Process of conducting research Flashcards
formulating aim or hypothesis
Most studies either have a general aim or a specific hypothesis. A hypothesis is a possible prediction that can be tested by collecting evidence to prove it true or false. E.g. we may suspect that family size affects educational achievement. If so, we can formulate a specific hypothesis as a cause-and-effect statement,
While a hypothesis is a statement about a specific relationship (‘A’ causes ‘B’), an aim is more general. It identifies what we intend to study & hope to achieve through the research. Often it will simply be to collect data on a particular topic, such as the way of life of a subculture.
formulating aim or hypothesis advantages
The advantage to a hypothesis is that it gives direction to our research. It will give a focus to our questions. Positivists favour a hypothesis as a starting point for research. This is because they seek to discover cause-and-effect relationships – e.g. that large family size causes underachievement. Using quantitative methods such as questionnaires, they formulate questions designed to discover whether & why these factors are linked.
The advantage of an aim is that it is more open-ended. We are not tied to trying to prove a particular hypothesis – instead we can gather data on anything that appear interesting about a situation. This can be very useful at the start of our research, when we know very little about the topic. Interpretivists often favour a broad aim rather than a hypothesis, since they are interested in understanding actors’ meanings, so the task is to find out what the actors themselves think is important, rather than to impose the researcher’s own possible explanations in the form of a hypothesis.
Operationalising concepts
before we can test it, we need a working or ‘operational’ definition of our key ideas – in this case, social class. The reason is simple: without a working definition, we won’t be able to count the number of working-class pupils who have or don’t have qualifications.
Now ‘social class’ is a fairly abstract concept so we need a way of measuring what class each pupil belongs to. This process of converting a sociological concept (such as class) into something we can measure is called ‘operationalisation’.
Operationalising a concept may seem straightforward, but a problem can arise when different sociologists operationalise the same concept differently. E.g. we might disagree about whether a routine office worker is working-class or lower middle-class. This can make it hard to compare the findings of different pieces of research.
Positivists are concerned to operationalise concepts because of the importance they place on creating & testing hypotheses. In contrast, interpretivists put less emphasis on operationalising concepts. This is because they are more interested in how the research participants see themselves that in imposing their own definitions of these concepts.
Pilot study
This involves trying out a trial version of the questionnaire or interview schedule (the list of interview questions) on a small sample. The basic aim of the pilot study is to iron out any problems, refine or clarify questions & their wording & give interviewers practice, so the actual survey goes as smoothly as possible. A pilot study may reveal that some questions are badly worded & hard to understand, or that the answers are difficult to analyse. This will allow a researcher to make any changes before the main study which will save time, money & the potential success of their research.
sampling
When conducting any research investigation, the researcher will need to consider who they are going to carry out the research study on. It would be impossible for a researcher to study the entire population (far too expensive & time-consuming) so a small selection of participants is chosen to take part in the study. This is known as a sample.
Target population:
The group of people that the sociologist is interested in researching in order to draw conclusions about their views or behaviour (e.g. single mothers or teenagers). Usually it isn’t possible due to the practical issues of time or cost to study the whole of the target population (unless it is very unique & small) which is why a sample is necessary.
Sampling frame:
A list of names taken from the target population from which the sample will be selected – e.g. the electoral register, the telephone directory, a school register. For a sampling frame to be accurate & representative, it must include all the potential members of the target population.
Sampling :
The process of selecting participants to study from the sampling frame. This can be done in a number of different ways which each have their own advantages & disadvantages.
Generalisability :
The degree to which the sample ‘represents’ (i.e. is representative of) the larger population, in that conclusions from the research on the sample can be ‘generalised’ to the wider population.
Bias sample :
One which does not represent the target population.
types: random
When a sample group is chosen completely at random
& everyone in the sampling everyone in the sampling frame has an equal
chance of being selected. E.g. All Y10 names put in a hat or into a computer & sample picked at random.
✓ Simplest of sampling types, cheap & easy to organise – either ‘names
out of a hat’ or a computer programme that selects names at random.
Everyone has a genuine, fair & equal chance of being selected.
🗶 Can still potentially lead to a bias sample (e.g. a lot of male nurses could
be randomly picked when there are a lot more female nurses) which will
make the sample unrepresentative of the target population so it will be
hard to make generalisations from the research findings.
systematic
Selecting every nth name from
the sampling frame which is not random. E.g. Take a school register & pick
every 10thname on the list.
✓ Simple, quick & cheap way of selecting a sample – just have to follow
the order the researcher decides. Also it is more precise than random
sampling if the sampling frame is large as it allows a more even spread
of participants.
🗶 Can lead to a bias sample as it is not entirely random & the same sort of
people might always being picked – e.g. if we take our sample from a list
of house addresses in a particular area & select every 4th house, this
would generate even numbers, which would give you a sample of houses
on one side of the road. It is possible that this may be the side that are
wealthier, e.g. middle-class, than those on the other side of the street.
stratified
Research population is divided up into relevant groups & a random sample is then taken from each of these groups to produce more representative data. E.g. Using the electoral role for Wallasey as a sampling frame & then dividing the list by gender & age before a random sample.
Developed to try & solve some of the problems with random sampling – no longer leaves everything to chance as relevant factors are taken into account assures that all subgroups of the population are proportionally represented in the sample.
Takes far more time to organise than a purely random sample which can also then be more expensive
QUOTA SAMPLE
like stratified sampling, this divides the
target population into subgroups (e.g. by gender, age, social
class, ethnicity, etc.) & then the researchers go out looking for
the right number (quota) of each sort of person required in
each category – e.g. married couples, male single parents,
female university students.
✓ Less time-consuming over stratified sampling & less
expensive as no sampling frame is needed.
🗶 Can lead to a biased sample which may not be
representative as the researcher’s own judgement can lead
to bias & a distorted quota
OPPORTUNITY SAMPLE
Sometimes called convenience
sampling. Involves researchers approaching anyone who is
available & willing to participate. E.g. selecting from passers
by in the street, a class of pupils in school or friends & family
✓ Quick & practical form of sampling that takes little or no
prior planning
🗶 Often not representative & the sample may be biased
due not being randomly selected – e.g. people out on the
street on a weekday afternoon will not include the huge
proportion of people in work or school.
VOLUNTEER SAMPLE
Sometimes called self-selecting sampling. When
participants have freely self-selected (volunteered) themselves to be part of
the study. Works through adverts in newspapers, leaflets, posters, TV & radio.
✓ Quick & practical compared to other techniques such as random sampling.
🗶 Volunteer bias can occur because people who self-select often have certain
social or personal characteristics that are different from those who do not.
They may be more educated, enthusiastic, motivated, social, etc. that what
is typical of the population. This can make the sample unrepresentative &
hard to generalise from.
SNOWBALL SAMPLE
Non-random sample where a researcher makes contact
with 1 person & then asks them to put them into contact with others to build-up
the sample. E.g. Finding 1 drug dealer willing to talk to you & asking them to put
you in contact with others. Used when it is difficult or impossible to obtain a
sample of people to research on. It is often used when the research is highly
sensitive/secretive/deviant (e.g. criminals, religious cults, etc.) or when the sample
is difficult to identity or is unusual (e.g. stamp collectors).
✓ Useful for gaining a deep insight into topics with no sampling frame which are
difficult to access – e.g. gang members, gamers, prostitutes, stamp collectors
🗶 Often not representative & can be biased due to being a small sample which
is not randomly selected – it relies on a very small network of people who have
come through recommendations which may be biased or incorrect.
Factors when choosing a sample
▪ The time allocated & funding available for the research study to take place.
▪ The methodological perspective:
- Positivists tend to favour research methods that produce quantitative data (e.g. questionnaires & official statistics) as these often use a large sample. This makes the findings more representative of the population, which allows for generalisations to be made, with the aim of establishing trends, patterns & relationships.
- Interpretivists, on the other hand, favour research methods that produce qualitative data (e.g. in depth unstructured interviews, observations & case studies). They tend to seek a deeper insight into people’s motives & the meanings of their actions, as these provide a more valid account. They are not interested in making sweeping generalisations about human behaviour from their results. Therefore, their sample sizes tend to be much smaller & less representative of the wider population.
