Social Psych Exam #1

Social psychology

How people think, influence, and relate to one another (TIR)

Kurt Lewin

Behavior = f(Person x Social Situation)

Main areas of social psych

1. Social cognition (thinking)

2. Social influence

3. Social relation

Social cognition

• The self

• Social beliefs & judgments

• Attitudes/behaviors

Social influence

• Conformity and obedience

• Persuasion

• Group processes

Social relations

• Stereotypes

• Prejudice

• Discrimination

• Attraction

• Altruism

Attributions

Explanations for the causes of behavior (e.g. why ppl are overweight)

External validity

How well the results can be generalized to real-world settings

Internal validity

How certain X causes Y (cause and effect); how much can we trust the results

Descriptive methods

• Naturalistic observation/observational research, survey research

• Higher in external validity, lower in internal validity

Experimental methods

• Lab and field experiments

• May lack mundane realism: how well it replicates everyday life (e.g. shocking ppl → not realistic)

• Experimental/psychological realism: how real the experiment feels to participants

r

• Correlation coefficient used for linear relationships (-1 to 1)

Random assignment

• Internal validity

Random sampling

• Generalizability

Research in the real world

• Research in the real world: 

  • Be a critical consumer of research → ask questions!

  • Research design (issue of causality), who and how sampled (generalizability), operationalization (validity of results)?

  • Ex. “BU Study Finds Plus Size Models Rising Presence on Social Media” 

    • A BU study examined IG activity among mainstream and plus-size models; main finding was that plus-size models had engagement comparable to mainstream models

    • When headlines go beyond data → the following was included in the press release: “Plastic surgeons may see an increased interest in procedures, including breast and buttock augmentation, that embrace a curvaceous plus size body type…” said Professor Vashi of BU

      • Soon after, this headline appeared → “Plus size models are driving a new plastic surgery trend: study”

        • HUGE distortion; not what study examined

Operationalization

• Measurement tool(s) in an experiment (e.g. scale that shows how satisfied someone is w/ their weight)

  • Look at previous studies to see their operationalizations (used before = most likely higher validity) 

Replication crisis (aka reproducibility/replicability crisis)

• Growing belief that the results of many scientific studies cannot be reproduced and thus might be misleading

  • Older popular studies → can’t always replicate OR effects aren’t as strong

  • Why is replication important for science? → can’t build on past research

Evidence for the replication crisis

1. The Open Science Collaboration

2. The Famous Marshmallow Test

The Open Science Collaboration (2015; Brian Nosek)

• Attempted to directly replicate 100 studies from several top-ranked journals w/ “high validity”

  • Replication rate was low (~36% of studies), and effects were even smaller

    • Even within the 36% replication success, effects were smaller than OG experiment

The Famous Marshmallow Test

• Original study → children who delayed gratification (defined by not eating a marshmallow to get a larger reward later) = better success, mental health, and physical health

  • Highly publicized, featured in textbooks, etc.

  • Replication only found a correlation ½ as strong

    • Relationship was substantially reduced after controlling for family background, early cognitive ability, and home environment 

    • The replication had a stronger study design → sample 10x larger than OG, more socioeconomically diverse (OG study sample was mostly rich and white) 

    • Some argue the data can be interpreted in different ways

Public costs of replication crisis

• Strong, confirmatory studies when they might not be

• When research claims to have strong effects but they’re not real at all, public trust → undermined

  • Bad b/c a lot of ppl already don’t trust science (e.g. vaccines) 

Scientific costs of replication crisis

• If scientists are basing off their research on studies that aren’t strong/real, the next set of studies will be faulty 

How researchers are addressing the replication crisis

• Greater transparency practices (e.g. planning to publish study in a journal → must pre-register)

• Improved methodology

  • Large enough sample size

• More independent replications

Harking

• Hypothesizing after results are known → not a good practice

Study findings often seem like common sense, but:

• Before seeing the results, people predict research results no better than chance

• Many beliefs that ‘seem’ correct are unsupported

  • E.g. children learn best when teaching style is matched w/ learning style → actually NOT true

Hindsight bias

• Something seems obvious after it occurs

  • Often makes it feel like the results were predictable or obvious

“But I know someone who…”

• Individual exceptions do not disprove overall patterns

• Patterns = average performance across groups

3 questions we’ll answer this semester:

• 3 questions we’ll answer this semester:

  • How does scientific truth and knowledge evolve?

  • Why do we think, feel, and behave the way we do?

  • To what extent is behavior shaped by who we are versus the situations we are in (person vs. situation debate)? 

    • People underestimate the power of situations on behavior