Jose 1
San State University School of Social Work ScWk 242 Spring 2008
Lab Exercise #4: Correlation and Multiple Linear Regression
Research Scenario 1a: Correlation You are a program manager in a large public child welfare agency located in an urban county. The programs you manage employ a total of 40 social workers who provide a variety of case management services to children and families involved in the child welfare system. As a program manager you are particularly interested in reducing employee turnover by preventing employee stress and burnout. You suspect that number of clients on caseload is related to employee stress levels. You ask your employees (N=40), to complete an employee stress survey that measures employee stress levels on a continuous scale with 0 = No stress, and 30 = High stress. The survey also asks employees to indicate the types of stress-reducing activities they regularly engage in, including such things as exercise. In addition the survey also contains demographic questions including race/ethnicity, age and gender.
2
Step 1: Creating the Variables
You will need a participant ID and 2 variables for this exercise: 1) Number of clients on caseload 2) Employee stress level
1) Directions for creating the number of clients on caseload variable:
In the "Name" field, type in the variable name: clients.caseload Use the tab button or the mouse to move the cursor to the "Label" field. In the "Label" field, type: Number of clients on caseload Leave the measurement level in the "Measures" field as scale. 2) Directions for creating the employee stress level variable
In the "Name" field, type in the variable name: stress.level Use the tab button or the mouse to move the cursor to the "Label" field. In the "Label" field, type: Employee stress level Leave the measurement level in the "Measures" field as scale.
3
Step 2: Entering the Data
Click on "Data View" and enter the following data Participant ID 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 Number of clients on caseload 50 54 60 45 42 61 65 63 43 52 51 49 67 39 42 64 54 51 60 46 42 49 66 53 65 54 58 48 54 68 67 41 59 65 41 50 60 65 Employee stress level 19 24 26 15 19 25 24 27 17 20 16 14 26 16 15 24 20 19 27 15 18 17 25 21 27 20 22 17 19 24 27 21 22 24 19 17 20 23
4 39 40 51 50 21 20
Step 3: Running the Correlation and Scatter plots TO GENERATE THE CORRELATION
From the top drop down menu, choose, "Analyze" Choose "Correlate" Choose "Bivariate" In the "Bivariate Correlations" dialog box, place the "number of clients on caseload" variable into the variables box Place the ""employee stress level" variable into the variables box Click OK and your correlation output will appear
TO GENERATE THE SCATTER PLOT
From the top drop down menu, choose "Graphs" Choose "Legacy Dialogs" Choose "Scatter/dot" In the "Scatter/dot" box, choose, "Simple Scatter" Click on "Define"
In the "Simple Scatter Plot" dialog box, click on "employee stress level" and place it in the "Y Axis" box Then, click on "number of clients on caseload" and place it in the "X axis" box o Click "OK" and the scatterplot will appear
5
Using an 8-Step Process for Hypothesis Testing
Once you have your output, use the 8-step process for hypothesis testing to describe and interpret your findings. Please refer to the research scenario and your SPSS output to answer the questions.
1) Identify the independent variable and level of measurement
2. Identify the dependent variable and level of measurement
3. State the null hypothesis
4. State the alternative hypothesis
6
5. Identify the appropriate statistical test, and alpha level
6. Present table of results (SPSS Output)
Correlations Number of clients on caseload 1 Employee stress level .821** .000 40 40 .821** 1 .000 40 40
Number of clients on caseload Employee stress level
Pearson Correlation Sig. (2-tailed) N Pearson Correlation Sig. (2-tailed) N
**. Correlation is significant at the 0.01 level (2-tailed).
7
27.50
25.00
Employee stress level
22.50
20.00
17.50
15.00
30.00
40.00
50.00
60.00
70.00
Number of clients on caseload
7. Describe results and decision to accept or reject the null hypothesis (use APA)
8
8. Provide s discussion of these results Include: Statistical significance Direction of the relationship Meaning and implications of results Limitations/future studies
9
Research Scenario 1b: Multiple Linear Regression Given the significant relationship that was found between caseload and employee stress level, you would like to learn more about the strategies that employees use to effectively cope with stress, including the role of exercise in reducing stress level. In order to help isolate the effect of exercise on employee stress level, you would like to statistically control for the potential confounding factors of number of clients on caseload, race/ethnicity, age, and gender.
Step 1: Creating the Variables
You will need to add 6 new variables to the dataset used in the correlation exercise: 1) Number of hours of exercise a week (continuous) Three dummy variables for ethnicity (Note: White vs. non-White not needed--excluded variable): 2) African American compared to Non-African Americans 3) Latino compared to Non-Latino 4) API compared to Non-API 5) Age 6) Gender: Female compared to male as baseline 1) Directions for creating the number of hours of exercise per week variable In the "Name" field, type in the variable name: exercise.week Use the tab button or the mouse to move the cursor to the "Label" field.
10
In the "Label" field, type: Number of hours of exercise in a week Leave the measurement level in the "Measures" field as scale.
2) Directions for creating the race/ethnicity indicator variables: a. African American compared to Not African American: In the "Name" field, type in the variable name: African.American Use the tab button or the mouse to move the cursor to the "Label" field. In the "Label" field, type: African American compared to Not African American Use the tab button or the mouse to move the cursor to the "Value Labels" o field. In the "Value Labels," click on the gray box and the "Value Labels" dialog box will appear. The Value Labels are:
0 =Not African American 1 = African American
In the "Value" box, type in 0, o Hit tab or use mouse to move cursor to the "Label" box Type: Not African American Click "Add"
11
Use the tab or the mouse to move the cursor to the ""Value" box o Type in 1 Hit tab or use mouse to move cursor to the "Label" box Type: African American o Click Add Then click "OK"
Change the "Measures" field to nominal
b. Latino compared to Not Latino Repeat the same process for the Latino variable using the information below: Variable Name: Latino Variable Label: Latino Value Labels:
0 = Not Latino 1 = Latino
c. Asian Pacific Islander compared to Not API Repeat the same process for the Asian Pacific Islander variable using the information below: Variable Name: Asian.Pacific.Islander Variable Label: Asian Pacific Islander
12 Value Labels:
0 = Not Asian Pacific Islander 1 = Asian Pacific Islander 3) Directions for creating the age variable In the "Name" field, type in the variable name: Age Use the tab button or the mouse to move the cursor to the "Label" field. In the "Label" field, type: Employee age Leave the measurement level in the "Measures" field as scale. 4) Directions for creating the gender variable (using male as baseline) In the "Name" field, type in the variable name: Gender Use the tab button or the mouse to move the cursor to the "Label" field. In the "Label" field, type: Gender: Female Use the tab button or the mouse to move the cursor to the "Value Labels" field. o In the "Value Labels," click on the gray box and the "Value Labels" dialog box will appear. Male is the baseline and is given the value label of "0" The Value Labels are:
0 = Male 1 = Female
13
Change the "Measures" field to nominal
Step 2: Entering the Data
Click on "Data View" and enter the following data
Participant. ID 1.0 2.0 3.0 4.0 5.0 6.0 7.0 8.0 9.0 10.0 11.0 12.0 13.0 14.0 15.0 16.0 17.0 18.0 19.0 20.0 21.0 22.0 23.0 24.0 25.0 26.0 27.0 28.0 29.0 30.0 31.0 32.0 33.0 34.0 35.0 Clients. Caseload 50.0 54.0 60.0 45.0 42.0 61.0 65.0 63.0 43.0 52.0 51.0 49.0 67.0 39.0 42.0 64.0 54.0 51.0 60.0 46.0 42.0 49.0 66.0 53.0 65.0 54.0 58.0 48.0 54.0 68.0 67.0 41.0 59.0 65.0 41.0 Stress. Level 19.0 24.0 26.0 15.0 19.0 25.0 24.0 27.0 17.0 20.0 16.0 14.0 26.0 16.0 15.0 24.0 20.0 19.0 27.0 15.0 18.0 17.0 25.0 21.0 27.0 20.0 22.0 17.0 19.0 24.0 27.0 21.0 22.0 24.0 19.0 Exercise. Week 7.0 2.0 0.0 4.0 10.0 3.0 2.0 0.0 2.0 3.0 6.0 12.0 1.0 5.0 4.0 0.0 10.0 5.0 1.0 17.0 5.0 7.0 1.0 4.0 3.0 3.0 2.0 7.0 8.0 2.0 1.0 5.0 2.0 0.0 7.0 African. American 1.0 1.0 1.0 1.0 1.0 1.0 1.0 1.0 1.0 1.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 Latino API Age Gender
0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 1.0 1.0 1.0 1.0 1.0 1.0 1.0 1.0 1.0 0.0 0.0 0.0 0.0 0.0 0.0
0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 1.0 1.0 1.0 1.0 1.0 1.0
34.0 35.0 27.0 25.0 39.0 45.0 53.0 45.0 32.0 33.0 25.0 27.0 29.0 35.0 36.0 32.0 45.0 44.0 47.0 48.0 21.0 23.0 25.0 61.0 55.0 57.0 43.0 34.0 60.0 45.0 42.0 38.0 37.0 32.0 35.0
0.0 0.0 0.0 1.0 1.0 1.0 1.0 1.0 1.0 1.0 0.0 0.0 0.0 1.0 0.0 1.0 1.0 0.0 1.0 1.0 1.0 0.0 1.0 0.0 0.0 0.0 1.0 1.0 1.0 1.0 0.0 0.0 0.0 1.0 1.0
14 36.0 37.0 38.0 39.0 40.0 50.0 60.0 65.0 51.0 50.0 17.0 20.0 23.0 21.0 20.0 10.0 5.0 0.0 1.0 3.0 1.0 1.0 1.0 1.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 28.0 56.0 51.0 23.0 45.0 1.0 0.0 0.0 1.0 1.0
Step 3: Running the Multiple Regression Test
From the top drop down menu, choose "Analyze" Choose "Regression" Choose "Linear"
In the Linear Regression dialog box, use your cursor to highlight "Employee stress level" and use the arrow to place it in the "Dependent" box. Then highlight the following variables and use the arrow to place them in the "Independent(s)" box: o Number of clients on caseload o Number of hours of exercise a week o African American o Latino o Asian Pacific Islander o Age o Gender: Female compared to male Click OK and your output will appear.
15
1) Identify the independent and control variables and their levels of measurement
2) Identify the dependent variable and the level of measurement
3) State the null hypothesis for the overall model and for each independent or control variable
16
4) State the alternative hypothesis for the overall model and for each independent or control variable
5) Identify appropriate statistical test and alpha level
6) Present table of results (SPSS output)
17
Model Summary Adjusted R Model 1 R .882a R Square .777 Square .729 Std. Error of the Estimate 2.010
a. Predictors: (Constant), Gender: Female compared to male , Latino, Number of hours of exercise per week, Age, API , African-Am, Number of clients on caseload
ANOVAb Model 1 Regression Residual Total Sum of Squares 451.119 129.281 580.400 df 7 32 39 Mean Square 64.446 4.040 F 15.952 Sig. .000a
a. Predictors: (Constant), Gender: Female compared to male , Latino, Number of hours of exercise per week, Age, API , African-Am, Number of clients on caseload b. Dependent Variable: Employee stress level
18
Coefficientsa Standardized Unstandardized Coefficients Model 1 (Constant) Number of clients on caseload Number of hours of exercise per week African-Am Latino API Age Gender: Female compared to male a. Dependent Variable: Employee stress level B 6.456 .254 Std. Error 2.872 .049 .575 Coefficients Beta t 2.248 5.225 Sig. .032 .000
-.358 .496 .152 1.278 .038 .385
.111 .841 .923 1.055 .032 .652
-.350 .062 .017 .120 .110 .050
-3.234 .590 .165 1.212 1.202 .590
.003 .560 .870 .235 .238 .559
7) Describe results and decision to accept or reject the null hypothesis (use APA)
When reporting multiple regression results, report 1) Significance of the overall model 2) Significance of independent or control variables
19
8. Provide a discussion of these results. Include: Statistical significance Direction of the relationships Meaning and implications of results Limitations/future studies