SCOM280 Exam 1

Communication

- Process of sharing meaning with others

- who (communicator) -> says what? (message) -> in which channel? (medium) -> to whom? (recipient) -> with what effect? (effect)

Theory

1) textbook definition: a formal statement of rules on which a subject is based, or an explanation of the relationship between variables

2) simpler definition

• way of making sense of the world

• organizes information

• sensitizes us to what is important

3) Various approaches to theory

• social scientific

• interpretive

• critical

Research

detailed or in-depth study of a subject to reach a greater understanding or to obtain new information about the subject

Method and methodology

1) method

• systematic technique/procedure used to do research

• each method has “rules” or principles to follow (“best practices”)

2) methodology

• the study of a particular method

• how we arrive at these “best practices”

Definition of ethics

1) textbook definition

• the actions, thoughts, values, principles, and communication practices for determining how to interact with others

2) dictionary definition

• principles of right and wrong

Ethical principles

1) respect for individuals (autonomy)

• a person has a right to make their own decisions

• informed consent: provides participants with study details, risks/benefits, informs them of their right to discontinue participation at any time, and researcher(s)’ contact info

2) beneficence

• researchers should act in the best interest of participants

3) Justice

• research should be fair and equitable

Institutional Review Board (IRB)

• reviews (and approves) research proposals

• responsible for enacting codes of conduct

• monitors research in progress

Privacy (confidentiality vs anonymity)

1) confidentiality

• researcher does not share names or information about participants

2) anonymity

• researcher does not know the names of participants

Deception/debriefing/confederate

1) deception

• giving false or misleading information to participants

2) debriefing

• if deception is used, researcher must explain true purposes of research study after participation is completed

3) confederate

• a person who secretly takes part in a research study in the guise of a participant; participants must be told in debrief

Paradigm

• an approach to research

• three major paradigms: social scientific, interpretive, and critical

Scientific method

theory proposed and/or revised → predictions made (hypotheses) → observations made (go in loop)

social scientific paradigm

• approach to research using empirical observations to test theories that explain and/or predict human behavior

• application of the scientific method to study human behavior (rather than natural)

  • assumes single, observable reality

empiricism & objectivity (social scientific paradigm)

1) empiricism

• can only research what we can observe/see

2) objectivity

• researcher should try to ensure that their personal biases do not interfere with the research and/or predictions

hypothesis (social scientific paradigm)

• educated guess (prediction) about the relationship between two or more variables

generalization (social scientific paradigm)

• search for patterns (or differences) that occur over different circumstances

• examples:

  • narrative vs statistical messages (organ donation); medium of presentation

  • speech anxiety and public speaking (difficulty of content)

the interpretive paradigm

• approach to research focused on understanding individuals’ subjective predictions

• assumes reality is socially constructed

  • individuals have distinct and equally legitimate interpretations of reality

• rejects social scientific ideal of objectivity

  • researcher is inseparable from the research context (subjectivity)

Critical paradigm

• approach to research focused on power, inequality, and social change

power (critical paradigm)

• power = ability to direct or control others

• coercive power: ability to punish behavior

• reward power: ability to reward behavior

• legitimate power: power from position (elected, appointed)

• persuasive power: ability to persuade

power in critical theories

• critical researchers typically examine power at the level of societal structures— who has power and who doesn’t; power and social inequality

• commonalities in critical theories:

  • idea that existing social structures are maintained by those in power (e.g., religion, education, media)

  • goals

    • expose power structures and social inequalities; advocate for social change

discourse (critical paradigm)

• discourse = totality of language use

• discourse constructs reality

research warrants

• warrant = assurance of results

measurement (social scientific warrants); conceptualization; operationalization

• accurate measurement is central to both the natural and social sciences

• conceptualization: how a construct is defined

• operationalization: how a construct is measured (i.e., open ended items, close-ended items, attitudes, behavioral intentions or observations)

Validity (social scientific warrants); content; criterion (concurrent and predictive validity)

• does the instrument actually measure what it says it does?

• content: does the instrument measure all of the necessary aspects of a construct?

• criterion: does the instrument effectively predict outcomes of a construct?

  • concurrent validity: criterion measured at the same time

  • predictive validity: criterion measured in the future

face validity (social scientific warrants)

• does a measurement make sense “on its face”?

• face validity DOES NOT provide convincing evidence of appropriate measurement

construct validity (social scientific warrants)

• does the instrument actually measure the theoretical construct?

• highest level of validity

• demonstrated by content validity and criterion validity

reliability (social scientific warrants)

• instruments should perform the same way over time

• intercoder reliability: how similar coders are in coding data; percent agreement, statistical measures (e.g., Cohen’s kappa)

• alternate forms: use of two or more instruments to measure the same construct

• test-retest: use of same instrument over multiple points in time (stability)

• internal consistency: individual items on a measure receive generally consistent responses (e.g., Cronbach’s alpha)

  • social scientists often use multiple items (and take the average) rather than a single item to measure a construct; helps eliminate error

Cronbach’s alpha scale for internal consistency (social scientific warrants)

α ≥ .90 = excellent

α ≥ .80 = good

α ≥ .70 = acceptable

α ≥ .60 = questionable

α ≤ .59 = unacceptable

_ is necessary for _

• reliability; validity

• a tool can’t accurately measure what its intended to if it doesn’t get consistent results

worthy topic (interpretive warrants)

• topic is interesting, significant, timely, and/or relevant to the discipline or society

• does the study address an important topic? (to the field/discipline, society, or groups/individuals)

rigor (interpretive warrants)

• appropriate use of theory, data collection, and data analysis

• consider:

  • is the theory(ies) I am using a good fit for this context?

  • is the sample appropriate for the purpose/RQ?

  • Are mt data collection/analysis techniques appropriate?

sincerity (interpretive warrants)

• openness as a researcher in discussing limitations

• a typical research article has four main sections:

  • introduction

  • method

  • results

  • discussion— nearly always includes a limitations and future directions sub-section

credibility (interpretive warrants)

• researcher dependably conveys expressed realities

• thick description: in-depth explanation; detailed, rich descriptions of experiences (e.g., use of quotes)

• triangulation: use of multiple datasets, methods, theories, and/or researchers to explore the same phenomenontra

resonance (interpretive warrants)

• using impactful cases/quotes to show your arguments

• transferability: ability to transfer results from one context to another

• aesthetic merit: how good is the writing?

  • good interpretive research is artistically and imaginatively written (e.g., reads like a good short story)

significant contribution (interpretive warrants)

• types of contributions:

  • theoretical: helps understand/explain a phenomenon

  • heuristic: prompts additional research or lines of questions

  • methodological: improves/changes how research is conducted

  • practical: real-world use/value

ethical (interpretive warrants)

• study conforms to ethical standards

• ethical principles:

  • respect for individuals (autonomy)

  • beneficence

  • justice

coherence (interpretive warrants)

• internal coherence: study achieves stated purpose

• external coherence: study fits within/connects to broader literature

critical warrants

• follow the same warrants as interpretive researchers, with a few caveats

• Reflexivity: critical researchers refrain from claiming they have produced a final, definitive statement of “truth”; encourage criticism of their conclusions (turning criticism back on itself)

• greater focus on power/social inequality

independent and dependent variables

1) independent variable

• variable that you predict will affect the dependent variable

• it is “independent” because it “stands alone” (not affected by any other variables)

• sometimes called the predictor variable

2) dependent variable

• variable that is being studied

• it is “dependent” because it “depends” on the independent variable

• sometimes called the outcome variable

research questions

• term used in two different ways

1) research question: focus of your study

• what you are trying to answer when you research a topic

• your research question should be based off existing research

• your research question should “move” the existing literature forward

2) research question: question in a study about a specific relationship (or lack thereof) between two (or more) variables

hypothesis (hypotheses and research questions)

• testable statement about the relationship between two or more variables

• derived from theory and/or existing research

directional vs non-directional hypotheses

1) non-directional

• difference or relationship exists, but direction or magnitude is not stated

2) directional

• difference or relationship exists, and direction or magnitude is stated

null and alternative hypotheses

1) null hypothesis

• there is no relationship between the independent variable and the dependent variable

2) alternative hypothesis

• there is a relationship between the independent variable and the dependent variable

• technically, formal statistical tests don’t find support for the alternative hypothesis but rather find lack of support for the null hypothesis (falsification)

data

• information collected in a systematic manner

• quantitative = numeric

• qualitative = non-numeric

types of data

• texts

  • written, spoken, performed, or symbolic messages

• observations

  • watching behavior in action

• self-reports

  • ask individuals to report about their own behaviors

• other reports

  • ask individuals to report about behaviors of others

population and sample

1) population

• the group of individuals/cases from which you want to collect data’

2) sample

• sub-group of population

• random sample: goal is to obtain a sample that is representative of the population; used to infer claims about population (generalization)

• non-random sample: useful when generalization is not necessary; easier and more flexible method of data collection

random sampling

1) simple-random sampling

• every individual/case in a population has an equal chance of being included

2) systematic sampling

• randomly choose a starting point in your data and then include every nth data point

3) stratified sampling

• identify mutually exclusive groups and then randomly sample from those groups

non-random sampling

1) convenience sample

• data is easily accessible

2) snowball sampling

• sample builds on recommendations from participants

3) purposive sampling

• when the focus of the study is a specific group(s)

4) quota sampling

• pre-determine categories and how much data you want in each category

conceptual and operational definitions

1) conceptual definition

• “dictionary definition” of a concept

• based on previous research— agreed-upon definition for a concept used in a study

2) operational definition

• specific methods, procedures, variables, and/or instruments used to measure concepts

Levels of measurement (NOIR)

1) categorical

• nominal: mutually exclusive categories; no logical order; no measurable/consistent differences between points

• ordinal: mutually exclusive categories; data that can be put in a ranked, logical order

2) continuous

• interval: measurable difference between data points; Likert and semantic differential scales

• ratio: measurable difference between data points AND a meaningful zero point (e.g., hours of true crime media consumed each week)

statistics

• a way of organizing, describing, and making inferences from data

descriptive statistics

• allow us to summarize data either numerically or visually

representing data (descriptive statistics)

• pie chart

  • categorical data; typically nominal

• bar chart

  • categorical data

• histogram

  • continuous data

• line chart

  • continuous data with multiple measurements (e.g., over time)

measures of central tendency (descriptive stats)

1) mean (M)

• average of scores; sum of the scores divided by the number of cases

2) median

• midpoint of a distribution; 50% of the scores above and 50% below the midpoint

3) mode

• most frequently occurring score

• unimodal = one mode; bimodal = two modes

• median + mode most resistant to skewness

measures of variability (descriptive stats)

1) range

• subtract the lowest score from the highest score in a distribution

2) standard deviation (SD)

• average distance between a score and the mean

• calculated by taking the square root of the variance

distributions (descriptive stats)

• how data is “spread” across all possible values

1) symmetrical

• data to the left and right sides of the mean are identical to each other (or close to it)

2) asymmetrical

• two sides of the distribution are not identical to each other

• skewness

normal distribution: “bell curve”

• a symmetrical distribution in which there is a single peak at the mean, and symmetrical distribution on both sides

• 68-95-99 rule

Skewness (descriptive stats)

• majority of the scores are shifted either to the right or the left of a distribution’s center

• positive skew: few points above the mean; tail to the right

• negative skew: few points below the mean; tail to the left

Kurtosis (descriptive stats)

• how “peaked” a distribution is

• positive kurtosis/leptokurtosis: very peaked distribution, outliers

• negative kurtosis/platykurtosis: flat; each score has same frequency

inferential statistics

• allow us to make conclusions (inferences) from a sample to a population

central limit theorem

• distribution of sample means becomes more normal and narrow as sample size increases

• practical implications:

  • as sample size increases, a sample is more likely to be more representative of the population (assuming random sampling)

  • allow us to draw conclusions (inferences) from a sample to the population

example of central limit theorem

• a poll finds 40% of women and 50% of men approve of Donald Trump

• Do men and women differ in their approval for Trump?

• it depends on the sample size— test of statistical significance is needed (inferential statistics)

Statistical significance

• probability that observed finding occurred by chance, assuming null hypothesis is true

• alpha significance level

  • p < .05 = less than 5% chance that observed finding occurred by chance

  • p < .01 = less than 1% chance…

  • p < .001 = less than 0.1% chance…

• NOT a measure of confidence (95% confident a finding is true)

Tests of difference

1) t-test

• detect mean differences between two groups

2) ANOVA

• detect mean differences between three or more groups

3) chi-square

• detect differences in frequencies (counts) between two or more groups

t-tests

• use when…

  • independent variable = categorical

  • dependent variable = continuous

  • you want to compare means to test for differences between two groups

• null hypothesis = no mean difference between two groups

• alternative hypothesis = there is a mean difference between two groups

types of t-tests

1) independent samples t-test

• there is no relationship between participants in the two samples

2) dependent samples t-test

• participants between the two samples are matched in some way (e.g., individual scores on a pretest and posttest; the variable (test score) is the same but tested at two different times

ANOVA (analysis of variance)

• use when…

  • independent variable = categorical

  • dependent variable = continuous

  • you want to compare means to test for differences between three or more groups

• null hypothesis = no mean difference between ALL groups

• alternative hypothesis = there is a mean difference between two or more groups

Why not run multiple t-tests?

• probability of making a Type I error increases as the number of tests increases

• t-test just compares one group to another, ANOVA compares all groups to each other

• ANOVA vs t-test

  • 2 groups: ANOVA = 1 test; t-test = 1 test

  • 3 groups: ANOVA = 1 test; t-test = 3 tests

  • 4 groups: ANOVA = 1 test; t-test = 6 tests

Type I and Type II errors

1) Type I error

• conclude alternative hypothesis is true (reject null), when it really is false

• False positive

2) Type II error

• conclude alternative hypothesis is false (accept null), when it really is true

• false negative

chi-squares

• use when…

  • you want to test the hypothesis that the distribution of frequencies differs between groups

  • you have variables that are all categorical

• null hypothesis = frequencies are equal across groups

• alternative hypothesis = frequencies differ across groups

Tests of relationships and prediction

1) correlation aka r

• examine the association between two variables

2) regression

• use one or more independent variable(s) to predict a dependent variable

test of correlation

• extent to which two or more variables are related

  • both variables need to be continuous (you can correlate a continuous variable with a categorical variable that is dichotomous)

• r

  • most common statistic to describe correlation between variables

  • numerical index ranging from -1.0 to +1.0

  • 0 = no relationship

• correlation ≠ causation

regression

• use when…

  • you want to predict a continuous dependent variable

  • you have one or more continuous independent variables (you can include categorical independent variable(s) in a regression if they’re dichotomous)

• goal of regression = predict dependent variable using one or more independent variable(s)

• R² = indication of how well independent variable(s) predict dependent variable (larger is better); coefficient of determination