Psychology deals with a broad range of areas - from the action of single hormones to factors that may contribute to genocide and the role of human relationships on cardiovascular health. The diversity of psychology means that many different methods are used to gather and analyze data. The methods researchers choose to carry out research are dependent on a number of factors, such as the purpose of the research, the characteristics of the participants, and the goals for generalizing or applying the findings.

It has been argued that qualitative research is not “scientific” because it is not built on the scientific method of the natural sciences. However, “science” as seen by some qualitative researchers should be defined in a broad sense as a systematic, rigorous, empirical task that must be carried out properly in order to produce trustworthy findings. There is an ongoing debate among qualitative researchers as to whether reliability (i.e. if the research methods and techniques used produce consistent data) and validity (i.e. that the research explains what it intended to) can be used in qualitative research. Some argue in favour and attempt to reformulate the concepts to fit better with the nature of qualitative research, while others claim that the very nature of qualitative research does not invite reliability and validity checks. Yet another position is introducing new concepts to describe validity and to ensure transparency and consistency of conclusions in qualitative research through alternative ways of checking results. The IB psychology program has adopted the latter position.

According to Ritchie and Lewis (2014), researchers can only come to understand behaviour through the participants’ interpretations. In experimental research, the researchers measure the effect of an IV on a DV.  The actual experience of the participants is often not seen as important data to gather in experimental research. Qualitative research, on the other hand, uses an interpretative approach which is based on understanding things from the view of the people involved.  The researcher should strive to be as objective and neutral as possible in the collection, interpretation, and presentation of qualitative data - and their research should be informed by the feedback from the participants. Ritchie and Lewis also argue that a combination of qualitative and quantitative methods may be necessary and helpful, but not all qualitative researchers agree with them.

Qualitative researchers are concerned with meaning - that is, they are interested in how people make sense of the world and how they experience events. They want to understand “what it is like” to experience particular conditions and how people deal with them. The objective of qualitative research is to describe and possibly explain events and experiences. This can lead to suggestions of how to overcome the problems that the research identifies. Qualitative researchers study people in their own environment, preferably in naturally occurring settings, such as schools, homes, hospitals, and streets.

Differences between qualitative and quantitative data

Psychological data come in many different forms, depending on the phenomenon the researcher is interested in. In quantitative research, the data are in the form of “numbers” that are easy to summarize and submit to statistical analysis. There are standard formats for data analysis, such as statistical tests. Quantitative data are meant for generalization beyond the sample from which the data were drawn.

Qualitative data are gathered through direct interaction with participants - for example, through one-to-one or group interviews, or by observations in the field. The data consist of text - for example, from transcripts or field notes. Textual data are open-ended and open for interpretation. The term used is “rich data”, which means that the research provides a depth of detail in their description of people, places, conversations, and so on. Because the data are rich, they are not easy to analyze, and there is no single way to approach analysis. Data may be analyzed with regard to existing theory or the data can generate theory.

Generalizing from qualitative research

Scientists normally want their findings to apply to other people than those who participate in a particular study. Generalizing findings from a study means that the results are relevant outside the context of the study itself. Some qualitative researchers do not find it relevant, but others argue that it is important that findings can be generalized. According to Lewis and Ritchie (2003), qualitative research could distinguish between the following forms of generalization:

• Representational generalization, which means that findings from qualitative research studies can be applied to the population from which the sample was drawn. A typical question could be if findings from interviews with IB students about how community service has affected their sense of internationalism. Representational generalization would question whether the findings could be applied to IB students in general.  Qualitative research normally involves small samples that are not selected to be statistically representative. This makes it difficult to generalize findings. However, if evidence from other studies confirms the findings, it is argued that generalization is possible.

• Inferential generalization, which means that the findings of the study can be applied to settings outside the setting of the study. This is also called “transferability."  When trying to determine the transferability of findings, the research context is important to consider. For example, if the study on IB students were carried out in a large national school in the UK, researchers would have to consider variables like the size and socioeconomic make-up of the school.  It may not be possible, for example, to generalize the results from a large, public school in Britain to an elite, private school in Moscow. In order to determine the transferability of the findings, more research would need to be done.

• Theoretical generalization, which means that theoretical concepts derived from the study can be used to develop further theory. The findings from a study might lead to inferences about how community service breaks down stereotyping or that a heightened awareness of environmentalism leads to the development of better communication skills. In that way, the findings from the study may contribute to wider social theory.

 Teacher only box

This is very much a TOK question. There are several ways in which students can discuss this question. 

1. It depends on what is being studied. Some behaviours may not be as open to influence as others - for example, the Stroop effect.

2.  It depends on how it is being studied.  Controls can be used to keep things objective - e.g. double-blind controls. However, in qualitative research, such controls are not common or usually highly effective.

3. This is one of the reasons that psychology has moved from using the term "participant" instead of "subject." The participant is supposed to know "as much as possible" about the study and has certain rights.

4. You may want to also have students consider whether this actually applies to animal research.

Ethical considerations in qualitative research

There are, overall, the same ethical issues involved in qualitative research as in quantitative research. These include informed consent, protection of participants from psychological or physical harm, respect for the participants’ privacy, and the right to withdraw from the research. However, there are special ethical considerations to be made due to the very nature of qualitative research. The characteristics of qualitative or field research usually include long-term and close personal contact with the participants, which may have implications for what the participants disclose to the researcher. It could also be the case that researchers become too personally involved in the problem under investigation and thus lose their objectivity.

Informed consent

The rule is that informed consent should always be obtained. This is stressed in all guidelines on ethical conduct in research. However, in some cases, where it would not otherwise be possible to study a phenomenon (e.g. use of violence in a street gang), ethics committees may allow exemption from the rule.

The participants should know that participation is voluntary. This is particularly important if the research is conducted by people who have some kind of relationship with members of the sample since participation could then be motivated by feelings of obligation. The researcher must provide the participants with sufficient information about the study, such as who funded the study, who will conduct the study, how the data will be used, and what the research requires of the participants - for example, in terms of time and the topics the study will address. It should also be made clear that consent can always be withdrawn. In cases where children aged under 16 years are involved, consent must be obtained from parents or legal guardians.

Protecting participants from harm

The researchers should take preventive action in all research, to avoid harming the participants. This is particularly true in sensitive research topics, such as sexual abuse, domestic violence, or alcoholism in the family. The researcher should not apply undue pressure on an individual to disclose something that they do not wish to share.  In addition, due to the nature of qualitative methods participants may disclose very private information that they have never shared with anyone before. This can happen because the interview situation seems like a friendly encounter, where the participant may feel comfortable and safe. After the interview, the participant may have second thoughts and wish to withdraw that information.  This must be respected by the researcher.

Prior to the interview, and before they agree to participate, the participants should have a clear understanding of the topics to be addressed. The researchers must approach sensitive issues through clear and direct questions so that participants are not drawn into irrelevant and sensitive details by mistake. If the participants show signs of discomfort, the researcher should be empathetic and consider stopping the interview. If the interview has dealt with emotional and sensitive issues, the researcher should try to return to less sensitive topics towards the end. It is not advised that the researcher should provide advice or counsel the participant, but he or she might provide useful information about where to find help if this is necessary.

Anonymity and confidentiality

Confidentiality means that research data will not be known to anyone outside the study. The researcher may have to change minor details in the report to avoid the possibility of participants being recognized. Anonymity relates to the way data are stored after the research. If interviews or observations have been videotaped and archived, it can be difficult to guarantee total anonymity, so these should be destroyed when transcripts have been made. If the researcher finds it necessary to archive non-anonymized data, the participant should give written informed consent.

The participants should be informed about the issues surrounding anonymity and confidentiality. The identity of the participants should not be known outside the research team, but in cases where sampling has involved a third party (e.g. managers, friends, teachers), this cannot be done, and in this case, the participants should be informed.  Focus groups are problematic in that it is difficult to guarantee confidentiality as the members of the focus group may - either intentionally or unintentionally - reveal information that was shared in confidentiality in the study.

 Teacher only box

For this question, students should consider generating a set of criteria for determining what a "poor ethical study" would look like.  And then they should be able to weigh the ethical concern of the three studies. Have students generate the criteria and then attempt to apply them in groups.  Then, looking at the result, how well do they think that their criteria worked?  What would then add or take out of their criteria?

Some potential criteria include, but are not limited to:

• Does the study lead to undue stress or potential harm to the individuals?
• What is the value of the research? Is there some benefit to carrying out the research?
• To what extent is the study potentially pressuring individuals to disclose information that they do not want to disclose?
• To what extent can the anonymity of the individual be preserved?
• Is this a question that will lead to an improvement of society and/or the lives of the participants?

Sampling techniques appropriate to qualitative research

Sampling methods in qualitative research differ from those used in quantitative research, where the purpose of sampling is to generate samples that are representative through random selection of participants. Qualitative research methods do not use samples that are randomly selected from a target population.

The samples used in qualitative research are much smaller than in quantitative research. The sample is not intended to be statistically representative.  A sample in qualitative research is chosen because it represents important characteristics of a population, and it is these characteristics that are the main concern in the evaluation of research.

The researcher can use a number of sampling procedures - for example, purposive sampling, opportunity sampling, and snowball sampling. All the sampling methods presented here rely on prescribed selection criteria. They all use small samples, but the sample can be supplemented during the research process.

Purposive sampling

Purposive sampling targets a particular group of people. The researcher decides early on which criteria should be used for sampling. These are based on the aim of the study and existing knowledge in the field. The participants are chosen on the basis of particular characteristics that will help the researcher to explore the research topic. This could refer to socio-economic characteristics (e.g. living below the poverty line), specific experiences (e.g. childhood abuse), occupation (e.g. nurse), or social roles (e.g. mother). All the features of relevance must be present in the sample. It may also be important that there is diversity within the sample - for example, in relation to age and gender.

There are different ways in which a purposive sample may be obtained.  In some cases, it may be a voluntary sample.  If a researcher wanted to study foreigners who work in the Czech Republic, then an ad in a local magazine may help to create the sample. If time is an issue, or if the researcher wanted to study foreigners who work for international IT companies in the Czech Republic, then a sample of opportunity might be used - that is, the purposive sample may be found at the Microsoft offices in Prague. Finally, in the case where the sample may be difficult to find - for example, foreigners who have been working and living illegally in the Czech Republic for at least ten years - then participants may recommend other people that they know who share this trait, to participate in the study.  This is a snowball sample. 

A problem with purposive sampling is that it often suffers from sampling bias - that is, the choice of the trait to investigate leaves important members of the population out of the research.  For example, if a researcher is carrying out research on women who left their husbands because they had a drug problem, the bias of the sample is that we don't have data from women with husbands who have a drug problem who haven't chosen to leave him. However, even though purposive sampling involves deliberate choices, it is not necessarily biased. It is believed that if the sampling process is based on objective criteria, and these are clearly documented and explained, the bias is limited.

Snowball sampling

This is a way of sampling which is used to get hold of participants without investing a lot of time and energy. The researcher simply asks participants in the study if they know other potential participants.

Snowball sampling can be used to locate people with characteristics that are rare or hard to find by having existing participants refer new participants. If a researcher is studying homeless people, it is not possible to have a list of all the homeless people in an area, but if the researcher finds one or two, they may help to locate others.

Snowball sampling also helps to build trust between the researcher and the participants. If a researcher is carrying out interviews with drug users, it may be difficult to get them to trust him and disclose information relevant to the study.  However, if a friend of the drug user recommends the study and feels that the researcher can be trusted, this helps to build a rapport more quickly with other participants.

The advantage of snowball sampling is that it is cost-efficient because the researcher does not have to use expensive and time-consuming techniques to locate participants. The major limitation of the method is that it is very difficult to avoid bias in the sample. There are also confidentiality concerns because the participants know the identity of other participants.

Opportunity sampling

Opportunity sampling - also known as convenience sampling -  selects a particular group of people who happen to be available. They are simply asked if they would like to participate. The researchers may want to study the effectiveness of a program for the prevention of antisocial behaviour in a youth group, using basketball training and coaching. The researchers follow the adolescents for a year, to see the impact the program has on young people. The sample represents the young people who are in the program but could also represent similar young people in similar programs around the country. In order to see if this is the case, however, similar research must confirm it.

Effects of participant expectations and researcher bias

One of the more important factors to consider in psychological research is that humans are not responding passively to research. Researchers need to be aware that research is an active process that requires reflection and interpretation of the data, the participant, and the research context, in order to avoid participant expectations or researcher bias. Participant expectations can be explained as the participants’ ideas about the research and the researcher which can affect the credibility of the data. If the participant feels they have to behave in certain ways in order to please the researcher, this will affect the value of the data in a negative way.

Researcher bias happens when the researcher’s own beliefs influence the research process.  For example, a researcher may "see what he wants to see" when carrying out an observation. This is known as confirmation bias.  This is why researchers often train teams to carry out the research, rather than carrying out the research on his or her own.  In addition, often double-blind studies are used. In this case, the researcher does not know which data is being presented for interpretation.  For example, in a case where researchers are carrying out a study of the effect of a self-esteem program in schools to help students develop perseverance when confronted with challenging tasks, the researcher may receive videos of the observations of students trying to solve a very difficult puzzle.  However, the researcher who is watching the video will not know whether the student he is watching has actually taken the course or not. This is done to avoid confirmation bias.

If it is assumed that “reality” in a research study is multiple and co-constructed, then we should be aware that participants who are asked to comment on the researcher’s interpretation of the data will not necessarily arrive at the same conclusions as the researcher. Some would argue that a “credibility check” could impose an artificial conformity on the analysis of the data. This would impact the meaningfulness of the findings. However, being aware of sources of bias is important. One way to minimize the effect of participant expectations and researcher bias is reflexivity throughout the research process.

Credibility in qualitative research

Credibility corresponds roughly to the concept of internal validity that is used in quantitative research. The credibility of research is established when the findings of the research reflect the meanings as they are described by the participants. This can be established by the researcher giving the interpretation of an interview to the participant to read. If the participant feels that the interpretation reflects what they said in the interview, then the research has credibility.

Qualitative researchers could also check whether their findings are credible by referring to other researchers' interpretation of the data.  In the case of an interview, this means having other researchers also carry out an analysis of the interview.  If the interpretations are similar, then we can conclude that researcher bias may not play a significant role in the findings.

Triangulation is a kind of cross-checking of information and conclusions in research. It involves the use of different perspectives, methods, researchers, and sources of data to check if the interpretation of data can be supported.  There are different forms of triangulation.

• Data triangulation is when a researcher collects data from different sources - hence, sometimes called source triangulation. By using this strategy, you can confirm the findings of one set of data with another set of data.  For example, I interview a student about the way she interacts with others.  I then interview both her English teacher and her father.  This gives me different sources of information that I can then compare and confirm the statements made by the initial interview.
• Method triangulation involves comparing data that come from the use of different methods. This could involve qualitative and quantitative methods. By using method triangulation, the researcher determines whether it was simply the choice of research method that led to the findings - or can we draw the same conclusions regardless of the research method?
• Researcher triangulation involves the use of several observers, interviewers, or researchers to compare and check data collection and interpretation.
• Theory triangulation involves looking at the data using different theoretical approaches - for example, a biological, psychological, and sociocultural approach.

According to some researchers, it is important in qualitative research to establish a set of strategies, which can increase confidence that research findings actually represent the meanings presented by the participant—that is, increase credibility. Triangulation can be one method used to do this.

Checking for understanding

HL: Understanding observations