|
Criminal Justice 605 |
Syllabus
Spring 2000
|
LEGAL NOTICE. Unless otherwise indicated, all materials on this page and linked pages at the rbtaylor.net addresses are the sole property of Ralph B. Taylor and © 1999-2000 by Ralph B. Taylor. All these pages were created by the author in his spare, discretionary time and not as part of required instructional activities, but rather as potential enhancements. Further, the preparation and storage of all these pages did not and does not involve Temple University resources in any manner. All users have the right to freely access and copy these pages provided that they: acknowledge the source, do not make changes on any pages, and do not charge more than copying costs for distribution. |
|
Click HERE for information
about upcoming paper presentations |
TIME: Thursday, 3:00 - 5:30; Gladfelter Hall, 5th Floor Conference Room
Instructor: R. B. Taylor
Office: 539 Gladfelter
|
Memos on different topics
3/16/00 Questions for Elliott |
|
Paper assignments |
|
Instructor |
R. B. Taylor |
|
Time |
Thurs: 3:00 - 5:30 |
|
Office |
539 Gladfelter |
|
Office Hours |
TuTh 11:30 - 1:00; Tu. 1:00 - 2:40; if these times to
not work for you call me up and we will set something up |
|
Lab |
5th Floor Gladfelter; available when building open; swipe ID card |
|
Contact |
215.204.7169 |
This course concentrates on a broad family of multivariate techniques: multilevel models or hierarchical linear models. Theseare appropriate when units of observation are nested within broader units, and sometimes even when that nesting is by time. You will learn the conceptual background for these techniques, the types of problems they are designed to address, the special terms associated with each, how each applies to a range of criminal justice research and evaluation questions, and, perhaps most importantly, how to carry out these analyses and interpret the resulting output. You will complete weekly readings and written assignments. Almost all written assignments will require completing a computer lab assignment. There will be an in-class final, written examination.
The Focus
Hierarchical Linear Models (HLM) or Multilevel models
(MLM) represent a significant advance in social scientists' ability
to understand how outcomes are affected by context, how individual
and contextual factors interact, how outcomes change over time, and
how to summarize results from a series of studies. We will be
concentrating in this course largely on the first two uses of HLM. We
will get to the other two uses of HLM - to investigate changes over
time, and to summarize studies -- if and as time permits. I am
optimistic. These models address a range of theoretical and
methodological issues relevant to criminal justice, sociology,
psychology, urban studies, education, and political science. The
issues include multilevel analysis, aggregation issues, contextual
analysis, and clustered samples. In simple, whenever the individual
units of study (e.g., students) are nested within a higher level unit
(e.g., schools), HLM is an appropriate, and some would argue the most
appropriate form of analysis.
There are questions about whether MLMs are just a "fad" right now in criminal justice and criminology research. I don't think so, for the following reason. First, the central problems in research and evaluation in criminal justice are most amenable to, and only amenable to, a multilevel approach. Offender or delinquency careers represent cases in point. In addition, the interaction between the lower level unit and the higher level unit (e.g., the officer and the police department) is fundamental to numerous theoretical and policy concerns.
Here are some examples of "units nested within
larger units" in criminal justice evaluation or research:
|
Level 1 units |
Level 2 units |
|
Residents |
Different Neighborhoods |
|
Police Officers |
Different Precincts |
|
Police Precincts |
Different Police Departments |
|
Cases Sentenced |
Different Judges |
|
Prisoners |
Different Prisons |
|
Sentenced Drug Offenders |
Different Drug Courts |
|
Juveniles |
Different Treatment Programs |
At the same time, I do NOT think MLMs are going to "solve" all or nearly all of our analytic problems. That would be asking too much. In fact I think many may become disenchanted with MLMs because they answers they provide may not be to our liking. We have many theory and policy ideas around the interaction between person and context. But MLMs may often find that these interactions are nonexistent or trivial.
In this class we use numerous examples from criminal justice, but also include a substantial number of examples from other disciplines. We will:
Learn the conceptual and methodological issues
addressed by HLM
The goals of this course revolve around three different sets of activities. First, we hope to learn the rudimentary conceptual frameworks underlying these analyses; what problems do the analysis identify, and how does it propose to address these concerns? Second, we want to see how these tools have been applied to actual research problems, in criminal justice and in related fields. Third, we want to have "hands on" experience with the techniques themselves, using available datasets.
We will be spending time both in the classroom and in the fifth floor lab. It is not clear at this time how much lab time we will have each week. But there will probably be some each week. In addition, you should try and plan it so that if class runs to 6:30, it does not create a problem for you. In past semesters with the same material I have found this is not an unusual occurrence.
I assume you understand the basics of OLS multiple regression, including:
variance
covariance
correlation
scatterplots
R squared
adjusted R squared
F test of R squared
b weights
standard errors of b weights
beta weights
t tests of b weights
constant
residuals
predicted scores
residual diagnostics
tests for linear vs. curvilinear impacts
coefficient of alienation
In addition, I assume you know your way around SPSS for Windows
You will be reading books, articles, and handouts. You
should find the following books in the Temple SWEATSHIRT and bookstore:
MAIN TEXTS
Ita G. G. Kreft, Jan De Leeuw (1998). Introducing
Multilevel Modeling (Introducing Statistical Methods) Thousand Oaks: Sage
Anthony S. Bryk, Stephen W. Raudenbush (1992). Hierarchical Linear Models : Applications and Data Analysis Methods. Thousand Oaks: Sage (ISBN: 0803946279)
SUPPLEMENTARY AS NEEDED
Lawrence C. Hamilton (1992). Regression With Graphics : A Second Course in Applied Statistics Monterey: Brooks/Cole Pub.
William D. Berry, Stanley Feldman (1985). Multiple Regression in Practice (Quantitative Applications in the Social Thousand Oaks: Sage.
We are going to be reading a good number of articles and handouts. There are different ways you get to these. The handouts you get directly off the website. Directions follow below. A couple of the articles you also can get directly off the website. Several of the articles I will put on departmental "reserve." I am going to get 3 copies of each of these, and put them in the graduate student office, down the hall. You may then take the readings and make copies of them as you need. Some of these readings you will be able to obtain through PROQUEST DIRECT but to do that you need to have a Temple PPP account, and know how to dial up and get into that from a remote location.
You will complete the assigned readings on a weekly basis and come to class prepared. You will complete and write up a number of analytical assignments; the purpose of each is to conduct a specific analysis, and write it up. These assignments will be due on average every two to three weeks. You will receive "credit" for turning in each of these on time. Each of these will build toward a broader, final paper. These papers will, hopefully, represent building blocks that can later be put together to form a paper of publishable quality on which everyone gets authorship. In order for such a project to work it will be necessary to people to carry out different analyses, and report on them clearly and succinctly, and to contribute directly to the final project paper. We will see how this works. If it gets too confusing, or turns out to be too demanding, we will "fall back" on everyone doing the same assignment, rather than people taking a piece of the final project.
These papers are to be typed, double-spaced with the due date your name and SSN at the top. These are not going to be graded, but you will get feedback on how to improve. Be sure to spell-check and grammar-check your papers.
The readings provide the needed conceptual background for carrying out the work assigned, and for understanding how MLMs are advancing scholarship in criminal justice and criminology. Thus it is important that you keep up with the readings. They provide not only examples that help us decode MLMs, they also provide excellent examples.
You will notify me beforehand if it is absolutely essential for you to miss a class. Given the amount of ground we must cover, a missed class may create a significant burden for your learning curve. If you do miss a class it is completely your responsibility to get all handouts, assignments, and so on, that were distributed.
Every now and then I will hand out some questions about the readings (articles) assigned for that week. I expect you, by the next class, to have attempted to have written out answers to some (the number will be specified) of those questions. Your answers should be typed, double-spaced, with both your name and SSN on the top of the page. To get credit for answering these questions, your answers should be turned in by the beginning of the following class. Be sure you have an extra copy for yourself. I am not grading your answers, but rather just giving you credit for trying to answer them. I am not sure for how many weeks we will do these.
Datasets and conceptual background
I am going to be spending most of
our time with one four city, multi-neighborhood data set on reporting
of drug crimes. Below are the links for the Davis, 4-city file. You
can download the SPSS file, and two codebooks, and a conceptual paper
I wrote outlining some analyses that might be interesting. One
codebook describes the citizen survey, and you will want to print
that out. The second describes a police nomination dataset. You may
want to scan that, but you need not print it out.
|
Description of file to be downloaded |
File name |
|
ICPSR 9925 4 city file (SPSS datafile) |
|
|
Codebook for survey (adobe PDF file) - you will want to print this out |
|
|
Codebook for police (adobe PDF file) - you will want to download and review this file, but may not want to print it out |
|
|
This is a WordPerfect document describing some conceptual issues to be addressed with the dataset, and classifying some of the variables that are in the dataset. Be sure to save this with a WPD extension. You CAN print this out with Word if you need to because there are no fancy graphics in it whatsoever. |
|
|
This is a review piece I wrote on the incivilities
thesis, including theory and measurement issues. Save it as an adobe
PDF file; the file will have SKogan's chapter, then mine. My chapter
starts on page 65 (28 of 51 in the file). The full citation is:
Langworthy, R. H. (Ed.) (1999). Measuring
what matters. Washington: National
Institute of Justice/Office of Community Oriented Policing Services.
To read the other chapters in the book go to: |
|
|
Updated 9925 data file, with new variables added; see handout in class. This also has a limited number of contextual variables added. THIS VARIABLE WAS CORRECTED AND RE-UPLOADED ON 2/2/00 |
|
|
An example SSM file, generated with TRANSYS The statistics for the file What the response file looks like These are the Level 1 and Level 2 data files that were used to create the SSM file |
|
|
For nonlinear models - download the 2 data sets, and the ssm file, be sure to put all of these on A drive. Last file is lecture notes |
|
|
These are files for the time model. You want to download these and place the data files and the ssm files on your C DRIVE IN A DIRECTORY CALLED 0413 - you will need to make this directory before you download NOTE - UPDATED 4/13!! |
Your grade at the end of the semester will be based on the following:
30 % Final paper (builds on earlier assignments)
30%Final examination. This will focus on the identification of an appropriate tool to use in a particular situation; and on interpreting results presented
20% CREDIT for answers on readings. I will have to make a judgment if your answers represent a credible attempt to answer the questions, but if they do, and are turned in on time, and are properly done, you get full credit.
20% CREDIT for written assignments, turned in on time, that build to the final paper. Again, I make a "credible attempt" decision, and beyond that, you get full credit.
GRADING POLICIES
1. Assignments are due on the date indicated. I reserve the right to lower the grade for assignments that are handed in late. The amount the grade is lowered will increase the longer the delay in handing the assignment in.
2. If you have an excuse for a late assignment I will take this in to account only if you notify me -- or try to notify me -- beforehand about the problem and I find your excuse for the delay to be a valid one (e.g., car accident), and you can provide documentation as needed.
3. We will discuss in class the nature of academic misconduct, including plagiarism. You are responsible for understanding the different varieties of academic misconduct. If I encounter solid evidence of academic misconduct I reserve the right to fail you on the assignment in question, and/or to assign you a failing grade for the course. I will try to state as clearly as I can the ways in which it is acceptable for you to cooperate with one another and network, and the ways in which it is not acceptable.
4. You do have a right to submit any assignment for regrading. You should state in writing the reason you think you deserve a higher grade, attach that to the original completed assignment, and return it to me. Your grade may go up, go down, or stay the same. I may consult with other faculty members as I deem fit.
5. Of course, on any grading issue that you and I are unable to resolve to our mutual satisfaction you do have available to you standard grievance procedures you may elect to pursue. I can tell you about those.
Lab
We will be spending time in the
lab, probably every week. You have access to that lab during regular
business hours, from 8:30 - 4:30, and can arrange to stay later if
you need to. I strongly recommend AGAINST working in there on the weekends.
If you have the student version of HLM at home, you want to be sure that you have the data files you need to work with the student version at home. Most important, you want to be sure you have used the TRANSYS utility to save your data files in SYSTAT format.
Load
This course ends up being somewhat
more demanding than some other graduate courses for some students.
In short, for some of you, this may "feel" like a four
credit or a six credit graduate course. Try to plan your weeks accordingly.
Software
For HLM we will use a specific program, put out by
Scientific Software International. The main web page for SSI is SSICENTRAL.COM.
This is a useful website, because you can look at the examples, and
get help interpreting HLM output. I STRONGLY ENCOURAGE YOU TO GO
THROUGH ALL THESE EXAMPLES ONCE WE ACTUALLY START WORKING ON HLM. The
department has bought a site license for this program. This means
that you need to do your work in the department on their computers if
you want to have access to the FULL version of the program
There is, however, a STUDENT version of the file that you can download for FREE. Go to this address to learn more:
http://www.ssicentral.com/other/hlmstu.htm
There are, however, a number of restrictions. Most importantly, you can not run very complex models, and your data need to be in ASCII or SYSTAT data format for things to work. From that page, here are the restrictions they note:
The student edition can run all the analyses the full version can in terms of models selected, statistical options and output.
Restrictions are, however, placed on the data used and the size of the model selected. The following restrictions apply in this edition:
The TRANSYS utility used for the importation of data is not included. The student edition will only accept ASCII, SYSTAT or SAS transport data files.
For a level-3 model, the maximum number of observations that may be used at levels 1, 2 and 3 are approximately 7500, 1700 and 60 respectively. Note that the restriction applies to observations in the case of the level-2 file, for example, and not to actual number of level-2 units to be included in the analysis.
For a level-2 model, the maximum number of observations at the two levels are 7200 at level-1 and 350 at level-2 of the hierarchy.
No more than 5 effects may be included in any HLM equation at any level of the model, and the grand total of effects can not be 25 or higher.
When these limitations are exceeded, an appropriate error message will automatically be displayed.
Sequence
The sequence of topics and readings, as best as I can
predict them at this point, appear below. All of this is subject to
change depending upon numerous factors, including el Nino. The
readings and assignments are for the week they are DUE.
You will see that next to each reading are some codes.
GO = copies from which YOU can make copies are
available in the Graduate Office down the hall. I have put 2 or 3
copies in there
PQD = you can get a full copy of this off of ProQuest
Direct. You need to have a Temple ppp account, and, if you are NOT
dialing straight into Temple, but are using another ISP, you need to
know how to set things up for a proxy server. You also will need
Adobe Acrobat READER (a free program) to print these things out.
RBT = available as a linked file directly off the
website. Click and it will ask you to download. Then fire up Adobe
Acrobat Reader
For the two main texts
K&DL = Kreft and deLeeuw
B&R = Bryk and Raudenbush
|
Class Date |
Topics / Readings handed out / Other |
|
1/20 |
Syllabus review; The presence of clustered data; The
main options for analyzing clustered data: focus on individuals (raw
or centered data), focus on aggregate units; add in contextual
variables or context identifiers. Understanding the ecological
fallacy; understanding the individual fallacy; What are the problems
with contextual analysis?; Micro-macro issues
In the lab: |
|
1/27 |
Interpreting contextual regressions READ:
LAB: building the contextual regression; the aggregate procedure |
|
2/3 |
More on interpreting contextual regressions; interpreting centered data; Understanding the limits of ANCOVA; the idea of fixed vs. varied slopes: ANCOVA vs. separate regressions for each unit Read
|
|
2/10 |
Assignment 1 Due: interpreting a contextual regression Introduction to HLM: The Level 1 model; Variances and Covariances; Variance Decomposition; Empirical Bayes estimates; "true" scores on the group means; HLM SUBMODEL 1: One-way ANOVA with random effects; HLM SUBMODEL 2: One-way ANCOVA with random effects Read:
LAB: Starting up HLM; Start work on ANOVA model |
|
2/17 |
HLM SUBMODEL 3: Random coefficients regression model Readings:
LAB: continue work on ANOVA model; one-way ANCOVA with random effects; working on the best Level 1 model |
|
2/24 |
Assignment 2 Due: Conduct and interpret a variance decomposition; complete and interpret a Level 1 model (One-way ANCOVA with random effects) HLM FULL MODEL: Intercepts and Slopes as Outcomes (IASAO) Read:
LAB: working on deciding whether or not to let some Level 1 slopes vary; selecting some Level II variables |
|
3/2 |
General Catch-up session on the full model; A Couple more examples READ:
LAB: finalizing selection of Level II models |
|
3/9 |
|
|
3/16 |
Assignment 3 Due: A "Full" model but with all fixed slopes Conceptual and empirical questions: Varying slopes at LI; predicting LI slopes with LII variables. What does this mean? |
|
3/23 |
Residuals: Doing Diagnostics; Variables in the file; Plots you want to generate; Read:
|
|
3/30 |
More on residuals |
|
4/7 |
Assignment 4 due: write up of residuals analysis Changes over time; implications for evaluation Read:
LAB: ? |
|
4/14 |
Introduction to the general probability model: GHLM Read
LAB: doing and interpreting a probability model |
|
4/21 |
The three level model Readings: TBA |
|
4/27 |
A host of last minute items: meta analysis via HLM; weighting; multiparameter tests
|
|
5/5 |
Final Examination; final paper due |
Sequence of topics for first class
The pervasiveness of clustered data
Conceptualizing different types of
clustering and the purpose of Stat II
Exercise in thinking about
different types of clustering
The typical approaches, outlined
individual level analysis
(ignoring the clustering)
Advantages of MLMs
The course goals and structure
Issues of academic misconduct
Description of the theoretical
issues to be addressed: the evolution of the incivilities thesis and
the problem of reporting drug activity
Description of the dataset
Description of the resources
available to you
what is on the RBT website
Break
LAB
pull down the main SPSS datafile