Research Ethics, Methodologies, and Data Analysis in Social Science

Ethics

rules and standards that regulate the conduct of a person or profession. Most proffesionall associations have formal codes of conduct.

Institutional Review Boards (IRBs)

Review research proposals to ensure that the risk faced by human subjects is minimal. Without approval, you can't publish research in a peer reviewed journal or get funding to support it.

Tuskegee Study

An unethical US public health service experiment where 3999 low-income African American men with syphilis in Alabama were deceived into believing they were receiving free treatment for bad blood. Researchers knowingly withheld treatment, including penicillin, to study the disease's natural progression.

Voluntary Participation

No one should be manipulated/forced into participating ex) students filling out surveys for professors shouldn't affect their grade.

No harm

principle that research shouldn't injure or embarrass the participant, or endanger their lives, friendships or jobs. Risk should be minimized

Informed Consent

Participants should have a full understanding of all possible risks, explained verbally and in writing. Participants must give consent.

Milgram Obediance Experiement

A study measuring the willingness to obey authority found that even when commanded to perform actions that conflicted with their personal conscience. Teachers believed they were delivering electric shocks ranging from 15 to 450 volts to a learner for incorrect answers.

Confidentiality

While researchers can identify respondents, they promise not to share information publicly. (In qualitative research it's common practice to use fictional place names to mask identities)

Anonymity

Neither the researcher/reader of the findings can identify the participants. More secure than confidentiality. No way to trace responses back to them.

Honesty

The principle that covers the analysis and reporting of data. Expected to admit mistakes and potential shortcomings.

Deception

Not revealing your identity as a researcher or not revealing true purpose of research. Is allowed but needs to be justified.

Asch Line Experiment

studied conformity and the effect of group pressure.

Politics of Research

deals with social implications and use of research. No formal codes of accepted political conduct. Should be aware of potential political implications of research/ interests that could compromise objectivity.

A research question should be:

Precise, Empirical, Practical. Spcoplogiclaly Relevant.

Quantatative Research

goal is to obtain a random sample that allows fro generalization to a population

Qualitative Research

goal is to obtain in-depth information about members of a population.

Example of ________________

Job loss → eviction → homelessness

Rent increases → eviction → homelessness

Health problems → job loss → medical costs → homelessness

qualitative research

Example of _____________

Job loss → 35% increased risk of eviction → higher likelihood of homelessness

Rent increase → higher eviction rates → homelessness

Health problems → medical costs → increased probability of housing loss → homelessness

Job loss → strong social support → lower probability of homelessness

quantitative research

Qualitative Research Sampling

smaller sample sizes, focus on information rich cases, more time consuming.

Purposive Sampling

based on specific characteristics relevant to the study, rather than randomly, in order to gain in-depth understanding of a particular group or process. based on collecting qualitative samples.

Convienience sampling

relies on whoever is available, useful in the early stages of research. Is nonrandom, so results can't be generalized to the population.

Homogeneous Sampling

Researchers choose participants who are very similar in important characteristics.

Theoretical Case Sampling

Researchers change or expand the sample during the study

Case Selection in qualitiative research

Process of choosing cases that allow researchers to deeply explore a research question. Researchers consider:

Within Case Sampling

A qualitative sampling strategy where researchers collect data from multiple people, events, or sources within a single case in order to gain a deeper understanding of that case.

Between Case Sampling

A qualitative strategy where researchers compare two or more cases along theoretically important dimensions to understand how differences in conditions affect outcomes. Cases often similar in most ways but different in one key factor.

logic of Explanation in Grounded Theory

Research approach where theory is developed from patterns, themes and categories found in data rather than starting with a fixed theory. Researchers repeatedly analyze data, using induction and deduction to build explanations in data.

Snowball (chain referral) sampling

A sampling method where initial participants recruit/refer additional participants, causing the sample to grow over time, often used to study hard to find or hidden populations.

Quota Sampling

researcher sets proportions of key characteristics in the sample to reflect population diversity, though it is not fully representative.

Ethnography

A method of naturalistic observation and participation that aims at understanding a group or society on its own terms. Seeks an emic understanding of the group rather than etic's understanding. Yields rich data but research risks influencing participants actions. Aims at description rather than explanation.

Informants

Someone within a group is willing to share their knowledge, To help the researcher understand the community and its daily life. Can also act as gatekeepers, allowing or limiting access to other members.

Fieldnotes

Detailed records made by a researcher while observing people in their natural setting. Let researchers see and think about events as they happen. Provide a richer picture of social life. Ethnographers also used audio or video recordings

Validity

About ensuring interpretations are accurate and trustworthy. researchers use member checks to let participants confirm findings.

Reliability

Refers to how consistently findings or codes can be applied. Can be checked by having multiple researchers observe same phenomenon.

Limitations of Ethnography

Only allows for the study of small groups. Groups may not be typical of the group being researched. Findings can't be generalized to the larger population. Researchers' presence and participation in a group can affect what they observe.

Qualitative Interviewing

explore a person's beliefs, attitudes, and experiences. They reveal how people understand their lives. Questions are flexible and can change during the interview.

Activiating Different Subject Positions

Interviewers can ask people to view the same topic from different roles. Everyone has multiple selves that we have in different contexts.

Intensive Interviewing

Guided, one-sided conversation. Explores participants' personal experiences with a research topic. Topics can be broad or narrow.

Participant Selection

Participants are chosen for firsthand experience with the topic. Goal is to understand participants' perspectives, meanings and experiences. Interviews follow up on hints, unanticipated topics.

Acquiring Preliminary knowledge of the Field

Prior to starting research, the researcher should know the literature and the community. Fluenecy in groups language helps design interviews. Informants and gatekeepers provide guidance and access. Early interviews with informants help improve questions and approach,

Social Categories and Identity

Race, class, and gender. May influence the interview, can't assume its irrelevant but theur relevance shouldn't be taken for granted.

Suspiscuois Insider Effect

Participants may change how they speak when they learn your researcher identity. Insider access can shift conservation.

Avoid the Morality Play Fallacy

Focus on interviewing people, not categories. Do not assume behavior or answers represent a group. Consider each participant's perspective individually.

Conversation Analysis

like ethnography but uses audio and recordings rather than field notes. The main focus is interaction not groups/individuals, rooted in ethnomethodology.

ethnomethodology

study of people's methods for creating a mutually intelligible social world. Focus on social practices. Indexicality, Reflexivity, Accountability.

Conversation Analysis

Grew out of ethnomethodology. developed a method for analyzing tape recordings of naturally occuring conversations.

Noticing in Conversation Analysis

observing interactions without preconceptions. helps patterns emerge naturally.

Conversation Analysis in vernacular actions

focuses on specific social actions and collects examples. Allows targeted study of recurring behaviors.

Sampling in Conversation Analysis

collecting examples of interactions rather than people. Researchers collect examples of social action, revealing general patterns in how actions are performed.

Five Core Features of CA (conversation analysis)

1. Analyzing utterances as features of action

2. Sequential analysis

3. Analyzing participant orientations

4. Treating interactional detail as a site of social organization

5. Analyzing both single and multiple episodes of interaction

Reliability and validity

Researchers use standard transcripts and recordings so others can check the data. Conversations will not be exactly the same each time they are recorded. Participants' responses show how they understood the previous turn. Publishing transcripts lets other researchers review the analysis.

limitations (CA)

Method focuses on observable interaction. Can't reveal people's inner thoughts / personal stories. Recordings also miss background details like identity, status or resources.

Describing Populations

Surveys study large populations using samples. Measure attitudes, opinions, and behaviors. Data can be reused but researchers can't change the original question.

Open Ended Questions

let respondents answer in own words but must be coded for analysis. capture more detail

CLosed ended questions

give fixed answer choices, easie rto analysis. more uniform

Contingency Questions

Follow-up questions asked only if a respondent gives a specific answer to a previous question. Makes survey relevant by asking questions applicable only to certain respondents.

Matrix Question

ask several questions using same answer choices (Likert scale). Makes responses easy to compare across items.

Pre-test

trying a survey with a pilot group before full administration. Researchers get feedback from respondents and revise questions based on comments.

Double barreled question

Question asking two things at once

Why should survey questions avoid jargon?

Respondents may misunderstand the question, reducing validity

Response-set problem

When respondents choose the same answer for all matrix questions

4 commone survey methods

Mail, face-to-face interview, telephone interview, online survey

One major issue with mail surveys?

Low response rates.

Nonresponse bias

people who do not respond differ from those who do respond

Disadvantage of face-to-face interviews?

Expensive and time-consuming.

Major problem with online surveys?

Representativeness (not everyone has equal internet access).

operationalization in surveys?

Turning abstract concepts into measurable variables.

reliability

Consistency of measurement

validity

Whether a measure actually captures what it is intended to measure.

experiment

study where a researcher manipulates a variable and observes the effect.

content validity

The variable manipulated by the researcher

dependent variable

The outcome affected by the independent variable.

post testing

Measuring the dependent variable after exposure to the independent variable.

Experimental group

The group that receives the experimental stimulus

Reactivity bias

participants change behavior because they know they are being studied.

double blind experiment

Neither the subjects nor the researcher know who is in which group.

randomization

Randomly assigning subjects to the experimental or control group. Makes groups similar and reduces bias.

Matching

Pairing subjects with similar characteristics across groups

Internal validity

Whether the independent variable truly caused the outcome.

External validity

Whether results can be generalized to the real world.

field experiment

experiment conducted in a real-world setting.

web based experiment

an experiment conducted online with digital participants

natural experiment

When real-world events create experimental conditions without researcher control