Download Research Methods: Deductive, Inductive, Qualitative, and Quantitative Approaches and more Exams Sociology in PDF only on Docsity!
Research Methods
Deductive Research -
- a deductive approach is aimed and testing theory
- characterized by a theory and hypothesis driving the collection of data
- Start with a theory and test it Inductive Research -
- an inductive approach is concerned with the generation of new theory emerging from the data
- characterized by data driving theory development
- start with the data and let the theory emerge Deduction - An approach that involves starting out with a theory - reasoning from a general idea to particulars; often associated with quantitative research. Data - Empirical evidence; information acquired. Empiricism or empirical investigation - Investigation based on observation, experience or experiment rather than on theory. Epistemology - The philosophical study of the nature and basis of knowledge - what is knowledge, and how do we know what we claim to know? Hypothesis - A tentative and speculative statement informed by theory about the possible relationship between two or more variables. Interpretivism - Interpretivism is one form of qualitative methodology. An Interpretivist approach to social research would be much more qualitative, using methods such as unstructured interviews or participant observation Interpretivists, or anti-positivists argue that individuals are not just puppets who react to external social forces as Positivists believe. According to Interpretivists individuals are intricate and complex and different people experience and understand the same 'objective reality' in very different ways and have their own, often very different, reasons for acting in the world, thus scientific methods are not appropriate. Intepretivist research methods derive from 'social action theory'
Intereptivists actually criticise 'scientific sociology' (Positivism) because many of the statistics it relies on are themselves socially constructed. Qualitative Personal Documents Participant Observation Unstructured Interviews Subjective Itneraction/Involvement/Rapport Feelings/Empathy Thick Description Individual Motives Humanistic Socially constructed, subjective, may change, multiple Objectivism - A philosophical position that argues for the independence of worldly phenomena from the observer, and the need to examine the world from a value-free perspective. Objectivity in social research is the principle drawn from positivism that, as far as is possible, researchers should remain distanced from what they study so findings depend on the nature of what was studied rather than on the personality, beliefs and values of the researcher (an approach not accepted by researchers in the critical, standpoint or interpretivist traditions). Positivism - prefer quantitative methods such as social surveys, structured questionnaires and official statistics because these have good reliability and representativeness. Positivists see society as shaping the individual and believe that 'social facts' shape individual action. The positivist tradition stresses the importance of doing quantitative research such as large scale surveys in order to get an overview of society as a whole and to uncover social trends, such as the relationship between educational achievement and social class. This type of sociology is more interested in trends and patterns rather than individuals. Positivists also believe that sociology can and should use the same methods and approaches to study the social world that "natural" sciences such as biology and physics use to investigate the physical world. By adopting "scientific" techniques sociologists should be able, eventually, to uncover the laws that govern societies just as scientists have discovered the laws that govern the physical world. In positivist research, sociologists tend to look for relationships, or 'correlations' between two or more variables. This is known as the comparative method Objective Detachment Trends/Comparisons
Cross-sectional studies are 'snapshot' studies of a problem or situation at a particular point in time. Examples of cross-sectional studies are an opinion poll of a sample of voters before an election or a comparison between one or more organisations of clerical workers' attitudes to the introduction of performance-related pay. Most surveys are cross-sectional in design, and many qualitative projects are cross-sectional in that, for example, interviews are conducted over a short period of time. Longitudinal -
- Longitudinal studies aim to monitor change over a period of time. For example, a group of people may be surveyed at one period of time and the same group of people will be surveyed again later in order to investigate changes or developments that have emerged during that time frame. Although, due to time constraints, it is usually not possible for students to conduct longitudinal research as part of their studies, it is possible to design an element of longitudinal research into student projects.
- 'Why Do They Go? Individual and Corporate Perspectives on the Factors Influencing the Decision to Accept an International Assignment' Experiments - All experiments seek to identify causal relationships between dependent and independent variables. The task of the researcher is to identify the factor(s) that caused the observed outcome. In order to pinpoint causal factors, you need to control the conditions under which the experiment is taking place so, typically, you might divide the population of study into experimental and control groups, then introduce or exclude potentially significant factors to the experimental group but not to the control group. You would then observe both groups over a period of time, and through close observation and measurement of the outcomes, you should be able to identify the effects on the experimental group of exposure to the potentially significant factor. Problems with Experiments -
- it can be difficult to achieve an exact match between the features of the control group and the experimental group.
- many experiments are conducted in artificial settings and it is questionable whether subjects are acting in the same way as they would act in their natural setting, or whether their behaviour and responses are influenced by the artificial conditions.
- it can be difficult for the researcher to be certain that they have the required control over the potentially relevant factors to be tested (the variables to be tested by exclusion from or introduction to the experimental group).
- some experiments might be unsuccessful if subjects had prior knowledge of the purpose of the experiment
Case Study - is a unit or element or part of some larger category. It could be a person or group of persons, or a situation or event. Or it could be a bank or a factory or a school. The idea of a 'case study' is to develop '... detailed, intensive knowledge about a single "case", or of a small number of related "cases"' (Robson, 2002: 89), and case studies can be conducted using a variety of methods - interviews, questionnaires, observation, documents or a mix of methods. Case studies may be longitudinal, involving observation at two or more points in an organisation's or event's life. Establishing the causal links over time might provide explanations of events. They are most useful in generating hypotheses. A detailed examination of one set of events can suggest a chain of causation that can be tested by spreading the research wider. Ethnography - The term 'ethnography' literally means the description of people, and it refers to the study of people and cultures. At the heart of ethnography is the view that we cannot understand people's lives, beliefs and behaviours unless we understand these things from the participants' own point of view. Ethnographers are required to spend lengthy periods of time in the field, sharing in the lives of the people they are studying and collecting data in ways that interfere as little as possible with the natural routines of everyday existence. The ethnographer needs to be flexible and responsive to changes in the setting as daily life unfolds and, typically, the ethnographer is data driventh in that data play a large role in informing and guiding them in their choices of what next to research and how to do it. Therefore, ethnography is associated with an inductive research approach, and observation, particularly participant observation, interviewing and spontaneous conversation are typical methods of data collection used by ethnographers. Ethnography is a form of qualitative research that is also referred to as the fieldwork method or the method of participant observation. However, ethnography is both a method for data collection and a method for analysis. The two are inseparably linked because in order to conduct ethnographic work, data collection and data analysis occur simultaneously; they are not separate and distinct phases of the research process. Ethnographic research can be used to generate theory and to test theory, and it can be undertaken for academic or practical and applied reasons. Surveys -
- use some form of quantitative measurement and are generally associated with the deductive approach -although most surveys do include qualitative analysis.
- through the process of creating and using their own knowledge, PAR aims to empower a group of people
- PAR sees as essential the processes of participation and collaboration PAR team should be voluntary Analysis of existing data - Often data have been collected but not used for the purpose of your research. If you are working for an organisation you may be given access to previous survey, census, taxation or other data that will help answer your research question. Although at Master's level you are not required to conduct original surveys or collect original data in other ways if there is sufficient information available to you to convincingly answer your research question, the same standards of validity and reliability should be applied to the data you use as to the data you collect. The same ethical issues about using personal information also apply. Secondary Sources - Existing data may already have been analysed by officials or by academics. Such analysis is described as 'secondary sources' and may make a contribution to answering your research question. It is unlikely that a simple review of secondary sources - a literature review
- would demonstrate sufficient thought and skill to gain a pass in a dissertation at Master's level, but most research does use secondary sources as a background to the questions raised and theory applied within the research enquiry. Literature Review - A literature review is often the most difficult part of a project to write, for two main reasons.
- First, it is likely that you will have read a considerable amount of literature about your topic generally, and from all this material you must then select what is relevant to your particular research project.
- Secondly, a successful review must clearly demonstrate the relationship between your own research concerns and the rest of the literature in your field of interest. Many projects are weak because there is insufficient linkage between the existing literature and the central research concern of the researcher's own project. Role of Literature Review -
- A literature review can help you to avoid repeating research that has already been done elsewhere by other investigators.
- It can also help you to locate explicit recommendations for further research (Saunders et al. 2000). Often, authors include a section on suggestions and recommendations for further research that you might use to help formulate your own research question(s).
- A literature review can help you gain insights into current and popular opinions about aspects of your research topic. You may find, for example, that aspects of your topic are considered 'newsworthy', and newspapers, radio and television, professional journals and trade journals have given coverage to it (Saunders et al. 2000).
- It can reveal the different methods and approaches previous researchers have used to research similar topics.
- It shows other researchers' empirical findings. Content of Literature Review - When thinking about the content of your review, there are three major aspects that need careful attention. These are:
- analytical evaluation of the literature
- relevance of the literature
- quantity of literature. A rigorous critical review of the literature requires you to move beyond description and to analytically assess and evaluate the significance of the existing research. How NOT to write a literature review - A critical review of the literature should be presented as a coherent, well-integrated thematic argument. Weak literature reviews often exhibit the faults discussed here:
- The review is presented as a random collection of views and perspectives on the topic.
- The review is presented as a string of quotations from authoritative thinkers in the field with few words of your own.
- The review is presented as a haphazard collection of different materials. Measurement - 2 reasons -
- • measurement enables us to be systematic and record our observations
- • measurement helps us to communicate our observations more clearly Social Measurements - social measurements are often more complex, and are much harder to compare between studies than physical measurements. Isomorphism - Two structures are said to be isomorphic when they exhibit the same structure from the standpoint of their basic concepts.
Likert scale is a special case as it may be considered an interval scale. Scale 1 = Strongly agree 2 = Agree 3 = Neutral 4 = Disagree 5 = Strongly disagree Ordinal Scale - An ordinal scale orders or ranks the categories of the variable, but the distances between categories are not equal and therefore cannot be compared. There is, as for the interval scale, no absolute zero. Exam grades, preference scales Binary scale - Binary scales are also referred to as dichotomous. Binary scales are used for nominal variables which can only exist in two mutually exclusive states (for example, male or female) or where there are a number of different but not mutually exclusive states and each state is recorded as 'present'/'absent' ('yes'/'no'). Nominal - Nominal scales simply give names to mutually exclusive characters or categories being scored for the variable, and measurement involves counting the number of subjects that fall into that character or category. A nominal scale for a social study might be, for example, three categories for type of work
- none (0)
- part-time (1)
- and full-time (2). Data Classification - Data classification attempts to relate the data types (quantitative, qualitative, categorical, discrete, and continuous) to the scales of measurement (ratio, interval, ordinal, and nominal), but there is some debate as to how this is done. Accuracy - Refers to how close the measure is to the real value. Character - Equivalent term to variable, the attribute we are observing, especially with reference to biological traits - e.g. wing length of a fly might be the character being measured. Continuous Variable - Variable that can take any value in a certain range, theoretically with an infinite number of points between units, but in practice, there is usually a limited number of points between units.
Counts - The number of observations or units or events of a particular type.
- Counts are tricky to classify and may be considered discrete or continuous depending on the circumstances. Derived variable - A variable derived from manipulation of other variables, as in the case of a ratio or a proportion. Descriptor - Equivalent to the term character. Discrete variable - Variable that can only take certain defined values, with no intermediate points between these values (equivalent to noncontinuous). Something that is counted Measurement - Measurement is the assigning of numbers to properties according to rules and is used to record and describe our observations. Mean (average) - A measure of position or central tendency for ratio and interval data: calculated by adding up all values and dividing by the number of observations. Normal distribution - Tests that do not assume the data have a normal distribution.
- Bell curve The Normal Distribution has: mean = median = mode symmetry about the center 50% of values less than the mean and 50% greater than the mean Parametric tests - Tests that assume the data have a normal distribution and are continuous and, when comparing samples, that they have homogeneity of variance.
- assume underlying statistical distributions in the data. Therefore, several conditions of validity must be met so that the result of a parametric test is reliable. For example, Student's t-test for two independent samples is reliable only if each sample follows a normal distribution and if sample variances are homogeneous.
measurement - i.e. ratio, interval, ordinal, or nominal. Sampling - Mixed methods = Validity - Concerned with whether what you are measuring represents what you want to measure. Sampling - Mixed methods = Variable - The attribute being measured or observed. Sampling - Mixed methods = Statistical inference - When estimating parameters or testing hypotheses, the process of estimating (unknown) population figures (parameters) from observed sample results in accordance with statistical principles. Types of interviews -
- Structured
- Semi-structured
- Unstructured Unstructured Interview - In unstructured interviewing, the term informant is more commonly used than 'interviewee' and 'respondent'. Unstructured interviews (alternatively called intensive interviewing, in-depth interviews, ethnographic interviews or research conversations) emphasise the active role of the informant and the flexibility of the researcher to respond to the direction in which the informant is taking the interview. The interviewer uses an interview guide - that is, a list of themes, topics, issues or questions that the researcher wants to be sure to ask about and wants the informant to talk about. When to use interviews -
- when you want to gain some deep understanding about aspects of people's lives from their own points of view. Interviewing can provide deep, detailed information and insights about people's experiences that cannot be accessed through the use of a rigidly structured, pre-formulated questionnaire. Interviews are the commonly preferred method when the research requires an investigation of emotions and feelings, which are complex matters and certainly not easy to quantify. Questionnaires without an interviewer will most likely lose much of the nuance and deeper information. Planning the use of interviews -
- Sampling
- Selection Info from pre-formulated questions - The information than you can obtain from your pre-formulated questions falls into four main categories:
- attributes
- attitudes
- behaviour
- beliefs Interview - Interview
- choose a method of administering interviews, from among the main types
- identify and explain the circumstances in which it is appropriate to employ qualitative interviews as a data collection technique
- plan research using qualitative interviews
- recommend a strategy for designing qualitative interview guides and apply this strategy to the design of such a guide for your own research
- identify, appreciate and describe the most important skills required to conduct qualitative interviews
- record and transcribe qualitative interviews
- identify and discuss ways to check the accuracy of qualitative interview data and be familiar with debates about this issue
- analyse qualitative interview data. Questionnaire - Questionnaire
- describe the main types of questionnaires, their modes of administration and the differences between them, and the major planning issues in using questionnaires
- plan and justify an appropriate method for administering your own questionnaire
- identify appropriate techniques for maximising questionnaire response rates
- choose appropriate strategies for conducting questionnaire interviews
- discuss all aspects of questionnaire construction and piloting
- outline the main ethical concerns in conducting questionnaire interviews
- identify the types of data required for your research and appropriate data analysis techniques. Designing a questionnaire -
- decide on the areas about which you need to ask questions
- write the questions that will address each of the areas or aspects that you have identified in step one. Writing questions is one of the most difficult tasks in designing questionnaires and ti is vital that you ask the appropriate type of question to elicit the information that you want.
7-State the statistical conclusion in the context of your research. Naturalism - Members of the setting must be left alone to get on with their normal, everyday activities, much as they would if the researcher was not present, so the researcher attempts to 'blend into' the setting. In this sense, ethnography is non-interventionist. For this reason, ethnographers do not rely on the traditional research tools of a positivist epistemology such as questionnaires, psychological tests, attitude scales and so on, which require researcher intervention to 'remove' those being studied from their natural, everyday interactions in order to take part in the research. Ethnographers enter into the social settings of those they are studying with as little disruption as possible, to gain the perspective of the insider. When to use ethnographic research - The major principle of ethnography is to understand whatever it is you are studying from the point of view of the people you are studying. This means that you must remain open and flexible to making changes to the focus of your research, based on your ongoing observations and research conversations with informants.
- importance may change over the course of your research, depending on your discoveries in the field. You may need to re-focus your ideas, re-prioritise them and add to them in order to make them meaningful and understandable in the same way as they are meaningful to and understood by the research participants. Your task is to properly discover what is relevant to your informants, in what ways, to what extent, and how this is the case. This may mean adjusting your original questions in the light of your data. Your task is not to create and then pursue your self-created matters of importance - this would be of artificial importance in the life-world of your participants. If you enter the research field with questions that are too rigid and finely tuned, you run the risk of missing what is important to your informants. Ethnographic Methods - Methods associated with ethnography Anthropological ethnographers often live amongst a group/society for a year or more, in order to learn about them. This fully immersive, long-term 'live and work' approach to ethnography has not proven popular within the field of usability. Part of the reason may involve cost, but it is also the case that anthropologists and usability practitioners are interested in different things. Anthropologists use ethnography in an attempt to fully understand as much as possible about an entire society. Usability practitioners are usually only interested in learning information that will support their reasoning on a specific design problem.
We would argue that deep, immersive 'live and work' ethnography is rarely required within the field of user-centred design. However, short ethnographic studies can be very useful for user-centred projects. For example: in order to understand the way in which a Merchant Bank trades and operates, a usability consultant might conduct an ethnographic study by working and socialising with its employees for a month. Individual methods which are available within an ethnographic study include: participant observation, interviews and surveys. All of these ethnographic methods can be very valuable in gaining a deeper understanding of a design problem. Usability practitioners often make use of these in order to develop their understanding of the relevant domain, audience(s), processes, goals and context(s) of use. When to use ethnography Ethnography is most useful in the early stages of a user-centred design project. This is because ethnography focuses on developing an understanding of the design problem. Therefore, it makes more sense to conduct ethnographic studies at the beginning of a project in order to support future design decisions (which will happen later 3 types of writing ethnographic notes -
- Thematic Organization - is a way of presenting material as a number of major components or topics or concepts
- The Chronology - this approach follows the passage of time of something that happened in the setting or to members of the setting
- The Natural History - this is an account that follows the passage of time in the field and explains, sequentially, the way in which the fieldwork 'unfolded' 3 ways of writing about yourself -
- The Realist Approach - The realist position is characterised by using the third- person voice; the ethnographer is largely absent from the analysis and the analysis is presented as a dispassionate narration of what happened, how it happened, who said and did what, and so on, in the setting.
- The Confessional Approach: this approach is gaining popularity in ethnographic writing, often as a deliberately chosen alternative to the dispassionate style of realist writings. In confessional' writing, the ethnographer writes about their personal emotions and experiences in relation to the research.
- The Impressionist Approach: impressionist accounts typically form parts or segments of writing within confessional writing or, less frequently, within realist writing, and they are usually written in an engaging and dramatic way that includes the voice of the ethnographer. Their focus is on something experienced by the researcher during fieldwork that is unique, memorable or notable rather than common or routine. Impressionist writing is usually about how the researcher experienced things at the time rather than the interpretations or understandings
Power of test More powerful Less powerful Used to compare Means Medians Ordinary Least Squares (OLS) - The most widely used method for estimating a linear regression model consists of minimising the sum of squared residuals from the fitted regression line, and is therefore called Ordinary Least Squares (or OLS in short). The OLS criterion consists of minimising the sum of squares of the residuals ei. The reason for squaring the residuals is twofold. Threats to validity in Qualitative Research -
- Interpretation: imposing a framework on the findings: the researcher may formulate the questions and the causal relationships in such a way as to obscure other possible realities.
- Theory: ignoring alternative explanations. At a higher level of abstraction, the whole theory about the way the particular phenomenon works may obscure other explanations.
- Reactivity: the researcher affects the answers. It is common for research subjects to be influenced by what they think the researcher wants to hear, or wants not to hear.
- Respondent bias: the selection of respondents may bias the sample and make invalid any statements about the population based on the selected sample.
- Researcher bias: the researcher may him/herself introduce bias into the data. Triangulation as a Solution to Validity Threats - One way to avoid some of these threats to validity is triangulation: looking at the question from more than one point of view.
- Data: choose more than one sample or data source for the same phenomenon
- Observer: to reduce observer bias, employ different observers, and different sorts of observers.
- Method: conduct the study with a mix of methods, especially a mix of quantitative and qualitative.
- Theory: be open to alternative theories that may provide explanations for the observed phenomena. Threats to External Validity -
External validity - also known as 'generalisability' - is the test of your results being achieved when your study is repeated with another sample? What are the threats to generalisability?
- Selection: findings are specific to the group selected and would not be repeated if another group were chosen.
- Setting: findings are specific to the context in which the study takes place.
- History: specific and unique experiences determine the outcome of the study.
- Construct effects: the constructs studied are specific to the group studied and would not apply to any other group. Margaret Mead Example - Threats to validity --- young girls made up fun stories rather than tell boring truths but the truth took decades to come out. Threats to internal validity -
- History: particular events that happened to the subject of the study which were different from all others. It is possible that the events studied are peculiar to the particular subjects and that the same cause would not produce the same effect in any other instance.
- Effects of pre-testing and piloting: in experimental research, it is possible that the responses are caused not by the events being examined but by the influence on the individuals of the pilot or pre-test stage of the experiments.
- Instrumentation: changes in the ways the variables are measured between two time periods. This is especially a problem when scales are used and the terms used to delineate the values (1-5 etc.) change between studies at different time periods.
- Mortality of sample. In longitudinal studies, the sample available at the beginning of the research may not be available throughout the study.
- Selection: non-random sampling. If the study claims that the sample represents the population, but the sample is not chosen randomly, this presents a threat to any validity of statements about the population.
- Ambiguity about causal direction: instead of A causing B, it may be that B is causing A.
- Diffusion of treatments: when control group gets access to the treatment of the experimental group. In experimental research, it is important that the control group is kept away from the factors available to the experimental group.
- Compensatory rivalry: where one group makes an extra effort to demonstrate its superiority. In management research, the performance of teams, units, factories, regions, is a common subject. Once teams know they are being studied, they may make extra efforts to ensure they are seen as successful.
