Cours 2

• Research design


  The overall plan or strategy for conducting a research study
• Main categories of research design

  • Exploratory
  • Descriptive
  • Explanatory
• Research design


  • Outlines the procedures and methods for collecting and analyzing data
  • Includes decisions about sampling, data collection methods, and data analysis techniques
  • The choice of research design depends on the research question, available resources, and nature of the research problem
• Exploratory research


  • Conducted when the researcher has limited knowledge or information about the research problem or topic
  • Goal is to gain a better understanding of the problem, generate new ideas, and develop hypotheses for further research
  • Involves a wide range of data collection methods, such as literature reviews, interviews, surveys, and focus groups
  • Data collected is typically qualitative and is analyzed using various techniques, such as content analysis, grounded theory, and thematic analysis
  • First step in the research process and can provide valuable insights for future studies
• Exploratory research examples


  • A psychologist conducts exploratory research through interviews with doctoral students and literature reviews to investigate potential factors contributing to increased stress and anxiety levels among college students
  • A researcher is interested in studying the experiences of first-generation college students but has limited knowledge of the topic. They conduct exploratory research through interviews with students and literature reviews to develop a better understanding of the topic
• Descriptive research


  • Describes and summarizes characteristics of a population or phenomenon
  • Focuses on what exists and how it exists, rather than why
  • Uses observational or survey methods to collect data
  • Involves statistical analysis to describe and summarize the data
  • Results can inform further research or guide decision-making

  
• Explanatory research


  • Investigates the cause-and-effect relationship between variables
  • Aims to explain or predict the relationship between two or more variables
  • Involves the use of experimental or quasi-experimental designs
  • May involve both quantitative and qualitative data collection and analysis methods
  • Results can help to develop theories or models that explain observed relationships between variables
• Forms of research


  • Experiment
  • Quasi-Experiment
  • Correlational Study
  • Laboratory vs Field Studies

  
• Internal validity


  • Refers to the extent to which a study measures what it is intended to measure
  • Results can be interpreted in a causal manner
  • Is affected by factors such as research design, sample selection, and data collection methods
• External validity


  • Refers to the extent to which the findings of a study can be generalized to other populations or settings
  • Is affected by factors such as sample representativeness, research context, and ecological validity
• Factors that still speak in favor of laboratory investigations:
• Operationalization


  • After determining which characteristics are to be recorded, operationalization determines how the variables are to be recorded
  • Selection of a data collection method
  • Determination of the measurement instruction and scale level
• Measurement


  • An assignment of numbers to objects or events, provided that this assignment is an algebraically defined mapping of an empirical relative to a numerical relative
  • A scale refers to an empirical relative, a numerical relative, and a mapping function that links the two relatives
• Types of measurement scales

  • Nominal
  • Ordinal
  • Interval
  • Continuous
• Nominal scale


  • Categorical measurement scale
  • Variables assigned to distinct categories or labels
  • No inherent order or ranking
  • Purpose is to classify or categorize objects or events based on their characteristics or attributes
• Ordinal scale


  • Measurement scale where variables are ranked in a particular order or sequence
  • Purpose is to measure the relative position or rank of variables
  • Differences between the variables are not necessarily equal
• Interval scale


  • Measurement scale where variables are measured on a scale with equal distance between each point
  • Variables can be quantified in terms of distance and can be added or subtracted
  • No true zero point on an interval scale, meaning that ratios cannot be calculated
• Ratio scale


  • Measurement scale where variables are measured on a scale with equal distance between each point
  • Variables can be quantified in terms of distance and can be added or subtracted
  • Has a true zero point, meaning that ratios can be calculated
• Examples of scale levels and possible statements

  • Equality/Difference: "A woman is not a man."
  • Greater/Smaller Relations: "2nd place is better than 3rd place."
  • Equality of Differences: "The temperature increases by 10°C." (from 10 to 20°C as well as from 80 to 90°C)
  • Equality of Ratios: "I am twice as heavy as you."
• Methods of obtaining data

  • Interviews
  • Questionnaires
  • Observation
  • Rating Scales
  • Simulations
  • Non-reactive procedures
• Interviews


  • Positives: Opportunity for in-depth exploration of topics, Ability to gather rich and detailed information from participants, Flexibility in questioning and adapting to individual interviewees, Potential for building rapport and trust with interviewees
  • Negatives: Potential for interviewer bias and subjectivity, Difficulty in analyzing and generalizing findings, Possible lack of generalizability to larger populations, Time-consuming and resource-intensive
• Interview example


  • Unstructured interview
• Questionnaires


  • Positives: Cost-effective and efficient way to collect data from a large number of participants, Standardization of questions and responses allows for greater comparability across participants and studies, Participants may feel more comfortable answering sensitive questions in a private setting, Ability to collect data anonymously or with minimal personal identifying information
  • Negatives: Limited opportunity for in-depth exploration of topics or follow-up questions, Potential for response bias, such as social desirability bias or acquiescence bias, Possible lack of context or nuance in responses, Difficulty in confirming the accuracy or honesty of responses
• Unstructured interview


  • Figure 1
• Questionnaires


  • Cost-effective and efficient way to collect data from a large number of participants
  • Standardization of questions and responses allows for greater comparability across participants and studies
  • Participants may feel more comfortable answering sensitive questions in a private setting
  • Ability to collect data anonymously or with minimal personal identifying information
• Questionnaires


  • Limited opportunity for in-depth exploration of topics or follow-up questions
  • Potential for response bias, such as social desirability bias or acquiescence bias
  • Possible lack of context or nuance in responses
  • Difficulty in confirming the accuracy or honesty of responses
• Likert type scales


  • Figure 2
• Observation


  • Direct observation of behavior in a natural or controlled setting
  • Opportunity to gather data on behavior that may be difficult to assess through other methods
  • Ability to collect data in real time and in the actual setting of interest
  • Potential for high ecological validity

  
• Observation


  • Possible reactivity of participants to being observed
  • Difficulty in establishing causality or determining the reasons behind observed behavior
  • Possible lack of generalizability to larger populations or settings
  • Difficulty in controlling for extraneous variables that may influence behavior
• Simulations


  • Ability to simulate complex real-world scenarios in a controlled setting
  • Opportunity to study and manipulate variables that may not be possible or ethical to manipulate in real life
  • Potential for high internal validity, as extraneous variables can be controlled
  • Possibility of collecting data on outcomes that may take a long time to occur in real life
• Simulations


  • Possible lack of external validity, as results may not generalize to real-world settings
  • Potential for participants to respond differently to a simulation than they would in a real-life scenario
  • Difficulty in simulating complex and nuanced situations
  • High cost and time investment in creating and administering simulations
• Non-reactive procedures


  Methods of collecting data that do not involve direct interaction with participants or manipulation of the environment, and therefore do not have the potential to affect or bias the data being collected
• Non-reactive procedures


  • Content analysis of written or recorded materials (e.g. books, articles, speeches, TV shows, social media posts)
• Sample


  A selection of units to be investigated, used to describe the population of interest when measurement of the complete population is impossible or too costly
• Good samples


  • Representativeness - resemble the population (target population) in as many (relevant) characteristics as possible
• Sample size does not equal representativeness
• Representativeness is particularly important for absolute statements (e.g., in election polls) but less important for causal hypotheses, where broad variability is more important
• In psychology, convenience samples are often used
• Representative samples


  A prerequisite for replacing complete enumeration surveys with sampling surveys
• Random sample


  Every object belonging to the population has the same probability of being selected in the sample