AP Psych Scientific Method: Research Methods & Data Practice

hypothesis

must be testable/falsifiable, an offered explanation for a phenomenon

experimental design

two groups (experimental & control), independent variable, dependent variable, random sampling, random assignment

experimental group

gets experimental manipulation, independent variable

control group

is left alone, NO experimental manipulation, dependent variable

Bobo Doll Study

occurred in 1961 (Bandura, Ross, & A. Ross), investigated relationship between observed aggression & imitated aggressive behavior in children by demonstrating said aggression on an inflatable clown named Bobo.

experimental group: children observing adults modelling aggressive behavior

control group: children not observing any behavior at all

single-blind study

one group of participants does not know a certain detail of the experiment they are in (ex: which group they are in)

double-blind study

both groups of participants do not know a certain detail of the experiment they are in

placebo effect

expected outcome based on prior knowledge or experience, like a manifestation

placebo control

occurs in order to ensure the only cause of the difference between the exp. & control groups is the experimental manipulation, a PILL with no actual effect

independent variable

manipulated/controlled by the experimenter, only difference, cause

dependent variable

measurement of the effect of the independent variable, effect

sample

a portion of people taken from a population as a representation of said population

population

a diverse, large group of people

random sample

a subset of a population in which every member of said population has an equal chance of being selected, ensures equivalence between characteristics, better generalization

random assignment

participants have an equal chance of being assigned to either group (exp. & control)

matching

grouping participants with a similar level of a certain trait

reverse causation

when one believes X causes Y, but in reality, Y causes X

conceptual definition

defining a concept in terms of other concepts

operational definition

explanation of how a variable or concept is used or assessed in a study

internal validity

the assurance that the only difference between the two groups (exp. & control) is only caused by the independent variable/experimental manipulation

external validity

the extent of how the results of a study can be applied to real life situations and/or generalized to populations

reliability

the assurance that the data is correct, if many trials were conducted, the results would stay the same

observation

examination

pros: abundance of info

cons: can’t apply to larger populations due to limitation of small samples, can’t generalize

survey research

lists of questions to be answered by research participants

pros: very convenient in collecting data from large samples, not time costly, can be done in 3 days (paper, online, orally/in person), easier to conduct generalization

cons: limited & subjective, less in-depth data, inaccurate answers

archival research

uses existing records in order to estimate possibilities/chances

pros: easy, collectible data to answer research questions, less time & money costly,

cons: no control over data at all, reliant on the past, no interactions w/ participants, data has to be modified in order to answer questions

Hogan Twins

girls conjoined at the head and are able to share sensory experiences despite being each their own person

clinical/case study

extreme focus/concentration on individuals under a particular, unique situation (ex: rare genetic disorder)

pro & con (at the same time): easy to conduct generalization for people who are rare, but cannot do the same for a more common population

generalization

the act of applying information to a population in order to gain a better understanding of said population

naturalistic observation

gathering data through observation of behavior in a setting that’s natural

pros: validity

cons: difficulty preparing & controlling (subjects), no control of behavior to be examined, costly in terms of time & money, need to be lucky

structured observation

examines living beings participating in certain activities (opposite of naturalistic observation)

observer bias

when observers are prone to projecting their ideals upon the end results of their data, expecting an outcome that caters to them

Jenkins Survey

asked how participants would treat certain people regarding their ethnicities in order to research about the backlash of the 9/11 attack on arab-americans

result: participants indirectly held some sort of prejudice against arab-americans

longitudinal research

repeatedly obtaining info over a long duration of time (ex: decades or even life-long)

pros: no worries over sociocultural aspects

cons: very time and money costly, uncertain results due to how much time is needed to come to a conclusion, consent & participation needed from all parties, if no commitment then discontinuation

cross-sectional research

compares multiple segments of a population simultaneously

pros: not too time costly

cons: limited by sociocultural aspects (meaning those who grew up in different times will be different socially & culturally)

correlational

relationship between two or more variables, said variables are affected equally & simultaneously

pros: excels in evaluating relationships between variables

cons: cannot discern causality/cause & effect between variables

does not equal causation b/c variables vary & do not necessarily establish causality

in order to establish causality, rule out any other variables (usually through a controlled study)

correlation coefficient

r, represents strength and direction of the relationship between the variables

closer to 1: stronger relation, more eligible for predictions

closer to 0: weaker relation, less eligible for predictions

positive: variables move in the same direction

negative: variables move in opposite directions

confounding/third variable

outlier, unexpected factor

illusory correlations

correlations that are imaginary & false, but thought to be true

confirmation bias

ignoring evidence that would tell one’s gut feeling is false

causation

when one thing causes another thing to happen

equals correlation due to an already existing relationship between variables based on cause & effect

controlled study

two equivalent groups of people although w/ differing experiences are compared, results are compared as well

epidemiological/observational studies

subjects observed over time, no ethical limits as long as the researchers aren’t forcing harm upon said subjects (subjects are doing it themselves or life is doing it to them)

measures of central tendency

single values that attempt to present descriptions of data sets by identification of central positions within said data sets

3 measures of central tendency

mean, median, mode

mean

arithmetic average

how to get: add all values in a data set & then divide by the # of values

pros: can be used with discrete data (occurs at certain points/intervals, countable & non-continuous) or continuous data (any value within a range, appear at any place across the continuum/continuous sequence), best for normal/symmetrical distribution b/c it includes all values

cons: susceptible to influence of outliers of which are abnormally large or small (skewed values)

median

middle score of a data set

how to get: arrange values from smallest to greatest and, if odd, take the middle mark. if even, take the two middle marks & divide by two

pros: best retains normal distribution (perfectly normal values, all measures identical) & strong resistance to influence of skewed values/outliers

cons: not so great in certain conditions when it comes to categorical data

mode

frequency of a value/score in a data set

how to get: identify the value that happens the most in a data set

cons: not unique, conflict occurs when two or more values share the highest frequency

measure of spread/dispersion

a method in order to describe variability in a sample and/or population

large spread: large differences, little representation

small spread: small differences, a lot of representation

has a strong relationship w/ measures of central tendency, provides idea of how well the measures of central tendency represent a data set

range

difference between lowest and highest scores of a data set

how to get/formula: maximum value minus minimum value

pros: sets boundaries of scores, instantly identifies when a value has broken a critical threshold when measuring targeted variable, useful in detecting errors when entering data

cons: has limitations

standard deviation

measure of spread/dispersion of scores in a data set

two of them: sample & population standard deviation

sample standard deviation

requirements: having only a sample, wanting to make a generalized statement towards a population of which the said sample is from

population standard deviation

requirements: having the entire population, having a sample of larger population, exclusive interest of said population, no desire for generalization

formula for sample standard deviation (tip to remember, s = sample)

formula for population standard deviation

null hypothesis

“devil’s advocate” position, suggests that whatever is attempting to be proven did NOT happen, that there is ZERO effect(s) (hence the definition of null as ZERO), UNACCEPTABLE, only find evidence to fight against it

high p-value = fail to reject null & can’t accept alternative

low p-value = reject null & accept alternative

alternative hypothesis

states the opposite of one’s hypothesis, that there is DIFFERENCE

p-value

chance of finding a difference within your study between variables

low p-value = high difference, rejects null hypothesis, accepts alternative hypothesis

high p-value = low difference, fail to reject null hypothesis, not accepting of alternative hypothesis

failing to reject = not necessarily accept, but still consider as possibility

one-tailed hypothesis

has a CLEAR direction, predicts direction of the effect (ex: positive or negative)

pros = researcher bias acceptable ONLY when biological activity/presence is measured

cons = reflects researcher bias

two-tailed prediction

direction is AMBIGUOUS/UNSAID, does not make decision over direction of the effect