WRITE-UP: BIVARIATE REGRESSION Null Hypothesis Scenario: The purpose of this study was to see if the number of nurses employed at a hospital Ho: There is no significant relationship between the independent variable (number of nurses could predict patient satisfaction. Researchers examined 20 hospitals similar in size and location. The amount of nurses working full-time for the hospitals were gathered from HR and patient employed) and the dependent variable (patient satisfaction) for a group of similar hospitals. satisfaction was averaged from surveys gathered over the past year. The patient satisfaction Data Screening survey ranges from 1 to 10 with 1 being not satisfied and 10 being extremely satisfied. The researcher sorted the data and scanned for inconsistencies on each variable. No data errors or Code Number of Patient Nurses Satisfaction inconsistencies were identified. A scatter plot was used to detect bivariate outliers between the 1 201 7.0 2 189 6.3 independent variable and the dependent variable. No bivariate outliers where identified. See Figure 1 for 3 249 8.4 the scatter plots. 14 211 7.1 15 200 7.4 Figure 1 6 192 6.5 7 160 5.4 Simple Scatter Plot 8 214 8.3 9 198 7,1 R2 Linear = 0.852 10 10 186 6.0 11 270 9.0 12 217 7.3 13 175 6.3 14 301 91 15 250 8.3 16 157 5.3 =1.26+0.03*x 17 198 7.0 Satisfaction 18 212 7.4 19 215 8.2 20 195 7.1 Overview The purpose of this study was to see if the number of nurses employed could predict . . Research Question 150 200 250 300 350 RQ: Is there a significant predictive relationship between the independent variable Nurses (number of nurses employed) and the dependent variable (patient satisfaction) for a group of similar hospitals?Descriptive Statistics Descriptive statistics were obtained on each of the Variables. The sample consisted onO participants. The number ofnurses at each hospital was obtained through HR. Patient satisfaction scores were obtained through averaging surveys obtained throughout the past year. Patimt satisfaction scores could range from 1 (not satised at all) to 10 (extremely satised). Descriptive statistics can be found in Table 1. Table 1 Descriptive Statistics Mnm'iu Maximo Std. N m m Mean Deviation Nurses 19 157 301 210.11 36.149 Satisfaction 19 5 9 7.23 1.113 ValidN 19 (listwise) Assumption Testing Assumption of Linearity The multiple regression requires that the assumption of linearity he met. Linearin was examined using a scatter plot. The assumption oflinearity was met. See Figure l for the brvariete sea plot. Assumption of Bivariate Normal Distribution The bivariate reyession requires that the assumption of bivariete normal distribution be met. The assumption ofbivaria'te normal distribution was examined using a scatter plot. The assumption ofbix-ariate normal distribution was met. See Figure l for scatter plot. Results A bivariate regression was conducted to see if the number of nurses employed at a hospital could predict patient satisfaction. The independent variable was the number of nurses employed. The dependent variable was patient satisfaction scores as measured by thePatient Satisfaction Sin-trey. The researcher rejectedthe null hypothesis at the 95% condence level wherelJ?) = 97.59,}: = .0001. There was a statistical relationship mm independent variable (it of nurses) and the dependent variable (patient satisfaction scores). See Table 3 for regression model results. Table 2 Regression indel Results Sum of Mean Model Squares gt Square F Sig. 1 Regression 18.984 1 18.934 97.888 <.001 residual .194 total a. dependent variable: satisfaction b. predictors: nurses the model effect size was extremely large where r=".923." furthermore r7: .852 indicating that approximately ofthe variance ofdependent variable canbe explained by independent efnurses variable. see table formodel summary. rue dei summary adjusted r. std. error of square estimate .923 .843 .440>