Ch 2 social

2 Research in Social Psychology

3 Overview

4 Three Main Goals of Research

1. Measurement and description

2. Understanding and prediction

– Can predict without fully understanding

– Can understand without being able to predict

3. Application and control

5 Two Main Types of Research

• Basic research

– Goal is to increase our overall understanding

– Some argue that the more basic the principle being studies, the less it matters who participates in the study in terms of sample representativeness.

• Applied research

– Focus is on applications and contributing to the solution of real-world problems

6 Two Main Types of Research

• Basic and applied are cyclical

– Basic research is often starting place for applied research

– Applied research can reveal new avenues for basic research

7 Developing Ideas

Beginning the Research Process

8 Where to start?

• Start observing and asking questions

• Search the literature (e.g., textbooks, journal articles, databases)

• Develop your research question(s)

9 Hypotheses and Theories

• Hypothesis

– An explicit, testable prediction about the conditions under which an event will occur

• Should be specific and falsifiable

– Based on observation, existing theory, or previous research findings

10 Hypotheses and Theories

• “But it’s just a theory…”

• Theory

– A well-established, working explanation of a phenomenon

• Ideally, theories:

–Are efficient and precise;

–Encompass all relevant information;

–Lead to new hypotheses, further research, and better understanding

•

11 Refining Ideas

Defining and Measuring Social Psychological Variables

12 From the Abstract to the Specific

• Conceptual variables

– Abstract or general variables

– Example: Personality

• Set of emotional qualities, ways of behaving, etc. that makes a person different from other people

• How can we measure something like personality?

13 Operational definitions

• Operational definition

– States specifically how the conceptual variable will be manipulated or measured

– Transforms the variable from the abstract to the specific

• Examples: Personality

– Use score on the HEXACO-PI-R (a well-established personality test)

– Count how often the individual engages in a neurotic behavior “in the wild”

14 Let’s look at some more examples

• Happiness -> count number of smiles

• Stress -> average hours of sleep per week

• Popularity -> “How many relationships have you had in the last year?”

• Relationship quality-> “How many dates have you had in the last year?”

15 Testing Ideas

Research Designs

16 Types of Research Methods

A. systematic observation

1. Case Studies

2. Naturalistic Observation

3. Laboratory Observation

4. Surveys

5. Archival Data

B. correlational studies

C. experimental research

17 Systematic Observation

• Allows researcher to describe and predict behavior

• Does NOT show causality

• Main types of systematic observation:

1. Case Studies

2. Naturalistic Observation

3. Laboratory Observation

4. Surveys

5. Archival Studies

18 1. Case studies

• A detailed description of a particular individual under study or treatment

• Example: Genie the feral child

– Raised in social isolation until the age of 13

– Did not acquire language abilities (supports ‘critical phase’ theories)

19 2. Naturalistic observation

• Observe behavior as it occurs in the natural environment

– Observing more than one person in some specific setting

– Measure behavior in a systematic way

• No real control over situation/setting

– Cannot establish causality

• Helps to form research questions/hypotheses

20 3. Laboratory observation

• Observe behavior within a controlled setting, such as a laboratory

• Researcher has much more control

– Control at least some of the variables (e.g., temperature, stimuli)

– Still no causality, but able to rule some things out

• Note, being in controlled setting may change people’s behaviors

21 4. Surveys

• Questionnaires and interviews

– Ask people about their experiences, attitudes, or opinions

• Note, self-report is not always accurate

– May not wish to reveal true thoughts/behaviors

– Memories for past thoughts or behaviors may not be accurate

22 5. Archival studies

• Examining existing records to investigate predicted trends or patterns

• E.g., Look at personnel records before deployment to examine factors that may influence resiliency to stress

• Still cannot establish causality

• Records may also be incomplete/insufficient

23 Systematic Observation

• Simply documents the rate or level of the phenomenon/behavior of interest.

• Does not tell us about the causal relationship

• Accuracy of these observations depends on:

– Sampling method

• Bad samples lead to bad observations

– Measurement method

• Bad measures lead to bad observations

24 Sampling Method Example

• Chicago Daily Tribune incorrectly projected Dewey as the winner of the 1948 presidential election

– Non-random sample

• Random sampling: selecting participants so that everyone in population of interest has an equal chance of being selected

25 “Black” vs. “African American”: What’s in a Name?

26 Measurement Method

27 Types of Research Methods

A. Systematic Observation

1. Case Studies

2. Naturalistic Observation

3. Laboratory Observation

4. Surveys

5. Archival Data

B. Correlational Studies

C. Experimental Research

28 Correlational Studies

• Correlation - a measure of how strongly two or more variables are related to each other

• Correlation coefficient (r)

– Can range from -1.0 to +1.0

– The closer value is to -1.0 or +1.0, the _stronger the relationship between the variables

29 Correlational Studies

• Correlation coefficient (r)

– Absolute value tells us the magnitude/strength of relationship

• E.g., +0.5 and -0.5 have the same magnitude/strength!!!

– Sign (+/-) tells us the _direction of the relationship

30 Correlation Coefficient

• Positive: variables change in same direction (both increase or both decrease)

• Negative: variables inversely related (change in opposite directions)

31 Advantages and Disadvantages of Correlational Research

• Advantages

– Great for looking at variables and relationships that cannot be manipulated or induced

• One very serious disadvantage

– Correlation is NOT causation !

• Directionality problem

• Third Variable problem

32 Explaining Correlations: Three Possibilities

33 Explaining Correlations: Three Possibilities

34 Explaining Correlations: Three Possibilities

36 Types of Research Methods

• A. Systematic Observation

– 1. Case Studies

– 2. Naturalistic Observation

– 3. Laboratory Observation

– 4. Surveys

– 5. Archival Data

• B. Correlational Studies

• C. Experimental Research

37 C. Experimental Research

• Allows us to control the situation and variables being studied

• True experiments are used to understand and establish causality

38 C. Experimental Research

• True Experiments have 2 important characteristics:

1. True Independent Variable(s)

2. Random Assignment of Participants

• Allow us to establish that changes are indeed due to the variable(s) of interest

39 Experimental Variables

• Independent Variables (IVs)

– Variable that researcher controls/changes/manipulates to influence DV

– The ‘treatment’ variable

– True IVs: can randomly assign participant to any level

• Dependent Variables (DVs)

– The ‘output’ or ‘outcome’ variable

– Researcher measures this to see if output depends on level of IV

40 Experimental Variables

• Subject Variables

– Think of these as special types of IVs

– Refer to pre-existing differences in participants

– Cannot randomly assign to levels, but we can group and make comparisons by subject variables

• E.g., compare females vs males, older adults vs younger adults, tall people vs short people

41 Experimental Variables

• Confounds

– A variable other than the IV that also changes across conditions

– Problematic: this variable may be responsible for changes to DV (rather than our IV)

• Calls our results and conclusions into question

• We want to minimize these as much as possible

42 Experimental and Control Groups

• Experimental Groups

– Group that is exposed to the manipulation/treatments

• Control Groups

– Group that is not exposed to manipulation/treatment of interest

– Typically serve as a baseline for comparison

– Good to have, but usually not a necessity

43 Laboratory and Field Experiments

• Laboratory setting

– Control over

• Setting

• Measure of participants’ behavior precisely

• Keep conditions identical for participants

• Field setting

– “Real-world”

• Advantage – people are more likely to act naturally

• Disadvantage – less control and cannot ensure identical conditions

44 Results and Statistical Significance

• Statistical significance – unlikely that results occurred by random chance

– p value of .05

• The probability of getting these results (if the null is true) is about 5 out of 100 times

• Replication

– Conducting experiment in different samples, using different statistical techniques

– “File Drawer Problem”

45 Internal Validity

• Degree to which we can be reasonably certain that the effects were due to our IVs

• Improve internal validity by:

– Controlling for potential confounds, using control groups, using “blind” studies

46 External Validity

• Degree to which we can be reasonably confident that the results would be obtained

for other people in other, but similar situations

– Concerns how generalizable our results are

– Improve our external validity by using representative samples, increasing our mundane and experimental realism

47 External Validity

• Mundane : extent to which experiment is similar to real world situations

• Experimental : extent to which the experiment is compelling to participants

48 Deception and Ethics in Social Psychology

49 Deception in Experiments

• Deception

– Withholding or providing false or misleading information to participants

• Confederates

– People who seem to be participants but actually are part of the experiment

• These are sometimes necessary

– Add to realism and create specific situations

– But how much deception should be ‘allowed?’

50 Ethics and Values in Social Psychology

• IRBs

– Ethical reviews are required for social psychology studies before they are implemented

• Informed Consent

– Give participants sufficient information to make an informed decision about

volunteering

• Debriefing

– Provide full disclosure at end of study

51 Ethics and Consent Online

• “To Facebook, we are all lab rats” (Goel, 2014)

– Mood manipulation study

• Manipulated content over course of a week (either positive, happy content or sadder, more negative content)

• Facebook argues that consent is given with agreement to terms of service

– But who actually reads those?

• Is it ethical to collect and use such data?

- While basic research seeks to increase our understanding of human behavior and is often designed to test a specific hypothesis from a theory, applied research focuses more specifically on making applications to the world and contributing to the solution of social problems.

- Experimental research allows researchers to determine cause-and-effect relationships—that is, whether the independent variable can cause a change in the dependent variable.

- Pros of using a self-report measure are that they give the researcher access to an individual's beliefs and perceptions. Cons of this measure are that they are not always accurate and can be misleading.

- A confound is a variable other than the IV that also changes across conditions. It is a serious threat to internal validity and, therefore, makes the issue of cause and effect in the experiment uncertain.

- Social psychologists attempt to replicate their results in order to reduce the probability that their findings are due to chance.