Psych GRE - Research Design, Statistics, Tests, and Measurements

• 1879
• founded first psychology laboratory

showed that higher mental processes could be studied using experimental methodology

• disagreed with Wundt fundamentally - Wundt believed that whenever you thought of something an image of that thing formed in your mind (there could no thought without a mental image)
• performed experiments to prove that there could be imageless thoguht

studied under Wundt; introduced mental testing to the U.S.

• in 1905, collaborated with Simon to publish the first intelligence test
• introduced the concept of "mental age" or the age level at whic a person functions intellectually, regardless of his actual chronological age

• first intelligence test
• use to assess the intelligence of French school children to ascertain which children were too mentally retarded to benefit from ordinary schooling

developed an equation to compare mental age to chronological age (which came to be known as the intelligence quotient, or IQ)

in 1916, he revised the Binet-Simon test for use in the U.S. - it was called the Stanford-Binet Intelligence test

• the first setp in research design involves indentifying a problem to study and forming this
• tentative and testable explanation of the relationship between two or more variables

characteristic or property that varies in amount or kind, and can be measured

how the researcher plans to define the varilables in the experiment so that the variables are measurable

variable whose effect is being studied and is the variable that the experimenter manipulates

the response that is expected to vary with differences in the independent variable

• true experiments
• quasi-experiments
• correlational

does not alter the independent variable (IV)

researcher uses random assignment and manipulates the IV

• does not use random assignment
• lacks sufficient control over the variables, and therefore, definitive statements on causal factors cannot be made

• researcher does not intervene at all in what is being studied, but observes what occurs naturally
• AKA: field study

subset of the population

each member of the population has an equal chance of being selected for the sample

assures that each subgroup of the population is randomly sampled in proportion to its size

miniature version of the population

• between-subjects
• matched-subjects
• within-subjects (AKA repeated measures)

each subject is exposed to only one level of each indpendent variable

• experimenter matches subjects on the basis of a varilable that he/she wants to control (if he/she wants to control for intelligence - he will pick the top 2 most intelligent, split them up, and split the pairs in this fashion until all subjects are assigned to groups)
• pairing ensures that both groups are approximately equal on the matching variable

• each subject is exposed to all the conditions (give all subjects all levels of IV)
• this allows the researcher to separate the effects of individual differences of a variable from the effects of the IV
• problem: people may just do better on the second test because they are more familiar with the test format - or they may do worse on the second test because of boredom

• half of the subjects are assigned to group 1 and then group 2; the other half are assigned to group 2 and then 1
• this is done because of the problems with within-subjects design (people may just do better on the second test because they are familiar with format or they may do worse on the second test because of boredom)
• all subjects will experience both levels in different orders

unintended independent variables

treating both groups equally in all respects except that the control group gets no treatment; the experimental group receives treatment

the control group is not necessarily similar to the experimental group since the researcher does not use random assignment

due to his/her expectations the experimenter might inadvertently treat groups of subjects differently, experimenter might let his/her expectations affect how the results are interpreted

• one way to control for experimenter bias
• neither the researcher who interacts with the subject nor the subjects themselves know which groups received the IV or which level of the IV

subjects do not know whether they are in the treatment or control group

refer to any cues that suggest to subjects what the researcher expects from them - if subjects have an idea what the researcher expectes, they will perform as expected

• a kind of demand characteristic
• when people are given a drug, they usually expect that the drug will be effective

tendency of people to behave differently if they know they are being observed

how generalizable the results of an experiment are

double-blinding

control groups

control groups

deception?

• descriptive
• inferential

• generalize beyond actual observations
• allow us to use a relatively small batch of actual observations to make conclusions about the entire population of interest

organizing, describing, quantifying, and summarizing a collection of actual observation

graphic represenations of how often each value occurs

• mode
• median
• mean

• value of the most frequent observation in a set of scores
• if all values occur with equal frequency, there is no _____

when the data has two modes

• the middle value when observations are ordered from least to most or from most to least
• not the halfway point of numerical values
• number in the middle of the ranking

the numerical halfway point between the highest score and the lowest score, the arithmetic average

• extreme scores
• the mean is the measure of central tendency that is most sensitive to these

• range
• standard deviation
• variance
• AKA dispersion of scores
• range (282)simply t

simply the smallest number in the distribution subtracted from the largest number

• provides a measure of the typical distance of scores from the mean
• square root of variance

• standard deviation ²
• description of how much each score varies from the mean

symmetrical bell-shaped curve

percentage of scores that fall at or below a particular score

• between -1 and +1 SD - 68% (34% between -1 and 0; and between 0 and +1)
• between -2 and +2 SDs 96% (14% between -2 and -1; and between +1 and +2)
• 100% between -3 and +3 SDs (2% from -2 to -3; and between +2 and +3)

• another way of calculating how many standard deviations above or below the mean your score is
• formula = score - mean / SD

• if you have a distribution of z-scores and calculate the mean and standard distribution, the mean of the distribution of z-scores will always be zero andthe standard deviation will always be 1
• this is true regardless of the mean and the standard deviation of the original distribution

t-score distribution has a mean of 50 and a standard deviation of 10

• a descriptive statistic that measures to what extent, if any, two variables are related
• two variables are related if knowing the value of one variable helps you predict the value of the other variable
• this helps us understand the relationship and degree of association between two variables
• allows us to mathematically specify how well we can predict the value of the second variable given the corresponding value of the first variable
• can only be between -1.00 and +1.00

a change in value of one of the variables tends to be associated with a change in the same direction of the value of the other variable

change in the value of one of the variables tends to be associated with a change in the opposite direction of the other variable

graphical representation of correlational data

attempts to account forthe interrelationships found among various variables by seeing how groups of variables "hang together"

• a tool researchers use to draw conclusions about populations based upon research conducted on samples
• helps the researcher decide whether the research hypothesis or the null hypothesis is true

the population mean is the same as the sample mean

if we reject the null hypothesis, the observed difference is ______

• the researcher decide what probability represents statistical significance before collecting data by establishing this...
• by convention, psychologists usually use 5% (most are willing to reject the nulll hypothesis only if they are very sure that observed differences are not due solely to chance)
• AKA alpha level

• rejecting the null hypothesis when it is true (there is no difference, but the researcher thinks there is a difference)
• significant results are due to chance

• accepting the null hypothesis when it is false (thinking there is not a difference when there is)
• AKA Beta

• t-test
• ANOVA
• chi-square test

used to compare the means of two groups

used to compare the means of more than two groups

tests the equality of two frequencies or proportions

• statistical procedure that can be used to make conclusions on the basis of data from different studies
• can be used to combine the results of these studies and come up with a more general conclusion

• assessing an individual's performance in terms of how that individual performs in comparion to others
• one problem: the population to whom the tests will be administered can, and often does, change

• AKA criterion-referenced testing
• concerned with the question of what the test tasker knows about a specified content domain

• consistency with which a test measures whatever it is that the test measures.
• if it is high, the test measures are dependable, reproducible, and consistent
• standard error

• an index of how much, on average, we expect a person's observed score to vary from the score the person is capable of receiving based on actual ability
• the smaller this is, the better

• test-retest
• alternate-form
• split-half

• same test is administered to the same group of people twice
• estimates the inter-individual stability of test scores over time

• tests reliability
• examinees are given two different forms of a test that are taken at two different times

• test takers take only one test, but that one test is divided into equal halves
• scores on one half are correlated with the scores on the other half

concerned with the extent to which a test actually measures what it purports to measure

• content
• face
• criterion
• cross validation
• construct - AKA convergent
• discriminant
• predictive

test's coverage fo the particular skill or knowledge area that it is supposed to measure

whether or not the test items appear to measure what they are supposed to measure

how well the test can predict an individual's performance on an established test of the same skill or knowledge area

testing the criterion validity of a test on a second sample, after you demonstrated validity on an initial sample

• how well performance on the tests fits into the theoretical framework related to what it is you want the test to measure
• AKA convergent validity
• in order to show a test has this , researchers also have to show that performance on the test is not correlated with other variables that the theory predicts that test performance should not be related to (discriminat validity)

• in order to show a test has this , researchers also have to show that
• performance on the test is not correlated with other variables that the
• theory predicts that test performance should not be related to.