Knowledge Check #1
Major components of a research article
1) Abstract 2) Introduction 3) Lit review/rationale 4) Methods 5) Results 6) Discussion 7) Conclusion
How to address low reliability
-Item construction: rewording qs, fixing -Length: make scale longer -Administrative procedures: how people are filling out questionnaire
When do you use a survey?
-Personality traits/states -Beliefs and attitudes -Knowledge -Behavior
Difference between sample and population
-sample: people/units included in research study -population: entire set of objects, observations, scores that have something in common (ex: all people in one age group)
Chronbach's alpha
0.90+ Excellent • 0.80-0.90 Good • 0.70-0.80 Respectable • 0.65-0.70 Minimally acceptable • 0.60-0.65 Undesirable • 0.60↓ Unacceptable
Threats to validity
Inadequate preoperational explication of concepts (measuring different ideas) Mono-operation bias (measuring once) Treatment effects (participants guessing the goal of study) Restricted generalizability (whether the procedure will work in different context) Levels of constructs (a scale measuring one thing or many?) Social threats to validity (guessing, social desirability)
Difference between independent and dependent variables
Independent variables: variable that is manipulated Dependent variables: variable that responds to the change of IV [DR IM]
Likert scales
Scale in which participants are presented with a declarative statement and asked to respond with a range of choices: strongly disagree, disagree, neither disagree nor agree, agree, strongly agree
Introduction
Attention-getter, links to topic and has thesis
Types of validity
Face validity - qualitative assessment, subjective Criterion validity - correlation between different items and production of another set of things Predictive validity - score of new measure predicting other measures Concurrent validity - seeing if scores are similar at same time Retrospective validity - relating previously measured criterion and relating it to new measure construct validity - measure based on theoretical construct
Conceptualization
Developing concepts and germinal idea MEASUREMENT: operationalization must match conceptualization
Semantic differential scales
Type of scale that asks participants to rate their opinions on linear scale that exists between two endpoints that have opposite meanings: Good/Bad etc.
Entering data into SPSS
Variable and data windows
Questionnaire
actually document used in a survey, containing questions and mental measures in an attempt to gain statistical information about a group
Conclusion
Summarize findings and demonstrate that main arguments are clearly explained
Assessing reliability
Test-retest: give once, then give at later date Alternate forms: testing different ways Split half: Splitting qs into half and half and ordering differently Hoyt: coefficient to figure out whether split half is reliable Chronbach's alpha: tells us if items are being answered the same way
Reliability
the accuracy that a measure has in producing stable, consistent measurements -whether or not people will answer Qs relatively similarly
research
the application of the scientific method to discover and interpret
Validity
the degree to which a test measures what it is intended to measure
null hypothesis
the hypothesis that there is no significant difference/relationship between two variables
epistemology
the study of knowledge Asks: How do you know what we know?
Nominal variables
categorical, mutually exclusive, exhaustive ex: political affiliation, sex
Ordinal variables
categorical, mutually exclusive, exhaustive, numerically-oriented ex: education level, letter grade
Concrete vs. abstract variables
concrete: stays the same (ex: year born) abstract: changes (ex: age)
Difference between cross-sectional and longitudinal survey design
cross-sectional: info from group of people at given point in time longitudinal: examines how people change over time
Subscale
hypothetical subdivision of mental measure
Survey
method of data collection for quantifiable information by asking group of people about their attitudes/perceptions/beliefs/etc.
two-tailed hypothesis
predicts that there is a significant difference/relationship, but doesn't indicate the nature of that difference/relationship
One-tailed hypothesis
predicts the specific nature of the relationship or difference
Abstract
previews article, gives away results in ~150 words
Methods
-provides enough detail to allow others to replicate study -transition between hypothesis and test of hypothesis
APA In-text citation
([author name(s), [year], [page #])
Results
-Describe your data (Stats plus statistics)
Discussion
-Discuss results -Limitations -Future research
Things to consider when constructing questions/scales
-Every item should reflect the construct. - Keep it short and unambiguous - Avoid leading items - Avoid double questions -Avoid always/never -Consider recall issues
Gap-based brainstorming
-context = studying something on FB and something on Twitter -sample = looking at different group of people -modern replication = take old study and replicate now -new concept = articulate idea and research -theory testing = extending an established theory
Units studied in communication
-individuals (1 person) -dyads (relationships) -groups (group management) -organizations (workplace dynamics)
Brainstorming heuristics
-introspective self-analysis -retrospectie comparison -sustained, deliberate observation -Jolting one's conceptions out of its usual ruts -Deductive reasoning -Reinterpretation of past research -Integrating multiple past studies
unit vs. item-nonresponse
-item non-response: 1 question not answered -unit non-response: when researcher fails to obtain any survey measurement on a sample unit
Scientific method
1) Theories (predict, falsifiable) 2) Predictions (hypothesis, think about new ideas) 3) Observations (fathering info/data) 4) Empirical generalizations (discussion of findings) [goes back to theory]
Steps to create a survey
1. Pick questions 2. Create clear instructions 3. Study design 4. Data processing and analysis 5. Pilot testing
APA References
American Psychological Association (scientific/social science) alphabet by author's last name. Use hanging indents. -generally: author, title, journal title, volume, DOI
Directional research question
Asks if there is a specific difference between two variables in a certain direction
Non-directional research question
Asks if there's a difference between variables
Operatoinalization
Assigning units of analysis to variables in order to quantify MEASUREMENT: operationalization must match conceptualization
You CAN have: [reliability/validity levels]
High reliability, low validity
Interval
Logical order, equidistant apart, no true zero ex: IQ
Ratio
Logical order, equidistant, true zero ex: GPA, temperature
Identifying Chronbach's alpha
Look for threshold, see if questions bring value up
Lit review/rationale
Looks at previous research, brings study from general to specific (funnel) and makes argument to test RQ/hypothesis
Factor analysis
Looks for patterns among participants' responses to each scale question in an effort to determine if participants respond in any discernable patterns
You CAN'T have: [reliability/validity levels]
Low reliability, high validity
primary vs. secondary sources
Primary sources = materials on a topic upon which subsequent interpretations are based (diaries, interviews etc.) Secondary sources = restatements/analysis of primary research
postpositivism
The belief that there is an empirical reality but that our understanding of it is limited by its complexity and by the biases and other limitations of researchers
alternative hypothesis
The hypothesis that states there is a difference/relationship between two or more variables
assessing source credibility
author, institution, peer-reviewed status to name a few
Differences vs. relationships
differences: degree to which one person/group are dissimilar from another person/group relationships: correspondence or connection between two variables