In this study we are analyzing the number of hospital beds per 1k population in Florida and trying to come up with some interesting findings which will help com-paring Florida with rest of the US and if possible will lead to a predictive model showing how different driving factors are related with the number of hospital beds per 1k population in Florida.
Summary Statistics and Explanation
• Number of hospital beds per 1k population has been calculated taking the ratio of Total number of hospital beds in a country to the total popu-lation of that country and then multiplying by 1000.
• For the basic comparison of Florida with rest of the nation, we took aver-age number of beds per 1k across all the country belonging to the state and plotted …show more content…
them in a chart for comparison.1
• The red bar in the chart denotes the average number of beds per 1k for Florida which is ~ 3. There are few states across the nation where this count is as high as 10, on the other hand few states have average num-ber of beds per 1k as low as 2.
Hospital Beds in Florida and the Nation
• We carried out hypothesis testing to find out whether there is any signifi-cant difference between the numbers of hospital beds per 1k in Florida and rest of the world.
• Our null hypothesis for this test was there is no difference between the numbers of hospital beds per 1k in Florida and rest of the world.
• Alternative hypothesis for this is that there are significant difference be-tween the average numbers of hospital beds per 1k in Florida and rest of the world.
• Firstly we performed F-test2 to check whether variance of these two groups are different or not, and we found that there are significant differ-ences between two variances.
• As our next step, we carried out a t-test3 to figure out whether there is any significant difference between two groups or not.
• In this test we found out that average number of bed per 1k in Florida is 2.62 and the same for rest of the nation is 3.52. Though it seems there is some difference in these two numbers, actually these measures are not significantly different with 95% confidence.
Predicting Florida’s Hospital Bed Density
• We started building our predictive model for Florida with 13 analysis vari-able as independent variables and beds per 1k as dependent
variable.
• As the first step of our modelling we tried to figure out whether there are any multicollinearity among the independent variables or not. The correla-tion matrix4 shows how all variables are related. We chose 90% as cut off and removed all the variables having high correlation. After this step we are left out with 4 variables as Population, Income, %Poverty, <9 school.
• Next we performed backward stepwise regression5 to find out significant variables with 95% confidence. But none of the variables came out to be significant.
We can conclude that reliable model cannot be built for beds per 1k for Florida
Conclusions
• We can conclude that number of beds per 1k does not differ significantly between Florida and rest of the country.
• Predictive model cannot be built for number of beds per 1k in Florida.