Crim 860

Assignment 3: Indexes and Scales

Last year’s 860 class decided to create a Likert scale dealing with views about the police and policing. For the first part of the assignment, each student sent me six (6) statements that we could potentially use in such a scale. Two were supposed to be “poor” statements, accompanied by an explanation of why the student believed them to be “poor” items. The remaining four statements were intended to be “good” items for inclusion in the scale; two were written in a “positive” direction, while the other two were written in a “negative” direction. After I received all the statements, l put them together in a survey format and each student was instructed to send a survey link to friends and family – a convenience sample. We ended up with 100 respondents among the class and others.

For this Assignment, you are being given access to the data that arose from that project ... ratings from 100 respondents on our preliminary pool of 67 items. You can find the SPSS file containing the raw data and a link access to item-by-item distributions at https://www.surveymonkey.ca/results/SM-tILuWguE3adiAdZOv1Dpww_3D_3D/ . Your job in this assignment is to send me a brief report explaining the choices you made along the way, a copy of your final scale, and an explanation of what you would do next, after you do the following:

a. We started with an item pool of 67 statements, roughly half of which were written in a pro-police direction, and the other half in an anti-police direction. You can see the full list of items and whether their author considered each one to be positively or negatively worded, at https:// Let’s say that we are hoping to end up with a scale that’s 12-ish items long.

b. Perform the next steps with your partner where you do successive iterations of computing a tentative total, computing item-whole correlations, and using that information to successively whittle down the pool of items to put together a tentative scale, explaining why you chose the particular set of items that you did.

c. Compute a split-half reliability (to pay homage to Likert) along with the Spearman Brown Prophecy and compute Cronbach’s alpha for your scale.

d. Question 69 asked respondents to self-report an overall attitude toward the police. How might you use responses to that question as a preliminary validation measure? Feel free to do the analysis if you know how.

e. Let’s say that we decide that the scale you create in (b) is one we will go forward with and subject to further analysis. What would you be looking to do before you would feel comfortable actually using it in a research situation, e.g., if VPD were to ask you to evaluate Vancouverite’s views regarding the police, or if you were now interested in looking at other correlates of those attitudes?

Please put your explanation in a file that follows the following naming convention: <Crim860_firstname_firstname_Scale.format> (where the two firstname slots are the names of you and your partner and “format” is whatever format your document is in – docx or pdf)

Your assignment is due by no later than 21 October.