Notes

P
access to finance and funding
time
access/opportunity
personal qualities and skills

E
informed consent
deception
confidentiality and anonymity
right to withdraw
avoidance of harm.. emotional/physical
vulnerable groups

T
reliabiliy
representativeness
valid



QUESTIONNAIRES

practical: quick and not time consuming, cheap,
geographicallly distributed large sample size so more representative
reliable: can be easily replicated/ connor 4000 across UK on w/c decisions on uni
ethical: can give informed consent, can withdraw when they please, can ensure anonymity (e.g. self-report surveys in crime)
can quantify this data and generalise: e.g. Rutter ( questionnaires enabled him 2 establish correlation between class size/numbers of staff to achievement)

low response rate + specific type of ppl who will sit down and actually fit it in.
less valid: detachment doesnt allow for rapport. (think oakley, dobash etc and feminist criticisms of "smash-and-grab" raids). pre-set questions are IMPOSITIONIST and reflect assumptions of researcher - "value-laden"
Lack of flexibility: stuck w those questions, cant explore a new area that emerges.
-lying, right answerism, etc highly likely in surveys


Interviews: unstructured?
time-consuming, plus, therefore less representative
interviewer effect ARCHER STUDY ON MUSLIM BOYS found the boys opened up more about racism when researcher was of asian descent than caucasian. (characteristics) AND hawthorne effectl, diminishing validity


rapport like dobash
more flexible
ethics - sensitive questions, vulnerable groups, but deffo give their consent
structured

more quick, practical, reach a lot of ppl. e.g. Yandwilmott sutdy of extended families in london so more representative
more reliable: can uses same pre-coded questions

OBSERVATION

overt
covert

valid - natural setting (ecological validity)
less reliable
less representative

going native
danger
ethics? covert morally questionable?

experiments
lab experiments
:( artificial - sociology not a science? not valid? r ppl in their natural settings?
:(not as represeantative?
:(hawthorne effect
:(psyhcological damage? for example, the bandura's controversial study on media influence on children/violent behaviour + no consent as usually have 2 cover aims of experiment e.g. milgram experiment

reliable!
variables are controlled
quantitative data


Field experiments
less artificiality - ecological validity - natural settings
valid and avoids hawthorne effect

no informed consent - rosenthal and jacobson on SFP. teachers lied to.
social characterisics
not reliable
not that representative
time consuming


Documents
High validity, rich in detail etc like Znaniecki used documents to study polish immigrants to USA
Cheap source of data
Saves time to collect data
Secondary nature minimises ethical issues
not representative some groups unlikely to produce personal documents etc (?emma griffins)
not replicable
not generalisable
written with an audience in mind- biased?

official stats
hard

soft

Stats are practical: accessible, cheap, quick. find out lots of info e.g. Census tells us abt declining secularisato in UK.
can quantify and can compare over time, across diff spectrums etc

prompts to research
reliable
representative

*census errors etc
*social constructionists: stats are not REALITY, but manipulatable. crime stats for e.g. product of overpolicing black ppl. marxists n feminists. etc etc domestic violence not taken srsly enough and not reflected in stats. marx crimes of powerful rarely appear in stats