Final Flashcards
What is research?
• A systematic process that answers a question and is free from bias
Explain the research process
- Identify research problem
- Specify research process
- Collect data, analyze & interpret data
What is a theory?
•Unified set of ideas that might explain a question(developed through research)
Inductive vs. Deductive Theory
- Inductive- starts with individual cases, the purpose is to develop overarching theories,(take individual cases build up to theory)
- Deductive- starts with overarching theories, purpose is to test overarching theories, (start with theory then test it)
What are variables?
logical set of attributes
Independent vs. Dependent Variables
- Independent= cause (what you change)
* Dependent= effect (depends on what you change)
What is a Paradigm and why is it important in research?
•Framework for how you see the world- it influences what research questions are asked, what results you expect to find, how you collect data
Positivist Paradigm
•You believe one truth exists and it can be found using scientific methods (also known as realist)
Post-Positivist Paradigm
•There is one truth and it can only be known imperfectly (also realist) *most scientists are this
Constructivist Paradigm
•There are multiple truths and each person has own version of reality (therefore own version of truth)
Constructionist Paradigm
•Multiple truths w/ own version of reality and everyone’s truth is shaped by social structure (feminist paradigms, race theory) (these ppl. Are critical theorists- knowledge is not value free, critical of truth)
3 purposes of reseach
- Explanation- why?
- Description- What, where, when, how (usually qualitative)
- Exploration- learn more about subject, to test feasibility of more extensive study, to develop new methods of collecting and analyzing
Nomothetic
•An attempt to identify casual factors of universal “laws”
Correlation vs. Causation
- Correlation- empirical relationship between two variables
* Causation- making/causing something to happen
3 parameters for causation
correlation, time order, non-spurious (not coincidental)
Unit of analysis
what or whom is being studied
sampling frame
•Representation of the pop. you are studying (requires you to consider criteria for who will be included in your study)
Quantitative Research
•The numerical representation and manipulation of observed and recorded data describing and explaining phenomena that these observations/data reflect
Pros and Cons of quant. research
pros: easy to analyze/summarize
con: can only tell us what happened through numbers
Quant. Methodologies
- Mode- most frequently occurring data point in set
- Mean- statistical average (add all data points together and divide by number of data points in each set)
- Medium- middle data point in rank order
Qualitative Research
• Collection of empirical materials that describe routine and problematic moments and meanings in individual lives
Pros and Cons of qual. research
pros: provides meaning and detail, offers place to start when researchers don’t know what to ask, flexible, cheap
cons: can be interpreted diff. ways, might be superficial (people talking/interacting), data can’t be applied wider
Qual. Methodologies
• Mixed Methods- using more than one method to answer question
mixed methods pros and cons
pros: address questions at different levels, develop new theories or data collection instrument, overcome weakness of single methodology,
con: takes time, discrepancies in data = challenges
Sequential vs. Concurrent
i. Sequential- quant then qual (qual results assist in explaining the findings of a quant study)
ii. Concurrent- two or more methods used to confirm and cross-validate findings (collected at the same time, to strengthen study)
Likert Scale Survey Questions
5-7 point scale used to allow individual to rate their attitude about a particular statement (neutral option, equality of scale)
Double Barreled Survey Question
• Asking for a single response to a question that has multiple parts
Contingency Questions
• Question intended for only some respondents (determined by response to other questions)(relevant, shorter=better, avoid negative items, avoid bias terms)
Survey question order
- Can influence how respondents answer questions
- Random order = chaotic
- Most interesting question first- initial question should be non-threatening
- Dull questions at end
Benefits of Pretesting your Survey
• To get out kinks of spelling, wording, question order, clarity of directions
Best practices when conducting surveys
• Be familiar with questionnaire, should be similar to pop being studied, follow question wording exactly, record responses exactly *ask probing questions
3 Components of ‘No harm to Participants’
- Respect for person- duty to protect sensitive/vulnerable pops. (Children, prisoners, intoxicated ppl.)
- Beneficence- research should benefit pop. being studied, benefit should be equal to suffering
- Justice- burdens & benefits of research should be shared fairly w/in society
Informed Concent
• Norm in which subjects base their voluntary participation in research projects based on a full understanding of the possible risks involved
Anonymity
• Guaranteed in a research project when participants can’t be identified by either the researchers or anyone else analyzing the data collected
Confidentiality
• Researchers identify a given person’s responses but promises not to do so publicly
Deception
• The act of making someone believe something that is not true
Debriefing
• Way of counter deception- interviewing subjects to learn about their experience of participation in the project & to inform them of any unrevealed
What are the obligations of Analysis & Reporting?
• Obligation to: Research subjects & scientific community
What is IRB? What is its role?
• Institutional Review Board- enforce ethical practices and procedures in research especially with human subjects
Population vs. Study Population
- Pop: the group we are interested in generalizing about
* Study Pop: the pop from which the sample is actually selected
Representativeness
- The quality of the sample in representing the pop you want o generalize about
- The data derived from the sample should be assumed to represent the pop
- Enhanced by probability sampling
Probability Sampling
• The general term for samples selected in accordance with probability theory (random sampling)
Random Sampling
each element has equal chance of selection
Simple Random Sample
units composing a pop are assigned numbers, set of random numbers are generated and units have those numbers are included in sample
Systematic Sampling
every #th unit is selected
Cluster Sampling
natural clusters are sampled initially, with members of each selected group being sub-sampled afterward
Stratification
• The grouping of units composing a pop into groups before sampling (can be used in conjunction with other sampling methods to improve representativeness)
Non-probability Sampling
• Any techniques in which samples are selected in some way not suggested by probability theory
4 types of non-probability sampling
o Purposeful- units included are selected on bases of researcher’s judgement about which ones will be most useful
o Snowball- each person interviewed may be asked to suggest additional ppl. for interviewing (field research)
o Stratification
o Quota- units selected into sample on pre-specified characteristics so total sample will have same distribution of characteristics assumed to exist in pop.
sampling bias
• Sampling approach in which some members of the study pop are less likely to be studied than others
Validity
how accurately the study/data reflects the concept it is intended to measure
4 types of validity
a. Face Validity- Does the study seem like a reasonable measurement of the variable? (someone unfamiliar with research so it doesn’t make sense to them)
b. Criterion Validity- When conclusion drawn from data set can be compared to criterion data collected later
c. Content Validity-The degree to which a measure covers the range of meanings included within a concept
d. Construct Validity-Degree to which a measure relates to other variables as expected within a system of theoretical relationships
Reliability
• Quality of measurement methods that suggest the same data would have been collected each time in repeated observations of the same phenomenon
Ways to improve Validity
- Triangulation
- Member checking
- Inter-coder agreement
- Memoing (notes to self throughout study)
- Time sampling
Ways to improve Reliability
- Memoing
- Triangulation
- Inter-coder reliability
Coding
identifying important info
Qualitative Data Analysis Steps (3 steps)
- Identify important info (coding)
- Categorize info (developing themes)
- Recognize Themes (conceptualizing: thoughts become more abstract/big picture, thinking about data as a whole)
Direction of Qualitative Data Analysis based on Inductive vs. Deductive Research
- Inductive: individual cases to larger theories
* Deductive: testing theories by looking at individual cases
Descriptive coding
• Little interpretation to store things known about the data item
Analytic coding
combining concepts to develop themes
Axial Coding
• developing themes, relating codes to each other
Increasing Validity & Reliability in Qualitative Data Analysis
replication of study
Univariate Analysis
• Analysis of a single variable for descriptive
Standard Deviation
• the amount of variability in a data set
Discrete Variable
• Variables whose attributes are separate from one another (gender, race, religion)
Continuous Variable
• Variable whose attributes form a steady progression (age, income)
Bivariate Analysis
• Analysis of two variables at once
Multi-variate analysis
• Analysis of multiple variables at one time
Standard Error
• Measure of the statistical accuracy of an estimate, the measure of uncertainty (large sample = small error, smaller sample = larger error)
Sampling Error
• An assumption that the data collected in perfect (error caused by observing a sample and not the entire pop. being studied)
Null Hypothesis
• Prediction that the two groups we are comparing are not significantly different from one another
Alternative Hypothesis
• Proposes there is a diff. between two variable being tested
Type I Error
• When we reject the null hypothesis even though it is true
Type II Error
• We do not reject null hypothesis, even though it is false
T-Test
• Analyze two means, comparison of two samples drawn independent of one another
P-Values & meaning of being less than .05 or greater than .05
- An obtained significance level
- Less: null hypothesis is rejected (there is real diff. in variables)
- Greater: the null hypothesis is retained (no diff. in variables)
