
1 Numerical and numerical
2 Categorical and categorical
3 Categorical and numerical
In this study, all variables have a numerical value and the correlation coefficients between each set of variables have been calculated. This means that the appropriate type of bivariate analysis is "numerical and numerical".
In this study, there are four possible effective "numerical and numerical" measures i.e. scatter diagrams, line of best fit, charts and correlation analysis.
Scatter diagrams are effective when you have individual occurrences. However, since we don't have individual occurrences for each kid, we could analyse the correlation coefficients given in table 2.0 (baseline), table 3.0 (32 week) and table 4.0 (52week). With the correlation (r) given, we could perform Z-Tests and T-Tests on the data to analyse the statistical difference between the three variables (baseline, 32 week and 52 week). This will enable us to know whether each data characteristic can be relied upon for effective results. This means measuring the differences in central tendency and comparing the ratio variables of each data set upon a time frame. If the statistical difference is too big, it means that the method is unreliable and volatile over time.
Alternatively, the correlation coefficients given could be plotted on a scatter diagram. The three data sets for the three periods (baseline, 32 week and 52 week) could be assigned different colors (Varmuza & Filmoser, 2008, p. 8). Next, they are plotted on a Cartesian plane to show the density of correla...