research methods I: unit 2 exam

between-subjects design

research between different groups of participants who are tested on different conditions or levels of an independent variable

between-subjects design: disadvantages

may require larger sample sizes, differences may have occurred due to individual differences, and groups must be equivalent

creating equivalent groups: random assignment

every participant have the same chance of being distributed into a random group

creating equivalent groups: blocked random assignment

participants are distributed evenly into groups to ensure there an equal number of participants in each group

creating equivalent groups: matching

participants are grouped into some subject variable of interest and are dispersed evenly across the experimental groups to ensure these variables are evenly distributed

→ these subjects variables of interest are often confounds that trying to be controlled for

conditions for matching

small sample sizes are not ideal, variable targeted must be related to the dependent variable, and measuring the matching variable must be feasible logistically

within-subjects design

the same participants experiences all of the conditions of a study and the difference is measured

within-subjects design: when to use

for smaller sample sizes, when individual differences may affect the results, and for when the differences of the same individual is essential (medication trials)

within-subjects design: order effects

when the order of specific sequences affect participant performance

progressive order effects: practice

when a sequence of conditions make proceeding trials easier from practice

progressive order effects: fatigue

when a sequence of conditions make proceeding trials worse from fatigue

carryover order effects

a specific order of a condition that may affect proceeding conditions than any other condition

→ e.g., participant who did AB did better than BA

controlling order effects: counterbalancing

altering the order of experimental conditions to reduce order effects

complete counterbalancing

every possible combination of conditions are tested

→ may not be plausible if there are many conditions

partial counterbalancing

a subset of all potential combinations of conditions are tested to ensure equal distribution of conditions across each position in a sequence

→ best for smaller sample sizes

latin square

table or square that ensures sequences of conditions are equally distributed across all columns and rows

reverse counterbalancing

reversing the order of a sequence of conditions and repeating it

→ e.g., ABCDDCBA

block randomization counterbalancing

assigning conditions in a sequence without repeating any conditions and starting a new sequence; ensures conditions are distributed evenly across all positions in a combination

e.g., ABCD → BADC → CDAB → DCBA

cross-sectional design

study design where different groups of participants are tested at one point in time

→ may be affected by cohort effects

→ less likely of loss due to attrition

→ requires larger sample sizes

longitudinal design

study design where the same group of participants are examined over multiple points in time

→ less likely to be affected by cohort effects

→ loss of participants due to attrition

→ a small sample is sufficient

cohort effects

people from peer/generational groups may have cultural/inherent differences from other groups

attrition

loss of participants who withdrew from the study

cohort sequential design

combination of longitudinal and cross-sectional design

follows different cohorts over a smaller period of time

→ minimizes attrition and cohort effects

experimenter bias

biases from the researcher that affect data

controlling experimenter bias: automated procedures

procedures that are automated by machinery or other means that do not involve the researcher

controlling experimenter bias: double blind procedures

neither the participant or the researcher know which group is the true IV and not the placebo

subject bias: hawthorne effect

participants will inherently behave differently when they are aware they are a part of research

subject bias: “good” subjects

subjects act purposefully/subconsciously act more cooperatively with researchers

subject bias: evaluation comprehension

participants are apprehensive when under research and tries to act ideally

subject bias: demand characteristics

when a participant slowly becomes aware of the study’s true hypothesis and begin to act differently in respect to the true hypothesis

controlling for subject bias: effective deception

proper deception to ensure participants don’t understand the true intention of the study

controlling for subject bias: manipulation checks

checking after the study (sometimes before) with participants to see if the deception was effective

controlling for subject bias: field research

conducting field research to avoid subject bias (observational studies help participants act naturally)

descriptive statistics

summarizes sample data into understandable metrics

inferential statistics

uses statistical methods to make inferences about the greater population given a sample

descriptive statistics: mean

measure of central tendency of the average of all points in a data

pros: representative of ALL data

cons: not resistant to outliers and skewness

descriptive statistics: mode

measure of central tendency of the most commonly occurring value in a data set

pros: useful for categorical data and is resistant to outliers

cons: not useful for continuous data and in small samples

descriptive statistics: median

measure of central tendency that finds the center value of a data set

pros: resistant to outliers in the data set

cons: not truly representative of data; not useful for inferential statistics

implications of outliers

outliers can skew measures of descriptive statistics and indicates there may be confounding factors

descriptive statistics: range

the minimum and maximum of a data set

pros: tells a researcher the range of a data set

cons: not resistant to outliers and overall not very useful

descriptive statistics: standard deviation

average distance of data points from the mean

pros: best measure of variability

cons: not resistant to outliers

descriptive statistics: variance

standard deviation squared

pros: essential for ANOVAs

cons: not very useful outside of that

descriptive statistics: interquartile range

middle 50% of data

pros: resistant to outliers and useful when there is a lot of spread

cons: not truly representative of the other 50% of data

skewness

a data’s uniformity and where the data tends to drag toward

positive skew

tail of the data is dragged to the right

negative skew

tail of the data is dragged to the left

normal curve

uniform bell curve of a distribution

kurtosis

weight of the tails of a distribution and how the data is distributed there

platykurtic

the distribution is fairly flat with the tails and center of the distribution being fairly close in value

mesokurtic

the distribution is balanced across its tails and center

leptokurtic

the distribution is heavily centered around the middle and the tails are light

null hypothesis

there is no difference or effect between an IV and DV

→ not what the researchers want

alternate/research hypothesis

there is an effect between an IV and DV

→ what the researchers want

alpha level

chance of committing a type I error

→ proportion of times where the null hypothesis was true and the differences occurred by chance

type I error

rejecting the null hypothesis when it’s true

type II error

failing to reject the null hypothesis when the alternate is true

systematic variance

variance that have occurred due to the IV affecting the DV

error variance

variance that have occurred randomly unrelated to the IV

→ confounds

file drawer effect

studies where the null hypothesis was failed to be rejected and kept away from publishing because of insignificant results

importance of replicating failed to reject null hypotheses

replicating the study could refine the measure methodology and obtain an effect from the IV

→ type II error

effect size

magnitude of the effect of the IV and not simply whether it had a significant effect

confidence interval

interval in which the true population statistic is estimated to fall within given a level of confidence

power

chance of correctly rejecting a null hypothesis in research

→ 1 - beta

tables

useful for when multiple precise measurements and data are to be reported

bar charts

useful to illustrate the difference between levels of an independent variable

histograms

useful to illustrate the distribution of data points

pie charts

useful for categorical data and when data adds up to a proportion

line graphs

useful for illustrating the interaction between two IVs

scatterplots

useful for illustrating the correlation between two variables

first code of 1953: hobbs committee

gathered to create guidelines for psychological research using input from previous incidents and researchers

first code of 1953: critical incidents procedure

surveyed psychological researchers for input on ethical guidelines to put into place for research

APA ethical principles of psychologists and code of conduct

many revisions; most recent being in 2002

addresses guidelines for practicing, researching, and teaching of psychology

principles of ethics: beneficence and non-malevolence

weighing risks and benefits of research, protecting participants from risks, and produce research for greatest good

principles of ethics: justice

ensure everybody is able to benefit from research and be treated fairly

principles of ethics: integrity

meticulously transparency and honesty

principles of ethics: respect for people’s rights and dignity

respecting people’s rights and their rights to confidentiality and anonymity

principles of ethics: fidelity and responsibility

professionalism and responsibility of actions

ten standards of ethics

resolving ethical issues

competence

human relations

privacy and confidentiality

advertisements/public statements

recordkeeping and fees

education and training

research and publication

assessment

therapy

institutional review board (IRB)

board of individuals that oversees research approval in terms of their ethics, risk to participants, and benefits

IRB reviews: exempt

studies that pose no discomfort to individuals at all

→ survey studies about economics, educational purposes, literature reviews, training purposes, etc.

IRB reviews: expedited

studies that pose minimal risk to participants or be subject to manipulation

→ deception studies, surveys about mental illness, etc.

IRB reviews: full

studies that are beyond minimal risk for participants in some way

→ medication trials

IRB reviews: no risk

no physiological or psychology harm or discomfort

IRB reviews: minimal risk

minimal physiological or psychological harm or discomfort

IRB reviews: at risk/beyond minimal risk

beyond minimal risk of physiological or psychological harm or discomfort

elements of consent

→ sufficient details about the study and what participants will be expected to do and how long of a commitment they will have to make

→ assurance of confidentiality, anonymity, and voluntary leave if one wishes to do so

→ contact information of researchers and whether participants would like to receive the results of the study

→ signatures

deception

manipulation of participants to falsely believe something to have them act naturally

special populations in consent

prisoners, institutionalized individuals, pregnant women, children, etc.

debriefing

after (sometimes during) a study is complete, a researcher informs a participants of the true nature of the study

dehoaxing

the true hypothesis and intent for the study is revealed

desensitization

negative feelings and/or stress caused by manipulation is accounted for toward the end of research

concerns with internet research

difficulties properly obtaining consent, difficulties properly debriefing individuals, and concerns over privacy and confidentiality in internet spaces

concerns with animal use in research

concern of animal welfare in research and whether or not their use is necessary

institutional animal care and use committee (IACUC)

IRB-like committee that sets ethical guidelines for animal research and determines whether or not research using animals are ethical/permissible

→ benefits vs. risks

→ some benefits to animals as well

requirements for animal care: expertise

all researchers and assistants must undergo training before handling or caring for animals

requirements for animal care: veterinarian care

animals must have access to humane care such as medical care, housing, food, water, etc…

requirements for animal care: euthanasia

animals who are in any amount of pain or suffering as a result of research should be euthanized for their own wellbeing

requirements for animal care: wild/lab animals

animal research with wild or lab animals must undergo their specific procedures

wild: ethical catch/release, treatment, and insuring minimal impact to environment

lab: ethical treatment and humane living conditions

scientific fraud

committing acts of fraud and malicious doing in scientific research with the intent of deception

plagiarism

purposefully stealing others work and claiming it as ones own