Field Methods Reviewer.pdf

Field Methods Reviewer Exam Details: Section Day Date Time Venue BSP Monday Sept. 30 9AM Classroom 84K BAP Monday Sept. 30 12:30PM Classroom 83 Exam Sources: Quizzes, Slides, Myers & Hans...

Field Methods Reviewer Exam Details: Section Day Date Time Venue BSP Monday Sept. 30 9AM Classroom 84K BAP Monday Sept. 30 12:30PM Classroom 83 Exam Sources: Quizzes, Slides, Myers & Hansen, Class Activities Exam Type: 80 Questions: 75 Multiple Choice, 5 True or False Topics TOPICS TOTAL 1 Research Ethics and Standards 6% 2 Nonexperimental Research Designs 26% 3 Quasi-Experimental Designs 9% 4 Sampling Methods 5% 5 Survey Design 40% 6 Correlation Analysis 14% TOTAL 100% Question Type  Nonexperimental Psychology Internal Validity If we can demonstrate that changes in behavior were caused by treatment. External Validity If findings can be applied outside the research setting. Phenomenology is the description of an individual's immediate experience. While this is often used in early psychological studies, the accuracy and objectivity is questionable. It also cannot explain behavior. It is often used with other research methods. Empirical Phenomenology relies on own experiences or data from other sources. It relies on data based on self-reflection on experiences relevant to the phenomenon of interest, descriptions of their experiences of the phenomenon, and accounts from culture and other research. Case Studies are used to study individuals. It is a source of inferences, hypotheses, and theories. It allows the study of rare phenomena such as deviant case analysis. This is often used in Clinical psychology. It may help develop therapy techniques. They are exceptions to ideas and practices. They have persuasive and motivational value. They can be compelling. Advertisers also use them to sell products. However retrospective data is reliant on memory, which is questionable. Field Studies such as Naturalistic Observation Studies is observing behavior in natural settings meant to answer specific questions. Archival Study is where researchers reexamine data. Qualitative Research Words as data rather than numbers. In qualitative research, bias can be used to help us. Survey Research is useful for gathering large amounts of data efficiently. Ethical issues in Survey research may involve Breach of Confidentiality and Sensitive Questions.  Constructing Surveys Closed Structured Questions have limited alternatives and are "easier" to quantify. Open-Ended Questions can be analyzed using content analysis or thematic analysis, which evaluates and categorizes the content of each answer. Designing Survey Items and Question Construction Keep items simple Keep people involved Avoid double-barreled (compound) questions Response choices must be exhaustive Content / Thematic Analysis Frequency Count Coding Categorizing Measuring Responses - Level of Measurement Nominal Ordinal Interval Ratio Selecting Levels of Measurement Psychological variables lead themselves to Continuous Dimension of traits, attitudes, and preferences. A semantic differential or Likert scale could be used here. Survey Items Start with easy and interesting items. Put demographics at the end. Utilize commonly used response options for sensitive demographic questions e.g. income range. For college students, age is not that sensitive yet. Word questions sensitively. Response Styles are people's tendencies to answer in a specific way. Willingness to answer decreases whenever people feel unsure about a question. Position Preference - answering a certain answer all the time. Manifest Content focus on the plain meaning of words. Response Acquiescence is when people mostly answer yes. Response Deviation is when people mostly answer no. Solve this by designing questions that make people think of the answer. Context Effects is when the position of a question influences how the question is interpreted. Solve this by including buffer items. Collecting Survey Data Questionnaires are a structured list of questions. Computer and Internet Surveys Response rate is 44% (Wu, Jiao, & Fils-Aime, 2022). Sending surveys to a clearly defined population positively impacts the online survey response rate. In addition, pre-contacting potential participants, using other types of surveys in conjunction with online surveys, and using phone calls to remind participants about the online survey could also yield a higher response rate. The use of incentives did not show a significant impact on the response rate of online surveys. Include a small gift. (If you have resources) provide incentives such as a prize e.g. certificate, tickets. Provide Simple and Clear Instructions Informed Consent: Protect confidentiality and identity. Self-Administered Questionnaires: Reactivity due to presence of observer, allow anonymous collection and fill up questionnaire in private. Group surveys may mean participants may be embarrassed or will not take your data seriously. Telephone Surveys Obtains high-quality answers. Calling costs money. People hate getting random calls. You need to train interviewers. Face to Face Interview Surveys Expensive in terms of time and money May be structured or unstructured interviews. Must have trained and respectable researchers, they can influence interviewee responses Establish rapport May be used for qualitative research Focus Groups: A group is facilitated by a facilitator and discusses a topic. Sampling Probability Sampling (See Slides) Nonprobability Sampling Correlation A correlational study measures the degree of the relationship between two traits, behaviors or events. Causation cannot be determined by a simple correlation. Pearson product moment correlation coefficient indicates the strength (weak vs. medium vs. strong) and direction (positive vs. negative) of a relationship. e.g. Positive vs. Negative Correlation Regression line is the line. of best fit The equation shows the linear relationship between scores r2 is the coefficient of determination. However, the correlation may be: Curvilinear (Anxiety & Performance) Range Truncated e.g. Limited by range Outliers Correlation does not imply causation (See: Spurious Correlations). Instead we may use the term 'predicts', “associated with”. Linear Regression Estimating a score based on the score of the other variables. There are only two variables. Factor Analysis A correlational procedure used to measure a large number of items and see the relationship among many traits or behaviors. Multiple Correlation When a variable can be predicted by many other variables. R uses information represented by many variables. Partial Correlation The influence of one variable held constant while computing the correlation between the other two e.g. controlling for confounding variables. Multiple Regression When you analyze the relationship of many variables together. Causal Modeling Path Analysis is a model of possible causal sequences using many related behaviors. Cross-Lagged Panel Design Quasi-Experimental Designs When you can't fully manipulate or control experiments, you can the effect of characteristics without manipulating them. HOWEVER, a causal inference cannot be inferred. Ex Post Facto "After the Fact" Studies Ex post facto studies look at differences between groups. However, the differences between groups are based on their characteristics. There is no manipulation and no random assignment. The preexisting differences are the "manipulation" because there are , things you SHOULD NOT manipulate: Death, divorce, illness. Nonequivalent Groups Design: Compare treatment conditions on pre- existing groups of participants. Cannot control the groups but can control the 'manipulation'. You can try to show the groups are as equivalent as possible. Longitudinal Design: Same subjects are measured over different points in time, usually for studying development. Cross-Sectional Studies: Different subjects are measured at a single point in time. Pretest/Posttest Design: Measuring subjects before and after the treatment. “Mag-aral kayo. [Please study.]” (Magno, 2024).