Complete the proceeddures in the attached file. Copy and paste the outputs onto a word document. Then answer the 12 questions listed using the outputs as a reference. Please read the instructions on the attached file carefully
PSYC 207 Assignment #4 (22.5 marks) Establishing significant associations (correlations) and Linear Regression INTRODUCTION Significant relationships/associations What if the researcher is interested in assessing the relationship between two variables? For example, there might be some interest in assessing the level of income with food security. What is the association there? Do people with higher annual salaries have higher food security? Or maybe the interest is in the relationship between exercise and happiness. Does the amount influence the level of happiness (e.g. more exercise = greater happiness)? When we assess relationships between variables, one of the statistical instruments we can use is the correlation coefficient. The type of variable measure you have determines the type of appropriate correlation coefficient to use. For example, assessing the association between two variables where the measure is continuous (e.g. interval/ratio), then a Pearson Product-moment correlation coefficient is used. Assessing the association between two variables where the variable is an ordinal measure, then a Spearman correlation coefficient is used. Tasks For this assignment you will using SPSS to assess differences and associations between variables, and then interpreting the findings. OUTPUTS Save the following outputs for Parts 1-2 by either using the export function or copy and paste to MS Word Part 1 Associations between variables (Correlation) For the correlation part of the assignment access the file ‘PSYC_’Bodner_Sleep_Income SP2021’ from Moodle. 1. Height and weight We are interested in establishing an association between height and weight. Our research hypothesis is that there is a positive relationship between height and weight. In other words, as someone gets taller their weight also increases. Procedure 1. Under Analyze Correlate Bivariate 2. For the Dependent List choose select the variables ‘HEIGHT’ and ‘WEIGHT’ and move to ‘Variables’ 3. Under ‘Correlation Coefficients’ select ‘Pearson’ 4. Under ‘Test of Significance’ select ‘Two-tailed’ 5. Make sure ‘Flag Significant Correlations’ is checked 6. Select ‘Options’. Check the ‘Means and Standard Deviations’ box 7. Select ‘OK’ 8. Save this output. Use the sub-title ‘Height and Weight Association’ to describe this output. 2. Age and weight We are interested in establishing an association between age and weight. Our research hypothesis is that there is a positive relationship between height and weight. In other words, as someone gets older their weight also increases. Procedure 1. Under Analyze Correlate Bivariate 2. For the Dependent List choose select the variables ‘AGE’ and ‘WEIGHT’ and move to ‘Variables’ 3. Under ‘Correlation Coefficients’ select ‘Pearson’ 4. Under ‘Test of Significance’ select ‘Two-tailed’ 5. Make sure ‘Flag Significant Correlations’ is checked 6. Select ‘Options’. Check the ‘Means and Standard Deviations’ box 7. Select ‘OK’ 8. Save this output. Use the sub-title ‘Age and Weight Association’ to describe this output. 3. Education and Income We are interested in establishing an association between education and income. Our research hypothesis is that there is a positive association between education and income. In other words, the higher the education the higher the income. Procedure We are comparing a continuous variable (INCOME) with an ordinal variable (EDUCATION). What we can do in some instances is recode our INCOME variable into an ordinal variable. Step 1 Recode INCOME Create a new variable wherein we recode Income into a class interval. Let’s say that someone earning an income between $10000- 39999 = 1 $40000- 69999 = 2 $70000-120000 = 3 See if you can remember how to recode like we did before on your own, but if you forget here are the steps: 1. Go to Transform menu Recode into Different Variables 2. Select ‘INCOME’ from the list of variables and move it into the Input Variable Output Variable box. 3. Select the name ‘INCOME_REC’ as your new variable and type it into the Output Variable box. 4. Under ‘Label’ type in ‘Recoded Income’ 5. Click the Change button 6. Select ‘Old and New Values’ button. This opens the Recode into Different Variables: Old and New Values dialogue box. 7. Select the first ‘Range; button under the Old Value dialogue box on the left; enter 10000 in the first box and 39999 in the second. Enter the value ‘1’ in the New Value box on the right. Click on Add. This creates an income range of $10000-39999 with a value of 1. 8. Do this again, but this time enter 40000 in the first box and 69999 in the second. Enter the value ‘2’ in the New Value box on the right. Click on Add. This creates an income range of $40000-69999 with a value of 2. 9. Do this again, but this time enter 70000 in the first box and 120000 in the second. Enter the value ‘3’ in the New Value box on the right. Click on Add. This creates an income range of $70000-120000 with a value of 3. 10. Be sure to label what each number means. Step 2 Assess association 1. Under Analyze Correlate Bivariate 2. For the Dependent List choose select the variables ‘EDUCATION’ and ‘INCOME_REC’ and move to ‘Variables’ 3. Under ‘Correlation Coefficients’ select ‘Spearman’ 4. Under ‘Test of Significance’ select ‘Two-tailed’ 5. Make sure ‘Flag Significant Correlations’ box is checked 6. Select ‘OK’ 7. Save this output. Use sub-heading ‘Education and Income Association’ to describe. Part 2 Linear Regression While correlation is helpful for showing the relationship between two variables (and in part, some explanation), in the real world there is usually more than one factor that explains things we observe. For example, obesity is considered to be a health issue in North America. Obesity is complex, involving multiple factors. One of the key measures for population level obesity is Body Mass Index (BMI), which is a ratio of height and weight with ranges between <18 (low="" bmi)="" to="">=30 (high BMI). Two factors that might contribute to obesity are calories burned per week (KCAL) and age. Recent research also points to skipping breakfast as a potential factor in weight gain. We want to know if each of these independent factors (KCAL, AGE and BREAKFAST) explain changes in BMI, independently and then together. Independently we can run these as separate correlations (Pearson); together we have to use Regression analysis. Regression analysis will let us know if each of these variables is significant to explain changes in BMI, and if so, which one has greater explanatory power. Your task will be to perform an independent correlation of each factor (associate KCAL and AGE with BMI), run a regression analysis with both KCAL and AGE in the model, and then interpret the output. PROCEDURE For the correlation part of the assignment access the file BMI_Regression_SP2021 Linear Regression 1. Under Analyze Regression Linear 2. In the Dialog Box, for the Dependent variable select ‘BMI’’ and move to ‘Variables’ 3. For the Independent variable, select ‘KCAL’, ‘AGE’ and ‘Breakfast’ 4. In the Statistics button, select the following: Estimates, Confidence Intervals, Model Fit, Descriptives, R-Squared Change, then press ‘Continue’ 5. Press ‘OK’ 6. Save this output (Title: BMI Regression) Hand in: 1. OUTPUTS (5 marks) - please ensure that the outputs have sub-titles to identify each Part 1 ‘Height and Weight Association’ SPSS output (1 mark) ‘Age and Weight Association’ SPSS output (1 mark) ‘Education and Income Association’ SPSS output (1 mark) Linear Regression SPSS output (2 marks) 2. QUESTIONS (17.5 marks) Please answer in full sentences. Use data from your outputs to buttress your answers (where needed). Please type your answers below the questions in bold font, please. Part 1 Correlation Review your findings from your data outputs for HEIGHT/WEIGHT and EDUCATION/INCOME associations. 1. Write out a statistical conclusion for each of your correlation findings (1 mark). How strong are the associations for each of the paired variables? Explain your answer (2 marks) 2. Calculate how much shared variance (%) there is between HEIGHT and WEIGHT. What does it menan for two variable ‘share variance’? (2 marks) 3. What can you conclude from your findings between EDUCATION and INCOME? (1 mark) 4. Can you actually conclude that increased education causes increased salary? Why or why not? (1.5 marks) Part 2 Linear Regression 5. Was the regression model significant? How do you know? Explain. (1 mark) 6. Review the output of ‘Model Summary’. If the model was significant describe the model fit. (1 mark) 7. Refer to the coefficients in the model. Which ones were statistically significant? Explain how you know this. (1 mark) 8. Write out a statistical conclusion for your regression analysis (2 marks) 9. Write out a research conclusion for your regression analysis (1 mark) 10. Of the B coefficients that were significant, explain how they can be interpreted with respect to the change in BMI. (2 marks) 11. Which one of these variables has a stronger effect on BMI? Explain why you believe so. (1 mark) 12. What is the function of multiple regression? In other words, why would a researcher use it? (1 mark) 1