Social Science Research Methods Notes

Sampling

• Definitions:
  • Sample: A subset of cases selected from a larger set/population.
  • Population: An abstract concept from which researchers draw samples and generalize results.
• Purpose of Sampling:
  • To create a representative sample of the population.
  • Requires mathematical methods (probability sampling) for accuracy.

Sampling Strategies

• Representative Sampling: Aims to mirror the larger population.
  • Includes nonprobability sampling (convenience, quota) and probability sampling (simple random, systematic, stratified, cluster, random digit dialing).
• Non-Representative Sampling: Serves purposes other than representation.
  • Uses nonprobability sampling (purposive, snowball, deviant case, sequential, theoretical, adaptive).

Types of Nonprobability Sampling

• Convenience Sampling: Selecting samples based on accessibility.
  • Advantage: Inexpensive and quick.
  • Disadvantage: May not represent the broader population.
• Quota Sampling: Identifying general categories and setting quotas.
  • Advantage: Ensures some diversity.
  • Disadvantages:
    • Limited to specific aspects.
    • Predetermined number of cases might not reflect actual proportions.
    • Relies on convenience sampling within each category.

Probability Sampling: Key Terms

• Sampling Element: The individual case or unit to be sampled.
• Target Population: The broader group to which conclusions are generalized.
• Sampling Frame: A list of all cases within the population.
• Sampling Ratio: The proportion of cases in the sample relative to the population.
• Parameter: A characteristic of the entire population.
• Statistic: An estimate of a population parameter derived from a sample.

Probability Sampling: Random Sampling

• Every sampling element has an equal chance of being selected.
• Sampling Error: The difference between the sample and the overall population.

Probability Sampling Methods

• Simple Random Sampling:
  • Procedure: Create a sampling frame and use a purely random selection process (e.g., random number table).
  • Central Limit Theorem: Repeated random samples will form a normal distribution, with the center resembling the population parameter.
  • Confidence Interval: Estimated range for the parameter using the sample.
• Systematic Random Sampling:
  • Procedure: Select every $x^{th}$ case in the sampling frame using a sampling interval.
  • Weakness: Can be unrepresentative if the data has a cyclical pattern.
• Stratified Sampling:
  • Procedure: Divide the sampling frame into categories and randomly sample from each category.
  • Use: When specific categories make up a small fraction of the population.
• Cluster Sampling:
  • Procedure: Multi-stage sampling where aggregated units are randomly selected, then samples are randomly drawn from those units.
  • Use: Covering broad geographical areas.
  • Probability Proportionate to Size (PPS): Adjustment in cluster sampling for unequal cluster sizes.
• Random Digit Dialing:
  • Procedure: Randomly selecting phone numbers from the list of all potential numbers.

Non-Representative Sampling

• Deviant Case Sampling: Selects cases that differ significantly from the dominant pattern.
• Sequential Sampling: Selects cases until no new information emerges.
• Theoretical Sampling: Selects cases to reveal theoretically important features.
• Adaptive Sampling: Uses multiple sampling stages (e.g., snowball followed by purposive) for hidden populations.

Qualitative vs. Quantitative Sampling

Survey Research

• History:
  • Domesday Book: An early census in England (1085).
  • Social Research Surveys: Started in 19th century England (e.g., Henry Mayhew's London Labour and the London Poor).
• Survey Questions:
  • Behavior: Frequency of actions (e.g., brushing teeth).
  • Attitudes/Beliefs/Opinions: Views on topics (e.g., political opinions).
  • Characteristics: Marital status, etc.
  • Expectations: Future plans (e.g., buying a car).
  • Self-Classification: Perceived social class.
  • Knowledge: Awareness of events/facts (e.g., election results).

Steps in Survey Research

• Preparation:
  • Develop hypotheses.
  • Determine survey type.
  • Write survey questions.
  • Define response categories.
  • Design layout.
• Physical Preparation:
  • Plan data recording.
  • Pilot test instruments.
• Sampling:
  • Define target population.
  • Determine sampling frame.
  • Determine sample size.
  • Select sample.
• Fieldwork:
  • Find respondents.
  • Conduct interviews.
  • Record data.
• Data Input:
  • Enter data into the computer.
  • Recheck entries.
  • Perform statistical analysis.
• Presentation:
  • Explain methods and findings in a research report.
  • Present findings to receive feedback.

Sources of Survey Error

• Respondent selection.
• Answering questions.
• Survey administration.
• Thinking errors.

Writing Good Survey Questions

• Use related questions.
• Core Principles: Avoid confusion and consider the respondent's perspective.
• Things to Avoid:
  1. Jargon, slang, abbreviations.
  2. Ambiguity, confusion, vagueness.
  3. Emotional language and prestige bias (association with respected figures influencing responses).
  4. Double-barreled questions (addressing multiple issues).
  5. Leading questions.
  6. Questions beyond the respondent's capabilities.
  7. False premises.
  8. Future intentions.
  9. Double negatives.
  10. Overlapping/unbalanced response categories.

Respondent Recall Issues

• Factors affecting memory of past events:
  • Topic relevance,
  • Intervening events,
  • Personal importance.
• Telescoping: Respondents may report recent events more readily and underreport past events.
• Techniques to Reduce Telescoping:
  • Situational framing.
  • Decomposition.
  • Landmark anchoring.
  • Bounded recall (for panel surveys).

Ensuring Honest Survey Responses

• Create comfort and trust.
• Careful word choice.
• Build context.
• Ensure anonymity.
• Technologies:
  • Computer-assisted self-administered interviewing (CASAI).
  • Computer-assisted personal interviewing (CAPI).
  • Randomized response technique (RRT).
• Social Desirability Bias: Respondents answering in socially acceptable ways rather than truthfully.

Knowledge Questions

• Respondents may exaggerate knowledge.
• Sleeper Question: Including fake events/people to check honesty.
• Types of Questions:
  • Contingency Question: Directs respondents to different questions based on their answers.
  • Open-Ended Question: Allows free-form responses.
  • Closed-Ended Question: Provides fixed response options.

Issues in Survey Responses

• Swayed Opinion: Incorrect answers due to social biases/sensitive topics.
• False Positive: Respondents falsely claiming knowledge.
• False Negative: Respondents concealing answers.
• Neutral Position: “No answer” or “No opinion” options.
  • Satisficing: Choosing neutral options to minimize cognitive effort.
  • Mitigation: Standard format, quasi-filters, full filters.
• Floaters: Respondents providing answers despite lacking knowledge.
• Recency Effect: Selecting the last answer choice when unsure.
• Selective Refusals: Refusing to answer due to sensitivity.

Presenting Survey Questions

• Use a spectrum of answer choices rather than binary options.
• Fairly present all alternatives.
• Wording Effect: How question wording impacts responses.

Questionnaire Design Considerations

• Length: Tailored to the respondent and survey type.
• Question Order: Arrange questions logically to minimize discomfort or confusion.
• Order Effects: Earlier questions influencing later answers.
• Context Effects: The setting/topic influencing interpretation.
• Funnel Sequence: From general to specific questions.
• Layout: Clear, organized, and uncluttered.
• Question Format: Survey form, checklist, bulletin, etc.
• Matrix Question: Combining questions with the same answer choices into a matrix.

Nonresponse

• Components: Location, contact, eligibility, cooperation, completeness.
• Leverage Saliency Theory: Varying importance of aspects (time, topic, sponsor) influencing participation.
• Tailoring: Adapting interview questions based on respondent cues.

Types of Surveys

Survey Interviewing

• Based on the naive assumption model (ideal, without communication problems).
• Conversational Interview: Collaborative model adapting to the respondent's understanding while maintaining research goals.
• Interviewer Role: Asking questions, explaining the study, and providing directions.
• Interviewer Tasks: Introducing oneself, asking questions, and accurately recording answers.
• Probes: Follow-up questions to get specific responses when initial answers are unclear.

Interviewing Bias

• Forms of Bias:
  1. Error from respondents.
  2. Unintentional interviewer errors.
  3. Intentional subversion.
  4. Influence of interviewer expectations.
  5. Failure to explore data.
  6. Influence of interviewer appearance.
• Cultural Meaning & Survey Interview
  • Positivist: interviewer must adhere to strict standards to avoid bias while maintaining a netural and uniform attitude.
  • Non-Positivist: human interactions are dynamic. Advocate for the collaborative encounter model where the researcher and respondent reach a collaborative agreement to generate responses of significant value.

Pilot Testing and Cognitive Interviews

• Pilot Testing: Testing questionnaires before implementation.
• Cognitive Interviews: Asking respondents about their thoughts while answering questions to evaluate and refine the survey.
• Ethics in Survey Research:
  • Privacy violation.
  • Voluntary participation.
  • Exploitation of surveys (pseudosurveys).
  • Misuse of survey results.

Nonreactive Research

• Methods not involving direct interaction with participants.
• Uses unobtrusive measures (techniques that don't intrude on those studied).
• Erosion: Measuring usage by assessing wear on surfaces.
• Accretion: Measuring by examining residue/waste.
• Follows a quantitative framework (conceptualizing constructs, linking constructs to nonreactive measures).

Coding and Measurement in Content Analysis

• Uses a coding system (procedure for transforming symbolic content into quantitative data).
• Measurement through structured observation (systematic and organized observation/documentation).
• Coding Aspects: Frequency, direction (support/oppose), intensity, size.
• Types of Coding:
  • Manifest Coding: Observing visible surface content.
  • Latent Coding/Semantic Analysis: Identifying underlying meanings/themes.

Coding Problems and Process

• Intercoder Reliability: Consistency across different coders.
• Content Analysis Steps:
  1. Formulate research question.
  2. Determine unit of analysis.
  3. Develop sampling plan.
  4. Construct coding categories and recording sheets.
  5. Check coding and intercoder reliability.
  6. Collect and analyze data.

Existing Statistics/Documents & Analysis

• Documents are usually available publicly.
• Social Indicator: Quantitative indicator of societal wellbeing used by policymakers.
• Secondary Analysis: Analyzing previously collected data.
  • Weakness: Potential mismatch with current research context.

Problems of and Concerns with Secondary Analysis

• Problems:
  • Flawed Unit.
  • Ecological Fallacy.
  • Validity.
  • Reliability.
  • Missing data.
• Ethical Concerns:
  • Inferences and Testing.
  • Content doesn't generalize to effects on readers.
  • Privacy and confidentiality.

Quantitative Data Analysis

• Before data can be analyzed, it must be organized for computer analysis.
• Need to quantify open-ended survey questions.
• Coding Procedure: Rules for assigning numerical values to variable attributes.
• Codebook: Document storing these coding procedures.
• Raw quantitative data must be organized and manipulated for academic usefulness.

Data Entry

• Data typically entered into a computer in a grid.
• Grid Components:
  • Data Records: Variables of a case (arranged horizontally).
  • Data Field: One or more columns in a database representing a variable's location information.
• Entry Methods:
  • Code Sheet: collect and transfer information to grid.
  • Direct Entry Method: Inputting data directly into a computer during research.
  • Optical Scan: Survey responses read by computer.
  • Barcode: Convert collected information to barcodes read by the computer.

Data Checks

• Random data samples (10-15%) verified to ensure validity.
• Methods:
  • Possible code cleaning/wild code checking: Checking impossible codes.
  • Contingency Cleaning: Verifying logical code combinations.

Types of Quantitative Data

• Nominal: Can be categorized.
• Ordinal: Can be categorized and ranked.
• Interval: Can be categorized, ranked, with equal intervals.
• Ratio: Can be categorized, ranked, equal intervals, and has absolute zero.

Data with a Single Variable (Univariate)

• Descriptive Statistics: Explains data patterns.
• Frequency Distribution: Distribution of cases into categories within a single variable.
  • Used to describe single variable data.
  • Can be shown through numbers or percentages.
  • Includes histogram, bar chart, and pie chart.

Measures of Central Tendency

• Reduce data from one variable to one value.
• Three Measurements:
  • Mode: Most frequent data.
  • Median: Middle data point.
  • Mean: Data average.
• Normal distribution:
  * where mean, median, and modus are within a similar range.
• Skewed distribution:
  * where there are larger gaps between mean, median, and modus.

Measures of Variation

• Dispersion illustrating data distance from its center.
• Measurements:
  • Range (largest - smallest).
  • Percentile (divides data into 100 parts). Concept similar to median.
  • Standard Deviation: Average collection distance from the mean.
  • Z-score/Standard Score: distance of data from the mean standard deviation unit.

Bivariate Relationships

• Analysis can indicate statistical relationship (covariation or statistical independence).
• Statistical Relationships:
  1. Covariation: variables have a relationship between them.
  2. Statistical Independence: variables do not have a relationship with each other.

Scattergram

• A graph of each case, in which:
  * Each axis represents a variable.

• Can describe interval and ratio variables but cannot describe nominal variable.

• Main characteristics of Bivariate Relationships:

  1. Form: Describes form of data, can include lines or curves.

• Direction: Linear relationship can be positive or negative.

• Precision: Data is more precise if Bivariate point are closer and less precise if they are more distant from the form represented in the data.

Bivariate/Contingency Table

• Can describe correlation among two variables with percentage.
• Based on the tabulation process.
• Can be in numbered form, or in the form of a percentage.
• Before graphing table, data is suggested to be grouped into short forms.
• percentage is calculated base on row or columns.
• Marginal total is the amount of data in one row /column.

Association Measurement Methods

• Measurements that express the relationship between 2 variable.
• Use the proportionate reduction in error/ knowing one variable reduces error in predicting another.
• Has 5 measurement methods:
  1. Lambda (λ):Measures 2 nominal relationships
  2. Gamma (γ): Measures ordinal relationships.
  3. Tau (τ): Measures ordinal relationships
  4. Rho (ρ): Measures interval or ratio relationships/ used in statistical study.
  5. Chi-square (χ2): Measures ordinal or nominal relationship

Multivariate Data

• Statistical Control: Alternative variables to explain relationships.
• In non-experimental research, statistical control uses a control variable to explain any impact on the relationship.
• If a control has no effect, than its a coherent relationship.
• If a control has an effect than its not a coherent relationship.
• measurements are good for indicating relationship between 2 variable.
• Net Effect: a variable independent of another.

Multivariate Data: Multiple Regression Analysis

• Data used in this form include ratio or interval studies.
• Usefulness:
  • Helps measure variables more accurately than data without independence measure
  • Explains variables effect to any other variable.

Inferential Statistics

• Used to test Hypotheses and to determine any large, broad range relationships outside of data that wasn't by simple Chance.
• Uses a probability theory.
• Significance:
• can find any relationship in smaller sample size
• used to estimated data scale
• Significance level used to measured any chance in relationship when not with the whole population.

Inferential Statistic Error

• Type I: can relate data that isn't together, in the process rejecting initial data.
• Type II: Can't relate proper data, when approving original data.

Field Research

• Focus to small communities whether social or ethnic.
• Focus on small size communities and ethnic backgrounds when studying topics.

Ethnography

• Field research emphasising descriptive culture to further understanding people groups internally for better study.
• Goal is to understand groups and their behavior from direct sources.
• Thick Description is data collected that studies and expresses the culture and social of groups of people.

Ethnomethodology

• Combines social aspects to knowledge , to expand relationships.
• A social normal is a weak standard that isn't to be taken seriously , therefore the only way to prove it is to violate the normal.

Study Logic

• Field research has integration based on a natural principle. People that are social follow the natural way of a natural society.

Stages in Research Study

• Requires preparation and adaptability.
• Must have a location in mind to start interacting.
  • Field Site: location and area of study.
    • Gatekeeper: A social expert who is aware of a areas environment.

Study Adaptability

• Field workers must study general events that are none controversial as a starter event.
• Access Ladder: field workers learn more about controversial events over time.
  • Field workers must adjust as they are on site to the area and cultural norms.
  • Don't immediately make relationships because of events being subject to quick change.
  • You can integrate by charm or trust, if you are trusted other locals will further give you more insight and information.
  • Freeze outs happen , where members don't cooperate with new comers.
  • Always maintain a look of interest at the location.
  • Be incompetent and ask questions to gain insight to cultural and social factors.

Data Recording and Collection

• Field data is what is remembered, recorded and looked into around the current area.
• Must look into local and verbal symbology to further integrate. Also to avoid coming off as rude or none respectful.

Interview Questions

• There are 3 types of interview questions.
  • Descriptive:
    • interview process to explore environment to further the research.
  • structural:
    • Can analyse data after a time of visiting
  • contrast:
    • Built on the contrast and data to verify. Focus on the similar and unsimilar of events.

Informants

• Informants are members of the community who discuss and further insight with outsiders.
• 4 characteristics: witness of cultural events, are currently in the field of events, willing to make time for research and unanalytical.

Reliability and Analysis.

• Internal consistency is key to inspecting a environment to see any key factors. Or to notice any attempt of deception.
• Field test and analysis helps make the process accurate and clear.
• Hardships to consider are:
  • Misinformation attempts, evasion and normal lies. Cultural backgrounds.
• Groups can perform actions of performance to hide or change original appearance, to the eyes and ears of visitors.

Validity in research and testing

• 4 types of validity:
  • Ecological: the reality of an event by making sure visitor doesn't interrupt.
    • Natural History: detail action of project
    • Member Validation: letting visitors confirm what the research as said.
    • Competent performance: make sure visitor is compatible to current events.

Ethics To think about

• Are cover up needed?
• keep it private.
• Avoid interaction with bad people.
• keep a good balance from outsiders and local.
• Can field work be a successful publishing?

Focus Groups

• Interview people informally by discussions.
• Advantages: expressions are encouraged and member feel like they are more in power.
  Focus groups are more flexible.

Limitations Of Groups

• Group can be to extreeme in behavior.
• A group can only study limited area.
• moderators can interrupt in group study.

Qualitative Data Analysis

• Concepts are general and can have multiple meanings, so evidence is important.
• New concepts must be updated over time.
• Must be categorized to better analyse.
• Goal is to look at general things in each aspect.

Coding of Groups:

• Coding involves adding labels to data and making concepts based on data categorization.
• Process:
  • Look at relevant date
  • Make relationships to each other.
  • Scan Qualitative studies

Memo Analysis:

• Is writing ideas for new studies ,as data is found.
  Creates new Hypothesis and new studies.
• Can better the relationships of other variables.
  • Ideal type can be theorised and be used as a standard during study
  • Approximation can be based on data for theories to be in effect. This allows for data adjustment in the event a new theory is found.

Analytical and Social comparisons and studies

• Comparing 2 theories to fine the real effect of a social or analytical scenario.
• Can use a analysis that helps picture a environment and study cultural domains.
• Combine theoretical description to description.
  • Can then be broken down to Time periods for study.

Data Study

• Method to make things better by using scenarios to perfect study bettered by past attempts.
  • Includes maps network and software for further study.

Software for studies:

• Software:
  • Text retrieval.
    • Text based processor.
    • Code bases
    • Theory bases
  • Compared Analyses
  • conceptual networks.

Similarities from past attempts

• can better detailed study.
• Study can come down to data.
• Can cause data error.

Differences that may cause error

• Data being numerical
• Different methods of data or environmental study
• Data study during the event

Social Studies:

• Reason for learning- Record the process for future study.
• Project descriptions for the future.
• Data presentation/ Discussion

Writing Process:

Goal - Learn to write and get to know more people.
Style - Be professional.
Tone -Be serious and to the point.

    ## Writing help:

-> helps break down and narrow ideas.
*Organize thoughts into categories to specify topics.
*Search the web- or Library to extend source and focus on a idea.
*Avoid plagiarism
-Paraphrasing with credits avoid copy and paste.

Writing Process:

1.Writing Prewriting.
*Outline your brain storm to have a organized citation..
2. Composing process
*Write down ideas as fast as you can and go back to make changes.
3. Rewriting process
*Edit by adding content, proofreading, correcting grammer/spelling/clarity and making improvements.

Data Process:

-> Summarize all important information relating all topics in a paper.
Present a statement and a testing process.

 ## Things to discuss:
* Data to hypothesis.
*Explain feeling/ make discussion
*Discuss alternatives that are possible

Qualitative Reports :

### Types:
* Field study can balance error and theorize with real study.
* historical- comparative.
    *Can take a story or describe data .
    *Or use graphs

Research study

• can determine the best to attempt the plan , goal and vision to conduct research.
• Can make contracts between two parties.
  Models for science and social relevance.