Please enable JavaScript.

Coggle requires JavaScript to display documents.

(Communication Research) - Coggle Diagram

- - - - Explore / Describe
        
        Used when a topic is not well understood
        
        Asks what exists, what happens, or how people experience something
      - Explain / Understand
        
        Looks for reasons, meanings, relationships, or processes
      - Predict
        
        Estimates what is likely to happen based on known patterns or variables
      - Identifies why the research is being conducted
    - - Why does this study need to happen?
      - Identifies an issue or gap that needs investigation
      - Connects to Literature Review
      - Leads to Purpose Statement
    - - What do we already know?
      - Identifies what we know, do not know, disagreements, theories, and gaps
      - Shows where the study fits
    - - What will this study do?
      - States what the study intends to accomplish
      - Quantitative
        
        Examine relationships
        
        Compare groups
        
        Test predictions
      - Qualitative
        
        Explore experiences
        
        Understand meanings
      - Mixed Methods
        
        Combines quantitative and qualitative approaches
    - - Research Question
        
        What do I want to know?
      - Research Hypothesis
        
        What do I expect to find?
        
        Testable prediction
      - Null Hypothesis
        
        Predicts no relationship or difference
      - Alternative Hypothesis
        
        Predicts a relationship or difference
    - - Who or what is being studied?
      - Entity about which conclusions will be made
      - Examples
        
        Individuals
        
        Couples
        
        Groups
        
        Organizations
        
        Messages
        
        Events
    - - Refines and specifies abstract concepts
      - Clarifies what a researcher means by a concept
      - Construct
        
        Abstract idea that cannot be directly observed
        
        Examples: trust, anxiety
      - Dimensions
        
        Specific facets of a concept
      - Indicators
        
        Observable signs representing a concept
      - Connection
        
        Concept -> Dimensions -> Indicators
    - - Turns concepts into something that can actually be observed or measured
      - Operational Definition
        
        Precisely states how a concept will be measured
      - Connection
        
        Conceptualization -> Operational Definition -> Measurement
      - Survey Research Connection
        
        Surveys often operationalize variables by asking structured questions
        
        Questions become indicators of the variable
        
        Good operationalization requires reliable and valid questions
  - - - Evidence gathered through observation
    - - Careful and deliberate observation
      - Variable
        
        Characteristic that can take more than one value
      - Attribute
        
        Specific value or category of a variable
      - Levels of Measurement
        
        Nominal
        
        Categories with no meaningful order
        
        Memory Aid: NAME
        
        Example: communication channel
        
        Ordinal
        
        Ordered categories
        
        Distance between categories is not necessarily equal
        
        Memory Aid: ORDER
        
        Example: satisfaction ranking
        
        Interval
        
        Equal distance between values
        
        No true zero
        
        Memory Aid: EQUAL INTERVALS
        
        Ratio
        
        Equal intervals + true zero
        
        Memory Aid: REAL ZERO
        
        Examples: age, time, number of messages
  - - - Consistency of a measure
      - Ask: Would I get consistent results?
    - - Does the measure actually represent the concept it is supposed to measure?
      - Ask: Am I measuring the right thing?
    - - A measure can be consistent but still measure the wrong thing
      - Reliability = consistency
      - Validity = accuracy of representation
    - - Extent to which the operational measure faithfully represents the concept
      - Important in Quantitative Text Analysis
        
        Researcher's coding categories must accurately represent the meaning of the text
  - - - Theory -> Observation
      - Start with theory
      - Develop prediction
      - Gather data
      - Test prediction
    - - Observation -> Theory
      - Start with observations
      - Find patterns
      - Develop explanation
    - - Deductive Coding
        
        Categories come from existing theory or research
      - Inductive Coding
        
        Categories emerge from patterns in the data
  - - - Total group of interest
    - - Subset studied
    - - How well the sample reflects the population
    - - Extent findings can apply beyond observed cases
    - - Random selection
    - - Not based on random selection
    - - Survey research involves
        
        Develop questionnaire
        
        Select sample of respondents
        
        Administer questionnaire
  - - - Survey Research
        
        Uses structured questions to collect quantitative data from respondents
        
        Often used to operationalize and measure variables
      - Respondent
        
        Person who answers the survey
      - What Surveys Can Measure
        
        Factual Knowledge
        
        What people know
        
        Cognitive Beliefs / Perceptions
        
        What people believe or perceive
        
        Affective Feelings
        
        Emotional evaluations
        
        Positive/negative, like/dislike, favorable/unfavorable
        
        Self-Reported Behaviors
        
        Past behaviors
        
        Present behaviors
        
        Intended future behaviors
        
        IMPORTANT: what people SAY they do may differ from what they actually do
        
        Traits
        
        Relatively stable disposition or orientation
        
        Example: general communication anxiety
        
        States
        
        Temporary or situation-specific disposition
        
        Example: anxiety during a particular conversation
        
        Communication Networks
        
        Who communicates with whom
        
        Can show how information and influence flow
        
        Demographics
        
        Age
        
        Education
        
        Socioeconomic status
        
        Race / ethnicity
      - Challenge of Asking Good Questions
        
        Memory Problems
        
        People may forget or inaccurately remember past events
        
        Social Desirability Effect
        
        People may give the answer they think is socially acceptable
        
        Response may differ from actual behavior
        
        Acquiescence Response Bias
        
        Tendency to agree or say yes
        
        Response Set
        
        Tendency to repeatedly choose the same type of response without carefully considering each item
        
        Wording Effects
        
        Small changes in wording can change answers
        
        Order / Framing Effects
        
        Earlier questions can influence how later questions are interpreted
        
        Cultural Bias
        
        Questions may not carry the same meaning across cultural groups
      - Good Survey Questions
        
        Clear
        
        Respondent should understand the intended meaning
        
        Short
        
        Usually better than unnecessarily long questions
        
        Single Focus
        
        Avoid double-barreled questions
        
        Double-Barreled Question
        
        Asks two things at once
        
        Makes it unclear which part the respondent is answering
        
        Avoid Negative Wording
        
        Negative wording may confuse respondents
        
        Avoid Bias
        
        Bias encourages a particular response or point of view
        
        Ask Questions Respondents Can Answer
        
        Do not assume knowledge they may not have
        
        Ask Questions Respondents Are Willing to Answer
      - Open-Ended Questions
        
        Respondents provide their own answers
        
        Can provide richer information
        
        Responses must often be coded before quantitative analysis
      - Closed-Ended Questions
        
        Respondents select from provided answer choices
        
        Easier to quantify and compare
        
        Response options can influence answers
      - Composite Measures
        
        Combine several indicators into one measure
        
        Often provide better representation of complex concepts than a single indicator
        
        Indexes
        
        Simple accumulation or combination of several indicators
        
        Scales
        
        Use an underlying intensity or structure among indicators
      - Likert-Type Items
        
        Respondents indicate degree of agreement or disagreement
        
        Example
        
        Strongly Agree -> Agree -> Neutral -> Disagree -> Strongly Disagree
        
        Frequently used in communication research
        
        May use several items together as an index
        
        Reverse Scoring
        
        Used when some items are worded in the opposite direction
        
        Reverses numerical values so all items point in the same conceptual direction
      - Semantic Differential
        
        Respondent rates something between bipolar opposites
        
        Examples
        
        Good -------- Bad
        
        Honest -------- Dishonest
        
        Common dimensions
        
        Evaluation
        
        Good vs Bad
        
        Potency
        
        Strong vs Weak
        
        Activity
        
        Fast vs Slow
      - Thurstone Scale
        
        Judges determine the intensity of potential indicators
        
        Attempts to create items with equal-appearing intervals
        
        Time-consuming and less commonly used
      - Guttman Scalogram
        
        Items have an order of intensity
        
        Accepting a more intense item implies acceptance of less intense items
      - Comparative Judgments
        
        Respondents directly compare phenomena
        
        Can examine
        
        Similarity
        
        Psychological distance
        
        Preference
        
        Rankings
        
        Paired Comparison
        
        Each phenomenon is compared with another
      - Question Order
        
        Order can influence responses
        
        Funnel Format
        
        Broad/open questions -> narrow/closed questions
        
        Reverse Funnel
        
        Narrow questions -> broad questions
        
        Coherence Order
        
        Group questions addressing similar topics together
        
        Contingency Questions
        
        Whether a question is answered depends on an earlier response
        
        Example: If YES -> answer the next question
        
        Counterbalancing
        
        Different respondents receive questions in different orders to examine order effects
      - Questionnaire Design
        
        Keep layout clear and uncluttered
        
        Provide enough space
        
        Use clear instructions
        
        Matrix Format
        
        Several questions use the same response choices
      - Pretesting
        
        Test questionnaire before the actual study
        
        Helps identify
        
        Ambiguous questions
        
        Questions people cannot answer
        
        Confusing instructions
        
        Design problems
      - Modes of Survey Administration
        
        Self-Administered Questionnaire
        
        Respondent completes questionnaire independently
        
        Face-to-Face Interview
        
        Interviewer asks standardized questions
        
        Telephone Survey
        
        Questions administered by phone
        
        Computer / Online Survey
        
        Technology used to administer and collect responses
        
        CATI
        
        Computer-Assisted Telephone Interviewing
        
        CAPI
        
        Computer-Assisted Personal Interviewing
        
        CASI
        
        Computer-Assisted Self-Interviewing
      - Strengths of Survey Research
        
        Useful for describing large populations
        
        Can collect many variables
        
        Standardization
        
        Same questions asked in the same way
        
        Useful for measuring attitudes, beliefs, perceptions, and self-reported behavior
      - Weaknesses of Survey Research
        
        Can be superficial for complex topics
        
        Limited context
        
        Can be inflexible
        
        Artificiality
        
        Measures self-reported behavior rather than actual communicative action
      - Secondary Analysis
        
        Uses survey data previously collected by someone else
        
        Advantages
        
        Faster
        
        Cheaper
        
        May use high-quality existing data
        
        Disadvantage
        
        Original measures may not perfectly fit the new research question
        
        Validity becomes an important concern
    - - Quantitative Text Analysis
        
        Systematic quantitative analysis of communication messages
        
        Examines content, function, form, or structure of messages
        
        Includes
        
        Content Analysis
        
        Interaction Analysis
      - Text
        
        Any behavior, object, or artifact involving symbol use
        
        Can include written, spoken, visual, or behavioral communication
      - Content Analysis
        
        Systematically analyzes the content of communication messages
        
        Useful when researchers want to study actual messages rather than people's memories of messages
      - Interaction Analysis
        
        Systematically examines actual interaction
        
        Looks for patterns in who communicates what to whom and how
        
        Useful for studying group communication and interaction processes
      - Why Analyze Text?
        
        Participants may not accurately remember message details
        
        Participants may be unable to report communication patterns
        
        Historical communicators may no longer be available
        
        Allows researchers to examine communication over time
      - Sampling Texts
        
        Probability Sampling
        
        Can be used when a sampling frame of texts exists
        
        Nonprobability Sampling
        
        Can be used when texts are deliberately selected
      - Units of Analysis / Meaning Units
        
        Researcher determines what portion of the text will be coded
        
        Unit choice should match the research question
      - Manifest Content
        
        Surface-level, directly observable content
        
        Examples
        
        Specific words
        
        Specific behaviors
        
        Frequency of an observable message
        
        Strength
        
        Usually easier to code reliably
        
        Memory Aid
        
        Manifest = what is plainly THERE
      - Latent Content
        
        Underlying or inferred meaning
        
        Example
        
        Overall theme or tone of a message
        
        Strength
        
        Can capture deeper meaning
        
        Limitation
        
        Requires more interpretation
        
        May reduce coder reliability
        
        Memory Aid
        
        Latent = meaning UNDER the surface
      - Coding
        
        Process of assigning units of data to categories
        
        Turns communication content into data that can be counted and analyzed
        
        Code Categories
        
        Operational definitions of the attributes being studied
        
        Should be
        
        Mutually Exclusive
        
        A unit should fit into only one category when categories represent alternatives
        
        Exhaustive
        
        Categories should cover all relevant possibilities
      - Deductive Coding
        
        Categories developed from theory or previous research
        
        Theory -> indicators -> coding categories
      - Inductive Coding
        
        Categories developed by examining the data and finding patterns
        
        Observations -> patterns -> categories
      - Coder Reliability
        
        Extent coders classify data consistently
        
        Unitizing Reliability
        
        Coders agree about where units begin and end
        
        Content Reliability
        
        Coders categorize the same units similarly
        
        Manifest content often produces higher coder reliability than latent content
      - Validity in Text Analysis
        
        Coding system must represent the intended meaning of the communication
        
        Representational validity is especially important
      - Strengths
        
        Studies actual communication messages
        
        Does not rely only on participant memory
        
        Can study communication over time
        
        Unobtrusive
        
        Researcher often does not affect the communication being studied
        
        Can often recode data without recollecting the original messages
      - Weaknesses
        
        Meaning may be difficult to infer
        
        Latent coding can be subjective
        
        Coding may be time-consuming
        
        Researcher assumes coding categories adequately represent message meaning
      - Connection to Survey Research
        
        Survey
        
        Asks people to REPORT their communication
        
        Text Analysis
        
        Examines the ACTUAL communication messages
      - Connection to Conceptualization
        
        Concept -> Indicators -> Coding Categories -> Numerical Data
  - - - Quantification
        
        Process of converting data into numerical form
        
        Needed so statistical software can analyze data
        
        Connection
        
        Observation -> Coding -> Numbers -> Statistical Analysis
      - Case X Variable Matrix
        
        Organizes quantitative data for computer analysis
        
        Rows
        
        Cases / participants / units of analysis
        
        Columns
        
        Variables
        
        Cells
        
        Value or attribute for a particular case on a particular variable
      - Coding Data
        
        Assign numerical codes to attributes
        
        Example
        
        Variable: response category
        
        1 = Strongly Disagree
        
        2 = Disagree
        
        3 = Neutral
        
        4 = Agree
        
        5 = Strongly Agree
        
        IMPORTANT
        
        Numbers used as codes do not automatically mean the variable is interval or ratio
        
        Level of measurement still matters
      - Codebook
        
        Document explaining
        
        Variable names / identifiers
        
        What each variable represents
        
        Numerical codes assigned to attributes
        
        Acts as a guide for understanding the data file
      - Data Cleaning
        
        Checks data for errors before analysis
        
        Possible-Code Cleaning
        
        Checks that only valid/possible codes appear
        
        Example
        
        If responses can only be 1-5, a value of 8 signals an error
        
        Contingency Cleaning
        
        Checks whether only cases that should have data actually have data
        
        Connected to contingency questions in surveys
      - Descriptive Statistics
        
        Describe and summarize the data collected
        
        Do NOT by themselves determine whether population differences or relationships are statistically significant
        
        Frequency Distribution
        
        Shows how often each value or category occurs
        
        Percentage Distribution
        
        Shows proportion of cases in each category
      - Central Tendency
        
        Describes the typical or central value
        
        Mode
        
        Most frequently occurring value
        
        Can be used with nominal data
        
        Median
        
        Middle score
        
        Half of scores fall above and half below
        
        Useful with ordinal data or distributions containing extreme values
        
        Mean
        
        Arithmetic average
        
        Add scores and divide by number of scores
        
        Sensitive to extreme values / outliers
      - Outlier
        
        An unusually high or low score
        
        Can distort the mean
        
        Median may better represent the center when strong outliers exist
      - Normal Distribution
        
        Bell-shaped distribution
        
        Mean, median, and mode are identical in a perfectly normal distribution
      - Dispersion
        
        Describes how spread out scores are
        
        Range
        
        Distance between highest and lowest value
        
        Interquartile Range
        
        Range containing the middle 50% of scores
        
        Standard Deviation
        
        Indicates amount of variability around the mean
        
        Small SD
        
        Scores are clustered close to the mean
        
        Large SD
        
        Scores are more spread out
        
        Very important descriptive statistic in communication research
      - Z-Score
        
        Standard score
        
        Tells how many standard deviations a score is above or below the mean
        
        Positive z-score
        
        Above the mean
        
        Negative z-score
        
        Below the mean
        
        Mean of a set of z-scores = 0
        
        Useful for
        
        Comparing scores measured on different scales
        
        Standardizing ratings from different measures or raters
      - Level of Measurement Determines Appropriate Statistics
        
        Nominal
        
        Central Tendency
        
        Mode
        
        Ordinal
        
        Central Tendency
        
        Mode
        
        Median
        
        Dispersion
        
        Range
        
        Interquartile Range
        
        Interval / Ratio
        
        Central Tendency
        
        Mode
        
        Median
        
        Mean
        
        Dispersion
        
        Range
        
        Interquartile Range
        
        Standard Deviation
        
        Connection to Inferential Statistics
        
        Interval / Ratio
        
        Can support parametric statistics when assumptions are met
        
        Nominal / Ordinal
        
        Often require nonparametric statistics
      - Subgroup Comparisons
        
        Describe different groups separately
        
        Can compare frequencies, percentages, means, and standard deviations
        
        Descriptive differences alone do NOT establish statistical significance
      - Contingency Table
        
        Table showing values of one variable according to values of another variable
        
        Often used for subgroup comparisons
        
        Independent Variable
        
        Grouping or explanatory variable
        
        Dependent Variable
        
        Outcome being described
      - Percentaging Tables
        
        Percentage Down
        
        Columns total 100%
        
        READ ACROSS
        
        Percentage Across
        
        Rows total 100%
        
        READ DOWN
        
        Memory Aid
        
        Percent DOWN -> Read ACROSS
        
        Percent ACROSS -> Read DOWN
  - - - Ask
        
        Am I describing data, testing differences, or testing relationships?
        
        What is the level of measurement?
        
        How many groups?
        
        Are groups independent or related?
        
        Are assumptions satisfied?
    - - Describe what the sample data look like
      - Frequency
      - Percentage
      - Mean
      - Median
      - Mode
      - Range
      - Standard Deviation
    - - Used to make judgments about differences or relationships beyond the immediate sample
      - Builds from the descriptive statistics introduced in Chapter 11
    - - Generally associated with interval/ratio quantitative data
      - Make assumptions about population distributions
    - - Generally associated with nominal, ordinal, ranked, or categorical data
      - Make fewer distribution assumptions
    - - p Value
        
        Probability of obtaining results this extreme if the null hypothesis were true
        
        Small p -> evidence against null hypothesis
        
        Common cutoff
        
        p < .05 = statistically significant
        
        p > .05 = not statistically significant
        
        IMPORTANT
        
        p does NOT tell you how large or important an effect is
    - - Shows magnitude or practical strength of a difference or relationship
  - - - Conceptualization -> Dimensions -> Indicators -> Operationalization -> Measurement
    - - Survey Research
    - - Open = respondent creates answer
      - Closed = respondent selects provided answer
    - - Complex concepts often need multiple indicators
      - Composite measures can improve representation of a variable
    - - Likert = degree of agreement with statements
      - Semantic Differential = rating between bipolar opposites
    - - Trait = relatively stable
      - State = situation-specific / temporary
    - - Respondent gives socially acceptable answer rather than necessarily reporting actual behavior
    - - Survey = asks people to report
      - Text Analysis = systematically examines communication messages
    - - Manifest = visible / surface content
      - Latent = underlying meaning
    - - Deductive = theory first
      - Inductive = data/patterns first
    - - Reliability = consistency
      - Validity = measuring/representing the intended concept
    - - Turning observations/data into numerical form for analysis
    - - Guide explaining variables and numerical codes
    - - Mean = average
      - Median = middle
      - Mode = most frequent
    - - How spread out scores are around the mean
    - - How many standard deviations a score is above or below the mean
    - - Nominal -> Mode
      - Ordinal -> Mode / Median
      - Interval or Ratio -> Mode / Median / Mean
    - - Descriptive = summarize the data you collected
      - Inferential = test differences/relationships and make inferences beyond the immediate sample