To: All Students In HTR 741 Winter 2013
From: Professor Michael R. Hunter
The following is the reading list for this entire course. For the custom reader start
page numbers with 1 on the page after each subsection’s cover page. The first few
words in the title of most articles should guide you. A “T” in front of a page number
means it is from the text book not my reader. While it seems like a lot it isn’t as it’s
been HEAVILY EDITED and is for the entire course.
Please read the articles starting with the Introduction and then in sequence
corresponding to what is covered in class and as per assignments.
1. INTRODUCTION: WHY SHOULD YOU STUDY RESEARCH?
Like anyone you have important questions: can school pay off, should I give it my
all, will a part time job give me everything grad school would, what car colour
should you buy, etc., etc. The following research results concerns such issues. As you
read think about if this research tells you anything you didn’t know, do you really
need research and on what possible personal and professional topics might it be
useful and available.
ITEMS 4 TO 8 ARE ALL THAT’S ON THE EXAM.
2. SAMPLING ISSUES
Pg. 1 Customer feedback…why get a 100% response and how can it best be done?
Getting a 100% response rate ensures that the information contained is statistically valid
and thus representative of the full population because it is a true random sample. You can
offer incentives to the participants.
Pg. 2 the Sample Size…Note the general differences between the theoretical versus
the actual ways of handling sample size. Why is there a difference?
The theoretical way of handling sample size would be to use the formula. You need the
variance of the population, the margin of acceptable error, and the confidence level.
Typically, the survey sample is maximized within the constraints of the survey design
(length, time, question amount, etc.) There is a difference because the formulas require
some information that most likely will not be found out until after the study has been
completed, and most surveys have multiple variables which may require different levels
of measurement than the formulas allow.
Pg. 3 SAMPLE SIZES…saves you from copying this down. Don’t memorize this!
T 131-132 What is quota sampling and why is it done?
A type of nonrandom sample, quota sampling is an improvement over haphazard
sampling. In quota sampling, a researcher first identifies relevant categories of people
(ex. Male and female, under age 30, etc.) then decides how many to get in each category.
Thus the number of people in various categories is fixed. It is done because the researcher can ensure that some differences are in the sample. (instead of having all those
interviewed be males under 20 for example)
T 132 What is purposive sampling and why is it used?
Purposive sampling is a type of nonrandom sample in which the researcher uses a wide
range of methods to locate all possible cases of a highly specific and difficult to reach
population. It is used in situations which an expert uses judgment in selecting cases with
a specific purpose in mind. It is used in 3 cases, it is used in unique cases that are
informative, to get to a specific target (prostitutes ex.), or when the researcher wants to
identify particular types of cases for in depth studies. Deviant case sampling – is a
special type of purposive sampling, in which researchers look for cases that differ from
T 133 What is snowball sampling?
Snowball sampling is a type of non-random sampling in which the researcher begins with
one case, then, based on information about interrelationships from that case, identifies
other cases, and then repeats to process again and again
T 138 Can different samples from the same population yield different results and if
Yes they can, this is due to the randomness of the selection, for example. You have 5000
marbles, half are red and half are white, you do a sample of 100 and find out that 35 are
red and 65 are white. You do it again and find out that 51 are red and 49 are white. Now
which of the two are correct?
T 145-146 What is a 3 stage cluster sample, how accurate would it be compared to
simple random sampling and what tradeoff is involved?
A cluster sample is a type of random sample that uses multiple stages and is often used to
cover wide geographic areas, in which units are randomly selected; samples are then
drawn from the selection. The advantage is he or she can create a good sampling frame of
clusters, and the elements within each cluster are physically closer to each other,
lowering costs. Stage 1 is a random sample of big clusters, stage 2 is random sampling of
small clusters within the big clusters, and stage 3 is sampling of elements within the small
clusters. Ex. A researcher wants a sample of people from Toronto, first they would select
streets and blocks, then houses within those blocks, then people within the houses. It is
less accurate however, each stage introduces sampling errors, there is a tradeoff between
cost and accuracy.
T 147-148 Why is RDD necessary and what are its positive and negative features?
RDD = Random digit dialing, it is a special sampling technique used in research projects
in which the general public is interviewed by telephone. It uses all possible telephone
numbers as a sampling frame. Three kinds of people are missed however, people without
landlines, people who have recently moved, and people with unlisted numbers. Very rich
people are also missed. Until the call is made, it is impossible to find out if a number is in
service or not, resulting in a lot of trial and error time wasting. 3. ANALYSIS
Pg. 1 THE PARALYSIS…What are the 3 “w”s of any business, how often should
they be employed and what is necessary for data analysis to lack lethargy?
The 3 w’s of any business are where, who, and what and when. Where refers to where
your potential participants are, and comparing them to areas with low response rates.
Who helps you understand what elements of your marketing plan need to be tweaked and
abandoned. What and when as in what do your players respond to and when do they
respond best, ex. Documenting what days are the most popular. For data analysis to lack
lethargy, it must be transferable into actual operating results.
Pgs. 2 to 6 As restaurant…As a manager how would you construct frequency
distributions, generally what could you use them for and how important would your
interpretive abilities be and why?
A frequency distribution gives a clearer picture of the spending patterns of customers and
may overcome the limitations and distortions inherent in the average check approach.
You place your price range on the x axis, and number of customers on the y axis. You
may notice things such as a cluster may occur because a menu item is particularly
popular. Peaks, clusters, and drops all directly relate to certain menu attributes. You may
discover that there is little price resistance at your location and that you should
experiment with dishes that generate more money.
Pgs. 7 to 11 A Chi-Square guide to save you from doing too much writing in the
T 93-94 What is spuriousness?
Spuriousness is a statement that appears to be a casual explanation but is not, because of
a hidden, unmeasured, or initially unseen variable. It occurs when two things are
associated but are not casually related because a third unseen factor is what relates them.
T 112 What is predictive validity and how can it be measured?
Criterion validity is where an indicator predicts future events that are logically related to
a construct, it can be measured by selecting a group of people who have specific
characteristics, predict how they will do on a certain question, and compare it to the
construct. Pilot testing the data, and therefore validating it.
T 315-321 What does a qualitative researcher organize data into, why is coding not
a simple clerical task, what does coding reduce and what is the one correct way?
Also, how many passes through the raw information are necessary and why and
how do concept maps help?
Pgs. 7 to 11 workbook… Chi Square class notes TO SAVE YOU TIME IN CLASS. 4. SPECIAL RESEARCH CONSIDERATIONS
Pg. 1 Fraud…What did this and might any research fraud cost? As well what causes
researchers to reach incorrect conclusions and how are these prevented?
This research fraud cost science and nature magazine to retract a very popular magazine,
some of the professors junior colleagues lost their publication output, his university
launched a criminal fraud case, he returned his PhD and sought mental health care, the
entire field of social psychology fell under suspicion. Incorrect conclusions can be
brought about by the role of media in hyping counter intuitive studies, the publish or
perish mentality, and the natural human bias. When the experiments are done, we still
have to choose what to believe, raw data helps.
*Pgs. 2 to 5 Ethics In Research…GENNERALLY what are the key issues/rights
involved in the ethics of research and to whom do they apply?
*Pgs. 6 only to the end of 8 THE ANONYMITY…When you become a manager
you’ll thank me for including this reading and putting it on our exam as you’ll need
to know for the exam: Why are questionnaires used on internal customers? Also, is
anonymity (or does the article mean confidentiality?) necessary and how is it best
achieved in such research? Finally, are open or closed ended questions
recommended and why/why not plus what’s the best way to help guarantee
demographic question integrity?
*Pg. 12 Question on…Should ethnic origin be probed and how complex is this and
Ethnic origin should be probed because that information is used by all governments at all
levels to plan programs, by medical researchers to trace genetic diseases, by educators to
monitor language training needs, and by ethno cultural groups. This is complex for
various reasons, the question itself is confusing due to the multiple outlooks you may
have on it. Canadian not being a true ethnic origin, people self-identify themselves with
being Canadian, but may be something else. Birthplace does not help due to ability to be
born somewhere, raised somewhere else, and currently live somewhere else. Ancestral
roots don’t help due to the vagueness and memory of the participant, people don’t know
how far back to go.
*Pgs. 13 to 14 (but only on pg. 14 up to the name “Schuman”. The use…What non-
response options can be used plus are and should they be used and why/why not?
Offering the participant an “I don’t know, not sure, not applicable” option is a common
method of non-response options, though adding these answers increases the proportion of
respondents who give a non-standard answer “floaters” , though these floaters have no
statistical impact. Sometimes there are patterns within the non-response options that as
long as they aren’t random, can yield additional data to the researcher. Some data that
you might gain from them would be that maybe the question is worded oddly? Or maybe
there is a genuine lack of interest in the topic. 2 Advantages of these options are: the
absence of opinion may identify a problem, where opinions are being expressed, they
may be more firmly based upon experience or knowledge. Non-response rates decreased when “don’t know” etc. are added.
T 43 What is scientific