BBH 310 Exam 2

variable

something that changes and can be measure

How many levels or values should a variable have?

2

construct

• theoretical event, outcome, person, etc that we care about

• Ex. hypertensions, stress

operational definition 

• set of procedures used when you measure or manipulate variables 

• make the construct real, quantifiable, and/or measurable

• Ex. mmHg

What two things does an operational definition include?

• what is observed

• what is measured

What are two reasons why we operationalize our variables?

• researchers want to focus on concrete terms for abstract concepts

• help researcher communicate their ideas

observable vs latent variables

• observable - physical variables that have obvious/direct connection to construct

• latent - psychological variables that are NOT directly observable, estimated from behavior

Examples of observable variables 

height or weight 

Examples of latent variables

happiness, mood, intelligence

reliability

• consistency or stability of a measure

• yields the same results EVERY time

validity 

• accuracy of a measure

• “measures” what it’s supposed to

reliability theory formula

observed score = true ability + random error

observed score

actual test score obtained from the measure you are using

true score 

someone’s real or true value for the variable being measured; not directly measured 

measurement error

difference between true and observed score

correlation

number that tells you how strong two variables are related to one another

How many values must you have to assess reliabiltiy 

two scores 

test-retest reliability

measure the same individuals at two different time points (type of reliability)

internal consistency

use responses from individuals at only one time point to examine their within-occasion consistency (type of reliability)

What are two statistical technique to use for internal consistency?

split-half reliability and Cronbach’s alpha 

split-half reliability

correlation of total score on one half of the questionaire with the total score on the other half of the questionaire

Cronbach’s alpha

correlation of each item on the questionaire with every other item (type of reliability)

interrater reliability 

extent to which at least two observeres agree in their observations 

What are two types of validity?

construct and criterion

construct validity

degree to which the measure reflects the construct its supposed to (what we typically think of)

criterion validity 

the degree to which measure corrrelates with a similar measure concurrently or in the future (what we use to compare)

What are the 2 types of construct validity?

• convergent validity - examines if a measure is related to another measure assessed by same construct 

• discriminant validity - examines if measure is NOT related to another measure that assess a different construct 

Example of convergent validity

wearable sensors tracking HR

Example of discriminant validity

wearable sensor for HR versus movement data

What are the 2 types of criterion validity?

• predictive validity - examining if scores on a measure predict behavior or outcome it is intended to predict 

• concurrent validity - extent of agreement between two similar measures taken at the same time 

Example of criterion predictive validity

using SAT scores to predict academic performance in college

Example of criterion concurrent validity

HR sensor relating to self-reported stress

interval scale

• numeric scale where distance between two points matter

• no true 0 point, ratios are impossible

• Ex. celsius

ratio scale 

• distance between two points matters

• has TRUE 0 point

• Ex. height, weight

Measurement scales that include continous variables 

interval and ratio scale 

Measurement scales that include categorical variables

nominal and ordinal scale

nominal scale

• groups with no order rank

• Ex. gender, ethnicity

ordinal scale 

• groups with rank order 

• Ex. SES

Pros to questionaries

access to personal day-to-day information

Cons of questionaires 

lack of clear units, biased or inaccurate 

confidentiality

partipants identify only know to search them

IV

independent variable - what’s being changed

DV 

dependent variable - what’s the outcome

What are the 3 major tenets of the Belmot report?

• autonomy - do ppl have free will to join or leave study?

• justice - is there an equal opportunity for human subjects 

• beneficence - does research hurt anyone? do the harms outweigh the benefits?

What are two risks in behavioral research?

loss of privacy and confidentiality

confidentiality

participant identity only know to research team

privacy

partipant information is disclosed only to research team and not to others

Who protects human research participants’ rights?

Institutional Review Board (IRB)

What was happened as result of the Nuremberg trial?

• Nazi human experimentation 

• Nuremberg code created - principle of ethical conduct of ANY human experimentation 

What was the result of the Helsinki Declaration?

provided guidelines for more medical research with humans

What was the result of Belmont report?

focused on behavioral research not just medical and provided ethical principles for both medical and behavioral research 

observational research 

• attempts to assess responses as they unfold in real life 

• no manipulation 

What’s the difference between quantitative vs. qualitative?

quantitative - focuses on variables that are numerical and can be counted (Ex. surverys, experiments)

qualitative - focuses on non-numerical data (Ex. text, video)

cross sectional study

data collected at one specific time point from different groups of people l

longitudinal study

data gathered for the same subjects repeatedly 

What are limitations of observational research? (3)

• unable to show cause and effect definitively 

• no control of participants or intervention

• many confounding variables 

What are the 3 major types of observational research?

• naturalistic - make observation in natural real-world environment

• systematic -

• indirect (survey, questionnaire)

Is naturalistic observation research quantiative or qualitative?

it can be both! primarily qualitiative description of observation, quantitative frequences of behavior occurence

What two concepts did Mehl et al. 2007 illustrate?

• natural observations get at important phenomenon than lab can

• ambulatory tech can provide access to rich/real world behavior

covert/concealed participant observation

• subject unaware you’re observing

• you immerse yourself in activity/event/behavior

covert/concealed non-partipant observation

• subject unaware you’re observing

• you observe participants from a distance without being involved

overt/non-concealed partipant observation

• subjects aware you are observing them

• you immerse yourself in activity

overt/non-concealed non-participant observation

• subjects aware you are observing them

• you observe participants from a distance without being involved

Which can be operationalized manipulated or measured variables?

both

What are 3 criteria necessary for demonstrating causal relationships?

• x and y are related 

• x comes before y

• confounding variables are ruled out 

What are pros of self-report?

• afforable 

• easy to deploy

• access to rich psychological experience 

What are the two major forms of systematic assessment?

• self-report

• task paradigms 

What are the cons for self-report?

• bias

• honesty

• inaccurate recall

When would you want to do a systematic observation?

• testing specific hypotheses

• want to count/measure something

When would you want to do a naturalistic observation?

• ethical situations

• gain qualitative results 

• new or unknown topic 

biases with observational research

• if participants knows someone observing → change behavior

• never able to control confounding variables

biases with survey research

• restricted to choices

• observer bias

What is the difference between self report and task paradigm?

• self report - ppl share their behaviors pertaining to whatever being asked 

• task paradigm - ppl engage in behaviors under more controlled conditions where stimuli and requirements are standardized 

How does reliability look like on a frequency distribution?

LESS variability, most volume in specific range

What is research integrity?

scientists ensure that data, research, intepretation are accurate and transparent

What are the good practices for research integrity?

What are bad practices of research integrity? (5)

• data manipulation

• hypothesizing after results are known

• fraud

• plagarism

• “p-hacking”

When is best to use self-report?

• want feelings, opinions, beliefs

• something cheap and effective 

What is the pros of task paradigm?

• cognitive testing

• decision-making

• stimulus rating

What’s the difference between probability and non-probability sampling?

• probability - all persons have equal probability of being chosen for sample

• non-probability - all persons DO NOT have equal opportunity to be chosen

What is the least bias probability sampling method?

stratified random

stratified random (2)

• list of two different groups, separately sample from each

• equal number from each group

simple random 

randomly select from list, more random than systematic 

systematic random

picking every _5th__ from list

cluster sampling

population divided into clusters, random sample of clusters is picked