MULTIPLE-SAMPLE STUDIES
INTRODUCTION
Vast majority of investigations employ only one sample, there are some that employ two or more. In general more than one sample serves to find out changes over time and facilitate comparisons as well as establishing causality, triangulation and hypothesis testing. Within these parameters, there are at least five types of study that deserve to be mentioned. These are
· experiments, which study two or more samples, in most cases simultaneously
· panel studies, which analyze the same sample at different points in time
· cross-sectional studies, which investigate different samples simultaneously
· trend studies, which study different samples at different points in time
· focus groups, which examine several natural or constructed groups
1 EXPERIMENTS
Experiments are a part of everyday life. Since the dawn of humanity people have experimented in many ways to ascertain the best ways of coping with everyday problems and to improve the quality of their life in general. For example cooks, husbands and wives, children, teachers and businessmen. This everyday practice was adopted and improved by scientists and became the standard method of discovering knowledge. Experimentation is more common in psychology than in other social sciences, it is currently being used by researchers of diverse affiliations and background. Experiments are almost exclusively used within a quantitative model. Qualitative experiments are practically unknown, and little has been written about them
1.2 Strengths and weaknesses of experiments
Strengths:
· Replication: Experiments are constructed in a manner that allows replication, assuming that repeating the procedure will not lead to different results.
· Prediction: Their structure and process, and their detailed and rigorous (precise) design allow reasonable predictions to a higher degree than many other methods.
· Causality: Experiments possess all the methodological parameters required for establishing causality.
· Precision: Rigorous (precise) planning and checking of the status of variables, and the validity and reliability of methods allow a high degree of precision in all steps of the research process.
· Convenience: The size of the samples, the detailed and accurate preparation of the experimental conditions, and the detailed research design make experiments a most convenient research method.
Weaknesses:
· Control: Depending on the circumstances, experiments may not allow the degree of control that is required to exclude unwanted or unintended influences outside the independent variables.
· Representativeness: Samples are usually too small to produce representative findings.
· Process: The research process is too technical and too artificial to allow generalizations.
· Ethics: There are cases in which conditions dictate that ethical standards are of secondary importance.
1.3 The nature of experiments
Experiments are used to find out the presence, type and degree of a causal relationship between two variables. An association is considered to be causal if it entails the criteria listed in the following Box.
| Minimum properties of causation An association is 'causal' if it is statistically proven to entail: · Correlation: A strong association between the variables. · Consistency: The association between cause and effect must be consistent. · Time order: The cause precedes (come first) the effect, and not vice versa, · Contiguity: Cause and effect must be contiguous (closest), that is close together in both time and space. · Exclusion of spuriousness: The relationship must not be spurious (fake); in other words, the effect is not the result of extraneous (irrelevant) intervening variables but of the causal variable alone. · Rationale: There must be a rationale that explains and justifies causality. |
| How do experiments work? The experimenter: · establishes and controls the experimental conditions (closure) · measures the dependent variable · introduces the independent variable (stimulus; cause) · tests the dependent variable · examines the dependent variable for changes (effects) |
Central to experimental studies is the systematic procedure employed to ensure that the relationship between independent and dependent variables remains intact and free of distortions. It is a task of the experimenter to
· maximize the effects of the independent variable on the dependent variable
· control outside distorting influences, and
· minimize the effects of non-systematic (chance) variables
Experiments adhere to strict principles of quantitative methodology, perhaps more strict than in other contexts. More obvious are these principles when referring to:
· the way in which variables are measured (must be consistent and methodologically sound)
· conditions under which variables are measured (to be held constant, especially through closure)
· the replicability of the study
· the comparability of the findings
· the expectation that repetition under controlled conditions is a basis for testing causality
1.4 Steps in experimental research
| The experimental research process Step 1 Selection of the topic and methodology. Asks: What will be studied? Which methodology will be used? Step 2 Methodological construction of the topic. Asks: What is the detailed definition of the topic? What are the variables? How will they be operationalized? What is the hypothesis? Step 3 Sampling and methods. Asks: How many groups will be constructed and tested? How will the respondents be chosen? Step 4 Data collection. Asks: What methods will be employed to collect the data? How will the methods be conducted? How will closure and other principles be adhered to? Is there a need for a laboratory or for assistants? Are computers or other equipment needed? How will they be procured? How will the data be collected? How will extraneous influences be controlled? Are issues of ethics being observed? Step 5 Analysis and interpretation. How will data be analyzed and interpreted? |
1.5 Experimental sampling
Characteristics of the sample
Selection of the subjects in an experimental design is usually undertaken by means of the probability sampling procedures. The experimental sample demonstrates the following characteristics:
· It includes two study groups: the experimental group and the control group. The experimental group is the one that will be exposed to the independent variable, namely to the stimulus; the control group is the group that is not subjected to the independent variable.
· The two groups must be checked for any systematic differences.
· The subjects in the experimental and control groups must be the same or similar with regard to the research topic.
· The selection of the subjects should be free from sampling bias.
Methods of sampling
There are at least three methods of sample selection in experiments depending on many factors, including theoretical/methodological considerations, experimental conditions, data required, nature of the research question and nature and type of subjects.
· Randomization: In this method, two subjects are selected randomly from a sampling frame, and one is placed in the experimental and the other in the control group. This process is repeated until sufficient subjects are selected.
· Subject matching: From the sampling frame, pairs of adequately matched subjects are chosen, and one is placed in the experimental group and the other in the control group. Again, this continues until the planned number of subjects is reached.
· Group matching: In this method, pairs of matched groups – that is, groups with subjects of the same or similar characteristics (e.g. average age, education, status, achievement or performance) – are chosen, and one group is used as the experimental group and the other as the control group.
1.6 Arrangement of experimental conditions
Experimenter should ensure that factors other than the independent variable do not affect the dependent variable. Where control of these factors is not possible, the experimenter should be in a position to assess the influence of such factors and to consider their effects when interpreting the findings.
‘Arrangement' usually refers to
· group composition
· the process of selection
· the structure of the setting in which the experiment is to be performed
· time of the test
· nature of the stimulus (independent variable) and its meaning for the subjects
· time between pre-test and post-test
· opportunity for manipulation or conditioning during the time between pre-test and post-test
· strict supervision
· instructions given during the experiment
1.7 The process of experimentation
Slop 1 Pre-test: This step is called pre-test because it takes place before the subjects are exposed to the experimental treatment. The pre-test includes measurement of the dependent variable, usually of both the experimental group and the control group. This measurement is taken as a basis for assessing the eventual changes of the dependent variable during the experiment.
Step 2 The test: During this step the subjects are exposed to the independent variable. This may entail taking part in a discussion, reading a book, taking a certain medicine, consuming an amount of alcohol, watching a film, attending a lecture or visiting a club. The nature and duration of the stimulus varies with the nature of the experiment and the independent variable. In some cases it might take only a few minutes, in other cases significantly longer.
Step 3 Post-test: After the experimental treatment has been completed, the subjects are tested again. This is referred to as post-test because it takes place after the experimental treatment. During this test the dependent variable is measured again in order to ascertain its present status. Following this, the results of the pre-test and post-test are compared and any variations in the results are recorded.
1.8 Analysis and interpretation of experimental data
The analysis and interpretation will concentrate on data collected through the pre-test and post-test. These data will be analyzed and compared and the differences computed, most likely by using statistical methods; an attempt will be made to establish a causal relationship between the variables in question.
1.9 Types of experiments in quantitative research
The following three types are the most common forms of experiments in the social sciences.
· Laboratory experiments: These are conducted in a laboratory, where all external factors can be controlled.
· Field experiments: These are performed in natural situations, such as city blocks, bars, meeting places of migrants, villages, churches and classrooms
· Demonstration experiments: Demonstration experiments may be field or laboratory experiments but with one group (the experimental group) only. Although they help to 'demonstrate', highlight or illustrate trends in or aspects of human behaviour, they are not true experiments because, first they do not contain a control group, and therefore no comparisons can be made, second, they do nor select subjects randomly, and third, there is no clear timing for the experimental treatment, since it can continue for as long as the subjects remain in the treatment.
Experimental designs
Social scientists use several types of experimental design. The differences between them lie mainly in the number of experimental and control groups employed in each experiment. The most common designs are listed below.
· Before-after design: This employs only one group, the experimental group, which is pre-tested, exposed to the stimulus and post-tested. The differences between the pre-test and post-test scores indicate the possible effects of the independent variable on the dependent variable.
· Classical experimental design: Here, two groups are employed, the experimental and the control group. Although both are adequately prepared for the experiment, pre-tested and post-tested, only the experimental group is exposed to the independent variable. This is the most common form of experiment.
· After-only experiment: This is identical to the classical experimental design expect that neither the experimental group nor the control group is pre-tested.
· Solomon two-control-group design: This design employs one experimental group and two control groups. In principle, one control group and the experimental group are treated as in the classic experimental design. The second control group is not pre-tested, but is exposed to the independent variable and post-tested.
· Randomized group design: Here, the design includes two experimental groups and one control group. This design is similar to the classic design except that it includes two experimental groups.
· Solomon four-group design: This design employs two experimental groups and two control groups. One experimental and one control group are treated as in the classic experimental design. The second experimental group and the second control group are not pre-tested but are exposed to the independent variable and are post-tested.
Social scientists are more likely to use field or quasi field experiments rather than laboratory experiments or multi group and the multi design experiments.
1.11 Field experiments
Field experiments (also known as naturalistic experiments) are conducted in natural situations, that is, in the field where the subjects live their lives: their home, school, street, shopping centre, church and so on. There are two types of field experiment: true field experiments and quasi-field experiments. The former require complete adherence to the principles of experimentation, while the latter do not.
The following is a quasi-field experiment: in a supermarket, one of the cashiers is advised, when handing change to the customers to add $20 to it. If the customer accepts the overpayment without comment, the cashier would re-count the change and return the proper amount to the customer. The researcher records the way in which customers react to the overpayment, and the number and type of people who accept or question it.
1.12 Social experiments
Most social experiments are examples of true field experiments. They are commonly used in educational, medical and psychological research, as well as by those working in the areas of social policy and programme evaluation. Characteristic of social experiments is that they choose their subjects by employing systematic random sampling procedures, and make no attempt to control experimental conditions and extraneous influences. Such control is considered impossible in social experiments.
Overall, social experiments are applied in the social context and compare social groups to ascertain whether certain factors have an impact on the community. For instance, low-income families-may be randomly chosen and assigned to experimental and control groups; the families of the experimental group are offered certain incentives while the families of the control group are not. The research question here is whether the families of the experimental group change in any way, and more importantly whether it can be assumed that incentives have an impact on low-income families.
1.13 Validity of experiments
Experiments are among the most valid instruments of social research. There are many difficult factors that are generally thought to raise questions about the validity of experiments. These are:
· Maturation: The change in the dependent variable may not be caused by the independent variable but rather by maturation of the subjects.
· Conditioning: Pro-testing might sensitize subjects and influence them to develop an interest in the experiment and respond atypically to experimentation.
· The history effect: Historical events might occur during the time between the pre-test and post-test, and might affect the responses to the latter.
· Changes in samples: These are changes due to mortality, spatial mobility or general unavailability of subjects and/or experimenters.
· Instrumentation: There is a possibility that changes in the dependent variable are due to changes in the nature of the tests during pre-testing and post-testing rather than to the effects of the independent variable (measurement decay).
· Interaction: Changes in the dependent variable might be caused by a combination of several factors, which may or may not include the independent variable.
· Sampling: Changes in the dependent variable might be caused by sampling problems.
· The Hawthorne effect: Changes might be caused by the fact that subjects know that they are being studied (also referred to as the reactivity effect).
· Modeling: The dependent: variable might: change because the experimenter expects the subjects to behave in a certain way and the subjects wish to please the experimenter.
· Ecology: Changes in the dependent variable might have been caused by the 'ecology', or the experimental setting and measurement.
This is not to be taken as a criticism of experimentation. Knowledge of such risk areas, which can affect the validity of experiment, empowers the researchers to construct an experiment on sound foundations so that distortions can be avoided.
1.14 Experiments in qualitative research
Nature of qualitative experiments
Semi-structured experiments of some kind have been employed occasionally in the social sciences, used in a systematic manner and within a qualitative paradigm. Researchers who employ them do so using their own understanding of what they consider a qualitative experiment. Overall, qualitative experiments are not seen as legitimate, logical or straightforward procedures that should be taken seriously. There are several reasons for this.
· It is argued that experiments are no longer as popular as they used to be.
· Qualitative methodologists do not seem to have made a serious effort to develop a suitable methodology for qualitative experiments.
Method of experimentation has over the years become synonymous with quantification, measurement, statistics, control, intervention, objectivity, standardization and similar qualities; 'to conduct experiments outside these parameters appears to be unthinkable. Quite simply, a 'qualitative experiment' seems to be a contradiction in terms!
Still, the need for qualitative experiments and the contribution they could make to social sciences has been recognized by some researchers who have either conducted experiments that are close to what we would call non-quantitative experiments, or made a contribution to the discussion of this topic, and described the manner in which qualitative experiments are conceptualized and practiced.
Rules of experimentation
By definition, a qualitative experiment: is expected to adhere to the rules of qualitative methodology, which are inconsistent: with the kind of experiments described in the previous section. Similarly, the object of study must be chosen in a manner that will be consistent with a qualitative model; in other words, the objects of study must be accessible to the tools of qualitative methods. Appropriate objects for qualitative experiments include individual or collective relationships and behaviour, and everything related to such relationships. Rules for doing qualitative research are summarized as follows:
· The researcher does not intend to test hypotheses but to develop new ones.
· The object is not defined rigidly but remains open and flexible, allowing future revisions and adjustments. More precise definitions are developed after the study is completed.
· The rules here are flexibility, openness and structural variability. This relates, for instance, to the choice of methods and sampling procedures.
· The emphasis of analysis and evaluation is on identifying commonalities rather than on stressing differences. The principle of dialogue is egalitarian (democratic) not critical/authoritarian. The process of analysis is a circular one, in which the author moves around the object of analysis throughout the experiment, looking for answers. To this is added the dialectic process of discovery: an answer raises a question, which demands a new answer, which leads to a new question which calls for an answer, and so on. This is continued until the research object is explained.
2 PANEL STUDIES
2.1 Introduction
The characteristic of panel studies are, first, that they study the same sample on more than one occasion; second, they study the same topic; and third, they employ the same methods.
The length of the time interval between the stages depends on factors such as the nature of the study object and the research purpose. Issues that change quickly may require shorter intervals than those changing only over long periods of time. When considering whether to employ a long panel study, there are a number of factors that must be taken into account.
| Where do panel studies usually go wrong? · Loss of subjects: This may be due to death, moving away, being unable to take part in the study and so on (the longer the study, the more likely it is that subjects will become inaccessible). · Conditioning: Subjects develop an interest in, and hence learn more about, the study object over time, and therefore are no longer representative of the population. · Instrument bias: Instruments may be employed in different ways in the various stages. Presentation, timing and conditions of data collection may vary. · Respondent bias: The subjects become accustomed to the research instrument and its questions, and are therefore better able to respond to the study's question in subsequent stages. · Study conditions: Data collection, even if performed in exactly the same way in the various stages, can be seriously affected by the personal and family conditions of the subject. Put simply, conditions of study can hardly be exactly the same for each respondent in all stages. · Costs: The longer the study, the greater the cost. |
2.2 Variants of panel studies
Panel studies are also used in other forms. In most cases, these variants are either not sufficiently different to deserve special attention, or are not 'studies' but tools or instruments used within other studies. Below are a few examples.
Cohort studies: Cohort (group) studies are panel studies that focus on cohorts; that is, they study the same cohort on several occasions. One may study all those born in a city suburb in 1990, conducting surveys every two years, recording the changes on a number of issues, and so identifying changes over time. The previous discussion on panel studies applies also in such cases.
Multi-stage studies: Here the researcher initially selects a number of respondents (primary selection units), and then draws from within that sample a smaller sample of respondents who meet the criteria for the study. The process of reducing the number of respondents by focusing on specific criteria and by forming new samples can continue for as long as it is required. For instance, provided that the size of the second sample was large enough to allow further treatment, the researcher could draw a third sample.
The main advantage of this sampling procedure is that it allows the establishment of a sample that is directly related to the research object. With every additional selection, the sample becomes more focused and more relevant to the research question, and the results can be expected to be equally relevant and more representative.
Multi-phase studies: In a multi-phase sampling procedure, the choice of samples and sub-samples is the same as in multi-stage sampling. First, the primary selection units are chosen; then a sample is drawn from them, and so on. However, in a multi-phase sampling procedure, each sample is adequately studied before another sample is drawn from it. Consequently, while in multi-stage sampling, only the final grouping is studied, in multi-phase sampling, all samples are researched. This offers an advantage over other methods, because the information gathered at each phase helps the researcher to choose a more relevant and more representative sample.
2.3 Cross-sectional studies
Cross-sectional studies employ samples from different sectors and compare them by using a set of criteria related to the theme of the study. Examples of such studies would include a study of workers in three different sectors of heavy industry to establish differences in working conditions and pay. In the context of descriptive studies, the purpose of cross-sectional studies is to establish differences between the sections. It can also produce data which will permit the establishment of causal relationships.
2.4 Trend studies
Trend studies are like panel studies in that they are conducted at different points in time; they differ, however, in that they do not use the same sample but different samples. Structurally and in methodological terms they are the same, and the points made about panel studies apply also to trend studies.
Advantages
· They avoid problems such as loss of subjects, conditioning and so on that are associated with panel studies.
· Selection of respondents is also easier, and entails much lower costs.
However, these advantages are countered by disadvantages, particularly regarding the comparability of the populations in question, comparability of samples, continuity in the application of method, and establishing causality. It is important to note, however, that when comparisons between groups over time are considered, trend studies are the only method available. In the example mentioned above, the employment of a panel study would have been most inappropriate; using the same sample at both points in time would have not answered the question of how young people choose a spouse because, by the time the study was repeated, the respondents would have no longer been young.
2.5 What is wrong with panel studies?
| Problems with panel studies Researchers employing panel studies face difficulties in · persuading respondents to take part in the study · maintaining the same structure and same criteria at each stage · avoiding or minimizing drop-outs · mortality, migration and change of residence · panel conditioning, whereby respondents gradually become interested in the research study and learn more about it, which may cause distortions in the research findings |
Nevertheless, these are very useful methods, and are widely used in the context of industry, household studies, service industries, sociological and psychological studies. Knowing their/weaknesses and taking precautions to prevent them from affecting the research can improve their value and usefulness in social research.
3 FOCUS GROUPS
3.1 Introduction
Focus group research can best be described as a loosely constructed discussion with a group of people brought together for the purpose of the study, guided by the researcher and addressed as a group. Due to the use of interviewing technique it is referred to as 'focus group interviewing,' and because it addresses the group rather than specific members, it is also known as ‘group discussion'. Given that more than one group is usually included in the study, and that each group constitutes a separate sample of respondents, focus groups can be seen as a multi-sample study.
Application of focus group
· Selection of people with a particular interest, expertise or position in the community
· Formation of the group by bringing these people together in the same venue
· Introduction of the discussion topic by the researcher, who acts as a facilitator and arbitrator
· Guiding the discussion, so as to address the research topic
· Encouraging discussion among the members of the group rather than between them and the researcher
· Observing and recording the discussion.
It emphasizes on discussion among the group members where the researcher acts as a facilitating observer (not an interviewer) is the essence of the method, and one that distinguishes it from a group interview.
The focus group method appears in two forms, the unstructured or semi-structured form and the structured form, which are employed by qualitative and quantitative researchers respectively
3.2 The purposes of focus groups
· As a pre-research method it can help to prepare the main study by providing sufficient information about the study object, about operationalization by defining indicators, and about preventing possible errors.
· As a main study it offers information about group processes, spontaneous feelings, reasons and explanations for attitudes and behaviour as adequately as any other method.
· As a post-research method it explains trends and variances, reasons and causes, attitudes and opinions.
In another form, group discussion is employed to bring about changes in the group and its members, as a result of the direction and intensity of the discussion. It can further provide valuable information about group processes, attitude changes and manipulation, the attitudes and opinions of group members, the group or the public, the effectiveness of certain methods and so on. It can also enable the testing of group strategies in solving problems (Flick, 1998). Finally, the study of focus groups can 'generate diversity and difference either within or between groups, and so reveal . . . the dilemmatic nature of everyday arguments' (Lunt and Livingstone, 1996: 96).
Overall, this method does not aim to analyze the group but rather to provide a forum that facilitates group discussion, to brainstorming a variety of solutions and to establish a mechanism of opinion formation. Within the context of the group, it is expected that, through mutual stimulation, a group environment will encourage discussion; increase the motivation to address critical issues; enable the facilitator to lead the discussion towards focal points and topical issues; and allow significant points of view to be presented in a real, emotional and summated form as spontaneous expressions (in other words, producing the opportunity for a controlled presentation of personal views). In this way, important information can be gathered in a relatively short period of time.
Within a qualitative research model, group discussion offers access to the construction of meanings while participants interact with each other within the group, the breadth and variation of chose meanings, and the way in which the group negotiates them. This is expected to occur when group members, first address, describe or explain issues introduced by the leader or a member of the group; second, compare different points of view, third, evaluate views and outcomes of discussion, and judge relevant arguments; and fourth, make decisions, draw conclusions or present alternative points of view, while trying to achieve a possible consensus.
3.3 The discussion process
The conduct of group discussion has been described in different ways. The basis of diversity is the nature of the underlying paradigm, although the nature of the topic plays a role in this context. Some (e.g. Mucchielli, 1973) contend that there are three major seeps in group discussion - warm-up, confrontation and relaxation - while others list six or even more steps. Lamnck (1993) refers to the steps: being a stranger, orientation, adjustment, intimacy, conformity and fading out of the discussion.
In more general terms, group discussion as a method of data collection unfolds in a manner parallel to that of our other models of social research. Nevertheless, there are some special points that mark tin's particular model. The following deserve special attention.
Choice of the group: The choice of groups varies significantly. In some cases, natural groups are chosen; in other cases, they are constructed by choosing appropriate respondents. The choice of the participants is equally guided by theoretical principles. Although structured selection is employed, it is more common for participants to be selected by the researcher according to their expertise and social attributes. Depending on the purpose of the study, in some cases the group is chosen to be homogeneous and in other cases diverse.
The size of the group is usually large enough to provide a basis for a reasonable discussion and to allow personal contacts among the participants, but not so large as to make such contacts difficult. In practice, the group size is between five and ten, although significant- deviations from this are not uncommon. Likewise, the number of groups included in a study varies. The research model the nature of the topic and the purpose of the study will determine the number of the groups as well as whether they will be addressed concurrently by using several facilitators, or consecutively. In practice, the number of groups employed in a study ranges from as few as ten to as many as 50.
Introduction of a goal-directed discussion: The discussion usually begins with a few general points to familiarize the participants with the group, moving on to a ‘discussion-generating question' related to the research topic. The group leader will intervene as required, directing the discussion to the research goals and keeping its course interesting and balanced, chat is, equally distributed among all members of the group. How the direction of the discussion will proceed from here is a question that depends on the nature of the underlying paradigm. Quantitative researchers will expect adherence to objectivity and distance between the leader and the members of the group, as well as strict and uniform procedures. Qualitative researchers will give little - if any - attention to these issues.
In physical terms, the environment in which the discussion will take place is chosen to be conducive to a comfortable and stimulating debate, without disturbances, Rooms are to be arranged so that they are not affected by noise or other forms of disturbance, and to facilitate eye contact.
Guiding the discussion: The discussion will be facilitated by the group leader in whatever way the situation requires. Discussions that are slow to start will be helped to gain momentum through additional questions, probes and other appropriate means; non-talkers will be encouraged to participate and those who dominate the discussion will be kept on track. Motivation, encouragement, stimulation and control will bring about a balanced environment that is conducive to group discussion. The nature of the mechanisms employed to ensure a fair and balanced discussion will vary with the underlying paradigm.
3.4 The leader:
The leader occupies a central position in the context of group discussion, and is normally a facilitator rather than a controller. The extent to which this method achieves its purpose depends to a large extent on the research context; within these parameters, the quality of this facilitator is very important. Leaders are generally expected to have the attributes listed in Box 8.6.
The leader's role will very much depend on factors such as the research topic, the nature of the group, and the underlying theoretical framework.
| The qualities of the facilitator Facilitators are required to have: · adequate theoretical and methodological knowledge of the research topic, and general intellectual capacity · experience with group work as well as the ability to guide the discussion effectively · the capacity to create an environment that will encourage involvement, control dominating participants and keep the discussion moving in the right direction · leadership qualities · the ability to develop a warm atmosphere among the members of the group |
3.5 Recording
There are many ways to record data produced in group discussions. Electronic recording is the most common. This has obvious advantages but may affect the readiness of some participants to speak, and requires a lot more time for viewing or listening to the tapes, and/or transcribing them. Manual recording is an alternative but it is equally limited by the innate difficulty of coping with intensive and multi-sided discussions. Having two leaders, with each of them taking notes in turn or one recording the data and the other acting as the facilitator, may be a solution.
The nature and content of recording depend on the nature of the study, available resources and the nature of the expected information. In one context the data may be quantitative and standardized. In other contexts it may be qualitative, entailing 'thick' descriptions.
3.6 Limitations of focus groups
The value of focus groups as a method of studying group discussion relies heavily on both the 'group' and the 'leader'; hence success depends very much on these two factors. It is therefore understandable that problems of this method noted in the relevant literature relate almost exclusively to these two points (Bergcr et. al., 1989; 339-44; Dreher and Dreher, 1991: 186-8; Flick et al,:1991; Mahr, 1995; Puris, 1995), which are listed below.
Weaknesses of focus groups
· Being in a group might make participants hide their real opinions, especially if their views can have effects on their personal life or professional career.
· Recording of the data can be problematic.
· Domination of the discussion by some members might affect the direction and outcome of the discussion.
· Some members may not participate in the discussion.
· There may be attempts to go along with the leader for many reasons (e.g. to 'get it over with', or to please a leader who holds an important position in the respondents' personal, political or professional life).
· There may be difficulties with keeping discussion on track.
· Group members may have reasons to offer a. collective front and deceive the leaders.
· The findings may not be representative.
Although these points are relevant and legitimate in their context, they should be taken as a reminder of tin: areas in which focus groups can 'malfunction' and not as discounting of the quality of the method. An experienced researcher will arrange the context and process of discussion so that distortions and problems will be prevented.
MAIN POINTS
· Studies often employ more than one sample. They may employ the samples at the same research stage, or at different stages.
· Experiments, panel studies and focus groups are examples of multi-sample studies.
· Experiments measure the effects of a stimulus on a subject in a controlled environment.
· Sampling in experiments is accomplished by means of randomization, subject matching and group matching.
· Experiments follow a set of steps. In a typical case experimentation involves a pre-test, a test and a post-test. There are several experimental designs.
· There are laboratory experiments, social experiments and field experiments.
· The validity of experiments depends on a number of factors, such as maturation, conditioning and instrumentation, the history effect, changes in samples, interaction, sampling, ecology, modeling and the Hawthorne effect.
· Longitudinal studies employ more than one sample within the same study.
· Panel studies and trend studies are two common types of longitudinal studies.
· Focus groups facilitate collection of data by means of group discussion.
METHODOLOGICAL DESIGNS
In social and behavioural sciences there are three types of methodological designs that are currently in use namely:
· Descriptive research design
· Exploratory research design
· Experimental research design
Descriptive Research Design
Descriptive research design aims at describing the current state of affairs at the time of study. This kind of research does not have a treatment or control group thus; it cannot measure and/or investigate the influence of any variable upon another. In other words all it allows you to do is to paint a picture for the reader of what occurred at a given time and space. It does not permit the researcher to deal with the question of why some thing happened at a give point of time.
(Salkind 1998)
Exploratory Research Design
Exploratory research design is suitable for investigating phenomenon about which hardly anything is known at the outset of the project. The researcher starts the project with a very preliminary notion of the object and context of the study. The provisional concepts that the researcher starts off with gradually develop and improve as the study progresses.
(http;//www2.uiah.fi/project/metodi)
Experimental Research Design
Experimental Research Design is a type of researcher which allows the researcher to investigate the impact of one variable upon another. It is different from both the exploratory and descriptive research designs in that it not only talks about what but also answers the why and how questions regarding the phenomena under study. Within experimental research design there are four types namely:
· True experiment design
· Quasi-experiment design
· Naturalistic experiment design.
· Natural experiment design
True Experiments
There are five steps to follow in conducting true experiments with the people.
· You need at least two groups, called the treatment group and the control group. One group gets the intervention and the other group does not
· Individuals must be randomly assigned either to the intervention group or to the control group to ensure that the groups are equivalent.
· The groups are measured on one or more dependent variables; this is called as pretest.
· The intervention is introduced.
· The dependent variables are measured again. This is the posttest.
The true experiments with the people are common in laboratory psychology and in the testing of new medicines. Laboratory experiments often produce results that beg to be tested in the natural world by anthropologists. When they are done outside the lab, experiments are called field experiments.
Quasi-Experiments
Quasi-experiments are most often used in evaluating social programs. Suppose a researcher has invented a technique for improving reading comprehension among third graders. She selects two third-grade classes in a school district. One of them gets the intervention and the other doesn't. Students are measured before and after the intervention to see whether their reading scores improve. This design contains many of the elements of a true experiment, but the participants are not assigned randomly to the treatment and control groups.
Naturalistic Experiments
In a naturalistic experiment, you contrive to collect experimental data under natural conditions. You make the data happen, out in the natural world (not in the lab), and you evaluate the results.
Natural Experiments
Natural experiments, by contrast, are going on around us all the time. They are not conducted by researchers at all—they are simply evaluated. Here are four examples of common natural experiments: (a) some people choose to migrate from villages to cities, while others stay put. (b) Some villages in a region are provided with electricity, while some are not. (c) Some middle-class students go to college, some do not. (d) Some cultures practice female infanticide, some do not.
Each of these situations constitutes a natural experiment that tests something about human behavior and thought. The trick is to ask "What hypothesis is being tested by what's going on here?"
No comments:
Post a Comment