Reference no: EM132254915

SPSS/Lab Mini project -

In this project, students will begin developing a variety of skills necessary for data analysis.

Building on the skills students learned in the first SPSS/Lab Assignment, students will learn how to:

• write hypothesis tests
• conduct 2 hypothesis tests for a two-category/group measure (independent t-test and Whitney-Mann U)
• conduct 2 hypothesis tests for a three-category/group measure (one-way ANOVA and Kruskall-Wallis)
• calculate 2 correlation coefficients (Pearson's and Spearman's)
• decide to reject or retain the null hypothesis

In this data analysis project, we will examine the five variables (community area, one independent variable and three dependent variables) that students chose for their first project.

Part I: Open your SPSS dataset and syntax files that you created for the Miniproject 1.

Ensure that the syntax you created for Miniproject 1 runs. If it doesn't, fix. The main reasons why this syntax won't run are:

• missing periods at the end of lines of syntax.
• misspelled variable names or command names.
• you've attempted to relabel or otherwise work with a variable that has not been created yet.

Part II: Construct new measures in this dataset by re-running the syntax you used to create these measures in the first lab/spss assignment.

First, create two new measures of community- one that has two groups and one that has three.

1. In order to do an independent samples t-test or Whitney-Man U, we need to test the differences between two groups. (Make your community variable into a variable with two groups.) There are several different ways to go about this... you could compare community areas that you would want to live in, versus those you would not. You could compare community areas with airports versus those without. You could compare community areas that touch the lake versus those that don't.

2. In order to do a one-way ANOVA or a Kruskall-Wallis test, we need to test the differences between three groups or more. (Make your community variable into a variable with three groups.) Same deal, you can do this with "north side", "south side", and "west side". Or something else.

Part III: Write 1 null and 1 alternative hypothesis statements for each relationship (a test of 2 groups, a test of 3 groups, and a correlation). Remember that you're basically going to conduct two tests for each hypothesis statement (a parametric test and a non-parametric test).

Hypothesis for two-category/group test of DV

They might look something like:

H1: The premature death rate will be statistically different in communities I don't want to live in compared to the communities I want to live in.

H2: The infant mortality rate will be statistically different in the communities I don't want to live in compared to the communities I want to live in.

H3: The breast cancer mortality rate will be statistically different in the communities I do not want to live in compared to the communities I want to live in.

Hypothesis statements for a three-category/group test of DV:

They might look something like:

H1: The premature death rate will be statistically different in at least one community are in Chicago (compared to the other two).

H2: The infant death rate will be statistically different in at least one community are in Chicago (compared to the other two).

H3: The breast cancer rate will be statistically different in at least one community are in Chicago (compared to the other two).

Hypothesis staetments for two interval/continuous or ratio level measures:

They might look something like:

H1: The hardship index and the premature mortality rate will be significantly correlated. H2: The hardship index and the infant mortality rate will be significantly correlated.

H1: The hardship index and the breast cancer rate will be significantly correlated.

Each of these will be assessed with two different tests, a parametric test and a non-parametric test.

Recall that parametric tests have requirements- usually that data are normally distributed, but there are other requirements. A non-parametric test does not have those requirements.

A major component of this assignment is evaluating whether each test produced different results that might lead to a different decision to reject or retain the null.

Part IV: Do hypothesis tests on two-category measures.

1. Independent Samples T-Test

Use SPSS to conduct to do an independent samples t-test. This can be done by clicking on "Analyze" → "Compare Means" → "Independent Samples T-test". Click on the dependent variable you made and click the arrow to move it into the test variable window. Then, click on the two-category variable you just made and the arrow to move it to the "grouping variable" window. You will need to click on "Define groups" to tell SPSS the numbers you want to compare. If you followed my example, it should be 1 and 2 or 0 and 1. Click Ok.

2. Mann-Whitney U Test

To do this, click "Analyze" → "Non-Parametric Tests" → "Independent Samples". Select radio button next to "automatically compare distributions across groups". In the "Fields" Tab, select your dv into the "test fields" window and the two category group you just made into the "Groups" window. Click "Run".

Part V: Do hypothesis tests on three-category measures (and community).

1. One-Way ANOVA

Use SPSS to conduct to do oneway ANOVA. This can be done by clicking on "Analyze" → "Compare Means" → "One-Way ANOVA". Click on the dependent variable you made and click the arrow to move it into the dependent list window. Then, click on the three-category variable you just made and the arrow to move it to the "factor" window. Click OK.

2. Kruskall-Wallis

To do this, click "Analyze" → "Non-Parametric Tests" → "Legacy Dialogues" → "K Independent Samples". Put your dependent variable into the test variable window and your newly created three-category measure into the grouping variable window. Make sure that the radio button next to "Kruskal-Wallis T" is selected. You will also need to define the groups. If you followed my lead (defining your categories as 1-3), you should be able to put 1 in the minimum window and 3 in the maximum. Click "OK".

Part V: Do correlation analysis on your original continuous measures.

1. Use SPSS to conduct to do a Pearson's Correlation Coefficient. Because the community area measure we have been working with in this project is nominal/categorical, we need to revert back to an independent variable you worked with on the first project. Correlate this measure with the DV you have been working with thus far.

This can be done by clicking on "Analyze"? "Correlate" ? "Bivariate". Highlight the dependent variable you made and click the arrow to move it into the variables list window. Do this with a dependent variable. Then, check that Pearson's are selected. Also, check that the "Two-tailed" test of significance is also selected. Click OK.

2. Use SPSS to conduct to do a Spearman's Correlation Coefficient. Because the community area measure we have been working with in this project is nominal/categorical, we need to revert back to an independent variable you worked with on the first project. Correlate this measure with the DV you have been working with thus far.

This can be done by clicking on "Analyze" → "Correlate" → "Bivariate". Highlight the dependent variable you made and click the arrow to move it into the variables list window. Do this with a dependent variable. Then, check that Spearman's are selected. Also, check that the "Two-tailed" test of significance is also selected. Click OK.

Part VI: Export Output file to Word, tidy, and interpret

This can be done by exporting your output file to a word document and then deleting all the extraneous output. Another option is to close the current output file and rerun your analysis on your syntax file, then exporting that "fresh" output, that doesn't have all of the mistakes and errors you've created along the way.

You should have 6 tables in this mini-project:

Table 1: Independent samples t-test

Table 2: Whitney-Mann U test

Table 3: One-way ANOVA

Table 4: Kruskall-Wallis H test

Table 5: Pearson's correlation coefficient

Table 6: Spearman's correlation coefficient

Your interpretation should focus on whether you should reject or retain your null hypothesis. And then a broader statement that speaks to your research question. You should also note if there were differences in your decision, based on your test. For example, you should include in your discussion that your t-test was not significant, but your Whitney-Mann was significant (and what that might mean).

Attachment:- Assignment File.rar