Analyzing Research Literature
QUANTITATIVE - starts with question (hypothesis). Look for statistics, random sampling, large samples, objective scoring.
QUALITATIVE - starts with general problem. Look for purposeful sample, generated questions, small samples, unstructured measures, intensive & in-depth, results in words not statistics.
TYPES
EXPERIMENTAL (RANDOMIZATION) - CAN DETERMINE CAUSE - ASSIGNS EXPOSURE)
CAUSAL COMPARATIVE (AKA QUASI-EXPERIMENTAL) (should note this as prone to causal link errors) - Associations between variables. The cause or consequence of differences that already exist. (2 or more groups and ONLY one variable). NON RANDOMIZED CONTROL TRIALS
Reliability (should be noted for quality)
test-retest reliability
internal consistency reliability
Cronbach's alpha (0.00-1.00) - above .75 are good
Validity
construct validity: extent it measure theoretical concepts you assume it measures.
content validity: what you measure is actually what supposed to be measuring. Is the content accurate?
criterion-related validity
correlation co-efficient (r= >.30 are good))
predictive criterion related validity
measure is related to outcome
Deductive
Inductive
epistemology: Multiple realities
purpose: reveal complexities, explore, understand, uncover meaning, generate theory, phenomena, unknown
natural context
format: interdependent, ongoing, dynamic data gathering & analysis
investigator: involved
epistemology: single objective reality
purpose: cause & effect, predict, explain, connect/correlate, describe, test theory
context: controlled
format: independent, gathering separate, analysis
investigator: not involved / neutral
Probability Sampling
Non-probability Sampling - used for either qualitative or quantitative
systematic: every "n"th person
cluster: subgroups as a whole are randomly selected
Simple random: drawing from a hat
stratified: random drawing of people from within subgroups
Snowball: the closest bring their friends
Purposive: intentional selection
Convenience: whoever is closest
Quota: need a certain number to work
true experiment - random assignment of participants (ENSURES CAUSE & EFFECT BY ISOLATING VARIABLES & REDUCING ERROR POTENTIAL)
Filtered
RCT
CASE CONTROL (2 groups with different outcomes - observational b/c look at the possible cause of difference between groups) (no exposure, observational) - Start with outcomes to try to find possible exposures.
COHORT (longitudinal study - risk factor studies, shared characteristic. (No exposure, observational) - start with exposures to see possible outcomes.
Article Synopsis
Evidence Synthesis
Systematic Reviews
Case Report / Case Series (Describes & explains individual unique case in a story)
SINGLE SUBJECT DESIGN / Single Case Design (subject serves as his/her own control). Behavior change in individual or small group. PROBLEM - can't generalize.
CORRELATIONAL (1 group multiple variables)
DESCRIPTIVE (statistical census - current status of variable)
Internal validity (cause & effect accuracy) vs. External validity (generalized to greater population)
Retrospective OR Prospective
survey research
archival
natural observation
CROSS-SECTIONAL (observational / no exposure)- group gets exposure & outcome at same time.. Snapshot in time.
Between Group (pre-test / post test)
Within Group (cross-over, repeat test)
Potential Errors
practice effect (repeated doing causes change)
carryover effect (earlier treatment affects later)
TYPES
LIFE HISTORY: chronologic, possibly retrospective reconstruction of a life by the individual or other significant relation. Meaning of life in context of life experiences.
NARRATIVE: derive meaning of person's experience by analyzing their stories. Analysis through questioning. Not linear.
ETHNOGRAPHY: extended study of organization or culture. Multi-data source, collected in the field. Analysis of meanings and application to broader scale.
PHENEMOLOGY: meaning of lived experience through narrative. No analysis. Just a recording or capture of experience at the individual level.
GROUNDED THEORY: development of theory from data. Theoretical sampling, abductive thinking, Data collection & analysis occur simultaneously.
Participatory Action Research: research in community focusing on participation and action.Collective inquiry. Occupational justice. People who form the design are they ones who experienced it. Participatory action wants to change the community.
MIXED METHOD occurs when types are combined