The research process, surveys and questionnaires Flashcards
research process
> before conducting research, they must be clear about how they will collect data.
1.sociologist identifies what they are interested in studying - topic chosen
2. clear hypothesis is established, allowing them to test a cause and effect relationship
3. any concepts that form part of the research are “operationalised” (defined, stating how they will be measured)
4. a “pilot study” allows method to be tried on small sample of people - allow for any errors to be corrected.
5. a “sampling frame” of people are chose to take part, this is hopefully “representative of the research population the study focuses on.
THEN RESEARCH BEGINS
Sampling
before research, sociologist identifies who their “research population” are - this refers to all people the research applies to - this is a high number, therefore sociologist chooses a “sample” of them hoping it is representative.
> sample is selected from an available “sampling frame” - list of people from which a sample of people are selected - school register for example.
> “sample” - the people selected from the “sampling frame”
Types of sampling
SNOWBALL SAMPLING - a researcher gains access to a group then uses member of the group to contact others, then uses the new group to contact others and so on - only way to get contact with secretive or deviant groups.
RANDOM SAMPLING - people in “sampling frame” have equal chance to be selected e.g. names out of a hat
SYSTEMATIC SAMPLING - people in sampling order are selected in order e.g. every 10th person in register
QUOTA SAMPLING - researcher go out looking for the right number of each sort of person required for study e.g. researcher may need 500 males and 500 females
STRATIFIED-RANDOM SAMPLING - sampling frame divided into groups e.g. by gender and then random sample is taken - ensures the sample is representative e.g. a study looking at boys’ attitudes in school, register would need to be divided into boys and girls, before a boy sample is taken.
Surveys
Social surveys collect standardized data from large amounts of people - e.g. the census in Britain -it is distributed to every household in England and Wales every 10 years and huge amounts of data are collected
Types of surveys
3 TYPES
1. FACTUAL SURVEYS - collect descriptive data relating to population
2. ATTITUDE SURVEYS - seek people’s opinions, thoughts and ideas
3. EXPLANATORY SURVEYS - begin to test any assumptions that researchers hold.
Closed-ended questions
> pre-coded, tick box, yes no questions - make it easy for sociologist to quantify data
sometimes a “likert scale” is used - allowing respondents to select from range of options, e.g. “strongly agree” “agree” etc.
closed- ended questions highlight patterns relatively easy
it is “time-efficient”
coding of responses people give is straight forward to spotting trends
Open-ended questions
> provide freedom to go into more detail
opportunity for respondents to expand upon an answer, enabling them to expand on initial comments
therefore data more likely to be valid as people can explain what they mean
Operationalisation of sociological concepts
> abstract concepts must be translated into concrete questions
for example, if the term “social class” is used, researcher must be clear with what this means in their study - so a precise measurement can be taken
without this, confusion may occur as they may have different definitions on terms - which would undermine validity
Postal Questionnaires
> these are mailed out to a sample of respondents - who are asked to complete it and return it.
EVAL GOOD - questionnaires can be distributed over a “large geographical area” - this helps obtain a “representative sample” of responses - how response rate is low
EVAL GOOD - allow respondent to take their time, in comfort at home - the privacy may help respondents to be honest when answering questions
EVAL BAD - there is no researcher present when answering questions, therefore if respondent doesn’t understand a question they cannot ask for reassurance - this can cause inaccurate data
EVAL BAD - the questionnaire may not be completed by “intended recipent” therefore may not be reflective of target sample.
EVAL BAD - people may be busy or not have grasp on English therefore not do the questionnaire and ignore it - waste.
Callender and Jackson (2005)
Fear of debt and Higher Education
used self-completion questionnaires to measure fear of debt associated with university study.
> researched attitudes of A level students in England towards debt and their decisions about whether to apply for uni.
> research involved “postal distribution” of 3,582 questionnaires to 101 sixth form and colleges - responses received from 1,954 students in 82 schools - 55% response rate.
> closed-ended questions - allowing easy analysis of trends found.
> concluded that “debt-aversion” was a social class issue - fear of debt was higher among low-income groups.
EVAL GOOD - VALIDITY - respondent responded to questions with peers, this allows them to clarify any confusement with someone else - increasing validity
EVAL BAD - VALIDITY - others being with each other can have negative impact also - their answers may be influenced by others rather than their own thoughts - this undermines validity
EVAL GOOD - questionnaires are “quantitative method” - as large amount of data can be coded and analysed scientifically
EVAL GOOD - cheap to produce - easier for sociologist to save money
EVAL GOOD - highly reliable - all respondents have identical questions
EVAL GOOD - findings are representative
EVAL GOOD - VALIDITY - nobody forced to do questionnaire - they can be done privately in their room - responses are treated confidentially - encourages them to be honest therefore maximum vailidity is achieved
what problems can be linked to questionnaires?
EVAL BAD - VALIDITY - interpretivists say they are problematic - as “bias creeps in” - sociologist creates questions are based on what they believe is important - this is called “researcher imposition” - responses cannot be valid
EVAL BAD - sociologist believes operationalising concepts will minimise respondent mis-understanding and invalid responses due to this - this is a “fallacy of judgement” - the truth of data sociologists believe that comes from questionnaires is no more than a “social construction”
EVAL BAD - VALIDITY - respondents can lie, forget and over-exaggerate answers - “role distance” between researcher and respondent prevents respondent asking about anything confusing - compromises validity.
Gary Crawford and Victoria Gosling (2004)
carried out a study into women’s participation in digital gaming
> based on questionnaire distributed to 392 undergraduate students at Sheffield’s two universities studying on sociology, social work or cultural studies.
> questionnaire included 30 short, mainly closed-ended questions
> researchers “admitted their research was not representative”
> 63% questionnaire respondents were female
> only 50% of women compared to 93% of men said they played a digital game in last three months.
