Вы находитесь на странице: 1из 56

ADV 344K- Advertising Research

First Day:
Learn the terms needed to have a basis understanding of research methods
and their role in advertising and PR. Complete research projects that will
teach you how to execute a few basic research tools.

About Kahlor: In journalism, corporate communications, non-profit


communications. Communicating about science and health topics.
Representations of race and gender in mass communication. (Kay-lor)

4th floor, 4.370


512-791-5726
kahlor@austin.utexas.edu

Joshua
ryooyuhosua@utexas.edu
OH: Thurs 10-11 a.m. and by appt. in 4.322B

Lecture based, provided with content not in the reading and supplement it.
Power points posted right before class. Download them and fill them in with
notes. iClicker Reef will be used to create in-class polls to keep us engaged.
Our phone is gonna be the iClicker $15 a year or something.
Canvas discussion posts about news articles that are related to adv/PR.
Check canvas often.
There are online quizzes, first one starts next week. These are entirely based
on readings. Due before class, it’s on chapter 3. They are 10 mins, 8
questions multiple choice. Remember the index and key terms. She drops
the lowest quiz. Usually open a week before.
Three in class exams and an optional final. See syllabus for dates. The
optional final is 12/15 at 7 p.m. It’s multiple choice plus some t/f and she
creates an online review.
There are seven research assignments of varying value. They make up 40%
of the total grade. Overseen by the TAs. Instructions become available about
a week before they’re due. Due dates are already on syllabus.
Research and participation in 2 research studies is mandatory.
Do readings and take notes in class. iClicker polls and discussion posts. Basic
studying and also the online quizzes.
September 4, 2018
Example of quiz: ___ are sources of external secondary data -> B. U.S.
census reports
Plan for today:
Examples of research in the workplace, research in Ad and PR, research
generally speaking
Research in the workplace
Example: let’s assume you’re a com intern working for a nursing home. Your
boss wants you to start a page for the nursing home.
The first question you ask yourself, what do I need to know first?
Who is on facebook anyway? Who is our target audience?
Otherwise, they would be using a valuable resource (the intern) for
something that didn’t make much sense.
Average age of Facebook user -> to be continued pewinternet.com
Before Facebook, look at the internet.

2
 Start with Pew
 Search internet use
 We see that 89% of Americans are online
 What about the other 11%?
o This may or may not matter for the Facebook page problem. Are
there any patterns to this demo? What if we see that a majority
of them were poor Asian immigrants? If your target market is
the elderly, you can see that 34% aren’t online. Less than a high
school education, 34% are also not online.
 That was just internet access, doesn’t speak to social media or
facebook use.
Social Media Fact Sheet
 Almost 70% of all US adults are on social media. 64% of 50 – 64-year
old are on social media. The average age was 40 something.
When we say social media, it’s more than Facebook. First, we look at
Facebook, but let’s see how it compares to others. We learn that about 2/3s
of US adults use Facebook and is popular among all demographic groups.
Are we gonna miss if we rely on Facebook? Looks like we wont.
Most Facebook users visit the site at least once a day and nearly half of US
adults get there new from there per day. 73% of adults use YouTube.
Instagram is at 35%.
 What are we to gain from a Facebook presence?
 Who would look for us?
 How to build traffic or members?
 Who makes decisions about living in a nursing home?
o People providing care at the time
o Who are they?
 Google “caregiver statistics”
These are secondary sources, because it’s data our org did not collect. Since
it’s secondary, you need to make sure that it’s credible. Note the citations in
the report (these are primary sources).

3
What we’re working toward is a SWOT. The dates of the sources matter.
Quick check
We cross check the secondary source with the primary source and look at
how to study was actually done.
The typical family caregiver… We found it on caregiveractionnetwork.com,
but it originated at the NAC.
Steps for locating, verifying data
 Find the original, primary source
 Look for the exact info you need
 Check their citations, and their research methods- how did they get
that info?
 Find the original citations and numbers
 Double-check all the numbers
So, what did we find?
The typical family caregiver is… More than 37% have children or
grandchildren under 18 years old living with them.
Do these people FB-> Do these 49-year-old women use FB?
Yes.
Generally speaking we use research as a tool for overcoming errors in
perception, reasoning and biases. Research leads us to data we can
rigorously evaluate, share.
Terms
Research – to investigate systematically
Sources – where we get info
Data – the info and facts we get
Secondary source – we found the info, but this source is not the original,
primary source for the information/fact
Primary source – this is who generated the information

4
Secondary data – data gathered for a purpose other than your immediate
study or project
Primary data – data collected specifically for a study or project; original data
Research is a tool for discovery
It’s a means to an end and the end is the solution
The resulting data might confirm what you expected or might not and teach
you something entirely new.
Research is science
Problem oriented
- Looks for a solution
Procedure driven
- Requires careful planning and execution
Empirical
- Based on observation, experimentation
Logical
- We can infer sound, consistent arguments that “make sense”
Community-based
- Shared with a network of researchers working to advance knowledge
Phases of adv and PR research
i. Determine research problem
ii. Select the research design
iii. Execute the research design
iv. Communicate the results
We will learn each of these steps across contexts.
Typical variables in investigations
- Demographics, lifestyles (for segmentation)
- Brand/Org awareness

5
- Brand/Org attitudes
- Consumer satisfaction
- Intention (purchase, donate, vote, attend)
- Past behaviors
September 6, 2018
SWOT assignment posted, due next week
Typical designs
- Exploratory – need to dig deep, usually qualitative
- Descriptive – usually based on numbers and statistical analysis, can
correlate multiple factors
- Causal – experimental to isolate cause and effect carefully
Applied vs. Basic Research
- Applied research is designed to answer practical, often immediate
questions
- Basic or pure research is driven primarily by theory or the desire to
build general knowledge – reaches beyond
For the first project we’ll focus on secondary data
We’ll rely on
- Syndicated research companies
- Media and consulting companies
- Trade associations
NEW SLIDES – Searching for data (Google, Mintel, etc.)
Starting point
- Start with background research
- Most of us start with google
- Book also talks about popular sources on page 54
- Gives us a general sense of the product, brand, company, etc.
- Helps develop specific questions to ask when we’re ready for market
data

6
Fenty Beauty
- This semester we will be researching Fenty beauty
- Before we look for any big secondary datasets on this topic we should
start by familiarizing ourselves with terms about the brand
Let’s Google
- Rihanna
o Last name Fenty
o Born in Barbados
o Humanitarian of the year
o Has 9 Grammy’s, 14 number ones on Billboard
o Fourth artist behind Elvis, Mariah Carrey
- Fenty
- Fenty Beauty
- Cosmetics
o Skincare products make up 37% of the cosmetic market, the
total revenue for 2016 was 62 billion ish, done by IBISWorld
o In North America in 2016 the market value of cosmetics was 93
billion with Europe being 114 billion and Asia Pacific being 141
billion done by Euromonitor
- Multicultural Cosmetics
- What else?
- What’s her brand portfolio?
- Other celebrities
Situation Analysis
- After you have a basic understanding you’re ready for the situation
analysis. We have the key terms and have a better idea of the
questions we’re after.
- Our key research question right now is: What is the current situation
for Fenty Beauty within the cosmetics marketplace?
- A situation analysis is the background research for a larger research
effort that will include primary research. It’s going to provide a starting

7
point for any additional research that we’re going to need to do. It’s a
step further, with hard data, and we’re gonna analyze it.
- It’s a snapshot of the sitch in which you are doing business at this
moment in time. What do we need to know before we do anything
else?
Reflection
- In our example from last class we looked for some basic secondary
data on internet use, social media use, caregivers, etc. We found the
secondary data and looked for the primary data within it.
- With more time we could have looked at more in-depth data- maybe
syndicated data to help us learn further. That was the beginning of a
situation analysis.
Show me the data: Mintel
- When we’re trying to go right to the data we can trust
- Look at the reminder of evaluating secondary data on page 52. We can
assume that most of the criteria are met when we use a well-known
data source like Mintel.
- But accuracy in the form of a date is key. How recent was the data
collected? Data from the last two years.
Visit the library website
- Go to the library website and enter what you’re looking for. Or we can
visit the Moody library website.
Mintel
- Mintel is syndicated consumer and market research data
- Expensive as hell but free if we use it through UT
- But if you try to get a Mintel report by going directly to their website
- The reports are over four grand…
- We’re focused on reports from the past two years
- There’s a search function within the report
- Other resources exist, like the marketing librarian website – slides
- Business course complete, BMI is good too

8
Business source complete
- Searches scholarly journals and mass media content: financial data,
books, trade magazine articles, industry reports, market research
reports, company profiles.
Remember what to look for
1. Purpose
2. Accuracy
3. Consistency
Incomplete
Ethnic hair care focuses on organic
Found a trade publication for drug stores
Some resources purpose is to point you toward a valuable resource
Take-away
- In the example on the last slide…
General Library help
- Ask a librarian, even has a chat function
SWOT
- A situation analysis is done to look critically at a current situation for
your product, company, etc.
- We typically draw from secondary data
- The actual analysis happens as you comb through available data and
look for clues
- The snapshot will inform your campaign or project as you move
forward
- SWOT is just one type of situation analysis tool with a memorable
acronym
o Strengths
o Weaknesses
o Opportunities

9
o Threats
- It helps organize the most important info when you need to process a
lot of information or data
- Look at the four-square slide
o The top, SW, are seen as internal to the organization. This means
that Fenty has more control over these aspects.
o The bottom, OT, are external factors and they have less control
over it
 Threats could be legislation hurting your product, etc.
 Opportunity could be something like the changing
demographics of the US. Fenty isn’t in control, but it’s
happening.
 Strengths could be Rihanna’s presence. If it all hangs
together well and tells a story, then it will be good.
It’s a process
- Start with the problem
o Generate some initial questions: Whose Rihanna, what’s Fenty?
- Information search
o Find lots of good info
- Organize info found according to the four-square
o I would call X an opportunity and here’s why
o I found this thing, and it could be a strength if you phrase it this
way and it could be a weakness if you phrase it this way. Pick a
side.
o Have you answered your initial questions?
o See any holes in the story that’s emerging?
- Patch together the SWOT
- Revise/finalize the SWOT to offer a cohesive snapshot where each box
has 4 – 6 important things
Information searching/organizing
1. Find as much info as you can
a. You have enough when you have too much

10
b. Go back and make sure the info is recent and credible
2. Sort the information according to themes that jump out
3. Toss out information that’s irrelevant even if interesting
4. Begin sorting the info to a draft SWOT table
5. Don’t make the info say more than it can
Sorting the SWOT
On slides
Research Problem
(On the assignment)
Strengths (internal, controllable)
On slides for info, all of them are
Note
Your assignment requires 4-6 points per square with
Caution for actual practice
SWOT analysis is subjective. It’s a guide.
September 11, 2018
- Assignment is due Thursday by class time
- Contact info on slides
- Research participation
Todays plan is searching for information using Boolean search logic and
media research
Searching a using Boolean search methods
- Searching the web and databases
- Offer tips on how
Boolean logic relies on AND OR and NOT. It’s recognized by search tools as a
way of defining a search term. George Boole made this
Seven steps to the search process:

11
One example search topic: Asian American consumers and luxury skin care
products.
- Step 1: write out specific information needed
o How much do Asian American consumers spend on skin care? If I
were to enter this into a search engine that isn’t as smart I
would edit out the unnecessary words.
o Get rid of the “how much do” and just have “Asian Americans
consumers spend on skin care”
o Do this with several key questions you need to answer.
- Step 2: expand initial search phrases by using synonyms and/or adding
an “s”
o Trends Asian American consumer
o Trends Asian American consumers
o These subtle differences unearth info
- Step 3: relevance and quality of results can be increased by adding a +
between multiple words
o Skin care + Asian American
o Asian + U.S. + skin care
o Asian living in America not same as Asian-American
- Step 4: conduct a search on several search engines
o Google
o Yahoo
o Bing
o Dogpile
o Pew
o Google scholar
- Step 5: Systematically read findings and reflect on the results
o Which search engines and phrases produced the best results?
o What’s missing? Where might you find it?
o Are there some sources to which you can return and search
more deeply?
o Did you discover new words or phrases that should be used in
additional searches?

12
 Like color cosmetics, that referred to a specific genre of
cosmetics.
- Step 6: revise original search phrases
o Use quotation marks
 “Asian American skin care”
o Narrow by eliminating words
 “skin care” -advertising
o Narrow by focusing on specific types of documents
 Asian American skin care trends:filetype:pdf
o Let search engine find synonyms
 Asian American skin care ~ trends
o Revise original search phrases
 Expand by increasing number of acceptable terms like
Asian American skin care OR
 Focus on and search one individual site “Asian American
skin care” site:webmd.com/
 Focus on web sites that use a key term in a page URL or
page title like Inurl: Asian American skin care Intitle: Asian
American skin care
- Step 7: search deeper, follow key links and references
o Use sites internal search engine
o Follow links to external sources
- Searching for info
o Play around with the 7 steps to see how they work
o Also remember to experiment with different key words
Media Research
Next up: media research is defined differently for this class and has
synergies with other classes. Why search the media landscape? Basically,
best practices. It’s not directly related to the SWOT, it’s a daily surveillance
practice you will engage in once you’re working with a specific brand, client
or organization.

13
Media research in this case refers to your research of media coverage
regarding your client… we are interested in media content as it happens.
The only way to know as soon as it happens is to be on it all the time.
Note on class synergies: in your other media research classes, you focus on
planning your paid media. This lecture is focused on earned media the
coverage generated by others that you get for free, both good and bad.
These are stories on the nightly news, tweets, facebook posts, brand
mentions, etc.
Her hope is that this will help us find media coverage then assess if you
should respond to it. But we’re not gonna plan the responses. We’re just
research based, and the strategies is HOW to respond. I’m teaching how to
survey the coverage so you know what’s going on.
Why monitor the media? Consider disparaging coverage of your
organization on twitter. How would you know it exists? What do you do
about it? Hyundai trolling twitter for negative customer experience.
- Hyundai has appointed a member of its customer connect team to
check twitter to seek complainers. The function is the same where
people have problems and we take care of them, the only difference is
now it’s public.
Why monitor the media?
- Celebrity endorsements
- Once selected you need to be constantly monitored for brand fit,
public behavior, controversy, etc.
Other reasons to monitor the media
- Spotting relevant trends
- What is your competitor doing?
- Other examples?
Trends
- Consider a new trend that you can share with your marketing team.
- Comfort is no longer a bad word…

14
- Fanny packs are in
Competition
- Monitor the media coverage to get a sense of what the competition is
doing. Whole food interns monitor media coverage of its competition.
- How often? Rapid story cycling and 24-hour media coverage means
you need to always be vigilant. In adv, some coverage might mean
pulling ads or rethinking your messaging.
How often
- Rapid story cycling and 24-hour media coverage means you need to
always be vigilant.
- In adv, some coverage might mean pulling ads or rethinking your
messaging
Best practices of media research
- 5 step protocol for insuring you’re informed
o Start internal, in person- talk to people within the organization.
What are you working on? What do you know? What should I
work on?
o Move to internal social media, something they have control over
whether it’s a group or a page
o Move to local, how does the media portray you? Not just
mentions of you, but comparables too
o Move to national, at least checking in several times a day with
national media- know what’s happening within the brand and
industry
o Then use tools, the social media aggregating tools introduced
Use Nexis or Lexis Nexis to search all the news. Must be a subscriber and we
have access to it on the slides
Then use the tools, once you have a sense of what has been going on, turn
on the tools. They give ideas, but not really useful for research.
The tools

15
- Social mention
- Google alerts
September 18, 2018
- Thursday might be an iClicker exercise
- Social mention, assignment in October. It’s a social media aggregator
Social mention and google alerts. Google alerts will help us track keywords
Spreadfast rounds up Instagram analytics and other social media sites
Qualitative methods
Characteristics of qualitative research, the different qualitative methods of
inquiry like observation, and participation.
Typical research designs
- Exploratory – need to dig deep, usually qualitative. We’re going to
explore through our research to find out the next steps.
- Descriptive – usually based on numbers and statistical analysis and it
can correlate multiple factors.
- Causal – experimental to isolate cause and effect carefully.
Overview of qualitative and quantitative research methods
Quantitative research has heavy emphasis on formal closed ended questions
Qualitative research is a collection of data in the form of text or images
using open ended questions. Smaller sample sizes.
Key difference
In quantitative research you have to know what you are looking for in order
to measure it. As qualitative I would see, how do people spend their time
online?
Qualitative research typically begins with an exploratory research question.
Purpose is to learn: how people make sense of their lives, experiences, and
the structures of the world. Qual is focused on exploration and not
measurement. It’s focused on the process of gathering fata, which is
interactive and dynamic. The researcher keeps track of her role in the

16
research, changes directions If the data suggests it. Focused on subjective
meaning
In ADV and PR
We use qualitative to explore phenomena like how people interact with
brands and organizations and what needs a product or service meets. Any
research questions that requires looking closely at individuals.
Qualitative research
The researcher is the primary instrument for data collection and analysis.
The researcher is very important. Build in subjectivity. This usually involves
fieldwork, going into the natural setting.
The process of qualitative research is inductive. Which means that the
researcher builds concepts and theories from the data. The researcher
embraces her subjective role.
Approaches to qualitative research
- There are several approaches
o Some forms of observation, but there is no preconceived
checklist.
o Ethnography
 Focused on observing and recording the human
experience and reality. Look for themes, assumptions to
construct a meaningful summary.
o Interviews
o Focus groups
- What distinguishes them is
o Level of involvement
o Whether conducted in the field versus artificially
- The first two are in the field (ethnography)
o Participant observation
o Non-participant observation
- Next two are not in the field but are still qualitative
o Intensive interviewing

17
o Focus groups
- Participant observation
o Researcher enters the field
o They try to see things as they happen but also becomes involved
o Book offers example of an ethnography of skydiving
- Developing and maintaining relationships
o Must be careful to manage relationships and there is a balance
between observing and participating
o Early interaction is sensitive
- Non-participant observation
o Consumer ethnography, kids tooth brushes
Netnography
- A substitute of ethnography
- The ethnography of online communities
- He wants to know who tweeted with that hashtag and how they
tweeted. He made netbase.
- Clips will be on the exam
- Who is this guy, what is he telling us about netnography, Robert
Kozinets
Is this research generalizable?
- No, the purpose I exploratory
Recording your data: taking notes
- Written notes are the primary means of recording data
- Tape recording/typing distracting in field work
Cannot take lots of notes while in the field since you are focused on
observing, so jot down everything
The brief notes are called jottings and they’re intended to serve the memory
when writing field notes later
Writing field notes should take longer than the observations. What
occupation does this remind you of?

18
Taking notes: careful note taking will yield a bigger payoff. Notes will suggest
new concepts, casual connections and theoretical propositions. Things to
include: the surrounding content, map of the setting, where individuals
were at different times. Notes should also include methodology and your
feelings and thoughts.
September 20th
Interviews, focus groups
Qualitative approaches distinguished by the level of involvement for the
researcher, whether the research is conducted in the field. Complete
observation then participation.
When we use observation and participation, we still focus on our [the
observers] interpretation of things.
Intensive [in-depth] interviewing uses open-ended unstructured
questioning. Seek in depth information on feelings, experiences and
perceptions. “Share with me your thoughts about fast food.” Just trying to
see what’s on their mind about fast food. We might probe their
understanding. Engages interviewees in a dialogue about what they mean
by their comments. Why choose this route?
Asking questions and recording answers.
- Intensive interviews
o Must plan questions
 In this example, we need to know that among the people
who go there, why do they go there and what kinds of
things do they think about when deciding to go there. It
was really skewed toward older white people. They had
ideas of things they thought might be a factor like parking
but they also had more open ended questions like, what
did you expect to get from this? The interviewers also
encouraged other questions.
o Questions must be short and to the point
o Tape recorders are more common than in field work

19
 Example is only in lecture. He and his wife made an event
out of it. His wife makes him go. His wife is educated.
What we found is that UT can be an intimidating place to
people who don’t have a college degree.
Focus groups.
- It’s like an intense interview for a group.
- Centered around a general topic.
- Relatively unstructured population. You could knock out a focus group
in 90 mins with ten people.
- Facilitator [aka moderator] encourages discussion among participants.
Why choose this route? You want to focus on what people are thinking and
feeling but you want to do that in a collaborative environment.
Focus Group: Mad Men
- First clip is the focus group then the next one is where the moderator
is going back to Don Draper to see what they learned in the focus
group and they disagree. Don had an idea and he thought he was
right, but he wasn’t. what did they expect the women would talk
about and what did they really talk about? What was the group made
of? What age were they?
- All women, in their early 20s, 22? Trying to earn trust with the focus
group. Three people watching, the interviewer is Faye. They’re talking
about beauty. What are the things we all do to keep ourselves
beautiful or enhance it? She primes then, “I have a routine.” Talking
about brushing her hair. One of the girls uses just water, another uses
a “night cream” and says that her ex didn’t appreciate her, her beauty.
That same girl was brought to tears by her insecurities, “does he look
at me that way? You can only do your best with what God gave you.”
Then the moderator came in and said, “what about makeup?” She
moved on to the next girl, Sara. One of the guys watching says, “My
strategy was right, they just want to get married and will buy anything
that helps.” Back to the crying girl -> she was with Don, the observer.

20
- Post focus group: Don is meeting with the moderator. Girl: turns out
the hypothesis was rejected. She said she tried routine, ritual, when all
they care about is a husband. Not a routine, not the makeup, just a
husband. So it’s not exactly about the product, it’s about the outcome.
- Kahlor: They thought that routine was really important to them. “Tell
me about your routine. I brush my hair x times…” Ultimately, the
conversation moved to beauty in service to the men in their lives. They
group of women were not diverse, about 22, and they were all
working in the same place. Race gender and race were all very similar.
How do you run focus groups?
- Group discussion of a topic for 1 to 2 hours
- Unrelated individuals
- Should contain between 6 – 10 people, book says 10 – 12 but that’s on
the large side
- Should run several focus groups, so you can compare them
- Generally the groups should be homogenous
o Similarity will bring a better level of comfort. Sometimes women
will differ to the men in the room, cultural norms. If you mix
people with different life experiences, there can be conflict and
it’ll be difficult to have like mindedness. Specific questions [in a
script, or moderator guide] guide the discussion. The moderator
also needs to be flexible and explore what comes up.
Salad dressing bottles (real):
- Gender, age, all of the above. Pretty much all women in their late 30s
to mid to late 50s. one man within the group. You can hear people
playing off of each other. Modern, then adding onto modern. We need
the synergistic communication. It offers portion control, it’s modern,
etc. The group was racially diverse, and it is because we don’t think
there’s a racial breakdown of preference for lids. They think that the
portion control would save you money. Portion control is the buzz
word there.
Salad dressing bottles (SNL):

21
- Three totally different people. Testing hidden valet ranch. Mellissa is
the frat guy and loved HVR. Melissa is eating a ton of it and the calm
guy, Mark, called out the taste perfectly. Mellissa is feeding off of
what Mark said and she’s isolating the other woman in the group.
Mellissa needs the cash, “game on.” She’s trying to come up with the
catch phrase. Sue said, “this could make even my husband eat salad.”
Mellissa saying everyone hates Sue.
- What went wrong? The facilitator offered money which made it a
competition. There was a lot of bias. The facilitator allowed one
person to dominate the conversation and offend the other members
as well. The group was too small.
September 25, 2018
Qualitative analysis
Analysis is how we make meaning from the data. If we’re doing quantitative
analysis, we use numerical representation of the response we gather so we
might locate averages and statistical relationships. We take in the numbers
assigned to the dots (1 – 7).
Just for comparison, this is the qualitative data that enables statistical
analysis. In qualitative analysis, we are working with texts and constructing
meaning more subjectively.
The texts are: transcripts of interviews, videos, field notes, pictures, videos,
printed articles, blogs, etc. You can see photographs and analyze them
qualitatively.
In hot science cool talks, we learned that they made a data out of the event.
In the raw qualitative data: interviews portion, it goes from interview to
respondent back and forth.
Qualitative analysis: focus on text. Exploratory and inductive: data drives
identification of variables of interest. You are the machine that drives
analysis. To fulfill the exploratory goals, analysis must be reflexive (keep re-
evaluating what you think you are seeking) and iterative (keep going
through the data):

22
Once through the data: what am I seeing here? Then again through the data:
am I still seeing that?
Data collection is also reflexive
The data analysis begins as data are being collected. When additional
concepts need to be investigated…
Basic guidelines for analyzing qualitative data
- Remember your driving research questions
- Don’t rigidly pursue your initial expectations
- Try to account for as much of the data as you can in your results
- Be very explicit when you report and interpret your findings
o Don’t assume client will see the same, explain.
Analyzing qualitative data videos (2)
- How to know you are coding correctly: qualitative research methods
o We code information as researchers. Elevate intuition into logic.
Do I code on emotion or historical record? Sequence or
frequency? Each option for coding is not equally important to
the research question. You could code a story as a memory, or a
sequence like a family trip, then boat, then swim. The coding
changed as the questions change.
o It depends on, what are the guiding questions of your research.
What was really important here. When you’re thorough in
memo writing, it’ll show if you were honest and transparent.
- Qualitative analysis of interview data: a step-by-step guide
o Part 1: a description of each step. Reading transcripts. Browse
and make notes of first impressions. Read the transcripts again.
o Part 2: start labeling relevant words, phrases, sentences, or
sections. Labels can be about actions, etc. You might decide
something is relevant to code because it is repeated in several
places, it surprises you, you’ve read something about it in
scientific articles, etc. You can use preconceived notions or not.
You are the interpreter, just make sure you are telling your

23
reader about the choices you make. Go through all the codes
and create categories. Your work now in comparison to the
previous steps is on a more abstract level.
o Step 4: Label the categories. First one can be adaptation.
Describe the connections between the categories.
o Step 6: find on YouTube
For us, we can reflect on data that has existed elsewhere. we talked to
youtubers about Fenty products and come up with some summary then
relate it to some report, which gives it some validity.
Constant comparison
Coding the data into two types of categories:
- Themes that the researcher derived from participants customs
- Theme’s that the researcher identifies as significant to the projects
focus-of inquiry
Telling the story
- At the end of the day, the data needs to tell someone something
cohesive.
- You write down your findings
- Next you create a narrative that ties together the themes that
emerged in the analysis and you tell the story of the data.
- Sprinkle the narrative with direct quotes to add authenticity. Similarly,
this woman said…
Example is class
Technology for analyzing qualitative data
- Atlas, NVivo, MAXQDA are popular programs but there are other
Summary
Analyzing qualitative data is
- Subjective
- Reflexive
- Iterative

24
- Results in
This is all exploratory research
September 27, 2018
- It is 60 questions, multiple choice and t/f
- Arrive early and fill seats in middle
- Get a scantron, some questions will be dropped
- The exam draws from the book and lecture: use study guide
- If you have blanks in ppt, email your peers
- SWOT grades are posted
- ALL students have the option to redo assignment, one per semester.
Jeopardy:
- Name four parts of a research proposal: do not forget an executive
summary at the beginning as well.
- The social media search tool is social mention, and sentiment data
- Watching someone user email is not netnography because
netnography is focused on community
October 4, 2018
- The current assignment is posted and has two parts, the photo and
the paper. In class assignment on Tuesday as well
Where were we, where are we going?
- Secondary research, when this pops in your mind know where to go to
get reliable secondary data. Consumer behavior is on Pew, market
data is on Mintel
- Also talked about media research and monitoring
- Qualitative research through observation, IDIs and focus groups.
Next
- Photovoice, content analysis and quantitative data

Photovoice- applied to consumer and audience research

25
- Often used to address social problems. This falls under public relations
specifically nonprofit public relations.
- Today: what is photovoice, why use it, history, goals, training and
examples then analyzing the data
What is photovoice?
- Participatory photography for social change. It’s qualitative research,
or research that tries to give voice to people in their own voice or
actions to understand phenomena.
- Uses photography to tell a story. Cameras are given to research
participants or they use their own to record their story in photos. You
would ask them to take pictures of something like, “take some
pictures of safety within your community.” In our assignment, that’s
the one we’re working on. Then we’ll look through the photos to
summarize what we’re seeing.
What’s it good for?
- Telling a story which words alone cannot do justice
- Enabling marginalized people to overcome verbal or written
communication barriers
- Allowing for a compelling visual summary of a problem
History of photovoice
- Developed 20+ years ago in an assumption that positive social change
needs to originate with the people- the organization came into the
communities. People can identify the issues of central concern to
them and they can express it in a way they couldn’t verbally.
Goals
- There are three goals for photovoice
o Enable people to record and reflect on their environment,
community, or culture- you ask them to figure out what those
photographs say. You engage them in the entire process.

26
o Promote sharing and critical dialogue among those represented
within the project- Columbia, South Carolina university looking
at a historically black community and the other way around.
o Produce visual images that will resonate with stakeholders
outside the project- they wanted officials to see what this
community has to offer.
o Video (a good summary video)- the plaza assisted living
photovoice project. Elderly people taking lots of pictures and
learning how to use a camera. They showcased the photographs
to family and friends. Why do they do photovoice? To create
beautiful images that they think everyone will enjoy. K: they can
share photographs with the kids, so they can see some art their
parents create and also share with the community.
Process to meet those goals
- Start with conceptualization of the problem
o K: the examples given actually have a problem driving the
projects, like disadvantaged communities not having positive
health outcomes. There’s lots of problems we can define to use
photovoice. In the nursing home, they wanted to give the
residents a voice.
o Identify the audience for that voice
- Facilitate training on how to use the equipment itself
o Devise a guide for taking pictures
- The community takes pictures
- The community analyzes pictures
- A story is crafted
- The story is told to the audience like the client, employees, policy
makers, donors, the media – whoever needs to hear the story
Enabling people to record their lives
- All photovoice projects have a facilitator (sort of like a moderator)
who:

27
o K: it resides with the community but they’re responsible what
happens
o Holds workshops throughout project, before during and after
o Builds trust within the project
o Makes sure that photographers and anyone photographed have
given written consent
o Facilitates the storytelling at the end of the project
Training
- Need to discuss:
o Power dynamics K: You need to think of the power of preserving
information and privacy
o Ethics K: if I were taking photographs of my neighborhood and I
was trying to show what makes my neighborhood scary then I
took an image of someone who can perceived as threatening…
o Potential risks to participants K: if I capture a photograph to
someone drinking underage at a party, I open them up to legal
liability. Do they understand that?
 How to minimize these risks
o How to give photographs back to community members to
express appreciation, respect, or camaraderie K: also going to
help them see appreciate it
Issues for discussion in group
- What is an acceptable way to approach someone to take their picture?
K: talk about how to take pictures of someone and why
- What kind of responsibility does carrying a camera confer?
Example: the following examples shows how a healthcare company, Kaiser
Permanente, used photovoice to explore the communities it serves.
- Why? K: If you picture St. David’s and the community in which they sit,
why might they want to know more about the communities they
serve? Why do they care? If there are a lot of kids in that area, they’re
going to build their hospital around some of these factors like having a

28
pediatric surgeon on call, bilingual or polyglot residents, etc. What are
the kinds of health needs communities have? What are the health
barriers that they have? It’s more in their interest to do prevention
than treatment. KP wanted to, for these reasons, to do a photovoice.
What are some important barriers to them to be more helpful? This
also builds good PR.
- In terms of the bottom line, unhealthy communities are expensive.
- Video: KP community health initiatives, safe place is to be active and
have healthy foods. Make the healthy choice the easy choice. Talk
about what’s going on behind the pictures. People want access to
healthier foods. They want to reduce the impact of crime, gang, and
drug activities. They want safe and welcoming parts and other safe
places for children’s to play. They want safer areas to ride their bikes.
Adam’s city high school students having no sidewalks- the long scary
walk home. Changes that increase biking and walking in future
development decisions. In Modesto, CA, they’re working with council
members and community members. Community health initiatives
where pictures send the message.
Is it generalizable? No.
What is it for?
- Giving a community perspective
- Empowering communities with a voice
- Sharing that voice or story with stakeholders in a compelling way
Raw qualitative data: photos
Analyzing photos
- Unlike other forms of qualitative research, the participants actively
analyze the data
- Typically brings the photographers together to view the photos as a
group K: brings a ton of people to talk about it

29
- Ask members of group to reflect on the analysis together – what do
we see in these photographs? K: you can print them and put them on
the wall and try to find common themes
- As a group, the participate:
o Select photographs for discussion
o Contextualize and story tell
o Code for issues, themes and theories
- This creates a story to be shared with the intended external audience
- One more example, George Green: photovoice is a discussion to
engage participants and creating community level change. They
participated with the Columbia housing authority. KKK story chased
them into the fire department by George Green. Funding is important
to do community research: Kirksey Foundation and internal grant to
the USC office that gave them credibility.
- In summary
o What is photovoice?
o Why use it?
o History
o Goals
o Training
o Examples two on the exam
o Analyzing the data
Strength is the likelihood someone is talking about you
Sentiment is the ratio of positive to negative
Passion is the likelihood that subjects will be talked about repeatedly by one
user
Reach is how many unique people are contributing to the discussion
October 11, 2018
- Exams can be viewed in office hours today. Be there 12:30 to 2:00
- Social mention assignment
- Photovoice is qualitative

30
- Research studies, there will be more
Content Analysis
- So far, we’ve covered how to find secondary data, remember Mintel
and Pew. And we put secondary data to use in the SWOTs.
- We talked about searching media coverage to learn about trends,
competition, sentiment, etc.
Transitioning to Quant
- Content analysis straddles both world of quant and qual, it’s focused
on texts. Related to qualitative in this way.
- But we analyze these texts quantitatively rather than qualitatively
Content Analysis
- Purpose is to find and quantify meaningful content in texts
- Typically, we analyze a collection of texts that represent a given time
frame like news articles from the past six months
Concepts in texts
- We call the object of study a text
- This includes news articles, video games, web pages, songs
- A concept is an idea or terms that summarizes similar observations or
experiences -> celebration of darker skin tones, violence, family,
profanity, and loneliness
- Might look for the presence or absence of these ^
- Typically, a quantitative approach
- That is, we find concepts and we count them then look for
relationships among those concepts using statistical analysis. The pre
and post release of Fenty- looked at coverage before and after
- In a scholarly setting, most content analyses are deductive- this means
it’s driven by theory, hypotheses, concrete expectations. We know
what we expect to see. Need to locate predetermined concepts to test
the hypothesis.
- In a professional setting, still deductive but driven by need to quantify
concepts that are meaningful, matched to business objectives.

31
Example: this example is called good clean fun, a content analyses of
profanity in video games and prevalence across systems and ratings-> are
some platforms more profane than others? Article on canvas.
- Good content analyses.
- Not a platform that is more profane
Coding is the process of finding and counting concepts.
- We have concepts, texts, and coding.
- Define concepts
 What concepts do we want to explore in the texts? A
concept is an idea or term that summarizes similar
observations or experiences. In this case, profanity
- Develop systematic coding scheme
 Develop a systematic means for locating concepts in the
data. What will we look for in these texts and how will we
know when we found it? Needs to be constructed and
given to the sample. The challenge is developing a reliable
coding scheme that will produce consistent results across
coders and texts. Frequency of the seven dirty words is
what they started with.
- Determine sampling strategy and collect the data
 Since you can rarely observe all content, must sample
from available texts. Must establish the universe to be
sampled like video games. Which games do you choose
and from which time frame, platform? NPD Group
provided sales data from x to x and it was x amount of
games from every platform.
- Practice on a subsample
 Good coding requires a ton of trial runs. Purpose for
practicing is to build agreement among coders. Need
intercoder reliability. Grading is a lot like coding. We test it
out together. Challengers are considering the
interpretations of the coders. You need to train people to

32
look for and see the same things. There are two layers to
text- manifest or surface variables and latent variables
which are completely underlying variables. Like how many
times a word came up.
- Execute full sample coding
 Code the full sample of texts. Report the results
meaningfully-> Generally, profanity was found in… Games
containing profanity tended… and it was not related to...
- Analyze the results to find relationships and patterns
 Analyze the results to find relationships and patterns ->
- Report the results meaningfully
Content analysis in practice
- To report on the tone of media coverage as negative or positive
- TO locate words used to describe our product
- To locate potential partners, key players
- See if a new campaign is generating product reviews
Resulting data
- Can help measure success, can inform needs of a new campaign, can
be used to link messages to audiences
October 16th, 2018
Experiments: causation and experimental design
We’re in the quantitative portion of this class
We will cover:
- Experiments
- Cause-effect
- Control
- Confounds
- Hypotheses
- Types of experiments
- Bobo doll experiment
- ABC news experiment

33
Overall types of research
- Exploratory (qualitative is often this)
- Descriptive (quantitative surveys often this)
- Explanatory (quantitative experiments are this)
Experiments
- Explanatory research seeks to identify causes and effects.
- Typically, we use an experiment when a relationship has been
observed and the researcher wants to explain that relationship- what
causes it.
- This is not new to you. Remember your school’s science fair? You have
to have an experimental group and a control group to see if carrying
or manipulating the independent variable impacts the dependent
variable.
Humans are complex
- Great challenge of explanatory social science research is that it’s
difficult to isolate a single factor or a number of powerful predictors of
any social phenomena.
- Humans are complex but you can find patterns. Experiments help
control some of the factors.
What we study- open university video about social psychology which is how
people interact with their environment/others. The cognitive social
perspective, critical social psychology 19/30. Interested in why people
behave the way people behave. Social psychology is how the mind is
structured by society- the science of the socially structured mind.
Experiments
- An experiment follows the scientific method
- An experiment is typically derived by a hypothesis and the experiment
will test the hypothesis.
- Causal effect is the change in one variable leading to the change in
other variable – other things being equal.
Terms

34
- Dependent variable a variable that varies due to the impact of the
independent variable. It depends on the value of the independent
variable.
- The independent variable is not impacted by the dependent variable
and that itself impact the dependent variable.
- Manipulation is the change or different you introduce in order to find
an effect. Causal relationship shows that an independent variable
causes a change in the dependent variable.
Criteria to establish cause-effect between two variables
First three: these are required
- Empirical association is a correlation between the two
- Time order of the independent variable where it preceded the
dependent variable
- Nonspuriousness is not due to other variables.
You must establish the above three to claim a causal relationship
Five criteria, cont.
- A causal mechanism that explains the relationship: why is this
happening?
- Specifying context for whom does this occur and in what conditions
Terms
- Control group is in an experimental design that receives either no
treatment or a different treatment from the experimental.
- Experimental/manipulation/treatment group is the one that receives
the manipulated stimuli
- Confounding variable is an unforeseen, and unaccounted for variable
that jeopardizes reliability and validity of an experiment’s outcome.
Control
- The key to an experiment is control
- Because the purpose is to explain a phenomena, one needs to be
reasonable confident that the explanation being studied is the only

35
viable explanation. True experiments have a control group and a
random assignment to create the test and control groups.
Confounds
- It’s important that one knows all factors in an experiment that means
to control for those confounds.
Theory, Hypotheses
- Theory is a coherent group of tested general propositions commonly
regarded as correct, that can be used as principles of explanation and
prediction for a class of phenomena.
- Hypotheses are predictions regarding testable relationships among
two or more concepts.
- Could be based on a theory. Hypotheses are usually directional
either positive or negative where x is positively related to y.
Hypothesis
- If an experiment is carefully conducted, the results usually either
support or do not support the hypothesis. An experiment can never
prove a hypothesis, it can only add support. We need to see a
relationship supported many times in many situations before proven.
Lab and field experiments
- Isolate relevant theoretical variables manipulate them and see if they
have an outcome. Control and measurement. Labs are an easy place
to do this. They manipulated the design variable where the room is
decorated or not. Significant difference between manipulations. Four
conditions. Video from Open University lab experiment and field
experiment. Group identity and individuality is acknowledged,
performance increases. Start with the question and move into
experimentation.
In the lab: Bandura’s social learning theory

36
- In social learning, Bandura states behavior is learned from the
environment via observation learning: Hypo- exposure to a behavior
causes replication of that behavior.
- Children witnessing an adult role model behaving in an overly
aggressive manner would be likely…
- Would this work when model is watched on TV too?
- Indep Var – witnessing aggressive behavior on TV
- Dep Var – level of aggression in play
October 23, 2018
- Next exam is November 1, on Tuesday. Review guide is published on
Thursdays- go over it then or Friday.
PowerPoint Notes
- Where we left off
- Experimental research on Thursday experiments to test
messages.
- Today
- Message testing ramped up a few slides from last week’s slide
deck. Begin biometrics.
- How do we know when a message work? We can use qualitative data
or quantitative data. We’ll go over both.
- Qualitative approaches are mostly common during message
development – exploratory. We can use focus groups and
interviews. We can show a concept and get feedback on that
concept. Video: focus group on John Kennedy’s focus on
advertising. Truth in advertising. What this is, is the concept (a
poster) then people giving them feedback on the book cover.
- Quantitative research is common to see a measurable impact.
This is good for larger scale message testing. This could be
surveys, paper surveys, interviews, online. Expose people to a
message and get their message using close ended questions. We
can use experiments, where we have a control group that sees
one message and the test group sees another- see if the test

37
group has an impact on our metrics. We can also use eye-tacking
and biometrics. Large samples, randomly selected can be
generalizable. This can be very important when you’re spending
a lot of money on a campaign.
- Message testing
- Sometimes research teases apart message elements (visuals,
sound, text) to see what work best. On the next slide we look at
a testing of visuals using DIAL testing. Positive and negative that
creates a numerical score.
- Visuals at the message: armed forces. Audience segmented
according to disposition toward joining the Canadian armed
forces. CMBTREADY were the group of young men and women
most interested in the combat role. Other categories were off
target, not combat ready, or combat maybe.
- Video: the multiple trend lines, people dipped at around the
fight terrorism mark. Why would have audience sentiment have
dipped when it comes to fighting something?
- Message elements… also channel
- Example: when we looked for information on social media from
Pew. What content really mattered here? The high use of social
media in the US across age categories. The next slide is
perceptions of trustworthiness by channel.
- Quantitative analysis allows for data summarization (descriptive stats)
- Can look at whole sample and subsample. You can see
underlying relationships using inferential statistics- these will be
covered in the next exam.
- Comparing subsamples based on demographic groups, attitude
and behavioral groups. iClicker poll on voting ad, then
vaccination ad. Following slides are data from moms who
viewed a vaccine PSA. The goal is to get moms to vaccinate kids
and what’s measured is the attitude toward vaccines before and
after. PSA characteristics- believability, importance, uniqueness.
I would summarize this as relatively positive on the believability

38
scale and relatively negative on uniqueness. The pre/post
change was significant-> believable, important, and relevant but
not unique.
- Measuring message impact
- The vaccine example was based on self-report measures. How
believable was this ad? But self-report can be biased? The social
desirability bias is how we deny undesirable traits, thin we have
socially desirable traits. Then demand characteristics when we
alter a response because we’re aware of experimentation.
Acquiescence bias is how we tend toward agreement.
- So? We use these measures of self-report. We just know limitations
and it can supplement with other types of data. Observation (earlier in
the semester), experiments (last week).
- Summary to this point
- Message testing can focus on specific message elements and the
channel can influence message outcomes. Message testing can
be qualitative or quantitative. The analysis of message testing
data can identify subsample differences and relationships
among message factors, attitudes and outcomes. Self-report
measures can be biased. One way to get around it is biometrics.
- Up next
- What is biometric research? Eye tracking? Etc.
- Biometric research measures voluntary and involuntary
physiological responses to stimuli like an ad concept or message.
These are not prone to social desirability bias.
- Eye tracking is the most common approach for studying
voluntary responses. Voluntary and it can track consumption
behaviors, responses to messages, etc.
 Consumption behaviors: watching shoppers in IKEA.
 Key measures, attracting attention. The first fixation is
when eyes are locked in one spot. Time to first fixation is
when we can then lock or fixate. Most fixations or where
eyes are fixed repeatedly. Focused attention on elements.

39
 Eye tracking on messages: is recorded without the
participant seeing the results. Videos like this one show
the results overlaid with the video of the viewer. Also note
how fast eyes move. Heatmaps use a graded color scheme
to show visual activity.
 Supplementing eye tracking data with interviews. Show
the eye tacking video and ask them to verbalize thoughts
and feelings they experienced.
 Facial coding is another term of physiological response
data. How Affdex facial coding works. FACS- facial and
head action units, indicators of emotional states.
October 25, 2018
- Look at all the slides. Study all of them with the notes.
Neuromarketing- using biometric research in the context of marketing
- More recently been used in advertising. How people respond and
process to ad information. How people are cognitively responding to
the message in front of them. If we use biometric data, they’re able to
control it in a way without or with less bias.
- Nielsen neuro pitch- consumer neuroscience learn and see videos.
Consumers don’t always do what they say they’re going to do. non-
conscious emotions drive decisions. These are consumer emotions and
responses. Consumers responses to ads, packing, in-store experiences,
and consumption at a granular level. Good to great communication.
Ad Council, public service advertising partnered with Nielsen. They’re
done consumer neuroscience studies. Media fragmentation is a
challenge for them. Avoiding pro-social bias. Use biometric research to
overcome biases in self-reporting.
- Biometric can be helpful for creatives- it can show you elements of a
creative strategy that pop with people. Have mint with blatant flavor
branding/packaging emphasizing fresh breath.
Two types: EEG, and fMRI

40
- EEG is electroencephalography and they have nodes to capture brain
activity. They all essentially do the same thing. It measures and
records the brain’s electrical activity. It provides insight into how much
of the brain while looking at its surface. Cannot pinpoint parts of brain.
- How an EEG works video: things one can get from an EEG episode.
Neuromarketing answers what happens when you watch a
commercial. The first graph is attention, the attention seems high
because the ad was interesting. The second graph is emotional
engagement with the ad. The third is memory retention. A good ad
primes your memory when it’s time to buy. Attention, emotional
engagement and memory are the three things to test with an EEG.
fMRI
- Functional magnetic resonance imagery. This is very intrusive and
expensive, but the data looks inside the brain. It’s looking at blood
flow and oxygenation. It produces cross sections, or tomograms. It
identifies specific activated and non-activated parts of the brain. It
uses blood flow and oxygenation.
Neuromarketing: concerns
- The external validity- there is no exact fit between bio-response and
sales. We can know that a certain part of an ad made people feel
some emotion, but we can’t relate that ad to buying behavior. This
equipment is very expensive. The equipment is very intrusive. Ethics-
can it override freewill? Does it distract focus who should concentrate
on facilitating well-reasoned decisions.
- Neuro Focus rap video: …
October 30, 2018
- Going to talk a lot about vocabulary
- Population
- Census
- Sample
- Elements

41
- Sampling frame
- Generalizability
- Target population
- Bias
- Video: concepts listed above. Sampling: Simple Convenience,
Systematic, Clustered. Sampling method depends on the nature of the
population and the resources available. We want all of the different
peoples/applies to be random. The sample should be split evenly as
well. There will always be sampling error. Simple random sampling,
convenience, etc.
- Simple random is theoretically the ideal method. You use
random numbers to decide which object is in the sample and
this can produce an unbiased sample. More practical when
population.
- Convenience sampling is convenient like taking 20 objects off of
a production line. They are often biased in some way. They can
have self-selection bias because people can choose to join.
- Systematic can be a good approximation of the random sample.
Pick 1 every 20 objects.
- Clustered, subgroups within a city. Cluster sampling can be more
convenient and practical.
- Stratified seems like clustered but the strata are chosen to
present different attributes within the population. Within each
group, a random sample is taken. Can lead to a good random
sample. You need info of the population.
- Population the entire group, set of individuals, or objects to which you
want to generalize. In this case, we’re talking about human
populations in which you want to generalize. A population can be:
news articles, advertisements, video games, and people. The
population should be defined because it’s the group in which you seek
to understand and from which you will draw your sample. What does it
represent? If I need to sample within the city of Austin, I’ll same the

42
whole city of Austin. If I want to sample the Latino community in
Austin, we’ll focus on the Latino community.
- In the video game they used sales data for their sampling frame. It’s a
list of all of the people in the target population we seek to understand.
The population is all video games. Stratified means you’re sampling in
layers to get to it. They focused on the most popular titles which
included the top 15 games for each platform. Continuing with their
sample, they had them play for 30 mins. They took samples within the
video games which made up their actual study sample.
- Human populations are often defined in terms of demographics,
geography, behaviors. If our target population are college students at
Southwestern, four-year universities. I could het self-reported students,
define the southwest by states, and exclude associate degree only
universities. Let’s say your employer is Pfizer and you want to
understand the advertising for pain relievers. The definition of the
target population is ads for pain relievers but that’s too broad, so we
have to break it down to something more specific. Any aspirin,
acetaminophen, ibuprofen-based… And define ads which would be any
paid-for message… Sample examines a subset of population a census
examines every member of the population. Census’ are difficult. It’s
only appropriate when sampling might eliminate important cases.
Credibility requires consideration of all members of population.
Population of interest is small and identifiable.
- Sampling is used more often than census in primary research. A
sample is a subset of a population. It’s a portion of the population that
is more practical to study. We could use a time frame to direct our
sample or geography by looking at the most populous cities in
America.
- Element is the individual unit of the population. In a human population
the element is a person, in a non-human population the element is
whatever you study.
- Sampling fame is a list of all characteristics of a population, we use this
to select the sample. It’s everything in the population.

43
- Generalizability is the extent to which your sample can be said to
characterize your population. One of the main shortcomings of an
experiment is that the results are not generalizable. In exploratory
research, generalizability is less important.
- In an unrepresentative sample, some characteristics are over
represented or under represented- this is sampling bias.
- Probability v nonprobability sampling. Probability relies on random
selection, truly random selection where everyone has an equal chance
of being selected. No one is favored. This is like flipping a coin. What
comes back, is probability. Nonprobability sampling is easier, cheaper,
but hard to generalize because not randomly sampled, so everyone
does not have an equal chance of selection.
- Probability means that the probability of selection is known and
nothing, but chance determines which elements are included. In
the coin flip, you have a 50% chance of being selected. A 1 out of
2 chance. In a census, the probability is 1 out of 1. As the
proportion you are selecting decreases, so does the probability of
selection.
- What do we need to determine before sampling? The population
size and the needed sample size. The needed sample size can be
determined with tools. We need a confidence level and
confidence interval. The CI, or the margin of error means that if
you use a CI of 4, and 47% picked the answer then the actual
percentage could be 43 – 51. They also tell you how sure you can
be. We can have 95% confidence level that this is the true
percentage. For the exam, we have to answer what is the
sample size needed for a 95% confidence level and a confidence
interval of 4 for the population of Austin.
- Types of random samples: Simple random sampling, systematic
random sampling and stratified random sampling. Don’t confuse
random sampling with haphazard. We mean a deliberate
selection process based on probability. Simple random is the
strongest probability sampling method and required procedures

44
so every element has an even chance of selection. Random digit
dialing is one way to select a random sample. We can purchase it
from organizations like Survey Sampling, SII, or Research Now.
Random number generators select numbers by chance, it is used
for drawing a random sample-> you need to assign each element
a number, this one is from Stat Trek.
- Systematic random sampling is a variant of it. The first element
is selected randomly from a list of people or from sequential files.
It only works if there is no bias reason in which they’re in that
sampling frame in a particular order.
- Stratified random sampling. If the population is a city of 10,000
people and the population tell us that the town is 70% white… If I
want to do a stratified sample that is proportionate we want the
sample to mirror the population demographics. If I wanted to do
a disproportionate sample then I would lower the white sample,
raise the Asian, Black, and Hispanic.
- Two problems are often cause for concern when drawing
random samples. If sampling frame is incomplete, the sample
cannot really be random. Always consider the adequacy of the
sampling frame. Nonresponse is a major hazard in survey
research, because non-respondents are likely to differ…
- Nonprobability sampling methods are used in qualitative
research, also in quantitative studies when researchers are
unable to use probability selection methods. Convenience
samples based on ease of location. No concern.
- Quota sampling is intended to overcome the flaw of a
convenience sample. It says I need to keep sampling until we get
to the quota of M to F. Quotas are set to represent certain
characteristics.
- Snowball sampling -> find interviewee then ask for three more
people.
- Variation and sampling error not watching it right now.
November 6, 2018

45
- Rest of class is focused on surveys and questionnaires
- We’ve done secondary research via Mintel, and SWOTs. Also done
media research. The highlights are daily monitoring practices like
social mention. Qualitative research have things like focus groups,
interviews, and observation.
Survey as a tool
- The survey is a tool to collect information or data from a sample of
individuals. We can collect quantitative data for statistical analysis.
That generally uses closed-ended questions to collect self-report
answers to questions.
Surveys are popular because…
- They’re versatile, efficient, and generalizable if random sampling is
used. It’s also good for measuring attitudes through well-tested
indices. We want to know how people feel toward whatever object we
need to know about.
Examples
- Some questions are designed to collect data related to experiences.
How long have you been a member? How likely are you to… First you
see what’s most likely to be the answers they give us.
- Other questions dig deeper for attitudes and behavioral motivations
and intentions. WARC example.
Survey research
- Tends to be descriptive v explanatory. It’s focused on describing
behaviors or attitudes and looking for relationships among variables.
But it can be used for explanatory purposes- to show cause and effect.
Advantages of survey methods
- Accommodate large sample sizes that can be generalized. Produce
precise enough estimates to identify small differences.
Disadvantages
- Questions are difficult to develop

46
Ways to administer surveys
- Person administered, in home, mall intercept, etc.
- Telephone administered with an interviewer. Elon university polls,
random samples, asking about presidential elections, have
quantitative data at the end.
- CATI, YouTube video
- Self-administered like mail.
Online surveys
- Quickest, cheapest, easy to include, not accessible by all, and non-
response is higher than other formats
More Survey Logistics
- Can’t be used for true causal tests. Can be costly if you want a random
sample. It’s a convenience sample using registered people in an
internet panel. Some bias as these users who register to participate in
research panels for pay. A truly random sample of 1000 people from
the general public costed $38,000.
Study Cost
- Price tag is impacted by questionnaire design. There is manpower
needed, material, length is expense. Data analysis is expensive.
Study Timing Requirements
- Telephone, email, and online generally lasts the least amount of time.
Interview length
- Personal contact essential for longer surveys to encourage people to
finish.
Response Rate
- Percent of valid sample. Calculation-> random sample = 1,200.
Remove the duds, - 150. Now down to 1050. Now those failing to
meet sample definition. – 250, down to 800. The rest that didn’t work

47
brings me down to 550. So we have 800 that are valid, and 550 that
participated so the response rate is 550/800 = 68.8%.
Improving Response rate in Online Surveys
- Well-written, motivating invitation, incentivize, links are functional.
Institutional Review Board
- UT requires us to have the IRB to ensure they do not need a formal
review.
Conceptualization
- In order to quantify or measure a phenomenon, we need to articulate
what it is- this is conceptualization. It’s the process of isolating the
concepts that we’re interested in. We learned about it in content
analysis. It’s an idea or term that summarizes similar observations or
experiences.
- Car example, understanding people’s attachment to their cars. Figure
out what attachment is. Operationalization is coming up with the way
of how to measure re something- what is the concept, how will it
surface and vary and how will we measure it with indicators.
- We go from concept, to variable, to indicator.
- Are there other ways to measure a concept? Yes, lots of different ways
to measure every concept.
Think aloud
- Asking about education with an open-ended question: What level of
education have you completed?
- Census: what is the highest grade of school you have completed, or
the highest degree you have received?
- Exhaustive and mutually exclusive. Exhaustive means they capture all
potential answers, and mutually exclusive means that there is no
overlap where one applies to them.
One item or more
- Variables that can be measured with single questions

48
- Common practice for demographics
- What is your gender? Male female other
- Public opinion polls also rely on answers to single questions
Why might a researcher not use one question to measure a concept?
Indexes/indices
- When several questions are used to measure one concepts and
responses are summed or averaged. Can measure things like
someone’s attachment to a product, etc.
Common types of closed-ended survey questions
- Likert scale questions
- Semantic differential options
Likert
- Strongly disagree to strongly agree
- Semantic differential questions can go from expensive to inexpensive
Scoring for online surveys
- Software assigns values to the data being collected
Four levels of measurement
- Nominal, ordinal, interval, and ratio
- Nominal is category, label, name with labels.
- Interval is the distance between attributes does have meaning
but no zero point. Likert scales are technically interval.
- Ratio is that there is an absolute zero point that is meaningful,
on a scale from 0 to 100, how angry are you?
Once you have the concept selected and figured out the right level of
measurement you need to consider the quality of the measures
Quality is validity and reliability. The validity of the measures related to, did
we measure what we intended to? Are operationalizations valid?
Measurement validity
- Three types are face (subjective or looks good)

49
- Content (captures all known dimensions)
- Convergent is when items share 50% or more of the variance in the
index)
- Discriminant is clearly distinct from other constructs
- Reliability is consistent scores over time and contexts. If a measure is
reliable, it is affected less by random error, or chance variation since
we can’t really measure
- Internal consistency is inter-item reliability
- The index of items or questions is reliable if it demonstrates
consistency across responses
- Interpret the Cronbach’s alpha rating using the following ratings
- Less than .6 is not acceptable
- .6 to .65 is undesirable
- .66 - .7 is minimally acceptable
November 13, 2018
- Guest speaker be here on Thursday, we meet next Tuesday
Guest speaker notes:
- Barbie was a cultural icon, 60 years old
- Mattel started making mistakes
- Barbie sold Pinktastic home for $25MM
- Mattel sales suffer, bad press
- BBDO joined early 2015
- Mattel needed business results and had bad habits- they hunger for an
outside opinion
- Complete brand restoration
- Plastic over purpose beforehand, so had to rethink what barbie is
- They sifted through Barbie research
- Dreamtopia ad, was all they did. Very childish product-oriented ad
- The effect was moms getting tired of the messaging, they had all the
data they needed they just wanted to hear anecdotally
- Ethnographies, etc. Mom’s would apologize to others for gifting them
- Breakthrough was 57 years old- lifelike doll

50
- Ruth Handler – through the doll, the little girl could be anything she
wanted to be. The fact that a woman has choices
- Mass marketing does not equal mass mattering – stop all the ads
you’re doing that make the Barbie image worse. You have to mean
something to moms
- The launch campaign of “imagine the possibilities”
- Can you help us with communications? They did social media,
primarily on Instagram tapping into #tbts. Then they launched the
imagine the possibilities piece. Lots of risk to do this. If you reject a
product, it’s very hard for a company to change their minds.
- What happens when girls are free to imagine they can be anything?
You can be anything ad. Massive risk and in terms of research, it’s hard
to test something like this. It’s impossible to project surprise. This ad
was completely on social media at the beginning then ended up doing
cutdowns for TV.
- They needed to have conversations with moms. Their media budget
was pretty small. Create something that mom would become the
media vehicle.
- Ad week jumped on it as the ad of the day, Ad age, ABC, Elle, LA Times,
HuffPost, everywhere. On social media it exploded.
- 19MM views in China, YouTube ad of the year. You need to change
actions, not minds. The “shares” are all the media $s you don’t have.
- 13% sales increase YOY. Holiday sales grew.
- Inclusivity: The day the inclusivity thing launched, they allowed a
Times magazine reporter to be in Mattel for 2 weeks. They found that
it was going to be the cover of Times magazine- Now can we stop
talking about my body? Was the headline.
- Evolution video: the new body shapes and colors ad.
- The launch of the bodies had no paid media behind it. It was the
number one current news story with 5.6 billion impression.
- Goldman Sachs said barbie is back.

51
- Full year later: NFL playoffs, now targeting dads. Moms and daughters
talking about how they played with dads. USA today: the real winner
of the NFL playoffs is Barbie.
November 15, 2018
- Everything is oriented toward this assignment
- Elementary quantitative data analysis: descriptive statistics
- This lecture covers making sense of the data
- What do you do with the data once you get it?
- We’ll talk about type of analysis, statistics in general, central
tendency, and frequency distribution
Types of analysis enables by survey data
- In descriptive analysis, we use stats to describe what variables look
like in our sample. In inferential analysis we look for relationships
among variables.
Stats
- A statistic is a numerical description of a population based on a sample
of that population.
- Useful, basic descriptive statistics include:
- Percentages: any proportion in relation to the whole. Can be
helpful when
- Frequency tables and distribution graphs: we can show
frequency as a table or as a graph. Table or graph.
- Measures of central tendency and variation
Looking beyond the counts
- Look for patterns in the data like actual frequency distribution gives on
perspective. The central tendency and distribution is another.
- Central tendency: usually summarized as the mode, median, and
mean + standard deviation
Choosing the most meaningful central tendency statistic

52
- Depends on: variables and their level of measurement like nominal,
ordinal, etc. Skewness of the variable’s distribution matters, too.
Skewed data
- Can have negative skew, means the left tail is longer and a positive
skew means that the right tail is longer, and the distribution is
concentrated on the left.
Mean = M or X(bar)
- Just an arithmetic average that is not used for nominal data. Best for
interval or ratio – where the numbers have real meaning (strongly
disagree…)
- Mean = sum of values across cases/number of cases
- On a scale of 0 – 10, please indicate how much you feel the following
emotions. A 0 – 10 scale is a ratio scale.
Median = Md
- The point that divides the distribution in half
- Not good for nominal data, no meaningful middle position. Only
used for skewed data like income. The median gives a sense of
the arrangement of values. Arrange the values.
Mode = Mo
- The most frequent value in a distribution or the probably average.
- A distribution can be bimodal, where it shows two categories with
fairly equal peak counts.
Measures of variation
- (check slide)
Range
- The range is a simple measure of variation. It’s the highest value minus
the lowest value and has all the possible values that might be
encountered in the given data. Can be misrepresentative by an outlier.
Standard Deviation

53
- Tells us how spread out the numbers are
Variance
- The average of the squared differences for each case from the mean.
You find the mean, for each number subtract the mean and square the
result, then work out the average of those squared differences
Normally distributed data
- Is bell shaped and has three standard deviations on each side. First
standard +- deviation contains 68%, second is 95%, third is 99.7%.
Qualtrics
- Has a credible URL and name (compared to survey monkey). Offers
some basic analysis and reporting of results. Start using this.
Likert items are interval data.
Innovativeness
- Ask important questions several times in different ways. 5 item survey
index, it is cool to own the first high tech…
Data
November 27, 2018
- Research participation due soon
- Part B1, innovativeness as a personality trait with 5 items to measure
it. The type of response options are Likert scale items with 7 points on
the scale. The items are the statements, response options mean
strongly disagree etc.
- For the screenshot show the evidence that you did it
- Answer items in order on a doc and provide all info that is requested.
Answer everything. Double or single space the text. Use 1’ margins, 12
pt font, no page limit. Attach a PDF or screenshot of the excel
worksheet after you get the means, etc. Throughout round to 2
decimal places except for p values which can be reported at 3 decimal
places.

54
November 29, 2018
Good presentation
- Confidence - 93
- Preparedness
- Conversational flow
- Engaging
- Clear
- Reliance on the presenter, not the ppt
- Knows the content inside and out
- Eye contact
On exam
Communicating research findings
- Establish credibility of research
- Effectively communicate findings
- Interpret those findings as sound and logical
- Serve as a reference for future strategic or tactical decisions
Research reports
- Client should be given a detailed description of
- Research objectives
- Research questions
- Background and relevant secondary data
- Description of the research methods
- Findings displayed in tables, graphs, or charts
- The research report or presentation must establish credibility
- Readers
 Executive summary
 Executive summary and the findings
 Entire report and appendixes
Format for marketing research reports
- Title page
- Table of contents

55
- Executive summary
- Research objectives
- Concise statements of method
- Summary of key findings
- Conclusion and recommendation
- Introduction
- Research methods and procedures
- Data analysis and findings
- Conclusions and recommendations
- Limitations
- Appendices
Really important to label everything clearly

56

Вам также может понравиться