Reference no: EM132303000

SPSS Analysis Assignment - 

This assignment consists of a series of tasks that generally should be completed sequentially. This assignment must be completed individually. In discussing the assignment with your peers (which should be kept to a minimum), your enquiries should be general in nature (about SPSS and not about the assignment questions) and should never reveal the solution to the assignment tasks.

Information and Tasks -

Data Information/Background -

For this assignment, there are two datasets that you will be using: Dataset 1 (SPSS_Assign_Pt2_Dataset1.sav) - You will use this dataset for Tasks 1 to 3. This is similar data to that which was used in the Part 1 assignment (important: you must use this new dataset, NOT your dataset from the Part 1 assignment).

The table below provides information on the variables in this dataset. The data was collated by a not-for-profit organisation named Gapminder, who aim to promote sustainable global development by increased use and understanding of statistics.

While the data is real, the experimental scenarios and research questions herein may not be factual and may have been manipulated for educational purposes.

Note: for the purpose of this assignment, id is not considered a variable, but is instead simply a unique country identification number.

Dataset 2 (SPSS_Assign_Pt2_Dataset2.sav) - You will use this dataset for Tasks 4 to 10.

You should save any changes you make to the dataset and output files as you go along. You will need to submit your final datasets (file extension .sav) at the end of this assignment. Your output (tables and graphs) files directly in SPSS (.spv extension) will also need to be submitted, but will not necessarily be marked (they're there for verification purposes only). The task descriptions below will indicate which tables and/or graphs are necessary for you to include in each section of your assignment document for your submission.

For Tasks 1 to 3, you must use SPSS_Assign_Pt2_Dataset1.sav -

Over tasks 1 to 3, you will be examining the relationship between hemisphere and life expectancy in 2007. To do this, you will be performing an independent t-test (ITT) with the hemisphere and lifeExp07 variables, and specifically investigating the research question: "Do countries in the Northern hemisphere have a higher life expectancy than those in the Southern hemisphere?" For the following tasks, use a one-tailed α = 0.05.

Task 1 - Independent t-tests (Thinking About the Research Question)

Answer these questions succinctly (use the following numbering in your submission; you must not re-write the question):

1. What is the independent variable (IV)? What type of variable (quantitative or qualitative) is the IV? If it is qualitative, state the number of levels (categories) in the variable.

2. What is the dependent variable (DV)? What type of variable (quantitative or qualitative) is the DV? If it is qualitative, state the number of levels (categories) in the variable.

3. Briefly explain (in 1-4 sentences) why an ITT is the most appropriate analysis for this research question. (Hint: think about the variables; you could also compare/contrast why other tests may not be suitable)

4. What is the null hypothesis for this research question? Write this in words only (statistical notation not necessary).

Task 2 - Independent t-tests (Visualising the Data)

Generate an error bar plot of the confidence intervals of the mean of lifeExp07 in each category of hemisphere. Insert this graph into your assignment submission.

Briefly describe (in 1-4 sentences) what this graph shows, with specific mention of:

• What the middle point of the confidence intervals demonstrates about the groups (i.e. what statistic does this middle point represent? Compare this between the two groups).
• What the overlap or lack thereof between the intervals suggest about statistical significance.
• What the difference/similarity in width between the two groups indicates.

Hint 1: Confidence level.

Hint 2: In thinking about the width of the CI, what are the some of the key determinants of the width of a confidence interval?

Task 3 - Independent t-tests (Hypothesis Test)

Perform an ITT to evaluate this research question. Assume the assumption of normality has been met and use the built-in test of homogeneity of variance to test this latter assumption (i.e. do not generate this using the Explore function).

For this task, you must insert your Group Statistics and Independent Samples Test output tables into your assignment document (remember to paste as an image). Also, you should provide commentary on your assumption (homogeneity of variance) and hypothesis test findings (see hint below). Ensure you state the value of the p-value for the ITT and not just its relationship to α (for example, if p = 1 × 10^-3, you should write it in this form, and not "p < 0.05"). State the p-value correct to at least 4 decimal places (even if you write it in scientific notation).

Given your statistical conclusion above, what can you conclude in relation to your research question?

Finally, briefly discuss (no more than 2 sentences) what some key differences (if any) in your conclusions or the output you generated would be if you had analysed this as a two-tailed test instead.

Hint 1: Confidence level.

Hint 2: use the pro forma on page 7 of the Module 3 SPSS Workbook to guide your answer for this task, and use the M3 lecture slides and/or refer to the Allen/Bennett text for examples of reporting your results. When describing assumption testing, you should briefly state the null hypothesis of the homogeneity of variance test, and how you've determined if it is violated or not. Be specific.

For Tasks 4 to 7, you must use SPSS_Assign_Pt2_Dataset2.sav

In tasks 4 to 7, you must use one-way analysis of variance (ANOVA) to examine the relationship between continent and life expectancy in 1977 (lifeExp77). In this analysis, assume all assumptions have been met and α = 0.05.

Task 4 - ANOVA (Thinking About the Research Question)

Answer these questions succinctly (use the following numbering in your submission; you must not re-write the question):

1. What is the independent variable (IV)? What type of variable (quantitative or qualitative) is the IV? If it is qualitative, state the number of levels (categories) in the variable.

2. What is the dependent variable (DV)? What type of variable (quantitative or qualitative) is the DV? If it is qualitative, state the number of levels (categories) in the variable.

3. Briefly explain (in 1-4 sentences) why an ANOVA is the most appropriate analysis for this research question.

4. What is the null hypothesis for this research question? What is the alternate hypothesis? Write these in words only (statistical notation not necessary).

Task 5 - ANOVA (Visualising the Data)

Generate an error bar plot of the confidence intervals of the mean of life expectancy for each continent. Insert this graph into your assignment document.

Comment (in only a few concise sentences/dot points) what this graph shows, with specific mention of what the middle point of the confidence intervals shows about the groups, any trends you observe and the results you expect to observe in the hypothesis testing.

Task 6 - ANOVA (Hypothesis Test)

Perform an ANOVA to answer your research question. For this task, only include the Descriptives and ANOVA output tables in this task submission.

Should you reject or accept the global (i.e. overall) null hypothesis? What does this conclusion mean in terms of the research question, and should you proceed with post-hoc analysis?

Task 7 - ANOVA (Post-Hoc Analysis)

Regardless of your conclusion about proceeding to post-hoc analysis in the preceding task, perform both Bonferroni and LSD post-hoc analyses (if you haven't already done so).

How many unique pairwise comparisons do you have for each post-hoc analysis (i.e. for each of LSD and Bonferroni)?

Complete the table in attached file (you should use the template table provided and delete rows as necessary). The name of the first comparison is provided as an example of how to fill in that column. The p-values should be reported to at least 3 decimal places (hint: if SPSS displays the p-value as ".000" to 3dp, what does this indicate?)

Do each your Bonferroni and LSD conclusions match, or do you have any contradictory results (where one indicates a significant difference, but the other does not)? If any contradict, suggest an explanation as to why.

What can you conclude about the relationship between these two variables in the research question?

For Tasks 8 to 10, you must use SPSS_Assign_Pt2_Dataset2.sav

Using the same dataset as above, the researchers now want to see whether the gross domestic product (GDP) per capita has changed from 1977 to 2007. As the data for GDP per capita is positively skewed, both the 1977 and 2007 data have been log-transformed to allow for parametric statistical tests. Use these log-transformed variables for the following tasks.

Task 8 - Paired t-test (Thinking About the Research Question)

Answer these questions succinctly (use the following numbering in your submission; you must not re-write the question):

1. What is the independent variable (IV)? What type of variable (quantitative or qualitative) is the IV? If it is qualitative, state the number of levels (categories) in the variable.

2. What is the dependent variable (DV)? What type of variable (quantitative or qualitative) is the DV? If it is qualitative, state the number of levels (categories) in the variable.

3. Briefly explain (in 1-4 sentences) why a paired t-test is the most appropriate analysis for this research question.

4. What is the null hypothesis for this research question? Write this in words only (statistical notation not necessary).

5. How many countries contributed any data towards this dataset?

Task 9 - Paired t-test (Assumption Testing)

List the assumptions of the paired t-test. Indicate which of these (if any) are testable using SPSS.

Perform assumption testing for this analysis. To test the assumption of normality (both graphically and numerically), you should use the Explore function. Insert the relevant output into your assignment document and comment on each - draw conclusions about whether the assumptions have been met overall.

Regardless of the outcomes of your assumption testing, suggest at least one alternative action that should be taken if the normality assumption is violated (note: you do not need to do this, just provide a suggestion).

Hint 1: there should be four assumptions (see the module lecture slides).

Hint 2: you will also need to use the Compute function as part of this task.

Task 10 - Paired t-test (Hypothesis Testing)

Regardless of the conclusions drawn from the assumption testing above, perform a paired t-test to test the null hypothesis. Include your Paired Samples Statistics and Paired Samples Test tables in your task submission. Also answer the following:

• What statistical conclusion can you draw?
• What does this mean in terms of the research question?
• The data from how many countries were included in this paired t-test analysis? Does this correspond to the number of countries you stated in Task 8 (Question 5)? Why/why not?

Task 11 - Finally Submit Your Assignment, Dataset and Output Files.

Attachment:- Assignment Files.rar