Research methods

Lab experiements

→ High degree of control by researcher, but more artificial

→ ^^**Positivists**^^ like as reliable + quantitative

* But will rarely use due to problems with them

→ ^^**Interpretivists**^^ find them invalid

\
^^**Milgram**^^ obedience experiment

Problems with lab experiments

→ ==**Artificiality**==; can results be translated to IRL? people react differently in artificial situations, invalid

→ ==**Identifying/controlling variables**==; only effective when ALL variables can be controlled, but hard to control social interactions

→ ==**Hawthorne effect**==; knowledge of being experimented on makes it invalid

→ ==**Ethical issues**==; lots of experiments have ‘blinds’ to conceal aims of research, cant obtain informed consent

→ ==**Limited application**==; labs small, cant study large-scale or past events/long durations

Field experiments

→ More true to life as takes place in real world

→ Cant control all variables

→ Sociologists will create/adapt a situation to their purpose

→ Participants **usually unaware**

→ Aims to **get some control, but more natural than lab**

\
^^**Zimbardo**^^ Stanford Prison Experiment

Advantages of field experiments

→ ==**Less artificial**==; set in real world

→ ==**Validity**==; people unaware of situation and in their normal environment, so act normally

Disadvantages of field experiments

→ ==**Less ctrl over variables**==; cant control everything IRL

→ ==**Limited applications**==; few situations can be turned into a field experiment

→ ==**Ethical issues**==; usually dont gain informed consent, would ‘give the game away’

Comparative method -- experiments

A ‘**thought experiment**’ identifying 2 similar groups with 1 different characteristic e.g. religion

→ ==**Avoids problems with experiments**== e.g. artificiality, ethics

→ Can study ==**past events**==

→ BUT @@**can't control variables**@@

* have you discovered true cause?

Durkheim -- comparative method

Compared suicide rates of otherwise similar Protestants + Catholics

What are questionnaires?

→ Social surveys distributed then returned

→ Answer pre-set questions, usually ==**close ended, pre-coded answers**==

\
The ^^**Census**^^

^^**Barker**^^ ‘making of a Moonie’

Why do positivsts use questionnaires?

→ Believe in **objective social reality**, scientific approach

→ Questionnaires are ==**replicable, quant. data easy to do on large scale**==

→ But ==**low in validity**==

Advantages of questionnaires

→ ==**Quick/cheap**==; dont need to train interviewers, responders complete themselves

→ ==**Large scale**==; can gather lots of data cheaply

→ ==**Easy to quantify**==; precoded close-ended Qs

→ ==**Reliable**==; easy to replicate + compare results between groups

→ ==**No researcher**==; no Hawthorne effect

→ ==**Hypothesis testing**==; can test hypotheses about cause/effect relationships between variables

→ ==**Little/no personal contact**==; can maintain detachment

→ ==**Representativeness**==; can collect from more people = more likely to be representativen

→ ==**Fewer ethical issues**==; respondents dont have to answer

Disadvantages of questionnaires

→ ==**May ask sensitive Qs**== which upset people

→ ==**Need to be brief**==; limits amt of info that can be gathered, ppl wont answer if its too time-consuming

→ ==**Respondent issues**==; have they received it? has the named respondent actually completed it?

→ ==**Low response rates**==; esp with postal questionnaires

→ ==**Unrepresentative**==; low RR distorts representativeness

* busy people may fail to respond e.g. w/c
* can increase by offering incentives/sending follow ups, but **adds to cost/time**

→ ==**inflexible**==; after finalising, cant explore new areas of interest

→ ==**lacks validity**==; no contact between subject + researcher, cant know of respondent interprets questions/answers the same way

* ‘occasionally’ may mean diff things to diff people

→ ==**people can lie**==; e.g. not knowing, giving answers to please/annoy researcher. affects validity

→ ==**imposing researcher’s meanings**==; via choosing questions + response categories, limiting answers, etc.

What are structured interviews?

→ formal interviews

→ use an **interview schedule** to ask interviewees Qs in same way

→ choose from limited list of answers

\
^^**Young & Wilmott** ^^to study the ext. family

Why do positivsts use structured interviews?

→ **Standardised**; easily quantified

→ Reliable + can produce large-scale representative data

Advantages of structured interviews

→ ==**Reliability**==; fixed list of Qs, easily replicated

→ ==**Representative**==; conducted quickly so can have a larger sample

→ ==**Cost**==; cheapest type of interview as quick + dont need to train interviewers

→ ==**Face-to-face**==; higher response rate + can explain research purpose

→ **Limited interviewer effect**; aka interviewer presence affecting responses, which is irrelevant here as contact is just asking/responding to fixed Qs

Disadvantages of structured interviews

→ ==**Lack of validity**==; interviewee cant explain their meanings

→ ==**Reliability**==; interviewer effect - diff in social characteristics/research setting, so cant exactly replicate

→ ==**Cost**==; more expensive than questionnaires

→ ==**Sensitive issues**==; not issues for these, as you need to build rapport

How useful are structured interviews?

→ ^^**Positivsts** ^^say more useful than unstruc, especially where aiming to obtain basic factual info

→ ^^**Feminists** ^^say patriarchal, as usually male interviewer in control. Hard for women to express experiences of oppression

Oakley -- feminism and interviews

There is a distinctively feminist approach to research

→ ==**Value-committed**==; takes womens side, gives voice to their oppression

→ Requires ==**involvement**== with lives of women they study

→ Aims for ==**equality/collaboration** ==between researcher/researched

What are unstructured interviews?

→ Ask open-ended Qs, no fixed set of Qs/answers

→ Qualitative; interviewee responds in words that are meaningful to them

→ **Free-flowing**; guided conversation

→ Can build a rapport

Why do interpretivists use unstructured interviews?

→ People can talk openly

→ Not restricted by fixed list of Qs

Advantages of unstructured interviews

→ ==**Highly valid**==; informal, so can build rapport + more likely to open up

* helps with sensitive issues

→ ==**cant impose ideas onto interview process**==; interviewees can reply in their own words + raise issues they value

→ ==**flexible**==; can ask follow-up questions for a truer picture

→ ==**in-depth responses**== as questions open-ended

Disadvantages of unstructured interviews

→ ==**Questionable validity**==; closer bond may mean respondent gives answers to please researcher

→ ==**Unreliable**==; cant recreate: Qs/responses are random

→ ==**Unrepresentative**==; time, so cant research as many people

→ ==**Sensitive issues**==; some people prefer anonymity of a questionnaire

→ ==**Cost**==; must train researchers, so higher cost

→ ==**Relevance**==; could waste time by ambling into irrelevant territories

→ ==**Group interviews**==; conformist answers/peer pressure

How useful are unstructured interviews?

→ Some validity problems

→ Best to **combine unstruc + struc techniques** into a ==**semi-structured interview**==

→ Standardised questions, open answers

Types of observation

→ ==**Participant**==; joins in with activities of group

→ ==**Non-participant**==; avoids direct involvement with group

\
→ ==**Overt**==; subjects know theyre being observed

→ ==**Covert**==; keep real identity/purpose secret

\
→ ==**Structured**==; systematically classifying behaviour into categories

→ ==**Unstructured**==; just recording what they see/experience in whatever way they can

Viewpoints on observation

→ ^^**Interpretivists**^^ unstructured PO

→ ^^**Positivsts**^^ structured NPO

Typical characteristics of participant observation

→ Observer finds a role within the group that lets them study behaviour

→ Record observation in **field notes**

→ Usually **years-long fieldwork**

→ Starting with open mind, **ideas emerge during study**

Why do interpretivists use PO?

→ Want to discover meanings underlying our actions

→ PO helps them do this by gaining understanding of their worldview

* Can check this via daily experiences, ==**see what they actually do not just what they say they do**==

→ ^^**Positivists** ^^claim it lacks reliability/representativeness

Advantages of PO

→ Highly valid due to…

→ ==**Naturalistic approach**==; observing group in natural setting acting normally (esp. if covert); likely to be accurate account of behaviour

→ ==**Authenticity**==; observing regular routines is often more authentic than questions

→ ==**Open research process**==; fluid/flexible, researcher can ‘go with the flow’ - other research ideas many emerge in process

→ ==**Offer insight**==; deeper understanding of behaviour, rather than just skimming surface

→ ==**Detail**==; recorded observations are detailed, feel true-to-life

→ ==**Can study closed groups** ==via covert PO

* e.g. cults, criminals

→ ==**Research opportunity**==; flexible, so when opportunities arise suddenly, researcher can seize the chance

→ ==**Flexibility**==; if new issues arise mid-study, can adapt PO accordingly

Disadvantages of PO

→ ==**Unreliable/unrepresentative**==; open-ended so cant be replicated, and can only study small groups, cant generalise

→ ==**Hawthorne effect**==; observer presence could affect group behaviour

* overt (presence of researcher) or covert (presence of new member)

→ ==**Going native**==; observer may over-identify with group, can affect their interpretations

* ^^**Ventakesh ‘gang leader for a day’**^^

→ ==**Interpretation issues**==; observer cant be certain theyve understood meanings in the same way as the subjects

→ ==**Deception**==; lack of informed consent

* could even ==**BE IN DANGER**== **(**^^**Humphreys**^^**)**

→ ==**Immoral activities**==; researchers may need to partake or risk blowing cover

→ ==**Hard to ensure anonymity**==; especially in small-group PO

→ ==**Getting in**==; may be hard to join

→ ==**Staying in**==; especially with covert obvs, may have to maintain false identity long-term

Overt and Covert PO compared

==**ETHICS**==

→ overt: problem of protecting identities

→ covert: lack of informed consent

==**ACCESS**==

→ overt: could be denied

→ covert: hard to gain entry, need similar characteristics

==**MAINTAINING MEMBERSHIP**==

→ overt: easy once accepted

→ covert: risk of blowing cover, could cause danger or end research

==**ASKING QS**==

→ overt: can ask Qs openly, increased validity

→ covert: direct Qs could raise suspicions, restricts observer

==**VALIDITY**==

→ overt: Hawthorne effect

→ covert: acts more naturally

How useful is PO?

→ Lack ==**reliability/representativeness**==

* But ^^**interpretivists**^^ more interests in validity, see PO as closest to being fully valid

→ May be ==**only way to study some groups** ==e.g. deviants

How does structured observation work?

→ Use an **observation schedule** to identify/measure patterns of behaviour

→ Researcher decides how to categorise behaviour in advance, codes so that data can be quantified + turned into statistics

e.g. recording frequency of event

→ Usually overt as difficult to conceal obs schedules

Why do positivists use structured observation?

→ Fixed categories = easily quantified

Advantages of structured obswervation

→ **Produces reliable data**; easily replicated by using same fixed categories

→ **Comparing data**; can quickly/easily produce quant. data by counting frequency of events observed

* Can directly compare different observations

Disadvantages of structured observation

→ ==**Use of categories**==; events may not fit any categories/may fit multiple

→ ==**Counting events**==; doesnt tell us their meaning

→ Different observers may ==**put same event in diff categories**==

→ Observer may ==**miss some events**== in an interaction

→ Only useful for ==**small-scale interactions**==

→ ==**Intensive** ==to carry out

What are official statistics?

→ Quant data collected by govt bodies e.g. ==**Census**==

→ ==**Hard stats**==; counts of events e.g. birth/deaths

→ ==**Soft stats**==; more easily manipulated e.g. crime stats

from the ^^**Office of National Statistics**^^

Why do positivists use official statistics?

→ Deliver large-scale, rep, quant. data with reliable methods

→ ^^**Interpretivists**^^ say too socially constructed

Advantages of official statistics

→ ==**Availability**==; already existing, cheap + readily available

* Some published regularly, can be compared
* Already categorised

→ ==**Representativeness**==; big samples e.g. Census

→ ==**Coverage**==; cover many important topics, especially which state are interested in

* ed, marriage, crime

→ ==**Prompts to research**==; provide starting points e.g. patterns in need of further investigation

* boys’ educational underachievement first discovered by official stats

→ ==**Background data**==; provide useful background material

* e.g. class/gender/ethnic makeup of a research group

→ ==**Comparability of data**==; quantative, can draw comparisons

→ ==**Reliability**==; same categories/modes of collection used every time, can be replicated

* BUT can be changed e.g. league tables

Disadvantages of official statistics

→ ==**Definition/measurements**== of concepts may differ from those of the sociologists

* stats measure class in terms of occupation, but some sociologists see this as based on ownership of property

→ ==**Reliability**==; may not be e.g. errors in recordings people completing forms wrong

→ ==**Social construction**==; official stats are socially constructed, result of social processes

* for crime to appear in official stats, it has to go through a process. @@**dark figure of crime**@@

→ ==**Political bias**==; ^^**Marxists**^^ say stats reflect r/c interests

* definitions used/how stats presented based on political choices
* @@**definition of unemployment changed 30x during Conservative govt**@@

→ ==**Male bias**==; ^^**Feminists** ^^say stats biased against women

* e.g. definition of ‘work’ in Census excludes unpaid housework

Atkinson -- official stats

Variation in what is labelled a ‘suicide’ based on coroners’ choices

→ SHows how these are socially constructed

How useful are official statistics?

→ Presented as social facts by ^^**positivsts**^^, but constructs by ^^**interpretivists**^^

→ Hard stats less socially constructed

→ Soft stats more

What are documents?

→ Secondary data made by individuals/groups, which sociologists use for research

→ ==**Personal docs**==; letters, diaries, memoirs, notes, albums

→ ==**Public docs**==; reports from govt/charities/businesses

→ Usually qual. data experssing beliefs, some have ==**statistical data**==

* @@**Black Report**@@ into health inequality

Why do interpretivists use documents?

→ Docs usually qualitative; researcher can explore meanings people can attach to events

→ Highly valid; freely expressed meanings of those who produce them

Advantages of personal documents

→ ==**Validity**==; written for personal purposes, genuine insight into people’s attitudes

→ ==**Cheap/fast**==; save time + free/cheap to access, but might be hard to access

Disadvantages of personal documents

→ ==**Unrepresentative**==; some groups (e.g. illiterate) arent likely to produce personal docs e.g. letters/diaries, so views aren't represented

* those with time/literacy may be overrepresented

→ ==**Benefit of hindsight**==; some docs are created post-event with benefit of hindsight e.g. memoirs, autobiographies

→ ==**Written with an audience/personal bias**==; e.g. letters written with audience in mind, may affect whats recorded

Advantages of public documents

→ Detailed

→ Cheap

→ Easy to access

Disadvantages of public documents

→ Content may be ==**selective/biased**== + reflect views of journalist/media owner

→ Authors aware that documents are publicly available

Historical public documents

→ Often only way to **study past societies**

* Especially if all survivors are now dead

→ Can be issues with interpreting

* MEanings of things change over time

→ Some historical docs are lost/destroyed; incomplete/unrep picture of the past

Content analysis and documents

→ Quant content analysis ==**measures amount of coverage given to issue**==

* News reports etc., amount of time/space given to particular POVs
* **doesnt tell us about meaning**

→ ^^**Interpretivists**^^ use qual content analysis to ==**examine meanings given to particular words/images**==

Scott -- usefulness of documents

Four tests of any document

→ ==**Authenticity**==; is it genuine?

→ ==**Credibility**==; can we believe it? is author sincere?

→ ==**Representativeness**==; how typical is the doc of a wider social group?

→ ==**Meaning**==; can we interpret author’s meaning correctly?

Examples of PO

^^**Barker** ^^‘making of a moonie’, overt

^^**Ventakesh** ^^‘gang leader for a day’, overt

^^**Whyte**^^ ‘street corner society’, covert

^^**Patrick** ^^‘glasgow gang observed’, covert

^^**Humpreys** ^^‘tea room trade’, covert