Phases 4b Prep

 

***If you complete the assignment during the lab hours, you will receive 200 points.

 

State your hypotheses.

 

Choosing the right model

 

Look into the summary table first. Summary table.

 

If all the variables are nominal or ordinal variables, go to Section A. (If the number of categories is more than four, you may want to reduce the number by recoding.)

If the dependent variable is nominal or ordinal and the independent variable is interval, recode the independent variable and create three to five categories. Then go to Section A.

If the variables are interval, go to Section B.

If the dependent variable is a nominal or ordinal, while the independent variable is an interval variable, the independent variable should be recoded into a categorical variable with two to five categories. Go to Section A.

If the dependent variable is an interval variable, while the independent variable is a nominal or ordinal variable, recode the independent variable and create two categories and go to Section B. 

 

*Copy and paste your output tables and figures. You will need them to complete Phase 4b. Make sure that your tables and figures are organized well and are not cut off.  Do not include the same table or figure multiple times. (You can double click the tables and figures and edit. You can copy and paste them as objects using the “paste special” function.)

 

*Points will be deducted for any wrong or redundant tables.

 

Section A. All the variables or nominal or ordinal

                  Or the dependent variable is nominal or ordinal and the interval independent variable has been recoded:

Bivariate Statistics

A) Recode the variables, if the number of categories is too many.  Produce a crosstab for your main hypothesis. (Do not forget to include column %. Produce test statistics (i.e., measure of association and test of statistical significance). If one of the variables is nominal and the other is ordinal, use a nominal measure. Since Lambda often underestimates the relationship and shows no significance, use Cramer's V instead of Lambda.  Cut and paste the crosstab table and the statistics table into Word.

B) Repeat the above for the alternative hypotheses (Both Rival 1 and 2)

 

Multivariate Statistics

A) Produce a multivariate crosstab. Produce test statistics. (here, you can treat one of the variables in your rival hypothesis as a control). Cut and paste the results into Word. The number of categories in the control variable should be about two to three. (It is fine to have more than three categories, but it will be more difficult to organize the results.)

Note: It is not necessary to repeat the above for the other control variable.

 

Note: Please print out your assignment and obtain an instructor's signature, when it is completed.

Section B. All variables are interval,

                  Or the dependent variable is interval while the independent variable is nominal or ordinal

        *If the independent variable is nominal or ordinal, recode the independent variable and create two categories first. Do not recode the dependent variable):

Bivariate Statistics

A) Obtain a scatterplot for X1 and Y and cut and paste it into Word. (SPSS: Graphs--Legacy Dialogs--Scatter. Move the dependent variable into the y axis box and move the independent variable into the x axis.) 

B)   Obtain a measure of association for X1 and Y. (i.e. Run Pearson's Correlation. Analyze--Correlate--Bivariate) Cut and paste the table into Word.

C)   Run a bivariate regression model for X1 and Y. Cut and paste the tables into Word

D) Repeat the above for the alternative hypotheses (X2 and Y, X3 and Y)

 

Multivariate Statistics

4) Run a multiple regression model.

 

Show your assignme to the instructor. If you make any mistakes, the instructor will let you know how to improve your assignment. You can redo your assignment and show the instructor the improved version. Repeat the process until everything is correctly done. Save your file and deposit the final version to Blackboard. Visit www.bsu.edu/blackboard  and go to POLS210. Find Assignment 1 and double click and upload your file as a WORD file. You will receive 100%. Blackboard allows you to deposit only once. If you did not complete the assignment during the lab hours, you must deposit your assignment to Blackboard by 5pm next day. The instructor will grade your assignment. (You will not be able to redo the assignment, even though your score is lower than 100%.)