Bio statistics
(12 pnts) Part 1: C Reactive Protein is a marker for inflammation. The CRP data used in this homework assignment come from a study investigating an intervention to improve caregiver’s attention to oral health in patients residing in a long term care facility in Manitoba, Canada. Data include baseline CRP and 6 week CRP. You will be using the BASELINE CRP data for this assignment.
The purpose of this exercise is to have you take a set of non-normally distributed data and transform it to make it normal. The output that you will create duplicates that which was presented in the PP/Panapto Transformation presentation, slides 44-57 but using this data set. You will also perform tests as presented in the Normality PP/Panapto presentation.
You will create all of the required output and then copy and paste it into the HW word document. If you have problems copying and pasting, right click on the graph or table and then click on Copy Special. Click on Image (JPG, PNG) and unselect all other options. This will give you the option to directly copy the picture into your work document. Additionally, I must be able to read your tables and graphs. Please size them appropriately once they are in the HW document.
The order of the graphs and tables pasted into the home work from the output is as follows:
Baseline CRP Histogram, Log10 Transformed CRP Histogram, Baseline CRP PP Plot, Log10 Transformed CRP PP plot, Baseline CRP QQ Plot, Log10 Transformed CRP QQ plot, Baseline CRP Box and Whisker plot, Log10 Transformed CRP Box and Whisker plot, Baseline Tests of Normality, Log10 Transformed Tests of Normality, Statistics Table.
These tables and graphs must be presented in this order. Please proof read your HW assignment before you upload it into BB.
Do not give me all of your output. I will not grade output.
Please be sure to type your answers in a font color that is NOT black.
1. Open CRP Skewed Data.sav
2. Go to Variable View
a. For variable Baseline_CRP type in the Label “Your name” Baseline CRP. You must have your own name appear in the Label to receive points for this exercise.
b. Once you have done this, return to Data View.
3. Create a graph of Baseline_CRP
a. Analyze>>Descriptive Statistics>>Frequencies
b. Click Baseline_CRP to Variables Box
c. Click Charts
d. Click Histogram and Show Normal Curve on Histogram
e. Click Continue
f. Click OK
4. Transform the data using Log10 as per the directions in the practice directions.
a. Name this new variable Log10CRP
b. Label this variable “Your name” Log10 CRP. Your Label MUST include your own name. If the label does not include your name you will not receive points for this exercise.
5. Create a graph of this newly transformed variable as in Step 3.
6. Create the Normal P-P Plots of Baseline CRP and Log10 CRP
a. Analyze>>Descriptive Statistics>>P-P Plots…
b. Click the variable to be plotted to the Variables box
c. Click OK
7. Calculate the output for Skewness and Kurtosis
a. Analyze>>Descriptives>>Frequencies>>
b. In the Frequencies box Click Baseline_CRP and Log10CRP to the Variables box
c. Check the box Display frequency tables
d. Click on Statistics
e. Click on Mean, Median, Mode, Std. Deviation, Variance, Skewness and Kurtosis
f. Click Continue
g. Click OK
8. Run the Kolmogorov-Smirnov and Shapiro-Wilkes tests to assess the data for normality. You will also obtain the Normal Q-Q plots and the Boxplots while doing this.
a. Analyze>>Descriptive>>Explore
b. Click Baseline_CRP and Log10CRP to the Dependent List
c. Under Display, the bubble beside Both should be clicked on
d. Click on Plots
e. Check the box for Normality Plots with tests
f. Click continue
g. Click OK
9. Calculate the Z-scores for Skewness and Kurtosis (page 184 in the Field text, or found in the Normality lecture.) Show your work.
a. Compare these calculated scores against the 99% Z-score (2.58)
i. Prior to transformation, was there skewness and/or kurtosis present?
ii. Following transformation, was there skewness and/or kurtosis present?
10. Make a final assessment of this transformation using all of the pieces of information obtained from all of these tests.
a. Did the tests and graphs show normalization of the data following transformation?
i. You must assess all transformed graphs and tests individually and indicate which found normality and which did not find normality.
b. Assess the Measures of Central Tendency for both Baseline CRP and Log10 Transformed CRP data. What do these measures, assessed together, say about normality of the data for Baseline CRP and for Log10 Transformed CRP data?
c. The overall decision: Did the transformation normalize the data? (Yes or No)
d. Would you use a parametric statistical analysis method to analyze this transformed data? (Yes or No)
i. Why or Why not?
**Many of the procedures provide additional output that you already may have. There are often many ways to generate the desired output from SPSS.
The purpose of running all of the processes as done above is to provide you with different approaches.
Parts 2, and 3:
(3 pnt) Part 2: Please write 3 questions that you would ask from this lecture. Your questions should include both the question and the correct answer with an explanation. These questions should reflect your own uncertainty as you progressed through the lesson.
(5 pnts) Part 3: Identifying data types is often very clear cut. It is easy to identify continuous data and discrete data. It is usually very easy to tell the difference between continuous and discrete data. Because parametric statistical analysis methods are used to analyze normally distributed continuous data and non-parametric statistical techniques are used to analyze discrete data, data type identification is very important. So…Historically, IQ score has been analyzed with both Parametric and non-parametric analysis methods. Additionally, IQ Score has been identified as being either discrete or continuous. Your challenge is to decide if IQ Score is Discrete or Continuous. You must select only one. It cannot be both. You will write a one paragraph (no longer than one page typewritten) paper supporting your identification of the data type of IQ score. State your reasons for your choice. You must back this up with references.
:
