Mejo 379 Midterm Review Guide - UNC Chapel Hill

Research

Inward-directed communication providing answers and opportunities.

Scientific Method

Systematic observations for research in AD/PR. Involves identifying a problem, forming a hypothesis, collecting and analyzing data, and drawing conclusions to create effective, evidence-based strategies.

Importance of Scientific Method

-Gains insights on competition, environment, internal operations and consumers
- Increases overall efficiency (saves $$)
- Helps:
Overcome major separations between firms and their publics
Improve internal operations
Adapt to change

Basic Research

Driven by a scientist's curiosity or interest in some specific question, done to expand human knowledge and conducted to increase understanding of basic principles

Goal of Basic Research

To expand frontiers of knowledge -- public and easily accessible, conducted over long time frame

Applied Research

Intended for a specific purpose or application often in particular proprietary situation

Goals of Applied Research

Scientific and purpose = solve problems. Practical, pragmatic, aids in decision making. Most AD/PR research = applied. Proprietary and not easily accessible to the public, conducted over a shorter period of time

Predictive Research

Uses current knowledge to forecast future changes. Establishes predictability -- helpful for businesses to predict customer/market changes, can predict changes in economy + environment

Subject

Person or object observed in research.

Operational Definitions

Defining a word for the specific purpose of the research

Independent Variable

Changes on its own, influencing dependent variable. The factor you change

Dependent Variable

Variable that changes as the result of some other factor (independent variable) and is measured in response to change

Intervening Variable

Explains the relationship between independent and dependent variables. Serves to explain how or why the effect happens as it operates between the independent and dependent variable

Discrete Variable

a quantitative variable that has either a finite number of possible values or a countable number of possible values

Continuous Variable

Can take any value within a range -- height, temperature

Controlled Variable

Factors kept constant during research.

Confounding Variable

Hidden factors affecting research results.

Hypothesis

Predictive explanation of an event.

Research Question

Indicates what the researcher seeks to know.

Hypothesis v Research Question

Hypothesis is an explanation or expectation of an event. Research questions indicate what the researcher wants to know, the highest priority of research

Null Hypothesis

Unexpected outcome differing from the prediction. Often easier to examine

Population

Well-defined group being studied.

Sample

Subset selected from the population.

Sample v Population

all members of the research population can be possible members of the sample, only a small section of the population = used for the actual sample

Research Plan

Strategy to achieve research objectives. Objective/goal
Seeking a strategy to meet the objective
Action plan needed on strategy implementation

Research Plan Development

Steps: define problem, method, execute, analyze, recommend.

Research Trade-offs

Balancing ambiguity, cost, and time constraints. There are always budget and time constraints, which impacts the research plan that will be developed. You pay for reducing ambiguity in decisions with time and money. Time: can dictate if your sample size is 'good enough' to feel confident about the decision, the sz/scope of a research plan. Money: primary research costs $. Time is also $. You have to understand the research limitations before finalizing the research plan so as not to waste.

Reliability

Consistency in measuring survey responses.

Validity

Accuracy in measuring intended variables.

Reliability v Validity

Reliability- the degree of consistency in measuring survey responses, provided there are no changes in the characteristic being measured (consistently in the same place, the data consistently comes out with similar results) vs Valid- you are actually measuring what you set out to measure

Research vs Intuition

Use research for insights, not just decisions. Determine if research is needed at all

Deductive Reasoning

Moves from known truths to conclusions. Involves specific logic and moves from something known to be true supporting an idea to some conclusion

Inductive Reasoning

Generalizes from specific observations. Makes general conclusions based on individual observations

Primary Research

Collection of new, non-existing data.

Secondary Research

Analysis of existing data and studies.

Primary v Secondary Research

Primary is conducting new research, secondary involves analyzing existing research

Types of Secondary Research

Includes published papers, databases, and government data. Census, Bureau of Labor Stats, Google Scholar / Book Search, Nielsen-online, Dynamic Logic

Advantages of Secondary Research

Low cost (unless syndicated), time for analysis, multiple viewpoints.

Disadvantages of Secondary Research

Limited specificity, limited knowledge on research methodology / data collection, conflicting viewpoints, lack of depth.

Syndicated Research

Comprehensive studies licensed to multiple parties. Provide info on media audiences -- uses: how many people were exposed to your message? What is the reach? How does this target audience use diff types of media? Helps predict future thinking and behavior

SEO

Search Engine Optimization. Natural website visibility in search results.

SEM

Search Engine Marketing Paid advertising to increase search visibility.

SEO v SEM

SEO - web engines find you
SEM - you pay to come up in certain searches

Pay-per-click Advertising

Purchasing keywords for search engine ads. When searched y consumer, post of a text ad at top of page or on side

Qualitative Research

Explores social meanings through interviews and observations.

Quantitative Research

Uses numbers to generalize findings statistically.

Qualitative Data v Quantitative Data

Provides deeper insights into emotions and motivations vs easy to generalize, objective, numbers-based

Advantages + Disadvantages of Qualitative v Quantitative Research

Qualitative data provides deeper insights into audience emotions and motivations but is harder to generalize, while quantitative data is easier to analyze and compare but may miss the "why" behind behaviors.

Source Credibility

Subjective measure of media and sender reliability. -- most often the subject of qualitative research

Concept Testing

testing new product concepts with a group of target consumers to find out if the concepts have strong consumer appeal. Uses interviews and focus groups before ad introduction.

Types of Qualitative Research

Includes personal interviews, focus groups, and ethnographies.

Focus Group

Small session for uncovering beliefs and perceptions.

Focus Group Advantages

Cost-effective, flexible, and ideal for consumer feedback.

Focus Group Disadvantages

Results can't be generalized; moderator skills matter.

Handling Qualitative Data

Stages include collection, analysis, and theory development.

Interpreting Qualitative Data

Insight = gained through observations and interviews,
findings may be highly personalized,
goal to gain in-depth info,
multiple investigators may be needed to collect + analyze findings

Presenting Qualitative Data

good reports should be truthful + insightful, containing both description and theory, should be a narrative account describing the study, its results, drawing theoretical insights and some practical applications

Interviews

One-on-one discussions for detailed topic exploration. Can be 15 minutes - an hour w/ opportunity to probe deeper. Types: personal/individual, dyadic, ethnographic, shop-alongs, mystery shopping (observational)

Interview Advantages

Ideal for sensitive topics; provides in-depth information. Quick + relatively inexpensive

Interview Disadvantages

Results can't be generalized; potential for bias; conducted in somewhat unnatural setting, lack of focus group synergy

Research Ethics

Ensures privacy and voluntary participation in studies.
Ethics: researchers are dealing with people - all have the right to privacy and to participate in a study only when they volunteer willingly and are not deceived in any manner. Researcher has responsibility to: exercise care in gathering and processing data and taking responsible steps to assure the accuracy of the results. Responses should not be fabricated / altered / discarded + researchers should not conceal info. Every research report should contain a full and complete description of the methodology

Milgram Experiment

a study of obedience by testing whether participants would follow orders, even when it caused apparent harm. Highlights ethical issues in research like the need for informed consent, the right to withdraw, and the protection of participants from psychological distress

Nominal Data

Categorical data without order or structure.
lowest scale of measurement, from statistical POV. Data = simply placed into categories w/o order, value, or structure → classes/categories are listed so all possible response options are mutually exclusive and collectively exhaustive, no mean scores may be calculated from nominal data

Ratio Data

Highest measurement scale with a true zero point.
the highest scale of measurement; often not used in social research. Response choices = ordered, with known/equal distance between choices, has a TRUE ZERO POINT Simplest examples of ratio scales of measurement: height, weight, age, length

Ordinal Data

Ordered attributes without standard distance between choices.
provides order of attributes or characteristics, report the order or rank of responses from smallest - greatest, best — worst, first - last. Provides opportunities for researchers to define attributes or characteristics in an ordered sequence. No standard distance between choices

Likert Scale

Measures agreement level with a statement (not a question) -- "satisfied"

Interval Data

Standard survey rating scale. Linear scale in which there is equal distance between choices; no true zero.

Reliability in Surveys

Degree of consistency of measuring responses as long as there are no changes in characteristics being studied

Standard Survey Rating

Interval scale with equal distances between integers.

Survey Creation Tips

Use established measures; avoid ambiguous wording. (Double negatives, double barreled questions, leading questions, loaded questions)

Likert vs Semantic Scales

Likert - Frequently used to measure attitudes and opinions, ordered set of responses from 1 extreme to the other (ordinal or interval msmt), way better than asking yes/no questions VS Semantic - ratings between 2 opposite adjectives, similar to a likert scale, no word-labeled points except for anchors, also way better than yes/no

Standard Deviation

Indicates how well mean represents data set. Used to tell how far away, on average, are the recorded figures from the mean. The smaller one = more desirable

Measure of Central Tendency

Includes mean (sum of/# of scores), median= middle response when all scores are lined up from least to greatest, and mode = response that happens most often in calculations.

Descriptive Statistics

Summarizes data through variables like mean.

Inferential Statistics

Tests relationships between hypothesized variables.

Descriptive v Inferential Statistics

Descriptive - summarizing the data through each variable, used to condense data (ex mean) VS Inferential - used to simplify and draw conclusions from data. When you want to test relationships between variables that have been hypothesized about

Confidence Levels

We want to be confident that our results are right. Indicates likelihood actual average is within range.

Correlation vs Causation

Correlation shows relationship; causation requires control.

Correlation: 2+ sets of variables are related to one another (or co-related). Empirical research can't prove that one set of variables causes the others to change, only that they change together --> causation cannot be meaningfully inferred from correlation

Lab Experiments

offer more control over experiments (vs field), they tend to maximize internal validity but do so at the expense of external validity bc the lab = extremely artificial, typically not resembling a natural setting

more natural environment (individuals = subject to more variables) → field experiments increase external validity