Political Inquiry Terms and Definitions
Research Design * Good research, bad research * Involves connecting theory and data.* Maximising leverage by using very few variables to explain many effects.* Reports on the degree of certainty of results.
* Shows true causal relationship, not just correlation. * Provides accurate data and accounts for other variables. * Determines if the relationship is unidirectional. * Purpose of research * To establish a relationship between two or more variables * To demonstrate that the results are generally true in the real world and not in just a particular context. To reveal whether one phenomenon precedes another in time, establish time order * To eliminate as many alternative explanations for the observed finding as possible * Choice of design * What affects * Is research exploratory, descriptive or explanatory * What are the practical limitations in investigating hypothesis. * Experimental design: * Classical experimental design, 2 groups, pre and post test, randomisation, * Simple post test: only post test * Repeated measure design, measure how long effect takes to start. Multiple pre and post tests. Multigroup design, more than 2 groups, can compare different levels of experimental variable. * Randomised Field experiments, applies logic of randomisation and variable manipulation * Non-randomised quasi-experiments. Purposeful selection, target a certain group. * Non-experimental design: single group, no control over assignment and application of IV, inability to measure DV. * Case study: small N designs. Provide detailed explanation * Comparative analysis: compare two or several units in relative detail * Focus group: gather information about reaction to certain IV. Surveys: large number of people measured to find causal relationships. * Aggregate data analysis: variables are averages or percentages of geographical areas, find causal relationships. * Longitudinal designs, time span. * Trend analysis: measurement on same variables at different time periods to examine changes. * Panel analysis: follows a group of participants. * Intervention analysis: measurement of change in the DV is observed and taken before and after. No interaction, mere observation. * Ethnographies: form of data collection through participant observation, interviews and questionnaires.
Field studies * Content analysis: textual analysis, study of recordings, written. * What they have in common * They all share the basic objectives of research design despite having different levels of internal and external validity. Using several designs together will cover each other’s shortfall. * They all attempt to draw sound conclusions supported by observable evidence * Terms * Causal vs spurious * Both show correlation between IV and DV, but in spurious the change in DV because 3rd factor caused changed in both. Causal is a direct relationship. 5 different relationships. Multiple causes without chain.
Multiple causes with chain. Multiple causes that affect DV, but are changed with the introduction of another variable. Spurious causality with antecedent variable. Chain causality with intervening variable. * Covariation * Demonstrates that the IV does in fact covary with DV. Not causal relationship yet. * Time order * Show that the IV precedes DV. Effect cannot appear before cause. * Alternative causes * Confounding factors. Factors that possible cause a change in DV as well. * Randomised controlled experiments * Experiments that allow the researcher to control the exposure to the IV through assignments to groups.
Selection and grouping all randomised. * Experimental design * The way in which the researcher controls exposure to test IV. 5 different designs. * Control group * The group of subjects that does not receive experimental treatment or test stimulus. * Pre-test * Measurement of the DV prior to administration of IV or experimental treatment. * Post test * Measurement of the DV after administration. * Internal vs external validity * Internal validity is the ability to show that manipulation or variation of the IV actually causes change in DV. * External validity is the ability to generalise from one set of research findings to other situations. History * A threat to internal validity. Events other than the experimental stimulus that occur between pretest and posttest measurements. * Along with maturation * Testing * When measuring the DV prior to the stimulus alerts the subjects of the research objectives. * Selection bias * Bias due to the assignment of subjects to experimental and control groups according to some criterion and not randomly. A threat to internal validity. * Experimental mortality * A differential loss of subjects from experimental and control groups that affects the equivalency of groups; threat to internal validity * Instrument decay A change in the instrument used to measure the DV, like different researcher conducting pretest and posttest. * Demand characteristics * Aspects of the research situation that cause participants to guess at the investigators goals and adjust their behaviour or opinions accordingly. Trying to “help”. * Simple post test * Similar to classical. Experimental group exposed but control not, only post test is conducted. * Repeated measure design * Contains several pre and post test measurements to know exactly how quickly the effect of the independent variable should be observed or how reliable pretest measurement of DV should be taken. Multi-group design * There are more than one experimental r control group created so different levels of the IV can be compared. Can involve both pre and post test or just one. * Field experiments * They are experimental designs that are applied in a natural setting. Adopts logic of randomisation and variable manipulation. * Non-experimental design * They are designs that do not follow the experimental model of two groups and have controlled exposure to IV. They are more practical to do but are not as strong in terms of establishing causal relationships. * Case study * Comprehensive and in depth study of a single case or several cases.
Provide detailed explanations * Comparative analysis * Comparing between two or several units in relative detail * Focus group * Often use to observe reactions to the introduction of the IV. Has a group of people who meet at a single location * Surveys * Measurement of DV and IV at the same time. Respondents report their exposure to various factors. No assignment to groups, examines groups basd on values of IV, measurement of DV to see differences between groups. * Aggregate data analysis * Variables that are displayed as averages or percentages, to find causality. Trend analysis and longitudinal design * Measurements on same variables at different time periods to see the changes caused by the IV on the DV.No manipulation of variables. Multiple measurements. * Panel study * A study that follows a group of participants where the same units are measured at different times. * Panel mortality * Refers to the participants in the panel study who drop out. Rate? * Intervention analysis. * Looks at the occurrence of the IV as an observation. Looking at the DV before and after IV. Works best when IV happens in a brief period in time and brief in nature.
Measure only before and after event. Literature Review * Purpose of Literature review, 7 * To see what has and has not been investigated * To develop general explanations for observed variations in a behaviour or a phenomenon * To identify potential relationships between concepts and to identify researchable hypothesis * To learn how others have defined and measured key concepts * To identify data sources that other researches have used * To develop alternative research designs * To discover how a research project is related to the work of others. * What constitutes a literature review It is made of different relevant articles that provide more insight into topic * It should help arrive at a good research topic * It should show what has and has not been researched * It provides a general explanation for variations in behaviour or phenomenon * It identifies researchable topics * It should help develop alternative research designs * Best methods for collecting literature * Using electronic databases like JSTOR, Web of Science and Google Scholar. * Searching by topic and key words, slowly limiting results, read the articles and find new words to narrow down the search. Searching by starting with a single article. Use data base to find more relevant articles or other articles written by the same author. Use the citations in that first article. Find articles that have cited the first article. * Best ways to write a literature review * It should rely on scholarly sources * It must relate directly to topic * Have to become familiar with as much of the research before selecting the final sources. * Summarising of relevant literature that focuses on over-arching topics rather than single articles. * Compiling all articles into something that makes sense. * Organising the topics ased upon the research question. * Identifying common themes or methodologies across the articles. * Discussion of conventional wisdom, illustrating how current politics has changed and identifying the flaws in past research. Sampling * Terms * Population * all case or observations covered by a hypothesis, all the units of analysis to which a hypothesis applies. * Sample * A subset of observations or cases drawn from a specified population. * Sample statistics * The estimator of a population characteristics or attribute that is calculated from sample data * Advantages and disadvantages of samples Advantages * It is cheaper and less time consuming as compared to using population * More convenient. * Disadvantages * They can be less accurate or more prone to error * Some studies do not use sampling, like case studies. * Population parameter * A characteristic or attribute of a population that can be quantified. * Estimator * A statistic based on sample observations that is used to estimate the numerical value of an unknown population parameter. * Element * A particular case or entity about which information is collected, the unit of analysis. When to use a sample * Practicality. When data from an extremely large population is required, it is impossible to interview or approach each and every subject. Thus sampling is require because although the sample statistics will not exactly equal the corresponding values, they will be reasonably close if sampling is done correctly. * Sampling frame * The particular population in which the sample is actually drawn from. * Random digit dial: purpose of and how and why it works * It is a procedure used to improve the representativeness of telephone amples by giving both listed and unlisted numbers a chance of selection. * It is used to overcome the problem of cell phone numbers which are unlisted * It works by randomly dialling numbers. * It works because it gives all numbers, whether listed or not a chance to get dialled. * Sampling unit * The entity listed in a sampling frame. Maybe same as an element or group. * Sample bias * The bias that occurs whenever some elements of a population are systematically excluded from a sample. It is usually due to an incomplete sampling frame or a non-probability method of selecting elements. Probability sample: types, pros and cons * Sample for which each element has a known probability of being included in the sample * Types: * Simple random samples: each element has an equal chance of being selected. * Pros: each element has an equal chance of being selected. * Cons * It is not truly random, small patterns of selection might appear. * Obtaining a list of the entire population is not possible, reducing the probability. * Systemic samples: elements are selected from a list at predetermined intervals * Pros It is easier to apply than simple random. * Useful when dealing with a very large population size. * Cons * May result in biased sampling: if elements on the list have been ranked according to a characteristic. The list contains a pattern that corresponds to the sampling interval. * Stratified samples: elements sharing one or more characteristics are grouped and elements are selected from each group in proportion to the group’s representation in the total population * Pros * Homogeneous populations, smaller sample seize is needed to achieve accuracy. * Cons Heterogeneous populations need a much larger sample to be accurate. * Proportionate samples: stratified samples where each stratum is represented in proportion to its size in the population. * Pros * Very representative of the population * Cons * Not good with heterogeneous populations * Disproportionate samples: stratified samples in which elements sharing a characteristic are under-represented or over-represented. * Pros * When the sample size is too small, can be used to increase it easily. * Cons * May not be representative of the population if weight factor is not used. Cluster samples: the sampling frame initially consists of clusters of elements * Pros * Used when no list of elements exists and to create one would be too expensive. * Reduces field work costs * Cons * There is greater imprecision. Samples are not representative of population. * Non-probability samples: types, pros and cons * Each element has an unknown probability of being included in the sample * Types * Purposive samples: when a researcher exercises considerable discretion over what observations to study. * Pros * Can learn more from carefully selected unusual cases. Cons * Not accurately representative of population. * Convenience sample: elements are included because they are convenient and easy for the research to select * Pros * Convenience * Large numbers easy to get * Cons * Unknown accuracy * Quota sample: elements are sampled in proportioned to their representation in the population. * Pros * Similar to proportionate sampling, but subjects chosen purposefully. * Cons * Selection bias * Snowball sample: respondents are used to identify other persons who might qualify for inclusion into the sample * Pros Good for relatively selective and rare populations * Relationship between samples and statistical inferences * Statistical inferences is the mathematical theory and techniques for making conjunctures about the unknown characteristics of the population based on samples. Making inferences about a population. * Trying to define more clearly what supportable means * Samples provide an estimate of population attributes and may be off from the true population parameter. The difference is the level of precision lost. * 3 types of errors in inference Expected values or sampling error * Expected value is the average value of a sample statistic based on repeated samples of the population. * Sampling error is the difference between a sample estimate and a corresponding population parameter that arises because only a portion of the population is observed * Standard errors * The standard deviation or measure of a variability or dispersion of a sampling distribution * Provides a numerical indication of the variation in sample estimates * Sampling distributions. A theoretical non-observed distribution of sample statistics calculated on samples on size N that, if known, permits the calculation of confidence intervals and the test of statistical hypothesis * Describes the mean, variation and shape of the distribution that is based on an independently and randomly drawn population. * It allows researchers to calculate the probability that sample statistics fall within certain distances of the population parameter. * Sampling error * the difference between a sample estimate and a corresponding population parameter that arises because only a portion of the population is observed * Standard error The standard deviation or measure of a variability or dispersion of a sampling distribution * Provides a numerical indication of the variation in sample estimates * Confidence * How much error this is in a sample. The degree of belief or probability that an estimate range of values includes or covers population parameter. * Sample distribution * Describes the mean, variation and shape of the distribution that is based on an independently and randomly drawn population. * Sample size. * 100 (11% +/-) * 600 (5% +/-) * 1000 (4% +/-) * 1500 (3% +/-) 4000 (2% +/-) Observations * Terms * Quantitative vs qualitative data * Quantitative data involves a large amount of data and its involves numeric manipulation * Qualitative data involves small number of cases, more in depth and it relies on quotations, comments, anecdotes and other written evidence to support arguments * Both seek to explain trends or patterns through systematically collected data. * Data collection, how do we choose the best way, characteristics of * How to choose the best way * Depends on the question * Depends on the sample Depends on which constitutes the best measures, validity of the measurements that a particular method will allow. * Depends on cost and availability, some are more observable than others, some less expensive. * Depends on the reactivity to the collection by the population. * Consider ethical implications. * Characteristics * Primary and secondary data * IRB, what role it plays and the significance of it * Institutional review board, a must for any test involving human subjects. * It is guided by 3 principles: respect for persons, beneficence and justice. The role it plays s to ensure that all research conducted will not harm any individuals, recognise that they are autonomous and there is distribution of benefits to participants. To ensure that no ethical boundaries are crossed. * Its significance is that it has removed all unethical research to be done, but this limits the way data is collected. Thus researchers must consider the benefits vs the burden of the research. * Observation, types and pros and cons: basic distinction is direct and indirect. * Direct * Allows researchers to view things in natural setting more often than laboratory. Field study or ethnography. Observation in lab gives more control over environment ‘ * Pros * Natural settings allow people to act normally, increases validity * Researchers can observe for longer periods of time * Lab can give more control. * Cons * Lab can also mean that subjects know they are being observed and thus alter behaviour, low validity * Natural setting cannot control external effects and variables. * Indirect * Observation of physical traces of behaviour, essentially detective work. Inferences are made based on physical traces * Two types of measures * Erosion Created by selective wear on some material. Looking at what has been eroded or removed, looking at what is left. * Accretion * Measures a phenomenon a manifested through the deposition and accumulation of materials. Seeing what has built up. * Pros * It raises less ethical issues than direct * Some materials or traces of them are more/less durable than other, making them more/less easy to measure, depending on the method. * Less obtrusive, much cheaper than alternatives * Cons * There are many threats to validity, prone to measurement problems * Can be difficult to make strong inferences * Participant Most field studies involve this, proverbial fly on the wall. Observing people for long periods of time. Assume a role or identity within the group. * Pros * Gain deep access into the group through informants * Natural setting * Observe for long periods so changes in behaviour can be studied * Has a degree of accuracy and completeness that other methods cannot provide. * Cons * It is not viable for every question, some things are just unobservable, like voting * Lack of control over the environment, inability to isolate factors * May be invalid or biased, going native.
Becoming part of the environment, perceived biased. * Difficulties with replication of study * Non participant * Same as direct. * Overt * Participants are aware of the investigators presence and intentions * Covert * Investigators presence is hidden or undisclosed. * Structured * Investigator looks for and systematically records the incidence of specific behaviours * Unstructured * All behaviour is considered relevant, at least at first, and recorded. * Ethical issues and how to avoid * Negative repercussions from associating with researcher because of the researcher’s sponsors, nationality or outsider status. Invasion of privacy * Stress during research interaction * Disclosure of behaviour or information to the researcher resulting in harm to the observed during or after the study. * How to avoid: ethical proofreading * Assume everything and all identities will be discovered * Look at actual words in manuscript * When describing potentially unflattering things go from general to specific * Be general about community at hand * Realise that data and research will be used again * Know what perspective and attitude is towards subject * Caution subjects multiple times Know study limitations and agreement are in advance * Have other people edit sensitive portions. Document analysis * Types of records, pros and cons of utilization * Two types of records * Running: produced by organisations rather than private citizens, carefully stored and easily accessed and is available for long periods of time. * Pros * Low cost in terms of money and time. * Accessibility * Covers more extensive period in time. * Extensive amounts of records * Many records are digitalised * Cons * At the mercy of those who keep the records, may or may not be biased. Questions on recording keeping practices, may not be kept properly. * Episodic: records that are produced and preserved in a more casual, personal and accidental manner. Diaries and memoirs. Important to political historians. * Pros * Used to illuminate phenomena rather than generate large sample * Provides different perspectives on certain phenomena. * Use qualitatively. * Cons * Gaining access to episodic records can be difficult, locating suitable materials is the most time consuming aspect of data collection * Many are not digitalised. Content analysis, procedures, pros and cons, intercoder reliability * Refers to the use of excerpts, quotes, and examples from recorded documents to support and observation. * Can be both qualitative and quantitative in nature. * Procedures * Decide on appropriate sample, what materials to include in the analysis. * Define the recording or coding units, decide what is actually going to be measured. * Choose categories of content that are going to be measured, deciding the recording unit. How to measure what you want. Code words? Devise a system of enumeration for the content being coded, a numeric system based on what is being coded. Code for number of times X appears. * Pros * Gives researchers access to subjects that would otherwise be difficult to attain. * Raw data are usually nonreactive, no undesired influences on behaviour. * Written records like newspapers have existed over long periods in time * They can be easily attained * It often enables us to increase sample size above what would be possible though either interviews or direct observation . * The cost of keeping records are borne by the keepers, it is low cost for those using them. Cons * Selective survival, record keepers may not preserve all pertinent materials, only selectively. Gaps may exist * Incompleteness. Gaps may exist due to fires losses of other types * Content may be biased. May be incomplete, selectively preserved, inaccurate or falsified * Unavailable to researchers because they are classified. * They lack a standard format because it is kept by different people. * Intercoder reliability * Refers to when two or more coders, using the same procedures, agree on how the content is analysed.
The more times the coders make the same conclusions the same way, the more claims can be made. * Written records, pros and cons * Pros * Can be used when other means like direct observation or interviews are not possible. * Able to use for large scale collective behaviour, cannot possibly interview all. * Raises fewer ethical issues than observation or interviewing. * No risk to individuals as long as they are not identified in records. * Cheaper than other forms. * Not affected by time or history, what is recorded cannot be changed. * Cons * Gaining access may be difficult, classified Records not kept properly, or purposefully changed Survey research and interviews * definition and difference * surveys are a method of data collection that acts as an alternative to experiments or simulations. * Instead of manipulating an independent variable to view its effects, survey design examines the relationship between variables, better at establishing correlation rather than causation. * Types of surveys, pros and cons, * Personal, face to face * High cost * High to medium completion rate * Potentially high sample congruence * Long to medium length * High data processing cost * Telephone Medium cost * Medium completion rate * Medium sample congruence * Medium to short length * High to low data processing cost * Mail * Low cost * Low completion rate * Medium sample congruence * Medium to short length * Medium data processing costs * Email * Low cost * Depends but low completion rate * Low sample congruence * Medium to short length * High to low data processing costs * Internet * Low costs * Depends but low completion rate * Low sample congruence * Medium to short length * High to low data processing costs * Group administration * Very low costs * High once group is convened completion rate Depends on group selection process, sample congruence * Variable length * High to low data processing costs * Drop off/pick up. * Very low costs * Low completion rate * Low sample congruence * Short length * Low data processing costs. * Completion and response rates, significance of * Refers to the proportion of persons initially contacted who actually participate. * When the proportion of the persons contacted who actually participate is too low, the ability to make statistical inferences about the population is limited * Too little responses, cannot make inferences. * Cost * Time and money. Have to balance between costs and completion rate. * Data processing * Answers of the surveys still have to be tabulated. All data need to be coded in a way that a computer can process them. * Problem with open ended questions. Have to be transcribed and coded. * A major reason for adopting internet and telephone surveys. * Sample-population congruence * Refers to how well the sample subjects represent the population from which they are drawn * Biases can enter through the initial selection of respondents or through incomplete responses. * The closer the congruence, the more representative the sample, higher validity. Length * Too long and respondents lose interest or start answering without care, get distracted. * Too short and not enough data is collected. * Well motivated participants can help with the problem. * Response quality * Cannot take the responses at face value. * Participants may not have the same interests or familiarity with certain topics. * People may be reluctant to express their opinions to strangers. * Busy people won’t answer truthfully. * Interview bias * Occurs when the interviewer influences the respondent’s answers, may have a larger effect on telephone surveys than in person surveys. Question types and wording, what to be careful for and why, pros and cons * Close-ended questions * Pros * Easier for people to answer and takes little time * Easier to sort and tabulate data * Sensitive issues are better paired with close ended questions * Cons * Force respondents to choose from a list they may not agree with * Single sided and double sided questions that can affect responses * Single side, agree or disagree with a statement * Two sided, gives two alternative statements to choose. Problems of recall versus recognition, can prompt answers that would have otherwise been blank. People don’t know the answer but recall when seeing the choice. * Oversimplified and distorted picture of public opinion * Open-ended questions * Pros * Allows respondents to state what they know and think. * Good for situations where likely answers are not known. * Cons * Respondents may respond too much or too little. * Problem with recording answers, tedious and difficult to code, interpretations vary as well. * Processing data is time consuming. * Wording Wording is important, question clarity is vital to get valid responses. Objectivity and clarity * Avoid * Double barrelled questions * Two questions in one. Sometimes participants might not agree with first question but is made to in order to answer the second. * Ambiguous questions * One that contains a concept that is not defined clearly. * Participants may interpret the question wrongly. * Leading questions * Reactive question. Encourages respondents to choose a particular response because the question indicates that the researcher expects it. * Giving what the researcher wants, validity is off. Push polls. * Question order and effects * The order in which questions are presented may influence the reliability and validity of answers. Participants any answer differently or stop when they see certain questions. Can solve response set. * Branching questions * A question what sorts respondents into subgroups and directs them into different parts of the questionnaire * Filter questions * A question that screens respondents from inappropriate questions. * Response set * Straight line responding. * May occur when a series of questions have the same answer choices.
Check the first few, then blindly check the rest. * Archives surveys, pros and cons * Existing surveys that have been designed in the past and are readily available to use. * Pros * Very cheap, for those with no access to funding. * Less time is spent designing the surveys. * More reliable and higher quality, widely used thus more reliable. * Well written questions. * Cons * The questions are not what you want. * Interviews, definition, pros and cons. * Interviewing is the act of asking individuals a series of questions and recording their responses.
May be face to face or over the phone. * Pros * Sometimes asking questions is the only way * Elites can provide valuable information, but only through interviews. Focused interviews. * Can provide more comprehensive and detailed information, rich variety of perspectives. * Excellent form of data collection in exploratory studies. * Cons * Can be difficult to administer * Have to take into account non-verbal cues and be willing to go off topic * Detailed note taking * face to face, strong necessity for interpersonal skills.