Assessing “Representativeness” of the survey
respondents
1) Were the home locations of the respondents to this survey spatially
random?
1.1 Open the ArcGIS and open ArcCatalogue.
1.2 Open ArcMap (Chose “a new empty map” option)
1.3 Either add surveylocations.shp and Sbcotractsdemog.shp or simply drag and drop these files from the ArcCatalogue to the ArcMap Window.
1.4 Visually examine the distribution of the points representing the Survey Respondent.
2) Were the locations of the respondents to this survey random with respect
to population density? (you do
have to do the analysis for this one)
1. In ArcCatalogue navigate to sbcontractsdemog and examine the Preview tab, change the option from Geography to Table (at the bottom of the window) you can examine the content of the attribute table.
2. To open the attribute table go to ArcMap, right-click on the Layer’s name and chose Open Attribute Table.
3. In ArcMap add data and chose (SpatialStatsClass/SbcoPopSurevy/TextlData/) SurveyResultsNeq724.dbf, right-click and chose open.
4. In both of the attribute tables, click on the header of the shape column to highlight it.
5. Right-click on sbcontractsdemog theme in the Content window of the ArcMap. Chose Joins& Relates -> join. BE SURE TO JOIN BASED ON SPATIAL LOCATION!!!! In the dialog window chose TO JOIN IF POINTS FALL INSIDE and make sure, that the other table is changed to SurveyResultsNeq483.
6. In the Attribute table if sbcotractsdemog check that it contains field from both tables. The last column titled Count indicates how many points (how many respondents) fall into one tract.
7. In this joint attribute table, click on Options-> Export and specify name and locations for the new .dbf file.
8. Open Excel software and chose Open Data Table. Under File type chose .dbf and navigate to your file.
9. Highlight all the columns containing the Answers to Survey Questions (P1_1 through P4_A8). In menu go to Edit->Replace and chose 6 to be replaced by . (period). Under the Advanced option make sure that you check box next to “Match entire cell contents”!!!
10. Scroll down the multiple rows “54” and “74” and delete multiple copies, leaving only one row per each. Save the changes.
11. Find column PERSONS and Insert new column titled PERSONS_Expected.
12. Sum PERSONS Columns.
13. Insert the following formula into the PERSONS_Expected:
E2/(total
of PERSONS)*588
14. Create New Column (Obs_Exp)Sq with following formula:
(Counts
- PERSONS_Expected)* (Counts - PERSONS_Expected)
15. Finally another column with following
(Counts
- PERSONS_Expected)* (Counts - PERSONS_Expected)/ Expected
16. Sum this column up….THAT’S YOUR Chi-Square!!!
Simple Demographic Comparisons
7) Based
on the responses to the question “Abortion should remain legal as defined in
Roe v. Wade”, are Democrats significantly more Pro-Choice than Republicans?
1. Open the Excel file SurveyResultsNeq724, highlight all columns with survey responses and go to Edit>Find>Replace and Find all “6” Replace with “.”, click on Options and check the box that says “Replace when value is entire field” so as not to change “P1_Q6” to “P1_Q.”
2. Save changes, Exit Excel.
3. Open JMP, click “Open Data Table” and choose “All Types” under “Files of type:” at the bottom. Find your Excel version of SurveyResultsNeq724 and open it in JMP.
4. In the far left column, change “PolParty” to Nominal (right-click on the icon to the left of the name, select nominal) and make sure that “P3_9” is continuous (the icon will be a “C”).
5. At the top of the window, select Analyze>Fit Y by X. Select “P3_9” and click “Y, Response”, then select “PolParty” and click “X, Factor”. Click OK. You should see a blank white graph.
6. Click the red arrow next to the heading and select “Means/Anova”. Your graph window should now show the interpretation of the data P3_9 by PolParty.
7. Evaluate Results.
8) Alo7ng
a similar vein, are women more Pro-Choice than men? (according to this survey)
1. Close “Fit Y by X” window from Question 7.
2. At the top of the window, select Analyze>Fit Y by X. Select “P3_9” and click “Y, Response”, then select “Sex” and click “X, Factor”. Click OK. You should see a blank white graph.
3. Click the red arrow next to the heading and select “Means/Anova”. Your graph window should now show the interpretation of the data P3_9 by Sex.
4. Evaluate Results.
9) In a
separate survey I found that women were more Pro-Choice than men and that
Catholic women were significantly more Pro-Choice than Catholic men. Is this true of the respondents to this
survey? How did you test that? If you did find the gap between Catholic men
and women significantly greater than the gap between men and women in general
what would a statistician call such a phenomena? If it were true, how would you explain it?
1. Close “Fit X by Y” window from Question 8.
2. At the top of the window, select Rows>Row Selection>Select Where, in the window that opens, select rows from ‘survey’ where Religion Equals 2 (Catholic). Click “Add condition” and click OK. You should now see that certain rows have been highlighted. Click on Rows>Row Selection>Invert Selection and the non-Catholic rows should be highlighted. Then go to Rows>Exclude/Unexclude. The non-Catholic rows should now be excluded.
3. Repeat steps 8.2 and 8.3. Your graph window should now show the interpolation of the data P3_9 by Sex for all Catholic respondents.
4. Evaluate Results.
10) Are
Republicans different than non-Republicans on the responses to any of the
questions about immigration.
1. Close “Fit X by Y” window from Question 9.
2. Go to Rows>Clear Row States. All rows should be deselected.
3. At the top of the window, select Cols>New Column… and name it ReporNot, Data type select “Numerical” and Modeling Type select “Nominal”. Click OK.
4. Select this new column, right-click and select Formula. In the “Functions” window, select Conditional>If. Then Fill in PolPtry into the first blank by selecting it from the “Table Columns” window. Select the condition by then going to Functions>Comparisons>a= =b. Type the number 2 into the next space. Then type the characters “R” into the next blank and “N” into the one below. The formula should basically read If PolPtry= = 2, “R”, else “N”.
5. Select Analyze>Fit Y by X, then choose P4_B1-5 for Y and RepOrNot for X. Do a Means/Anova for each chart.
6. Evaluate Results.
11) Is
there any relationship between “Religiosity” and responses to the question “The
earth has a finite supply of natural resources such as water, arable land, etc.
which imposes a limit on the number of people which can sustainably live on
it.”
1. Close “Fit Y by X” window from Question 11.
2. Make sure relgAct is continuous data.
3. Select Analyze>Fit Y by X, Set P1_15 for Y, relgAct for X. Click OK.
4. Go to the red arrow, select Fit Line.
5. Evaluate Results.
Factor Analysis
12) Run Factor Analysis…
1. Start up JMP 5.1
2. Select open data table
3. Locate the survey results, this will depend on where you have chosen to save this document, and highlight it so it will appear under ‘File name’
4. Highlight and delete columns P4A_1 through P4A_15
5. Click on Analyze
6. Highlight Multivariate Methods
7. Click on Multivariate
8. Highlight all questions by using the shift key to highlight multiple items
9. Click on Y,Columns
10. Multiple warning will appear, just click ok
11. Click OK
12. An alert will appear, click Continue (twice)
13. A multivariate chart will appear, click on the red triangle in the upper left hand corner next to Multivariate
14. Highlight Principal Components
15. Click on ‘on correlations’
16. Click on the red triangle next to Principal Components/Factor Analysis
17. Click on Factor Rotation
18. You will be asked ‘How many rotated factors?,’ type 5 and click OK
19. Click on the red triangle next to Principal Components/Factor Analysis and click on save rotated components
20. Analyze the chart and determine which questions have a factor contribution score greater than or equal to .40 (this applies to both positive and negative figures).
21. Identify the questions in each factor which have a value of or greater than .40 and create a ‘name’ for this particular group (for example, ‘save the wheals’).
13) Do Statistical Tests…
1. Select Analyze from the tool bar
2. Select Fit Y by X
3. Highlight
sex and click on X,Factor
4. Highlight
Factor 1-5 (by holding down the Ctrl
key while clicking on the factors) and click on Y,Response
5. Click OK
6. Click on the red triangle in the upper left hand corner and select Means/Anova/Pooled t
7. Record any significant values in the chart with an asterisk (*) and the actual value of significance.
8. Repeat step 1-7 for all other categories in the table (Political Party, Religion, Religiosity, Income, Education, and Race/Ethnicity)
9. Provide a detailed explanation for the significant differences found in each factor.
Spatial Analysis
15) Test for significant Differences/variation for all of the factor
scores (1-5 )and the population density and percent non-white of the
respondents home locations. If you find any significant differences provide an
explanation.
1. Open table with survey responses and factor scores in ArcView.
2. Open this attribute table and combine it with the SBcoBlocksDemog table by shape.
3. Summarize the table by the Key column. This should generate another table with Key and Count columns.
4. Join these tables by Key column.
5. Export this file to in .dbf format. Do not lose this file.
6. Open the file you just exported with Excel. Save as a text file.
7. Open this text file in JMP. You should have the factor columns as well as all of the geographical information from the Blocks coverage such as "population density, percent white, black, asian, Hispanic, etc.
8. Create a new column titled PctNonwhite. Enter a formula for the column by right-clicking and selecting formula. Select “Persons” from the window on the left, then select the division sign from the middle, then select the persons column again as the denominator. Click the Persons in the numerator and hit the minus sign form the middle, then select “white” from the window on the left. Click so that entire formula is boxed in and then click * and type 100. Your formula should now read
[(Persons-White)/Persons]*100
9. Go to Analysis>Fit Y by X, Y= Factors 1-5, X=Pop Density
10. Evaluate results.
11. Go to Analysis>Fit Y by X, Y=Factors 1-5, X= PctNonwhite.
12. Evaluate results.