HAS945- Introduction to Biotatistics

Overview

A data set has been made available through eLearning: go to the HAS 945 Moodle page, then to the section Assessment Overview. There you will find the data in an SPSS data file called surgery.sav. This file contains data for 141 babies who underwent surgery at a pædiatric hospital. The following variables have been included in the data set: 

Part 1: Categorical Variables

1. Open the data set in SPSS. Check that the total number of rows matches what you expected.

2. In the Variable View window, ensure that the Measure has been set correctly for each variable. Provide a list of variables with the appropriate Measures in your report.

3. Recode the variable place to create a new variable, named place2 with label “Place of birth recoded”:

4. Recode the variable gender to create a new variable, named gender2 with label “Gender of baby recoded”:

5. Create a contingency table of the two new variables, place2 and gender2. Include the output in your report.

6. Conduct a χ 2 hypothesis test for independence between place2 and gender2 at the α = 0.10 level of significance. State the relevant null and alternative hypotheses and report your findings. Interpret the p-value obtained from SPSS in the specific context of this dataset.

Part 2: Comparing Means

1. Create a table of Descriptive Statistics for birthwt by prematur. Include this table in your report. Comment on any differences in the statistics between premature babies and those who were born at full term.

2. Create separate histograms of birthwt for premature and full-term babies. Comment on the shapes of these distributions.

3. Create side-by-side boxplots of birthwt by prematur. Comment on any outliers or other unusual features.

4. Conduct an independent samples t test to compare the mean birth weights between premature and full-term babies. State the relevant null and alternative hypotheses and report your findings, using a 5% significance level. Interpret the p-value obtained from SPSS in the specific context of this dataset. 

Part 3: Linear Regression

1. Create a scatterplot of birthwt against gestatio. Comment on any patterns that are evident. Would you expect these variables to be negatively correlated, positively correlated, or uncorrelated?

2. Fit a linear regression model. Explain how you chose which variable should be the Dependent variable and which should be Independent. Include a histogram and normal probability plot of the standardised residuals in your report.

3. Report the test statistics and p-values for all relevant hypothesis tests, using a 5% significance level.

4. Provide 95% confidence intervals for the regression coefficients, β0 and β1.

5. State the mathematical equation for the line of best fit. Interpret these results within the specific context of this dataset. 6. Calculate ˆy for a baby with gestational age of 36 weeks.