Com 300 Exam 2 Study Guide
Com 300 Exam 2 Study Guide COM 300
Popular in Methods of Communication Inquiry
Popular in Communication Studies
verified elite notetaker
This 19 page Study Guide was uploaded by Nick Bahoric on Wednesday March 30, 2016. The Study Guide belongs to COM 300 at Michigan State University taught by Ronald Tamborini in Spring 2016. Since its upload, it has received 12 views. For similar materials see Methods of Communication Inquiry in Communication Studies at Michigan State University.
Reviews for Com 300 Exam 2 Study Guide
Report this Material
What is Karma?
Karma is the currency of StudySoup.
You can buy or earn more Karma at anytime and redeem it for class notes, study guides, flashcards, and more!
Date Created: 03/30/16
Com 300 Exam 2 Study Guide FULL I=Identification U=Understanding A=Application B=Identification+Application ?= Not Specified LECTURE 10 Measurement Quality Quality of Observations ? Depends on quality of measures Two Methods for Estimating Quality B Reliability Consistency of measure Validity Accuracy of Measure I. Reliability Consistency of measure A. Dimensions of Consistency for Reliability (2) A Stability Consistency across time Challenging if variable changes over time (ex. attitudes, weight) If the thing that you are measuring does not change across time, every time I use my measure to make the observation I should see the same thing (ex. every time you step on a scale) Equivalence Consistency across measures When all of the measures are measuring the same thing When they are not measuring the same thing then the measuring technique does not have consistency. If it’s not consistent across measures then it’s not reliable. B. Assessment of Reliability (3 types) U TestRetest Designed specifically to measure consistency across time (expect to see same result each time) Alternate Forms two “parallel” tests measuring same concept given at two different times Attempts to assess both consistency across time (stability) and consistency across measure (equivalence) Internal Consistency best measure of equivalence (consistency across measures) Alpha statistical measure (1 means perfect reliability, 0 means no reliability) Alpha is a function of Interitem correlation The length of the test Number of items II. Validity Accuracy of measure A. Types of Validity (3) U Content Validity Measuring all concept dimensions Criterion Validity Predicting behavioral outcomes Construct Validity Accurate and complete measurement B. Assessment of Validity (4) ? Content Procedure Specify construct’s content dimensions Design items representing content Experts judge content validity Check intersubjective agreement among judges Predictive Approach (pragmatic validity) Concurrent prediction If I give you a breathalyzer it’ll tell me you’re drunk right now but if not you’ll be drunk tomorrow Future prediction SAT/ACT tests Known Group Method Comparing those known high and low on construct Discriminates based on how we would expect people to score Convergence Techniques Similar results using multiple measurement methods LECTURE 11 Designing Measurement Instruments I. Appropriateness of questionnaires A. Problems with use of questionnaires ? Person does not know answer Person is unwilling to tell answer Person is unable to give answer B. Reasons for use of questionnaires ? Best or only way to get measure II. Administering the questionnaire I A. Modes of Data Collection Oral Written B. Response Rate Problems Total nonresponse Absence of any information about a given element or elements, and item The researcher is unable to locate an element for questioning Located elements are unwilling to respond to questions Certain elements are incapable of responding due to factors such as illness and communication deficits (ex. deafness or inability to speak English) The researcher misplaces or otherwise loses one or more pieces of collected information Item nonresponse Researcher got some but not all of the information required from certain elements A respondent’s unwillingness to answer personally sensitive questions (ex. requests for information about personal income or sexual preference) A respondent’s inability to answer questions because they are poorly constructed, being obscure or uninterpretable III. Designing effective surveys I C. Types of Survey Questions (2) Closedended Rating, ranking, and inventory questions Forced choice, information seeking device asking respondents to choose their responses from a list supplied by the researcher Openended Highly structured or less structured (e.g., interview, focus group) Freeresponse scaling technique Invites respondents to supply unstructured answers to questions or to discuss the subject(s) named Contingency and Filter Items Item that asks you to answer subsequent items differently depending on what your answer might be (ex. if you answered yes to 3, skip to 7) D. Question construction (what to do and what to avoid) Should be clear and unambiguous All words and phrases should have precise and easily understood meanings Should be simple and brief Words should be chosen for simplicity and ease of understanding Should not be leading Leading questions contain emotionally loaded language that encourages biased responses (ex. do you agree with the generally accepted view that television programs contain too much explicit violence?) Should not be doublebarreled A doublebarreled questions asks for a single response to several questions Should be nonthreatening Questions that ask respondents about personally or socially sensitive issues (ex. sexual behavior, drug habits) are potentially threatening E. Question Sequence General ordering pattern to make use easy Funnel question pattern begins with broad questions followed by progressively narrower or more specific ones Inverted funnel pattern narrowly focused questions are followed by more general ones Group topically related questions Questions that follow a logical sequence are always easier to answer than questions having no coherent order Easy questions first Easy questions motivate respondents to continue a questionnaire, whereas difficult or timeconsuming ones may be discouraging or frustrating Avoid creating response bias Response bias tendency of some respondents to react to all closed questions the same way regardless of content This may be countered by mixing questions of different lengths and types Techniques to avoid response bias Projective Techniques Something ambiguous (ex. ink blot tests) Structured Indirect Methods Answers given about other people reveal things about themselves (ex. questions about drugs) Randomized Response Technique Person answering the question is led to believe that the person asking doesn't know what their response means LECTURE 12 The Nature of Research Designs I. Types of Research Designs A A. Quantitative or Qualitative research Quantitative Observe things and represent observation in numbers Collecting observations Translating into numerical form Statistical analyses Inferences based on statistics Qualitative Represent specific qualities Critical inspection to synthesize observations Verbal conclusion about behavior B. Interpretive or Functional research Interpretive Attempts to discover what meanings exist and how they are reached The way individuals make sense of their world through communicative behaviors Studies of people's ordinary conversations including the meanings and actions associated with talk Functional Attempts to determine cause and effect relationships Typically regard meanings as the starting point for drawing conclusions about the antecedents and consequences of human communication Methodologies that assess the behavioral and attitudinal effects of various communication forms (ex. media programs and public discourse) C. Experimental or Naturalistic research Experimental Manipulation of variables Involves the prestructuring or manipulation of the research environment and the observation of people's reactions Naturalistic No manipulation of variables Involves observing and recording ongoing communication behaviors during the course of "normal life activity" D. Laboratory or Field research Laboratory Researchers bring communicators into a controlled environment to observe their verbal and nonverbal behaviors Field Takes place in the communicator's natural environment E. Participant or Nonparticipant research Participant The investigator contributes actively to the communication process being observed Nonparticipant Outside spectator who does not enter into the target communicative interactions in any way F. Overt or Unobtrusive Research Overt Subjects are aware of the researcher’s presence Unobtrusive Subjects are not aware that they are being tested G. CrossSectional or Longitudinal research CrossSectional Observing at one single point in time Longitudinal Observing repeatedly at different points in time H. Basic or Applied research Basic Research that's designed to test basic theory (basic theory not intended to address a specific problem) Applied Research that's conducted to test the logic of a theory (the best way for me to address a problem) II. The Validity of Research Designs B A. Research Progression Effects (4) As our study progresses, these are things that can threaten the validity of our observations History Incidental environmental events occurring during data collection that alter the beliefs and behaviors a research participant ordinarily would exhibit Maturation Changing physiological and psychological processes that affect the beliefs and behaviors of research participants during the course of a study Mortality The loss of research participants as a study moves toward completion Statistical Regression Least likely of the four to occur Suggests that sometimes people randomly score differently at time one than they would have normally Extremely high and extremely low scores tend to regress toward the mean score of all research participants B. Reactivity Effects (4) Demand Characteristics Refer to all the situational cues that implicitly convey the true nature of the researchers hypothesis to the participants These situational cues include rumors, past experience as a research participant, the research setting itself, the behavior of the researcher, and research procedures Evaluation Apprehension The participants desire to present a positive selfimage to the researcher, or at least to provide no grounds for a negative one Participant wishes to give socially acceptable responses when participating in research Researcher Effects • Characteristics of the researcher affect outcomes of study Biosocial: gender, race Psychosocial: perceived level of anxiety, warmth, need for approval, etc. Test Sensitization Heightened intrapersonal awareness that prompts respondents to deliberate more carefully about the problems at issue than they normally would If we know we're being tested or if we've already been tested once C. Sampling Deficiency Effects (2) ? Sample selection problems Applies mostly to survey research Deal simply with problems associated with the people that you choose to be in your study Sample assignment problems Apply only to experimental research Begin after you have already chosen the people to be in your study Only relevant to experiments III. Approaches to Conducting Research in Communication ? LECTURE 13 Experimental Research I. Characteristics of Experimental Research U A. Allows study of an independent variable’s impact on a dependent variable Focus on independent variable that can be manipulated Creates different levels of conditions of the independent variable B. Control of extraneous variable Allows view of manipulated variable’s effect with no extraneous variable Control increases internal validity This means that the effect observed is due to the independent variable and no other extraneous cause Two types of experimental control Hold extraneous variable constant Random assignment (not random sampling) C. Two defining characteristics of a true experiment Random assignment (randomization) To create equivalent groups Manipulation of the independent variable II. Threats to Internal Validity (5) ? A. Maturation Natural change during procedure B. History Unique events during procedure C. Instrumentation Effect due to change in measuring procedure D. Mortality Effects due to subject dropout E. Selection Nonequivalent groups; the effect due to subject attribute differences III. Dealing With Threats U A. Selection is the most serious threat Randomization helps control this threat B. Other threats harder to eliminate Need careful attention to design and procedure LECTURE 14 Experimental Research (continued) IV. Operationally defining independent variables ? A. Continuous or Discrete variables Continuous Variable on a continuum Goes from low to high (age, how religious you are) Discrete Categorical differences Can only be one type or another (religion, gender, movie genres) B. Stimulus or Organismic variables Stimulus Can be and is manipulated (violence someone is exposed to Organismic Can’t be manipulated (age) C. Single Factor or Factorial design Single Factor One independent variable Factorial Two or more independent variables D. AllStimulus or AllOrganismic or MixedModel design AllStimulus Uses random assignment to all cells AllOrganismic Uses purposive assignment to each cell MixedModel Uses random assignment for stimulus cells and purposive assignment for organismic cells V. Assessing Experimental Effects U A. Main Effects The effect that one independent variable has when averaged across levels or conditions of all other independent variables Refers to the unique impact of each independent variable on a dependent measure B. Interaction Effects Changes in the effect of one independent variable when looking at different levels or conditions of another independent variable Reflects the joint impact of two or more independent variables operating in tandem 4 Types Ordinal The first independent variable produces significant differences at only one level of the second independent variable Not parallel so there’s an interaction Come together at one point (ordinal) One line is parallel to axis (ordinal) Disordinal The first independent variable produces significant differences at all levels of the second independent variable Not parallel so there’s an interaction Neither line is parallel (disordinal) Lines do not come together at one point (disordinal) Lines can cross and still be disordinal Symmetrical The first IV produces opposite effects at different levels of the second IV (note: the lines cross when graphed) Nonsymmetrical The first IV is not reversed at different levels of the second IV (note: the lines do not cross when graphed) C. Laboratory and Field Experiments/ Validity Tradeoff in control and artifact Experiments are higher in internal validity and lower in external validity Internal validity Asks whether sample main and interaction effects are attributable solely to the independent variables selected External validity Addresses the question of whether sample experimental effects can be generalized to parent populations and to everyday social settings Laboratory experiments generally achieve a higher degree of internal validity than do field studies because intervening variables can be controlled better in the laboratory than in the field Field experiments often have more external validity because field researchers operate in natural social settings Uses of the Experiment When goal is causality, control, and internal validity VI. Simple Experimental Designs (3) U A. PretestPosttest Have two random groups. Give a pretest to both. Manipulate one group and the other group serves as a control group with no manipulation. Give post test to both groups R O1 X1 O2 R O3 O4 B. Posttest Only Start with two random groups. Manipulate only one group and the other serves as a control group. Give a posttest to both groups. R X1 O1 R O2 C. Solomon FourGroup Design There are four randomized groups. One group has a pretest, manipulation, and posttest. One group has a pretest, no manipulation, and a posttest. One group has no pretest, manipulation, and posttest. One group has no pretest, no manipulation, and a posttest. R O1 X1 O2 R O3 O4 R X1 O5 R O6 VII. Complex Experimental Designs (3) ? A. Independent Groups Design Characterized by the random assignment of different research subjects to each separate cell of an experiment B. Repeated Measures Design The same group of randomly selected experimental subjects participates in every cell of the experiment C. Mixed Design Different subjects are randomly assigned to the treatment levels associated with at least one independent variable, whereas the same subjects participate in the treatments associated with a second independent variable VIII. Procedures for Conducting an Experiment (5) ? A. Selecting an experimental setting B. Operationally defining and validating variables C. Selecting subjects and assigning subjects to experimental cells D. Administering the experiment E. Analyzing experimental results IX. The Experiment and Contemporary Research U A. Criticisms of the Experiments Artifact and limited scope We create situations that are not representative of real life B. Three Uses of the Experiment Examining Causality Control Internal Validity LECTURE 15 Survey Research (Nonexperimental) I. Characteristics of Surveys (4) I A. Relies entirely on organismic (attribute) variables B. Collects data in natural setting C. Uses structured set of questions D. Conducted for three purposes (exploratory, descriptive, and explanatory research) II. Assessing Causal Relations A. Experiments are often better than surveys ? Surveys often have inadequate internal validity B. Criteria for inferring causality ? Covariation Both variables change together Time order Cause precedes effect in time Nonspuriousness Outcome cannot be explained by other causes C. Criteria of Surveys (3) I Good at establishing covariation Sometimes establishes time order (often does not) Inadequate for establishing nonspuriousness D. Surveys and Nonspuriousness U You never know all other possible causes (spurious variables) Those known can be held constant or measured and controlled statistically Controlling is called elaboration Elaboration increases confidence III. Strengths of Survey Research I A. Feasibility B. External Validity C. Indication of Causal Relations IV. Types of Survey Research Design U A. CrossSectional A large representative sample All gender, race, and income all different parts of our population, measured at one point in time, can estimate relation of two variables within a population Can estimate incidence and distribution of variables, or relation of two variables within a population X O1 B. MultipleSample (or staticgroup) Compares people from different population, or differing on a static feature (e.g., gender) Observes the strength of static feature’s relation to attribute variables of interest X O1 O2 C. Longitudinal (panel and trend studies) Data from subjects are collected at two or more points in time Observing change in two variables at different time points helps establish time order Panel and Trend studies Panel studies Use the same subjects at each point in time Trend studies Use a different sample from the same population at each point in time O1 X O2 LECTURE 16 QuasiExperimental Research I. Characteristics of Quasiexperimental designs ? A. Cannot assume equivalent comparison groups (random assignment is not possible) B. Steps taken to deals with resulting problems II. General Approaches to Deal with QuasiExperimental Problems (2) U A. Multiple measures comparing intact groups Nonequivalent pretestposttest control group design O1 X1 O2 O3 O4 Threats to validity Selection, maturation, mortality Nonequivalent pretestposttest control group design with multiple pretests O1 O2 X1 O3 O4 O5 O6 Threats to validity potentially identified by comparing O1 to O2 with O4 to O5 Selection, maturation, mortality B. Measures at different points in time within the same group Interrupted time series designs Change in level or intercept O1 O2 O3 O4 O5 X1 O6 O7 O8 O9 O10 4 4 4 4 4 6 6 6 6 6 **Productivity (making cars) before and after new policy** Change in slope or drift O1 O2 O3 O4 O5 X1 O6 O7 O8 O9 O10 4 5 6 7 8 12 16 20 24 28 Problems with time series Maturation, history, instrumentation, mortality Seasonal trend Knowing time point for measuring change Lengthy data collection (50 data points) III. Methods of Collecting Data ? LECTURE 17 – Sampling I. Basic Sampling Concepts U A. Element The basic unit sought (e.g., person, message) B. Sampling unit Elements considered for sampling (e.g., MSU students) C. Observation unit The person providing information on the element D. Population The group of elements you wish to estimate (e.g., all college students) E. Sampling frame The list of sample units from which you select (e.g., MSU student directory) F. Sample A subset of elements selected from the population G. Statistics and parameters Numerical characteristics of the sample and the population H. Sampling error The extent to which the sample statistic deviates from the population parameter II. Sampling and the Probability Theory U A. Random samples from heterogeneous populations allow estimates of the true population characteristics The degree of accuracy is a function of the sample size and the homogeneity of the population III. Uses of Sampling (3) I A. Increases external validity B. Helps assure that research findings (on average) will not differ from true population by more than a certain degree Called the “margin of error” C. Helps guarantee all types of elements (e.g. people) are included in desired proportion IV. Types of Sampling (2) A. Nonprobability Sampling ? Cannot specify each element’s chance of being in sample More convenient and economical Representative when population is homogeneous Okay when goal is not to estimate population values but to examine possibility of causal relations B. Nonprobability Sampling Types (3) B Convenience Each available element is included until desired size Quota Convenience sampling used to select from specified categories Purposive Specific strategies used to pick each sample element (e.g., I'll pick this one because X) C. Probability Sampling ? Each element’s chance of being in the sample is specified Thus, you can specify chance that sample does not differ from population by more than a degree D. Probability Sampling Types (3) B Simple Random Each element has an equal chance of inclusion Stratified Random Random sample is selected from different strata (can control % of each stratum) Cluster of Multistage A sample is drawn from a large group (e.g., for convenience), and then a simple random or stratified random sample is drawn from the first sample V. Selecting a Representative Survey Sample (5 steps) ? A. Defining the Population Clarify in whom are you interested B. Specifying Sampling Elements Clarify the "things" you want to study C. Securing an Adequate Sampling Frame Make sure the frame represents the population Threats (3) Missing elements Elements that are missing (ex. Women not in phone book) Foreign elements People that aren’t really part of the population Duplicate elements People listed more than once D. Selecting a Sampling Method E. Determining Required Sample Size Calculate the minimum sample size based sample size and population homogeneity First determine sample size (n) Then calculate final sample size (n’) LECTURE 18 – Using computers in research (on exam?) A. Data entry Creating data files B. Data definition Creating control files C. System files D. Procedure command structures Creating procedure files E. Computer analysis training For use in individual research exercise #3 Good to know: A. Criteria for Causality Covariation Both variables change together Time Order Cause precedes effect in time Nonspuriousness Outcome cannot be explained by other causes B. Criteria for Formulating a Conceptual Definition Denote essential qualities The definition should tell us all the parts of it Noncircular Should avoid tautology (thought of as using the word to define itself Clear and Precise Clear Primitive terms Precise Boundary conditions C. Criteria for Evaluation of an Operational Definition Taps the richness of the concept (measures complexity) Allow for standardization through concreteness Allow for replication Match the concept with a good numerical scale D. Intervening Variable E. Extraneous Variable Notes: denominator inc, answer dec participation in decision making > sense of personal achievement> vocational commitment income>
Are you sure you want to buy this material for
You're already Subscribed!
Looks like you've already subscribed to StudySoup, you won't need to purchase another subscription to get this material. To access this material simply click 'View Full Document'