Compare and contrast the use of focus groups with surveys.  Your answer should include a discussion of issues related to epistemology and ontology as well as validity, reliability and generalizability.

To illustrate your argument draw on examples of research about poverty or health.

Introduction

In order to conduct valuable social research, we need to balance the strengths and weaknesses of the methodologies available to us. We need to consider what is the question that we want to answer and how we want to answer it. What is our purpose and what methodology makes more sense to achieve it. How do we justify using one method over the other? What is our reasoning behind it? These are some of the questions we have to answer before we start involving ourselves in research. They would help us choose between qualitative research, quantitative research or a mixture of both. We also have to concern ourselves with more specific questions: Is it easier to generalize with one method than the other and why is it important for it to be generalizable? Are our indicators valid and reliable?

The objective of this essay is to compare and contrast surveys with focus groups, highlighting their weaknesses, strengths and principal differences. This isn’t to say that one method is better than the other but rather to detail their characteristics in order to simplify making a decision. In order to accomplish this, we will go from the general to the specific. First, comparing and contrasting qualitative and quantitative methods in terms of ontology, epistemology, theoretical perspectives and rigor; second, we’ll proceed to define, compare and contrast the two methods we already mentioned employing the concepts discussed in the first section; subsequently, we’ll illustrate the differences with an example of sexual health research in adolescents; and finally, we’ll share some last thoughts about this discussion.

Qualitative vs. quantitative research

Both qualitative and quantitative research try to understand the real world, the difference lies in the methods they use to grasp it. Quantitative research tries to codify and quantify the social world and qualitative research tries to describe it focusing more on details and description than in numerical indicators. In order to decide which way we want to study the social world, we have to know which lenses are at our disposal. Which is why we’ll dedicate this section to examining ontology, epistemology and multiple theoretical perspectives and how they can help us justify our employed method of research.

Crotty (1998) describes Ontology as ‘the study of being’. Is there a reality that’s independent of our interpretation? Does knowledge exist even if we’re not conscious of it? These are the questions that concern ontology (Ritchie and Lewis 2003). Furthermore, a realist would answer ‘yes’ to the first question, there is a world outside of our description of it. On the contrary, idealism argues we can only perceive reality through our social constructions (Ritchie and Lewis 2003). Additionally, epistemology can be defined as the way we know what we know and how we explain how we know it (Crotty 1998). For Bryman (2016) it is an issue of what can be considered acceptable knowledge.

In this respect, quantitative research is usually considered more focused in the ontological position of objectivism. This means that researchers are only discovering a truth that was already there, the discovery is not subject to interpretation or influenced by the researcher’s pre-conceived beliefs. There exists a separation between facts and the interpretation of those facts. Positivism, an epistemological view, is directly related to this idea. Positivism is based on a scientific approach assuming that reality is objective; resulting in a hypothesis that can be tested, measured and proven correct or not. These hypotheses are based on empirical findings instead of speculations; hence, the results can be considered verifiable knowledge. This idea of objective truth has been criticized in three aspects, 1) there will always be some influence from personal feelings and experiences (Babbie 2013), 2) there is no way to prove a theory by using a sample, since if there’s only one observation where the theory isn’t true, the hypothesis would be regarded as false (Gray 2004) and 3) this paradigm doesn’t always reflect our irregular and uncertain world (Crotty 1998).

For qualitative research, the ontological view more significant is constructionism, which Crotty (1998) puts simply as ‘there is no meaning without a mind’, which translates in people constructing different explanations for the same question. The subject starts creating knowledge as he interacts with the world (Gray 2004), in other words, he is not discovering knowledge but constructing it. As positivism is linked to objectivism, the theoretical perspective linked to the epistemological constructivism is interpretivism. This approach, contrary to positivism, considers impossible to conduct research that isn’t based on cultural and historical interpretations. It criticizes the use of natural sciences for social research, which should be focused on the explanation and interpretation of human actions (Bryman 2016).

Furthermore, we need concepts that help us evaluate the measures we employ to test our theories, to be able to confirm we’re uncovering the objective truth. This is where validity and reliability become relevant. Validity, in general, refers to how trustworthy is your research and if the measure you’re using really measures the concept (Bryman 2016), this can be divided between internal and external validity. You can affirm a study has internal validity when there’s evidence of causality, which means you can assure that changes in your dependent variable are caused by changes in your independent variable; while external validity refers to how generalizable the intervention is (Sullivan 2010). Quantitative methods are considered a better instrument for both external and internal validity as this definition focuses on measurement of variables. To respond to this, qualitative researchers (Lincoln and Guba 1985; Bryman 2016) evaluate credibility instead of internal validity. This criterion focuses on how acceptable or credible are the researcher’s findings to the group he studied. It takes into account that there isn’t a reality independent of our interpretation but rather different constructions of social reality, thus the researcher tries to confirm if he captured that social world correctly.

The sampling methods in the two methodologies that are being compared are the reason they differ in their success at external validity. Qualitative research uses small samples that are selected because of their uniqueness while quantitative involves a random sample that is statistically representative. Big sample sizes are important considering that when the variable has a normal distribution, the bigger the size of the sample the closer the mean will be to the population’s mean (Gilbert and Stoneman 2016). This allows the assumption that the same results would be obtained if the intervention were to be replicated in a population that shares the same characteristics as the sample studied. Qualitative data focuses in exceptional cases and representative samples aren’t common, which can result in a more detailed description of the situation but also makes it harder to generalise outside of the sample and jeopardizes the external validity of the findings (Gilbert and Stoneman 2016). This is not to say that quantitative research is always generalizable, as the data can be collected from non-random samples, but rather than it is more generalizable than qualitative. As was the case with credibility, the term of external validity has been adapted to fit qualitative methodology using the term transferability (Lincoln and Guba 1985). Researchers should detail their database deeply to make it easier for others to decide if the experiment can be transferred to other contexts.

Equally important, the methodology is reliable when you obtain the same results each time you repeat an experiment (Babbie 2013), reliability is also concerned with consistency and stability. Given this definition, subjectivity can be a problem and therefore is an issue for qualitative research, considering that the attitude or pre-conceived beliefs of the interviewer might influence the answers. Identically, it can contradict constructivism as explained by Ritchie (2003): ‘The ‘constructivist’ school, for example, argue that there is no single reality to be captured in the first place so replication is an artificial goal to pursue’. The parallel concept would be ‘dependability’, where an auditor examines the process of the inquiry (Lincoln and Guba 1985). It is important to explain that there can be reliability without validity but the contrary isn’t possible. You can have an instrument that measures something consistently but doesn’t measure what it was supposed to. But if the instrument is inconsistent then it can’t be measuring what it is supposed to (Black 1999).

We agree with the authors (Morse et al. 2002) that argue that validity and reliability shouldn’t be substituted in qualitative research by Guba and Lincoln’s (1985) trustworthiness (credibility, transferability, adaptability and confirmability). Every research independently of the methodology being used should be held to rigour standards and, as Morse et al. explain, because trustworthiness is checked after the study is completed there’s a risk that it would be too late to correct the detected problems.

In summary, we can describe quantitative research as more inclined to realism, positivism and objectivism. In turn, this makes it more adequate when your goal is to describe variance in a population (Gilbert and Stoneman 2016), generalise, and explain causal effects rather than describe differences between groups in a population. Qualitative research, contradictorily, doesn’t believe there is only one truth and rejects the use of natural sciences in social research. Accordingly, its philosophical approach to research questions differs from the former as well as its capacity to demonstrate validity and reliability. We will explore all these concepts more deeply for two specific methodologies in the next section.

Focus groups vs. surveys

In this section, we’ll compare and contrast focus groups and surveys. We’ll proceed to define both of them and then we’ll describe and analyse their differences in social research. A focus group is a small group of between six and ten people who express their view about a particular topic that has been tightly defined by the researcher, who acts as a moderator (Gilbert and Stoneman 2016). Surveys, on the other hand, use questionnaires to collect information that then will be codified. Surveys can be asked by the interviewer or they can be self-administered either by post or online (Babbie 2013).

First, let’s compare internal validity. In surveys, it is easier to determine the internal validity of an indicator. You can control the environment and the rest of the variables collected to test out the causality between your independent and dependent variables. As a matter of fact, the codification of the variables and the representative sample allows for statistical tests of causality (e.g. regressions). But the validity of the findings might not be as expected if the meaning of the question isn’t clear. This could be the case if the list of answers is biased (e.g. unequal number of positive and negative options) or there are two questions in one. Contrastingly, focus groups don’t have a clear determinant of causality, even if they get the same answers or explanations from the participants in the groups they can’t guarantee that the correlation wasn’t brought about by exogenous factors. As mentioned before, we can comment on the credibility of the findings rather than the validity. Unlike in interviews, participants in focus groups can be challenged by other individuals and this coupled with the permissive environment and less directive role taken by the researcher results in a more realistic account of the participant’s thoughts (Krueger 2014; Bryman 2016){Bryman, 2016, Social Research Methods;Krueger, , Focus groups : a practical guide for applied research;Krueger, , Focus groups : a practical guide for applied research}.

Second, we’ll discuss the external validity of the methods. Focus groups consist of a homogeneous group that is selected because they have something in common. They require a balance where the group is homogeneous enough so everyone feels comfortable to speak but with some heterogeneity to collect diverse opinions (Acocella 2012). Some authors (Greenbaum 1998; Krueger 2014) don’t consider this a problem as they argue that focus groups are not intended to be generalized in the first place, as the sample isn’t selected randomly. In contrast, surveys are mostly conducted with the purpose of generalizing and are conducted in large samples of people that are randomly selected. One threat to the ability to infer from this sample are non-respondents which contribute to non-sampling error (Sullivan 2010) as the sample might no longer be an adequate representation of the population. This happens when some people fail to answer some or all the questions in the survey; and can be induced by surveys that are too long, too complicated or when items aren’t arranged in a logical manner.

In addition, we will discuss reliability in the two methods we’re examining. Focus groups can encounter multiple problems that will result in biased answers and thus won’t reflect the participant’s truthful opinion. This includes the risk that a person will change its opinion after listening to the rest of the group talk, especially if most of the group seem to agree on the topic (Bryman 2016); shy individuals that aren’t open to sharing their feelings; a group where participants have different backgrounds or social status might make some participants afraid of expressing themselves; while people with a more dominant status take control of the discussion and intervene more frequently; and participants that will make-up answers to portray themselves as rational and give socially-desired responses (Krueger 2014). Additionally, the transcription of the recording might be difficult, with people talking over each other, participants that don’t speak loud enough and participants that have similar voices. Moreover, the opinions that will endure are the ones that are agreed on by most of the group, not capturing the thoughts of the minority. Finally, different interviewers in focus groups might ask questions or intervene in different ways, which can result in biased answers. This is a risk that is eliminated in surveys as everyone receives the same questions.

But this isn’t a problem exclusive of qualitative research. For example, a survey where the questions aren’t crosschecked or aren’t too clear might confuse the person interviewed, resulting in an unreliable answer. If a respondent doesn’t understand a question there is no one to clarify what it means in the case of self-administered surveys. Focus groups interviewers have the opportunity to ask participants to elaborate an answer or explain better what they meant (Bryman 2016). Following-up is one way of assuring the reliability of the answers obtained; we will expect to receive the same answers for the same group of people if we repeat the experiment. Because surveys predominantly consist of close-ended questions, we would expect more reliability in their findings. On the contrary, focus groups incorporate opinions that could have been influenced by the person’s situation that day and that could change if asked with a different group of people.

Another contrasting point is convenience. Both methods can be conducted online but they present some differences. Research conducted online has become popular for the following reasons:

• Participants can answer them whenever they have the time (Bryman 2016);
• There are no geographical constraints;
• They can be more inexpensive for the research group because they don’t have to spend as much time and they don’t have to incur in the cost of travel;
• They are easier for some participants as they don’t have to spend time sending their answers by post or transporting to focus groups or interviews meeting points;
• They can feel more comfortable answering sensitive or personal questions. They won’t feel intimidated or ashamed by sharing their responses in front of an interviewer or others Respondents.

For focus groups there are two types of online methods: Real-time or synchronous and asynchronous (Bryman 2016). But they can result in an environment that’s harder to control. The moderator can’t be available all day and if the discussion goes off on a tangent, it can be burdensome for the researcher to bring it back to the topic. This isn’t a problem for online surveys because there isn’t the need for constant intervention from the researcher or interviewer. Additionally, they can send all the surveys at one and you can automate data entry, eliminating the possibility of human errors. This can make them quicker for the researcher to administer and analyse than focus groups. In the negative side, probability sampling is more difficult when conducting online surveys as the respondents must have to be selected by a common interest (Bryman 2016). Besides, in both post and self-administered surveys, there’s the chance that the person it was intended for isn’t the one answering, this will be a threat to the validity of the data (Gray 2004). Moreover, the sample can be biased as people with access to the Internet can have a higher social status and education than their counterparts; and you might be unable to capture everyone, as illiterate people won’t be able to answer.

Finally, there’s an important difference in the ‘deepness’ of the findings. Surveys need to simplify their questions so they can be appropriate for all the participants but they will be omitting some of the participant’s opinions. Babbie (2013) suggests that for complex topics, surveys can be too superficial. In this sense, qualitative data can be richer. The researcher can try to understand how and why there are differences in the studied group; participants have the chance to refine their views, build on what others are saying and deep their own and others commentary (Ritchie and Lewis 2003). There is a trade-off between the unambiguous, precise and robust findings of the surveys and the deep understanding of a topic that results from focus groups.

Surveys vs. focus groups: Sexual health

In order to illustrate what has been discussed above, we’ll be using research concerning sexual health in adolescents in focus groups and surveys. All the studies were carried out with the purpose of identifying the reasons for the increase of sexually transmitted infections (STI) and teen pregnancy. For focus groups, we selected two research articles on the topic (Roberts et al. 2005; Bay-Cheng et al. 2011). In one of the articles, the groups consisted of 43 girls from 14 to 17 years that formed seven focus groups. In the second article, there were four focus groups, eight females and eight males from 15 to 17. Both works selected agreed that because it was a sensitive topic, focus groups were ideal because of the less directive role of the researcher and the similarity between the participants made it easier to share their opinions and feel mutual support. None of the researchers used a random sampling selection technique; one advertised in a local newspaper while the other selected two schools and then selected students based on their communications skills, listening skills and interest in health-promotion activities. This translates in the studies not being generalizable as they can only represent the sample of people interviewed, especially in the second case where the participants all came from the same two schools. This also represents biased results, which harms its external validity. In the second article, the schools chosen were centrally located and had strong academic achievements. It could be argued that they might be more open-minded and knowledgeable and consequently might not represent other social classes or students that don’t have access to a better education. The authors of both studies also recognize that the participants might be embarrassed to be honest about their experiences and opinions and might feel pressured to give socially desired answers; even when they’re assured of the confidentiality of their answers, they still have to communicate them in front of strangers of their peers.

For the survey example, we chose an article that designed a sexual education intervention for eight graders to analyse its impact on sexual health (Grose et al. 2013). They administered a survey before and after the intervention. The total of students in that grade was 148 but only 95 (64%) had their parents consent to participate, the intervention took place in a school in ‘an economically and ethnically diverse’ area. Participants were between the ages of 12 and 15. A correlation analysis was used to examine the relationship between the variables collected (gender ideology and sexual knowledge and contraceptive beliefs) for boys and girls separately. Similarly to our previous example, this research couldn’t be generalized as the sample wasn’t randomly selected and isn’t representative of the population. The sample is biased seeing that only students with their parent’s consent were allowed to take part. Nonetheless, since the students were interviewed before and after the intervention, changes resulting from the second survey can be regarded as occasioned by this intervention. This proves causality and therefore is an indicator of internal validity.

There are two important points we’ll like to call attention to in the examples chosen. First, quantitative methods aren’t inherently more rigorous than qualitative; we can have articles, as the one mentioned, where the findings are not generalizable. It is worth emphasizing that the findings from this research can still be relevant and can still make an impact even if it is just helping us understand the problem a little better. Second, while the studies were very similar in their goal, the findings present some differences. As we mentioned before, surveys helped us establish the relationship between the variables while focus groups can highlight the interaction between the group and how they construct meaning collaboratively.

Conclusion

After deciding your research question, one of the most critical decisions will be to choose which method will help answer that question better. This essay depicted some of the main characteristics of quantitative and qualitative methodology as well as surveys and focus groups specifically. We can conclude that if we want to answer our question with an underlying philosophical approach that points to a reality with different interpretations, qualitative methods are more appropriate. Accordingly, focus groups capture a realistic view of participant’s interpretation of a topic, as well as their joint construction of answers. We obtain this complex analysis in exchange for results that can’t be generalised and that lack validity and reliability. On the other hand, with surveys we can obtain more objective results funded in hard sciences, but that might only show a simplistic and unrealistic view of the social world. And although surveys can be considered more rigorous, it isn’t necessarily the case; they’re still at risk of not accomplishing consistent or accurate indicators.

Finally, it is important to recognize that the research methods could complement each other; we could get a representative sample and initially conduct a survey to test the theory; and afterwards select participants for focus groups that will help us ‘to see the issue through the eyes and hearts of the target audience’ (Krueger 2014).

References

Acocella, I. (2012) ‘The focus groups in social research: advantages and disadvantages’, Quality & Quantity, 46(4), 1125-1136, available: http://dx.doi.org/10.1007/s11135-011-9600-4.

Babbie, E.R. (2013) The practice of social research Earl Babbie,13th ed.. ed., Belmont, Calif.:Belmont, Calif. : Wadsworth Cengage Learning.

Bay-Cheng, L.Y., Livingston, J.A. and Fava, N.M. (2011) ‘Adolescent Girls’ Assessment and Management of Sexual Risks: Insights from Focus Group Research’, Youth & Society, 43(3), 1167-1193, available: http://dx.doi.org/10.1177/0044118X10384475.

Black, T.R. (1999) Doing quantitative research in the social sciences : an integrated approach to research design, measurement and statistics / Thomas R. Black, London:London : Sage.

Bryman, A. (2016) Social Research Methods,5th edition ed., Oxford University Press.

Crotty, M. (1998) The foundations of social research : meaning and perspective in the research process / Michael Crotty, London:London : Sage.

Gilbert, G.N. and Stoneman, P. (2016) Researching social life,Fourth edition / ed.

Gray, D.E. (2004) Doing research in the real world / David E. Gray, London:London : Sage.

Greenbaum, T.L. (1998) The handbook for focus group research / Thomas L. Greenbaum,2nd ed.. ed., Thousand Oaks ; London ; New Delhi:Thousand Oaks ; London ; New Delhi : Sage.

Grose, R.G., Grabe, S. and Kohfeldt, D. (2013) ‘Sexual Education, Gender Ideology, and Youth Sexual Empowerment’, Journal of Sex Research, 51(7), 1-12, available: http://dx.doi.org/10.1080/00224499.2013.809511.

Krueger, R.A. (2014) Focus groups : a practical guide for applied research / Richard A. Krueger, University of Minnesota, Professor Emeritus; Mary Anne Casey, Consultant,5th edition.. ed., Los Angeles : SAGE.

Lincoln, Y.S. and Guba, E.G. (1985) Naturalistic inquiry / Yvonna S. Lincoln, Egon G. Guba, Beverly Hills, Calif. ; London:Beverly Hills, Calif. ; London : Sage Publications.

Morse, J.M., Barrett, M., Mayan, M., Olson, K. and Spiers, J. (2002) ‘Verification Strategies for Establishing Reliability and Validity in Qualitative Research’, International Journal of Qualitative Methods, 1(2), 13-22, available: http://dx.doi.org/10.1177/160940690200100202.

Ritchie, J. and Lewis, J. (2003) Qualitative research practice : a guide for social science students and researchers / edited by Jane Ritchie and Jane Lewis, London:London : Sage.

Roberts, A.B., Oyun, C., Batnasan, E. and Laing, L. (2005) ‘Exploring the social and cultural context of sexual health for young people in Mongolia: implications for health promotion’, Social Science & Medicine, 60(7), 1487-1498, available: http://dx.doi.org/10.1016/j.socscimed.2004.08.012.

Sullivan, E. (2010) Research methods for public administrators / Elizabethann O’Sullivan, Gary R. Rassel, Maureen Berner,5th ed.. ed., New York:New York : Pearson Longman.