Module 4: Week 4 Introduction
Overview
You will explore how when designing criminal justice research, sample size comes to mind. You will also
explore how some practitioners believe it is necessary to survey 100% of a population; however, you will
discover that solid statistical analysis avoids the use of a 100% sample size research design.
Learning Outcomes
Upon successful completion of this module, you will be able to:
Describe the strengths, weaknesses, and uses of self-reported data.
Explain probability sampling and non-probability sampling.
Identify five sources of internal invalidity.
Be sure to complete all Introduction, Learn, and Apply items in Module 4: Week 4 - Sampling, Design,
and Self-Reporting by 11:59 p.m. (ET) on the assigned due date
Read
Bachman & Schutt: Chapters 5-6
Mosher, Meithe, & Hart: Chapter 4
“Exploring Demographic, Structural, and Behavioral Overlap Among Homicide Offenders and Victims”
(PDF File)
“The Recidivism Rates of Female Sexual Offenders Are Low: A Meta-Analysis” (PDF File)
Watch
“Sampling”
Transcript: On behalf a Liberty University is Helm School of Government. Welcome. I'm Dr. Jo Cox, the
Criminal Justice Program Director for all the residential criminal justice programs at Liberty University. Welcome to
welcome to our stadium. Just a couple of weeks ago, last week, and the stadium was completely full of people at
this time. We had several thousand students in the middle of the field that you may be able to see. Every single one
of these bleacher seats was felt up. We had over 30000 people at the stadium at the same time. And what I want to
talk to you about is not just a great stadium and a great place to go to school that you've already chosen. Just want
to talk to you about statistics, saying that we cringe about sometimes and particularly sampling. And really just
look at we have a bunch of people that could be here or were here last week. How do we use sampling if we want
to do some type of survey research? And certainly if we wanted to look at 30000 people sitting in a stadium, will be,
we could talk to everyone. I'm, we had enough people. We could get a group of researchers out. We can have some
questions. We could ask every single one of those people questions. However, unless we had thousands of
researchers, we will be here for an awful long time. Just think if you had five minutes worth of questions and you
have 30 thousand people, how long would that take to survey them? So sampling research, what is sampling
research? Why did we do sampling in research? And part of it is cost effectiveness. Let me just read some stats to
you and just bring this to a criminal justice perspective. If we have a town of, let's say 20 thousand adults and we
want to survey all 20 thousand adults, want to use a 100 percent population size. And we have 20 questions. And
by tiny due an introduction and say goodbye and asked 20 different questions. It takes about 20 minutes. 20
thousand people times 20 minutes is 400000 minutes. That's 6,666 hours. If one person did that working 40 hour
weeks, that's 166 weeks to get that survey research done in a town of 20 thousand. Well, let's say we have a team
of ten surveyors that go out and do the same work. While still that 16 weeks to complete. Say you pay those folks
about $12 an hour to do the survey research for you. And that's just the research alone, not the manipulation of the
data. You would pay over $79 thousand in fees just to get the survey research done, either one researcher or a
group of 10 researchers. So if you're a police department or anybody in the criminal justice field, how in the world
can you afford that if you're a small police department or even a large fleet department, is there a different way to
do it that yields potentially the same results? And if rely upon statistics and good statistics, I think we can do that.
So let's use the same example, 20 thousand people and we say, I just want to use a 20 percent population size.
Well, if I take 20 percent of that population, I actually drive my cost down from $79 thousand to $15 thousand.
Again, using good statistics and good statistical research, I can draw the same conclusions potentially. But let me
talk to you about a couple of pitfalls. When we look at this entire population. If you can just imagine all of the seats
filled up, there's some issues that may come about. So we have a population. Now we might have a sample size and
a sample size might mean 20 percent. So we take 20 percent of the 30000 people that may be here or there some
issues that may come up. Certainly. If we look at this field And the people represented in this field, Liberty
University's a conservative Christian university. So if we do survey research at a conservative Christian university
with friends, family, parents, of the students that are here. Can we use that data to apply on all populations
throughout the United States? Can we use that data to apply to other conservative Christian universities? We also
have what we call combo here, ten to 12 thousand students getting together three times a week for a worship
service. If we do survey research at Canva and kids 17 to 21 years of age, can we draw conclusions from that
convoke and our survey research to apply to everyone 17 to 2122 years of age throughout the United States. And
we may say now as a special type of population, for other issues, we may be able to say yes. So a couple of issues
that make that may come about. We talk about how do we generalize, sample generalization. And if we have
certain special populations, conservative Christian university, it could be the opposite as well. We have to be very,
very careful that we don't crop, draw conclusions from sample research, survey research of this population and
apply it to other populations. Here has to be further research done. Show that we can take survey research from
one population and apply it to another. So please be very, very careful if you're a police officer, you're a police
leader. You want to do survey research. You go to one particular neighborhood and do that survey research. Is that
neighborhood representative of your entire community? King, the conclusions that you draw from survey research
from that one neighborhood, be generalized to all the neighborhoods in your community or in your random
sampling? Do you need to use a sample of all the different members in the community? Are all the different
neighborhoods in your community and have a random sample drawn from that. So you can draw some
generalizations. That's something that can be very, very dangerous if you're not careful and don't realize that, don't
just go to one place. You may have to go to multiple places. So I really just wanted to talk to you today about survey
research that we don't have to use 100% sample size. When we can use solid statistical research that we can use a
certain population of people in survey those as cost saving. As time-saving. We can still draw conclusions from it.
However, we have to be very, very careful and the conclusions that we draw and how we apply those conclusions.
Powered by TCPDF (www.tcpdf.org)
