Reference no: EM132285730
RESEARCH METHODS IN PSYCHOLOGY - STATISTICS REGRESSION Data Analysis Assignment
A large marketing company wants to improve the way it assesses candidates for job interviews. So far, it has adopted a battery of questionnaires to measure IQ, Social Skills and Leadership. Based on previous data from a sample of workers, analyses have shown that each of these scales contributes to predict Work Performance (the latter was assessed through a standardized scale). The company is now considering the possibility of revising the battery by examining the role of two potential new scales, one measuring Motivation and the other measuring Metacognition. To this aim, workers are asked to fill questionnaires measuring IQ, Social Skills, Leadership, Motivation, and Metacognition. In addition, Work Performance was measured as dependent variable. All scales were scored so that higher numbers indicated higher levels of the variables.
Before you start, you need to create a sample of data for your own analysis and save it to disk as follows:
Create your own personalized sample of data to set as follows:
1) Download "Regression Coursework Dataset.sav".
2) Open the file in SPSS.
3) Go to Transform, and select Random Number Generator.
4) Under Active Generator Initialization, tick Set Starting Point, and make sure that Random is selected; click on OK.
5) Go to the Data view, and choose Select Cases on the Data menu.
6) Click on the third button down "Random sample of cases" and then click on the Sample button.
7) Click on Approximately and enter 20 for the % of cases to select, then Continue.
8) Set Output to Copy selected cases to a new dataset, and enter a name for the saved dataset (e.g., FrancescoRigoli.SAV). Click OK to create the sample. You should find that it has around 400 cases.
9) Finally, use File - Save As to save the dataset on your drive with your name in the title (e.g. soniadata.sav). Click on Save.
You are now ready to run the analysis. If you need any help with creating your sample, please ask for help.
The assessed work begins here. When asked to include SPSS output in your Word document, select the table or figure, and right click for the menu - Copy Special - then make sure Image is ticked. You will then be able to paste the material into Word using right-click and "Keep Source Formatting". (This setting stays for the session once set).
TASKS - Create a Word document to answer the following:
a) The main research question is whether adding the new scales (Motivation and Metacognition) to the previous battery (including IQ, Social Skills, Leadership) helps predicting Work Performance. Run the appropriate Regression analysis in SPSS to answer this question. Indicate which Regression method you have chosen. Describe briefly this method and explain why you have chosen it.
b) Based on your SPSS analysis above, Copy and Paste the following three Tables into WORD:
- Model Summary
- ANOVA
- Coefficients
c) Report all the main statistics that describe the performance of the Regression model as a whole (as reported in the SPSS tables). Briefly describe the meaning of each statistic. If you have estimated more than one model, answer this question for the model estimated last.
d) Based on your analysis, can you infer that (in the population) the model is useful for predicting Work Performance? And why? If you have estimated more than one model, answer this question for the model estimated last. In answering this question:
- Explain what the null hypothesis is;
- Explain what the experimental hypothesis is;
- Indicate which statistic is adopted for this hypothesis testing. Report the value of this statistic and explain how this statistic is calculated (relying both on the general theory and on the information provided by SPSS for this specific dataset). Also, explain why this statistic is appropriate for testing the null hypothesis;
- Report one or two statements which describe the results of this hypothesis testing using the appropriate format (use alpha threshold = 0.001);
e) Considering each predictor individually, can you infer that, in the population, that predictor contributes to predict Work Performance beyond other predictors in the model? Answer this question for each predictor. If you have estimated more than one model, answer this question for the model estimated last. In answering this question:
- Explain what the null hypothesis is;
- Explain what the experimental hypothesis is;
- Indicate which statistic is adopted for this hypothesis testing. Report the value of this statistic (for each predictor) and explain how this statistic is calculated (relying both on the general theory and on the information provided by SPSS for this specific dataset). Also, explain why this statistic is appropriate for testing the null hypothesis;
- Report one or two statements which describe the results of this hypothesis testing using the appropriate format (use alpha threshold = 0.001);
f) Write the equation for predicted Work Performance (if you estimated more than one model, do this based on the model estimated last).
g) Rank all predictors based on how much each influences Work Performance. Describe the meaning of the statistic you are using for this judgement and report this statistic for each predictor. If you have estimated more than one model, answer this question only for the model estimated last.
h) Use SPSS to look at the correlation between all pairs of variables. Copy and Paste the obtained table into word.
- Is there any predictor which does not contribute to the Regression Model (based on the analysis for point e above) although it shows a significant Pearson correlation with Work Performance (use alpha threshold = 0.001)? If any, report the results of the correlation hypothesis testing for this predictor (using the appropriate format discussed during lectures). Why a predictor with significant Pearson correlation can be eventually excluded from a Regression Model?
- Consider predictors that contribute to the Regression model (based on the analysis for point e above). Is there any predictor showing a significant Pearson correlation with Work Performance (use alpha threshold = 0.001) for which the correlation has opposite sign to the regression coefficient? If any, report the results of the correlation hypothesis testing for this predictor (using the appropriate format discussed during lectures). Why these apparently puzzling results emerge in this dataset?
i) Double-check whether the assumptions of Regression are met for your Regression analysis. Also produce graphs when appropriate. If you have estimated more than one model, answer this question only for the model estimated last. When analyzing each assumption:
- Describe the assumption you are considering.
- State whether the assumption is met and motivate your statement.
- When evaluating an assumption based on graphs, please rely on visual inspection. Based on visual inspection, please explain why you think an assumption is met or not.
j) Finally, consider the main question of the research study. Does adding the new scales (Metacognition and Motivation) to the previous battery (including IQ, Social Skills, Leadership) help predicting Work Performance? How much? Motivate your answer and report the appropriate statistics produced by SPSS (using the format discussed during lectures). In conclusion, based on your Regression analysis, would you suggest the company to revise the assessment method? How?
Attachment:- Assignment Files.rar