research methods and data analysis - exam 2

when doing learn, only answer w term

naturalistic observation

observation of events at they occur in natural settings

field research

research done in natural, real-life settings outside the lab

archival research

studying information from existing records, made in natural settings

which of these is NOT a type of field research

• surverys

• case studys

• archival research

• applied researcd

applied research

archival research

studying information from existing records, made in natural settings

surveys

directly asking people questions in natural settings

case studies

making extensive observations of an individual or small group

program evaluation

conducting evaluations of applied procedures in natural settings

field experiments

conducting experiments in natural settings in order to understand causal relationships among variables

FALSE

T/F - for naturalistic studies, hypotheses are drawn before

ethology

study of organisms in their natural environment

ex: observation of chimpanzees produced new knowledge about social nature of chimps

for normative development and made stages

What did PIAGET use CASE STUDIES for?

made psych clinic to study children and developed psychoeducation

What did WITMER use CASE STUDIES for?

develop consciences theories

what did FREUD use CASE STUDIES for?

exploratory research

research conducted to gather more information about a problem and then make hypotheses → can retest using hypotheses + higher level constraints 

low-constraint research 

____ allows for creativity, thinking by analogy, conceptual leaps, allow wide-rangin search, and look for patterns

TRUE

T/F- naturalistic observation allows one to familiarize with events

low constraint feasibility

-easier thing to carry out

-show whether something is possible

generalizability

extent to which research findings are acceptable to the outside world

information gained from low-constraint research

-identifying new information

-negating a general proposition

-identifying contingencies

-qualitative research methods

contingency

a probabilistic relationship between 2 variables, but can't be proved

ex: ethologist Yinbergen observed that the parent gull provided food when chick pecked spot on bill

NOT

INCREASES

**Contingencies do ___ state X causes Y, only ____ probability

qualitative

low constraint methods are ____ research (qualitative or quantitative)

goals of qualitative research

-describes and analyzes real-life functioning

-understand actions from subjectives perspective

how low constraint research negates a general proposition

use a low-constraint study to find even 1 counterexample to dismiss a general idea

ex. Man is the only one to make tools

---one naturalistic observation of apes disproves this

DOES NOT

low constraint studies _______ establish a general proposition

CANNOT

DESCRIPTIONS

low constraint studies ____ make theories or causality, only makes _____

research problem statement

a probabilistic relationship between 2 variables, but can't be proved

unobtrusive observer

research observes the behavior of participants without participants being aware 

ex: hidden camera is ultimate example

participant observer

-any researcher gathering data whole being an active part of study

-no longer naturalistic as they are impacting the situation

advantage of participant observer

researcher is able to tell hypotheses by creating situations

measurement reactivity

participants behaving abnormal as they know they are being observed

reactive measures

measures that are prone to distortions when you know you are being watched

nonreactive measures

measures that are NOT prone to distortions when you know you are being watched

habituation

decreasing responsiveness with repeated stimulation

ex: Strum observed baboons → initially agitated → habituated → calm

unobtrusive measures

measures not obvious to the participant so they are less influenced by measurement reactivity

ex: counting visitors as they come into the building

coding

classifying aspects of the participant/their behavior into predetermined research codes

coding example

classifying actions into categories

archival records

-any data source for events tthat have occured

-ex: printed narratives or numbers, photos, recordings

what archival records do

-detect problems

-predict events

-track diseases

strengths of archival records

useful for large scale and long term trends 

limitations of archival records

the researcher had no part in data collection = cannot put any constraints

sophisticated studies

complex research projects that move beyond simple observation

selective survival

a problem with archival research

some records survive, some don't

-bias on what data remains

content analysis

identifying categories of events from records

representativeness

how closely a sample resembles the population

Lack of random sampling

Why is representation a limitation in low-constraint studies?

sampling

process of drawing a sample from population

the more people = the greater the representativeness = the greater confidence in generalizability 

DISTORTED

OUTSIDE

sampling can be ____ by variables ____ researcher’s control

casual inferences

-establishing that one variable really does cause another

-cannot draw these from low-constraint

confounding

casual inferences can be drawn if all ____ variables are controlled

ex post facto fallacy

draw causal relationships from observation

there is no support for it

ex: most hockey players are canadian → being canadian leads to playing hockey

1. covariation of events

2. time-order relationship. 

3. alternative explanations have been ruled out 

what 3 things are needed to establish causation?

experimenter reactivity

action by researchers to influence participant response

experimenter bias

experimenter impact on observations or recording of observations

confidentiality → mask identity 

consent → needed most of the time

How do low-constraint studies keep confidentiality and get consent

correlational research methods

-observing the relationship b/t two or more variables without manipulating them 

-quantifies the strength and direction of the relationship

functions of correlational research methods

1. predict future events 

2. provide data that can negate a theory, but not prove it 

FALSE

T/F: Correlation = Causation

differential research methods

-studies that impact of a naturally existing non-manipulated variable on a dependent variable by comparing two+ groups

-IV: classification variable

-DV: behaviors measured in the different groups 

types of differential research methods

qualitative (gender, political party, etc)

quantitative (age, IQ score, # of years of education)

cross sectional design

a design that compares the performance of people of different ages of at different times in history

cohort effect

shared life experience of a group influences their behavior from other ages/cultures/groups

longitudinal design

follow a group over a time to observe changes

longitudinal design disadvantages

takes a LONG time to complete

time series design

-type of longitudinal design 

-multiple measures taken before and after manipulation

independent samples

asking the same question over time to a new group of people

confounding variables

2 variables vary at the same time

--can't keep both consistent

-makes it unclear what caused the DV

artifact

what you think is the effect of the IV is actually the effect of some 3rd variable that is confounding

more structure

create multiple groups

limit confounding variables

more control procedures

Why is differential research higher constraint than correlational?

when it is illogical or unethical to use random assignment

When to use a differential over experimental?

looking at more than 2 variables for correlates

What are multivariate correlational designs?

regression equation?

a more specific problem statement

demographic variables

Attributes of subjects that are collected to describe the sample.

Ex. age, gender, marital status.

experimenter expectancy

experimenter bias

seeing the results you want to see

influences behavior

filler items

things in a study not to actually measure, but to draw the participant away from the real measure

gets more authentic results

moderator variable

a variable that modifies the relationship between other variables

ex. gender, culture, ethnicity

cross-cultural research

differential research to study and compare cultures

SMALLER

having a two tailed vs one tailed test makes power ___

80

A study should have ___% power to be worth doing

theoretical signifigance

study’s ability to help us understand the factors that affect the dependent variable, rather than showing noticeable differences

1. comparing groups that differ on ONE variable

2. use several comparisons across different groups

rules of tumb for theoretical signifigance (2)

increase

increase

____ effect size by ____ the predicted difference between population means

take a sample that has less variation

how do you decrease standard deviation in order to increase effect size?

ways to increase effect size

larger sample size

-smaller SD

-one tailed test

-less extreme significance level

-increase difference between population mean and experimental mean

arguments for effect sizes

comparable to other studys

meta-analysis

arguments against effect sizes

a small sample size can have a large effect size by chance

arguments for statistical testing

can show confidence in data

arguments against statistical testing

interpretation that no stat. sig = no effect

results of the dwyer study

found people were more distracted and dissatisfied if phones were used in social interaction, causes distraction

limitations of dwyer study

study 1: can't tell if results are from you using a phone vs. friends using a phone

study 2: reverse causality, increase phone use by having survey on phones

study 1 - experimental

study 2 - correlational

statistical power

probability that the study will produce a statistically significant result if the research hypothesis is true

effect size and sample size

2 things that power depends on

less

null is more extreme, and therefore harder to get 

more extreme the significant level → the ___ power

WHY? 

greater

more participants = smaller curve = more power

the smaller the SD, the ___ power

WHY? 

strength of relationship among variables

what does correlational research assess?

1. predict future events 

2. provide data to backup/refute theories 

functions of correlational research

type 1 error

rejecting the null when its true ( false + )

type 2 error

failing to reject the null when its true (false -) 

t - test

statistical measure

the difference between TWO mean scores