Chapter 11 – Research Methodology
Designing Psychological Research
A variable is something that can be changed, such as a characteristic or value. Variables are generally used in psychology experiments to determine if changes to one thing result in changes to another.
Independent variable (IV) – In an experiment, it is the variable that is controlled and manipulated by the experimenter. For example, in an experiment on the impact of sleep deprivation on test performance, sleep deprivation would be the independent variable.
Dependent variable (DV) – The DV is the variable which is measured by the researcher as a result of the manipulation of the IV. For example, in the experiment that researches the impact of sleep deprivation on test performance, test performance would be the dependent variable.
Extraneous Variables – These are generally unwanted variables that can have an impact on the relationship between the IV and DV. Examples of extraneous variables in our study relating to the how sleep deprivation affects test performance would be: age; gender; academic background. There are two basic types of extraneous variable:
Participant variables – This relate to individual characteristics of each participant that may impact how he or she responds. Examples are mood, anxiety, intelligence, awareness.
Situational variables – These relate to things in the environment that may impact how each participant responds. Examples are the temperature of the room, outside noise.
A hypothesis (plural hypotheses) is a precise, testable statement of what the researchers predict will be the outcome of the study. There are two main ways of expressing a hypothesis:
The Null Hypothesis – States that there is no relationship between the two variables being studied (i.e. the IV does not affect the DV). Any results found are due to chance.
In a study investigating the effect sleep deprivation might have on test performance, the null hypothesis would be: There will be no significant relationship between sleep deprivation and test performance.
The Alternative Hypothesis – This is also called Experimental Hypothesis when the method of investigation is an experiment (e.g. a laboratory; field; natural or virtual)
The alternative (or experimental) hypothesis states that there is a relationship between the two variables being studied (one variable will have an effect on the other). In a study investigating the effect sleep deprivation might have on test performance, the alternative hypothesis would be: Sleep deprivation will significantly affect test performance.
An experiment involves the manipulation of the IV to see what effect it has on the DV, while attempting to control the influence of all other extraneous variables.
Laboratory Experiments – These normally take pace in a controlled environment, such as a laboratory. The researcher deliberately manipulates the IV while maintaining strict control over the extraneous variables through standardised procedures
- Strict controls and procedures mean other researchers can easily replicate lab experiments and check for reliability
- High control over extraneous variables implies cause and effect can be studied without any influence from other unwanted variables.
- The artificiality of the setting means participants might not demonstrate real-life behaviour – this dramatically reduces ecological validity.
- Demand characteristics are likely.
Field Experiments – These experiments are the same as laboratory experiments in terms of how researchers treat the IV, DV, and extraneous variables. However, the laboratory experiment is swapped for a real-life setting such as a school, town centre or hospital.
- Ecological validity is higher than in a lab experiment due to the ‘real world’ setting.
- We can assume there is less bias from sampling (participants do not have to be brought into the lab) and demand characteristics (if subjects are unaware that they are being tested).
- The researcher’s control over the environment is reduced in the real world. Therefore, extraneous variables are more likely to confound the results and greatly reduce validity.
- More time consuming and thus can be more expensive.
Natural Experiments – Where the change in the experiment is not brought about by the researcher. The change would have occurred whether the experimenter was interested in it or not. The IV occurs naturally. Example: To investigate the levels of anxiety in phobic and non-phobic people. In this study, nobody induces phobia in the participants. It simply exists.
- Great ecological validity
- Great generalisability.
- We can assume there is less bias from sampling (participants do not have to be brought into the lab) and demand characteristics (if subjects are unaware that they are being tested)
- The cause and effect relationship of the IV and DV can be more easily affected by confounding variables.
- Participants cannot be randomly allocated to the experimental or controlled condition.
Steps of a Research
- Research Aims and Selection of Topic – Identification of a problem that will be studied.
- Reviewing Existing Evidence – Conducting a literature review (finding out what is already known about the chosen topic) as drawing on the ideas of other sociologists helps to clarify the issues and in making decisions on how to proceed.
- Hypothesis Setting – Theory or explanation at the start of a research that the research is designed to test.
- Choosing a Method – Researcher needs to evaluate the type of data he/she wants to collect and which type will give evidence which can prove/disprove the hypothesis.
- Pilot Study – Small scale test of a piece of the research project to test for problems and ways to improve research methodology. (Doing so saves money and time in the future during larger-scale researches).
- Sampling – Taking selected members of the survey population (all those to whom the findings of the study will apply). Usually taken so that the research is representative (researcher can claim findings apply to all members of the population). To be generalizable, the sample has to be a cross-section of the population.
- Research is carried out and Data is collected to be analysed
- Hypothesis is proven true or false and findings are usually published.
NOTE: Sometimes sampling isn’t required. E.g.: Many countries have a census (social survey carried out by the government) to get information about EVERY single person in the country. Censuses collect information about the whole population, not a sample and therefore, have generalizable findings in regards to that country.
Sampling Frames and Types of Sampling
Sampling Frame – A list of members of the (survey) population from which the sample is chosen.
E.g.: of sampling frames:
- Electoral rolls: List of everyone who is registered to vote along with their address. Problem: wouldn’t contain anyone below the age of 18 (however is ideal for researches concerning only adults).
- Telephone directories: Give addresses as well as telephone numbers and are usually easily available. Problem: list only one member of the household, do not provide information about other members at the same address, do not list people without telephones or those who have chosen not to be included.
- School registers: Lists of children in school with information about their gender, age, etc. Problem: available only to genuine researchers, require permission from those in authority (such as headteacher).
Types of Sampling
- Random Sampling – Everyone in the sampling frame has an equal chance of being chosen. (E.g.: Can be done by drawing names from a hat). Random samples are not always representative.
- Stratified Sampling – Sampling frame is divided into categories (e.g.: boys and girls) and then, a random sample is taken from the categories. (AKA Stratified random sample)
- Systematic Sampling – Regular pattern in choosing the sample. Not random as other names in the frame have no chance of being chosen.
- Cluster Sampling – For survey populations spread over large areas, certain areas are chosen as sampling frames from which random samples are taken.
- Opportunity Sampling – Taking people who are available at the particular moment as the sample. (E.g.: Researcher stopping people on the street and asking them questions. Not random as people not present do not have a chance to get chosen and researcher has a hand in who to choose.
- Quota Sampling – Finding a certain, pre-decided number of people who fit certain characteristics required by the researcher.
- Snowball Sampling – Finding one respondent and getting them to put you in touch with one or more others.
NOTE: Last three sampling methods do not have a sampling frame.
Ethical Issues in Implementing a Research Strategy
- The participants should not be harmed – includes, but not limited to, physical and mental harm. Participants should not feel distressed, angry or upset. May happen if they’re asked about something that disturbs them. Can be protected by confidentiality
- The participants’ informed consent must be obtained – respondent must agree to take part after fully understanding what is involved (purpose of the research, when and where the findings may be available and what they may be used for). Participants reserve the right to withdraw and the researcher shouldn’t try to persuade them in doing otherwise.
- Researcher should not invade the participants’ privacy – Participants reserve the right to refuse to answer particular questions as they may invade their privacy. E.g.: questions regarding their earnings or religious beliefs.
- The participants should not be deceived – Researcher may present their research as something different from what it is by lying or not fully disclosing the true nature of the research in order to get the participant to answer more naturally. If deception is used, the participant needs to be debriefed at the end of the research (told about the true nature of the study).
- Participant’s name, address, etc should be kept anonymous and confidential (it shouldn’t be possible to trace a particular individuals’ answers from the published findings.
Research Methodologies in Sociological Investigations
Self-Completion/ Postal Questionnaires
- An example of quantitative data.
- It is a primary source of data.
- Involves a pre-set of questions that the respondent answers and returns to researcher.
- Questions will most likely be ‘closed’ questions.
- Answers will be limited to such responses as ‘yes’, ‘no’, ‘sometimes’, ‘unsure’, or may take the form of factual information,
- e.g. how many rooms do you have in your house?
- Can cover a wide sample. E.g.: by selecting from postal code areas.
- Relatively cheap to administer – cost of stamps, questionnaire publication, etc.
- Low cost as it can cover a large number of people for small amount of money.
- Only needs minimum involvement of researcher and hence, cannot influence answers given. Time only needed in drawing up questionnaire, sending it and analysing results.
- Responses are usually easy to quantify because of questionnaire construction. This is especially the case with pre-set questions.
- Good for obtaining factual information.
- Respondents may like the anonymity of a postal questionnaire and therefore may be more honest in giving answers to questions.
- Respondents can complete the questionnaire when convenient.
- High in reliability.
- Can get a poor response rate because people forget to send it back or lack motivation or incentive to do so.
- Postal questionnaires are usually limited in terms of length of questionnaire and type of questions that are asked. If too lengthy or too complex then respondents will not take time to complete it.
- No way of exploring issues.
- Respondents are unable to clarify any points they are unsure about.
- Relies on respondent’s ability to read and write.
- Relies on respondent’s ability to understand the questions.
- Low validity.
Examples of use:
- Attitude surveys. E.g.: Research into TV violence.
- Lifestyle surveys. E.g.: Research to find out people’s consumption habits.
- Quantitative data.
- It is a primary source of data.
- Carried out face-to-face by a researcher.
- Researcher asks a set of pre-set questions.
- Questions will tend to be ‘closed’ so as to demand a limited response.
- Ensures a good response rate, as questions have to be answered there and then.
- If questions are pre-set then the results are easy to quantify.
- Good for gaining factual information.
- Respondents would not have to be able to read or write to take part in any study.
- Respondent can ask for clarification if they have not understood something.
- Researcher may be able to build a rapport with the participant, gaining trust and hence getting more valid answers.
- Can be costly as it involves face-to-face contact.
- Need the researcher to be there to carry out the interview.
- Can be time-consuming for researcher and respondent.
- If pre-set questions are used these cannot be explored (only clarified if something is not understood).
- People may not answer honestly. This may because they are too embarrassed, or they give an answer that they think the researcher wants to hear – social desirability.
- Researcher may influence the answers given through their own social characteristics or through interaction with the participant.
- High in reliability.
- Socio-economic status and political views.
- Ethnicity and attitudes to ‘stop and search’ policing.
- A method for obtaining quantitative data.
- It is a primary source of data.
- Used to gain statistical information that can be used to represent wider populations
- Involves a pre-set of questions that respondents answer.
- Questions will most likely be ‘closed’ questions and will be standardised – every respondent is asked the same questions.
- Most answers will be limited to such responses as ‘yes’, ‘no’, ‘sometimes’, ‘unsure’, or may take the form of factual information, but some scope for open answers.
- Efficient and practical way of collecting information from a large number of respondents.
- Very large samples and coverage made possible.
- Statistical calculations can be made to measure reliability, validity, and statistical significance. Amenable to the collection of a wide range of information
- Relatively easy to administer format allows researcher to focus on directly relevant information.
- Reliant on respondents being honest, motivated, able to respond and remembering accurately. Hawthorne effects.
- Not appropriate for studying complex social behaviour where an academic understanding may be required. Results may be superficial and anecdotal.
- Answers may lack depth and may not adequately reflect qualitative aspects.
- Although sample is often randomly gathered, respondents are normally selected, which reduces reliability and validity.
- Often used to study modes of behaviour, values, attitudes, and beliefs.
- Often takes the form of content analyses.
- A method for obtaining quantitative data in the form of statistics.
- Often involves systematic and in-depth examination of a film, documentary or newspapers.
- Involves categorising an aspect or incidence of a particular behaviour or use of language and recording frequency of occurrence.
- Allows in-depth analysis of materials not normally subject to such detailed analysis.
- Cheap and easy to conduct.
- Makes use of readily available household equipment.
- Reliable in nature.
- Does not involve people as respondents so avoids arising ethical issues.
- May be time-consuming and pedantic to conduct.
- Information may only be applicable to the resources under investigation.
- Information may be difficult to qualify.
- Difficult to allocate different materials to different categories.
- Quantitative data produced – doesn’t show correlation or causation, why a media text is the way it is, or how it affects the audience.
- Studies of how stereotypes are reinforced by characters in a TV drama such as the BBC 1 programme River City.
- Studies of TV news coverage of current events.
- Quantitative Data
- Secondary source as the researcher is using existing data.
- Statistics used would be those gathered by government, police, health authorities, etc.
- Often used to analyse trends in social behaviour.
- Statistics must be treated with care, as all statistics require interpretation.
- Good for quantitative studies. E.g.: How many crimes are reported each year.
- Can save researcher a lot of time as information has already been gathered.
- Low cost.
- May be a good indicator of a general trend of a particular social behaviour.
- Some statistics gathered from a wide representation of the population.
- May be biased because of the way information was gathered. The researcher has no control over this.
- People may lie in official statistics. For example, it is estimated that 1 million people did not complete Census forms in 1991 because of Poll Tax issues.
- It may be difficult to use statistics for comparison between different time periods. This is because indicators and criteria may change between time periods. For example, statistics on socio-economic status.
- Trends in violent crimes.
- Socio-economic status and health care.
Qualitative Interviews – Unstructured, Semi-structured, Focus group and group interviews
- Qualitative Data
- Primary Source
- Researcher has a number of broad topics/general areas to cover with interviewee
- Questions would be ‘open’ questions.
- Respondent is allowed to elaborate on any of the areas covered.
- Allows the researcher to explore issues in an in-depth way.
- The researcher is not restricted to pre-set questions.
- The researcher can clarify points and explore particular points.
- Good for ascertaining meaning, feelings, motives, etc.
- Detailed and valid data on point of view of respondents, who can say what they really think.
- Can lose track of the purpose of the interview.
- The interviewee may digress into irrelevant information.
- Can be difficult to quantify results, as much of the data may be descriptive.
- Can be time-consuming for the researcher and respondent.
- High cost because of high researcher involvement.
- May be difficult to compare answers given by different individuals.
- Not reliable as they are difficult to replicate.
- Respondents may be affected by interviewer bias or interviewer effect.
- Studies that explore causes of marital breakdown.
- In-depth studies of poverty and how it affects
- Qualitative Method
- Primary Source
- The researcher observes the social behaviour of others
- Records what he/she observes either at the time or as soon as possible after the event.
- The researcher has to take what she/he sees at face value and interpret what is observed.
- Good for describing ‘natural’ behaviour – if the individual/group being observed is unaware of the researcher’s presence.
- Good for gaining an in-depth picture of social behaviour.
- Needs a high input from the observer in terms of time.
- Costs are high, as researcher needs to be there all the time.
- Difficult to quantify behaviour.
- No way of checking details or exploring issues further.
- There may be bias on the part of the researcher in what he/she sees.
- Ethical considerations related to individuals/groups being observed without their knowledge.
- Social behaviour in public places, e.g.: Racial prejudice on public transport or sharing behaviour of children in playground at school.
- Carried out periodically over long periods of time
- Researcher usually uses same group of respondents – panel studies – and the panel members are interviewed on a regular basis
- Shows how people’s lives change over time
- Possible to see what factors may have brought about a change in people’s lives over time
- As respondents are committed, high chance data collected is valid
- Commitment of time, money and effort required from researcher for a long time
- Inevitable drop out (sample attrition) as people die/move away/ do not want to participate.
- May change the participants – Hawthorne effect – as they may start thinking about the aspects of their life they are questioned about more and may act differently as a result.
- British Social Attitudes survey
- British Crime survey
- Qualitative Data
- Primary Source
- Researcher becomes a participant in the group/situation he/she wishes to observe.
- Researcher’s presence will probably be unknown to those being observed or may only be known to one or two key individuals.
- There are three stages to participant observation – ‘getting in’, ‘staying in’, and ‘getting out’.
- Gives an in-depth picture of social behaviour.
- Can give a realistic picture of social behaviour.
- Is good for exploring issues of feelings, meanings, interactions and processes.
- High in validity
- High involvement of researcher in terms of time – Researcher has to be in the situation.
- Costs are high because of high involvement of the researcher.
- Can be biased.
- ‘Hawthorne effect’ – the presence of the researcher may change or influence the situation or the behaviour of those he/she is studying.
- Can be dangerous – For example, participant observation into gang behaviour.
- Can be biased because the researcher becomes part of what he/she is studying.
- Difficult to quantify results – Data tends to be descriptive.
- Difficult to record – If the researcher is part of a group, writing down details may be impossible.
- May be difficult to generalize findings – Findings may only apply to a particular situation or group.
- Reliability is low
- For covert observation, research needs to have characteristics allowing them to “fit in” the group.
- May lose objectivity if they come to identify strongly with the group
- Has to spend a lot of their time and energy maintaining their cover rather than gathering information initially – applicable only for covert observation
- Behaviour of drug users.
- Classroom behaviour.
- Qualitative data.
- Secondary source.
- Uses existing data such as diaries, letters, personal accounts.
- May be found in personal collections, published form, government archives, libraries or museums.
- Provides evidence for in-depth accounts, case studies, or to give a historical perspective to a particular study.
- Can give insight to a particular situation or period in time.
- Good for looking at society from a particular individual’s point of view.
- May be the only source of information about a particular society, event, etc.
- May support other evidence. For example, statistics on disease in the early part of the century may be supported by personal documents from physicians of the time.
- May be biased. For example, the author may be aware that someone would read his/her account.
- If the person is no longer alive, then there is no way of checking his/her account.
- Personal accounts only say what the person wanted others to know – they do not tell us what is missed out.
- If some documents are in private collections it may be difficult to get permission to use in research. This may also be the case with government documents that may be subject to laws regarding confidentiality and time lapses before disclosure.
- They may be subject to data protection legislation.
- The authenticity of some documents may be questioned if authenticity cannot be proved. For example, the case of the ‘Hitler diaries’ that proved to be a hoax.
- Studies showing the changing role of women throughout the centuries.
- Studies that examine changing social structures, e.g. feudalism /capitalism.
- A method for obtaining qualitative data
- Often involves systematic and in-depth examination of a single event or case over time.
- Involves detailed study, data collection, analysis of information and reporting of results.
- Often carried out to glean specific information and understanding rather than to test hypotheses.
- Allows in-depth analysis and understanding of particular cases.
- May generate ideas and hypotheses for future research.
- May complement the use of other methods such as interviews and observation.
- Very time-consuming and demanding of researcher.
- Information may only be applicable to the case under investigation.
- Information may be difficult to collate.
- Information may be difficult to quantify.
- Studies tend to be of individuals, events.
- Allows researcher to support quantitative data with qualitative providing a research with reliability and validity
- Can be used to cross-reference researcher’s interpretations to other data collected to check for accuracy.
- Can provide a balance between methods.
- Using several methods is time consuming and expensive
- Researcher needs to be skilled in various research methods
- Positivist and interpretivist researches are based on very different ideas, so it may be difficult to combine them
Types of Data
- Tends to deal with numerical data
- Low involvement of researcher, e.g. in terms of time and face-to-face contact
- High number of people being researched
- Examples include: postal questionnaires, structured interviews, surveys, official statistics
- Tends to deal with descriptive data
- High participation by researcher in terms of time, face-to- face contact
- Low number of people being researched
- Examples include: in-depth interviews, non-participant observation, participant observation, case studies, personal documents, visual resources
Sources of Data
Primary sources – This is when new data is gathered by the researcher
This would include:
- participant observation,
- non-participant observation,
- case studies,
- structured and unstructured interviews,
- postal questionnaires
Secondary sources – This is when the researcher uses existing sources of information
This would include:
- official statistics
- non-personal documents,
- visual resources
Reliability and Validity – Reliability refers to whether the research can be repeated or replicated – is it consistent? Researchers can ensure that the research is reliable by controlling the variables including extraneous variables, using a standardised procedure, using scientific measures which produce quantitative data as this is easier to compare for consistency. Inter-rater reliability refers to when two or more researchers agree on what they see. This increases the reliability.
Validity refers to whether the study reflects the truth. Researchers can ensure validity by using a natural environment and tasks which means that participants are less likely to show demand characteristics and provide socially desirable responses. Ecological validity is a type of validity which refers to whether the setting is natural for example a setting which participants are familiar with will have ecological validity whereas a lab experiment which is unfamiliar will not have ecological validity.
- The most ethical observations are overt observations because the participants are aware that they’re being observed and can choose whether to take part. The least ethical observations a covert observation because participants are unaware that they’re being observed and so cannot provide consent
- The most reliable observation is a structured observation because the event is planned and staged which means there can be more control over the variables. The least reliable observation is a natural observation because this is in a natural setting which means it is harder to control the variables especially extraneous variables.
The most valid observations are natural observations and covert because it is a natural setting that the participants are familiar with and they are unaware that they’re being watched. This means that the observation has ecological validity because the participants are more likely to show truthful behaviour. The least valid observations are structured and overt observations because the participants are aware that they’re being observed and might show demand characteristics. Also, artificial settings will reduce the ecological validity of the observation.