Research Methods in Applied Social Psychology

Common Sense

• Emphasize the problem’s severity and prevalence to motivate change

  • Much of what social psychologists study might seem to be common sense

Social Psychology vs. Common Sense

• Scientific research methods help us test hypotheses (which may be derived from common sense) and confirm/disconfirm

• Hindsight Bias -- Results seem far more obvious after the answer is known

Norm Theory

• Suggests people are powerfully affected by social norms

• Descriptive Norms

• Injunctive Norms

• Dynamic Norms

Descriptive Norms

• What most people do/what is normally done -- Norm of IS

• Something that is common

Injunctive Norms

• Consensus about what you should do/what is normally approved of -- Norm of OUGHT

  • Something that is socially acceptable

Dynamic Norm

• A norm that is communicated as changing -- Behaviour is becoming more/less common

  • People pay attention to change and can be powerfully influenced by it

Pollution Examples

• Research reveals that emphasizing all the littering leads to more littering -- It backfires

• Showing lots of litter may show the problem, it might also convey a descriptive norm

  • People tend to behave according to what others are doing

Using Norm Theory to Develop Interventions

• In a clean environment, people were less likely to litter even if they saw someone else doing it as compared to a littered environment

• Descriptive vs. Injunctive Norms

How Common Sense Might Result in Backfire

• Common sense emphasizes the problem’s severity and prevalence to motivate change

  • Lacks structure and supported research to help inform people

The Problem with Social Psychology

• The possibility that measurement of a variable changes that variable

  • Example -- Measuring floor tile erosion in a museum had no bearing on the people who once stood there, but if you measure exhibit popularity by standing there with a clipboard and stopwatch, you might scare off some museum patrons and possibly obtain a biased set of responses about the exhibit

Variables

• A property of a person, an object, or an event that can vary in quantity (amount or degree) or in quality

• Independent Variable

• Dependent Variable

Independent Variable

• The variable the experimenter expects to have an influence

• May be measured

• May be manipulated in an experiment

Dependent Variable

• The variable the experimenter expects will be influenced

• Does not get manipulated

Definitions

• Construct

• Operationalization/Operational Definition

Construct

• The conceptual or hypothetical explanatory variable (or latent variable that cannot be directly observed

  • E.g., love, happiness, interest, morality, power

Operational Definition

• The specific way in which a researcher measures a variable

• Translate this construct into a meaningful measure

  • e.g., how do you operationally define power?

    • Body language, selecting a bigger or smaller seat, etc.

Construct Validity

• A measurement procedure that reflects the meaning of the variable accurately

• Does the operationalization provide a valid measure of the variable of interest?

Research Designs

• Purpose is to identify key variables in a social phenomenon and to examine relationships among variables

• Four types:

  • Correlational

  • True Experimental

  • Quasi-Experimental

  • Descriptive

Correlational Studies

• Involves measuring variables and determining the association between them

• Correlation can rage from -1 to 0 to +1

• Correlation between variables can be positive or negative

Correlation and Causation

• Correlation does not necessarily indicate causation

  • A third variable might account for the relationship between the two variables

  • Reverse order of causality -- X causes Y rather than Y causing X

True Experimental Design

• Manipulation of IV

• Random assignment

• The goal is to determine cause (IV) and effect (DV) -- If only the IV differs between conditions, can more confidently conclude that changes in the IV caused changes in the DV

Random Assignment

• Equal chance of any person being assigned to any level of the IV -- Individual differences should be evenly distributed across conditions

Problems with Experimental Designs

• Feasibility and Ethics

• Internal Validity

• Extraneous Variables

• Can be artificial

• External Validity

• Ecological Validity

Feasibility and Ethics

• Deception -- Can only occur if debriefing is possible

• First do no harm principle

• Delivering different benefits -- Deny the control group helpful treatment

  • Wait-list Control Group is a solution

Internal Validity

• The extent to which one can be certain that it is the IV alone that has produced differences in the DV

• Must rule out the possibility that extraneous variable might have produced the differences obtained

  • If extraneous variables might have produced the differences observed, that is a threat to internal validity

External Validity

• The extent to which research findings can be generalized beyond a particular study to other people, times, settings, regions, etc.

Ecological Validity

• The extent to which the experiences and behaviours we observe in research are genuine in the sense of being similar to those that ocur in everyday life

Experimental Designs -- Research Settings

• Lab vs. Field Studies

  • Lab is more artificial but more easily controlled, harder to control all aspects of a situation out of a lab

• Ecological Validity

• External Validity

Quasi-Experiments

• Often unethical to randomly assign each research participant to different experimental conditions

• Resemble experiments in some way but lack one or more key features

  • IV may occur naturally or be manipulated -- Not always by experimenter

  • Lacks random assignment

  • May or may not have a control/comparison group

Types of Quasi-Experimental Designs

• Pre test-Post test Design

  • With Non-Equivalent Control Group

• Interrupted Time Series Design

Pretest-Posttest Design

• Initial observations are taken, the treatment is implemented, and observations are taken again so that a comparison can be made between what happened before and after the treatment

• Non-Equivalent Control Group

  • A number of extraneous factors may have led to results, posing threats to internal validity

Internal Validity Threats in Quasi-Experimental Designs

• Can be found in experimental and quasi-experimental designs

• Confounds

• Resentful Demoralization

• Differential Attribution

• Diffusion

• Experimenter Bias

• Demand Characteristics

• Maturation

• History

• Repeated Testing

• Instrument Change

• Regression to the Mean

Confounds

• An unmeasured or uncontrolled third factor distorts the true relationship between IV and DV

Resentful Demoralization

• Participants become angry, discourages, or uncooperative after learning they were deprived of a favourable treatment

Differential Attrition

• Participants drop out of study groups at unequal rates for different reasons

Diffusion

• Participants in a control or comparison group learn about an intervention/treatment form the experimental group and adopt those same behaviours

Experimenter Bias

• Researchers unconscious expectations, beliefs or actions skew a study’s results, making it unclear whether the outcome is caused by the IV or by the researcher’s influence

Demand Characteristics

• Subtle cues or hints in a research study that let participants figure out the hypothesis/what the researcher expects them to do, causing them to change behaviour

History

• An outside, uncontrolled event occurring during a study that influences participant’s behaviour or the DV, making it unclear if the change was caused by the IV

Repeated Testing

• Participants score or behaviour on a later test changes simply because they have already taken the test

Maturation

• Natural changes in study participants over time can cause changes in the results instead of the experimental treatment

Regression to the Mean

• Participants who score exceptionally high or low on initial measurement naturally trend closer to the group or population average on subsequent measurements

Instrument Change

• Changes in results cause by modifications or inconsistencies in the measurement tool, equipment, or observer rather than the experimental treatment

Descriptive Studies

• Observe variables and summarize the observations by using descriptive stats

• Measurement of variables without manipulation

• Summarize measurements of variables

Qualitative Research

• A means for exploring and understanding the meaning individuals or groups ascribe to a social or human problem

Difficulties with Quantitative Methods

• Only allow you to look at a few variables at a time

• Not well suited to generating theory -- Identifying what the key variables are

  • Can only measure what was asked -- Specific questions

• Do not give research participants a voice in describing their experience

Qualitative Methods

• Inductive in nature

• Can assist in generating theory, identify themes and variables to study further

• Gives participants a voice

Qualitative Methods -- Limits

• Social desirability

• Sometimes smaller samples -- Less generalilzable

• Limits of awareness -- Misperceptions of causality

• Researcher bias could shape response

Mixed-Method Design

• Quantitative and Qualitative methods

• Qualitative to inform theory, provide richness, give participants a voice

• Quantitative to collect larger samples, more controlled, more generalizable

Random/Probability Sampling

• Done with survey and Quantitative research

Random Assignment

• Used to place individuals in experimental conditions

Purposive Sampling

• Used to select information-rich cases

• Aim to recruit a sample when participants represent a wide range of characteristics

• Done with Qualitative research

Formulating the Research Questions

• Some examples of questions that lend themselves to qualitative research:

  • What are the major pressures and stresses that students encounter while attending university?

  • How do you prepare for the risk of heat waves and other extreme weather events?

  • How has social media and online dating changed the way people think about seeking romantic relationships?

  • What do you like and dislike most about your political party? your opponent party?

Data in Qualitative Sudies

• Direct quotations

• Detailed descriptions

• Records

Qualitative Interview

• Most common type of qualitative research methodology

• Interview to find out things we cant observe directly

  • feelings, thoughts, intentions

  • Behaviours that took place when we weren’t there

• Task of the interviewer is to make it possible for the interviewee to bring the interviewer into their world

Data Analysis

• Read transcripts

• Identify themes -- Open coding

• Organize themes -- Into a model, hierarchy, or process