Please enable JavaScript.
Coggle requires JavaScript to display documents.
Non-experimental Research (Descriptive (Things to note (Observer's…
Non-experimental Research
Aim
- to describe/explained/predicting current existing characteristics
Types
Descriptive
Relationship
Comparative
Correlational
Survey
Other
Content analysis
- analysis of written/spoken record; purely descriptive (themes, patterns & conclusions based on objective)
clearly defined response categories and blind observer (bias, only look at things that match your expectations)
Archival research
- x collect data, study existing records (census data, court records, published research articles)
Case history
- single/ few cases (developing of disease)
Meta-analysis
- define variables → review past researches (around 500) → meta-nalysis (comparing results)
qualioty of research/ studies that use dofferent methods
Ethnography
- go into a cultire and immerse into them then observed (live there)
Ps observation
- problem of reactivity; blind observer (single/double blind)
Descriptive
Quantifying behavior methiods
Frequency
- how many times
Duration
- how long
Interval
- e.g. in every 2 min
single event VS behavior sequence (e.g before having eye contact, what they do?)
OD of behavior: specific, unambiguous, simple, in line with research objectives
Sampling
Time
- in 5 min observed then go
Individual
- only focus on 1 person/ particular case
Event
- in the particular surrounding, focus on 1 region (observed oni 1 behavior)
describe behaviors (observational study)
Reliability
- consistent across researchers
Percent agreement
Cohen's kappa
Pearson's r
Things to note
Observer's bias
- mood, past experiences, values, approve of bhvs
Over-interpretation of behavior
- they have these bhvs bcuz
Conclusion about relationships should NOT be drawn
Observational techniques can be experimetnal or non-experimetnal
- through equipment/ Bolby's attachment
Natuaralistic onservation
- unobstrusive
use of special equipment- recording
habituationi Ps comfortable/get used to taht you at thr
other indirect measures- x directly in touch with Ps (x interfer/ contact; they x knw you are observing)
actual behavior; time consuming and expensive
Relationships
Comparative
examining differences on the DV between 2 groups of Ps
Correlational
Simple correlation
examine relationship btw 2 variables
no causation link
low reliability
- difficult to identify relationship (Cronbach's alpha)
Restriction of range
- difficult to identify relationship (due to convenience/ researcher's objective), e.g. age/ IQ range
large sample sixe/ many variables- inflate relationships
Prediction
simple VS multiple
Significance of correlation coefficient
statistical siginificance (r)- p value
Practical significance (r2)- how meaningful is ur sig results
Survey
cross-sectional VS longitudinal
- cross onli collect data at 1 point of time
Steps in designing survey
wrintg items
- open-ended, restricted (yes/no, MCQ), partial open-ended (other, list out), rating scale (smiley face), Likert
Assemble questionnaire
- Coherent and visually pleasing; demographic 1st/last (priming, answer based on stereotype, 4get), continuity issue (randomized/not- > drop out, switching attention → < consistency; good: check simply answer); first question should be engaging and interesting
Response rate
- low = limit usefulness
Administring questionnaire
Mail
- nonresponse bias (x response but have meaningful result/insight)
Internet
- similar background, confidentiality
Telephone
- quick, > willing to disclose
Group-administered
- efficient; peer pressure/ discussion
Ethical issues
- Anonymity (no way to knw their identity) & Confidentiality (able to trace but choose to keep)
Reliability
- test-retest (same group, different time), split-half (split the survey half; level difficulty different)
increased no. of items, standardized administration procedures, minimal scoring errors, appropriate and well-written items
e.g. for children - level of understanding
Validity
content
- represent all facets of the construct
construct
- measure what it measure
criterion-related
- related to an outcome
concurrent
- relate with current/lattest same scale
Predictive
Sample
representativenessi covering the construct/ scope you interested
Random VS Non-random