1/43
Looks like no tags are added yet.
Experimental design
Gold standard of research that is a way of observing cause and effect
Classical experimental design
Independent and dependent variables (before for baselin), experimental and control group, random assignment, and pre/post testing
Double-blind
Researchers and participants don’t know
Threats to internal validity
History (might amplify or suppress), maturation (Growing/aging), testing (pretest might affect), instrumentation (need to measure the same), statistical regression (extremes will normalize), selection bias (might be different)
Rule out threats
Do our best, random assignment, shorten time frame
Construct validity
Actually representative of the real world (low internal validity if controlled)
External validity
Would it be observed the same in the real world
Types of classical experiment
Post-test only- do if you don’t want to tip off
Factorial Design- more than 1 experimental group
Quazi-experimental non-equivalent group design (match people according to factors chosen)
Survey
Give good information (limited answer options)
Qualitative method
Interviews (conversation with predetermined question
Types of interviews
Structured- predetermined questions
Simi-structured- standard questions, but can ask a few additional
Unstructured- unprompted conversation
Substantive frame
Topics covered in an interview (questions after)
Challanges
Getting access to participants, time, encouraging participation
Conducting interviews
Face-to-face is best, but do phone or online (rapport is key)
Rapport techniques
Conversation, actively listen, short/clear questions, respect, and positive end
Focus groups
Interview in a group setting (group think, dominent members, and difficult to discuss sensitive matters)
Recording data
Notes as the interview occurs (usually record)
Analyzing
Look for patters (key words/ themes)
Field research
Observe phenomenon in the real world (understand what we can’t ask)
Options
Full participant- not revealing you are a researcher
Participant as observer- make it known you are researcher
Complete observer- when you are not apart of the situation
Recording
Notes, photos, audio (establish criteria you are looking for
Strengths
Flexible, detailed info, natural setting (more external validity)
Weakness
Small sample and not always representative (issues with reliability)
Secondary data
Data collected by others: agency and content analysis (doc),
Agency records
Can be used to explain as well as for applied studies (policy effects)
Content analysis
Any type of study (how messages are being portrayed)
Agency record types
Published- public
Nonpublic- need request
New data- they collect for you (can reduce validity)
Nonpublic
Many agencies have entire data sets not made public, but there are issues with underrepresentation
New data
is cheaper, but it is difficult to get staff to cooperate and may no be reflective or designed for research
Secondary data issues
May not get all of the right data or variables
Content analusis
Systamatic study of messages in communication (look for patterns)
Inter-rater relyability
All researchers agree on what they see
Applied research
When we look at policies and evaluate them (process or outocme)
Fidelity
Is it operating the way it should be
Impact acessment
Is the outcome good or negative
Steps for applied research
1.) Determine if you can do an assessment
2.) Identify goal of program
3.) How will we measure (consider issues in context)
4.) Specify variables
Designs
random assignment (if eval before) or non-equivalent group design
Political context
Need to stay objective, identify concerns, and educate on research
Ethics
Research that conforms to the standards of a given group
Ethical standards
1.) No harm can be done (benifits must outweight the cost)
2.) Voluntary participation (need informed consent)
3.) Anonymity (double blind) and confidentiality- at least confidential
4.) Deceiving subjects (can’t lie, but don’t reveal too much)
5.) Analysis/reporting- ethical obligation to science and society
Promoting compliance
Internal checks, Belmont report, code of professional ethics, and IRB
Institutional Review Board
All agencies that conduct research have one (have to submit application); is more complicated if you collect own data
What do they ask about
Harm. data collection, questions, sharing data, benifits (might have to go through 2 IRBS)
Guidlines
Informed consent that informs about the study (kids assent), inform not harm/benifits (must have prisoner rep)