Psych 434 Final

What do the parameters a, b, and c represent in item characteristic curves?

a=discriminablity parameter(steepness) b=difficulty parameter(P=.05) c=guessing parameter(bottoms out) Theta is the ability level.

%variance due to error

reaming so 100- % variance to due to true score

Index of relability

correlation between obtained score and theorertical true score

Estimated true score

find z score then mulitply by index then add (+)z and subtract (-)

How did Esquirol Contribute to the development of psychological testing?

french psychiatrist, known for classifying different mental disabilities, based it on people language disability. (history or clinical )

Confidence interval

if 68% cI then it just one SD, if 96% then you mulitply sem by 2 ex True score(+/-) Sem

A researcher reports that 10% of the variance in test scores depends on error variance. What does this Mean?

%due to error is 10%, %variance to true score=90% rtt=.90 subtract 100-10

What kinds of questions best illustrate the concept of criterion-related, concurrent validity?

Diagnostic test Does this test results represent the future or now?

Relablity cofficent(rtt)

% of variance due to true score(relablity corretled directly)

What is the difference between test sensitivity and test specificity?

(test sensitivity)test can pick up disorder present, ( A/(A+C), proportion of those with the outcome/diagnosis/condition with the risk factor/symptom/test result- a correct decision( retrospective epidemiological)starting from the conclusion and going backwards ex shows ADHD in EEG Abnormal. -Test specificity- when test cannot pick up disorder when not present. (D/B+D), proportion of those without the outcome/diagnosis/condition without the risk factor/symptom/test result- a correct decision(retrospective), true negative, ex don't have ADHD(respective) +EEG shows normal. Mulit martix (X axis outcome, Y axis is risk factors,symptoms) A True postive B False Postive C False negative D True negative

Explain the concept true-score confidence intervals. How are these intervals used? Why is this an important concept?

-the confidence interval is the possibility of difference between true score and observed score Add or subtract from the true score and the SEM -shows test has error in it.

A researcher reports that 85% of the variance in test scores depends on true variance in the trait measured. What does this mean?

15 PERCENT is error reliability coefficient= .85 True= variance 100-85=15%

What proportion of Y scores falls between plus and minus one standard error of estimate from the Y score that is predicted for a particular X score?

34% of the measures fall between the mean and plus or minus one standard deviation. 68% to 96% confidence interval

What would be a good illustration of the concept of convergent validation?

A grade in a english course and test grade based on a final exam= good relation, high correlations between the related items. measures of the same trait should be strong (Same-trait, different-method correlations are in bold and called the validity diagonal.).

In what kinds of situations would a test developer be concerned about face validity?

A test developer should be concerned about face validity when the test taker takes the test very seriously. Face validity of public can have influence and have importance. Popular media, To draw attention

Suppose that a personality test with four different subtests was correlated with a measure of success in college. How would you determine which personality subtest would be most accurate in predicting college success?

Accurate= validity coefficient, validity increases, increases accuracy, decreases error -.8 is better than +.7 absolute zero

How did Galton contribute to the development of psychological testing?

Another grandfather of psychology. Believed in the nature side, which means he believed what we inherit is from parents and none from the environment. Developed test focused on nature, book called hereditary genius. Selective breeding- Believes you can form any group with similar background. Evolutionary theory/ Comparative psychology- teaching language skills to other species other than humans ex personalities.

As the reliability of a test decreases, approaching zero, what happens to the standard error of measurement?

As the reliability of a test decreases, approaching zero the standard error of measurement increases as does can get as high as the standard deviation of the entire test. As the reliability of a test increases, moving away from zero the standard error of measurement decreases.

Why is it important to know the base rate of a condition in a particular population?

Base rate is important to know because the validity of test will give you accuracy to test have to beat 95% Test has to be better base rate opposite of maximum criterion proper probability. condition with a relatively low prevalence or base-rate is bound to yield high false positive rates, which exceed false negative rates Base rate= (A+C)/ (A+B+C+D)= proportion of those in the population who have the outcome/diagnosis/condition= change max criterion- 95% Base rate- 5%

How Do The various kinds of correlation coefficients align with scales of measurement? Examples of each of the following kinds of correlation coefficients-Biserial(rb)

Both contributed variables(ration/interval), 3rd coefficient, but one variable is dichotomous(arbitrary), nominal ex high/low anxiety and test score abitrary dictomy-split you decide

In a multitrait-multimethod matrix, what do different trait--same method correlations (i.e., the "c"s) represent?

C's represent different trait same method relates to validity coefficient= divergent validity(discriminate) Not correlated

How does skewness influence measures of central tendency?

Central tendency measures the mean, median and mode. Skewed Distribution- rarely uses mean and mode uses median. Means tend to be skewed so it can have inaccurate mean.(means pulled off in skewed)

What is generalizability theory? How does it differ from classical test theory

Classical theory - dividing error/ true score into error(undifferentiated) reliability,Xo=Xt+ E,Relizidual error variance due to true score/ different components of error score= undifferentiated generalizability theory-(G-study)-allows us to separate the effects of multiple facets and to adjust measurement strategies accordingly, multi-facet, different sources of error. reliability ,separate, different sources of error, the target is the focus of the study, rating each person on biligerance. Analyse data is to used analysis of variance, as if these items are generalized to all items that can be asked. D-study- decision study-focuses on the # items of the test( increases the amount of items, increased variability, increases reliability, increases correlation coefficient, item response curve, items cronbach's alpha= internal consistency. Takes results of error + estimates would get if i had different amount of observed items.

How did the theory of evolution and the development of comparative psychology influence the history of psychological testing?

Comparative psychology- teaching language skills to other species other than human. Ex Human Vs Aps, Figuring out certain species from another species- test of individual species,Cognitive function and personality(individual differences) Involves charles drawin Allows us to look at the personalities of other species and humans.

Describe two kinds of difference scores. What are some of the reliability concerns regarding difference scores?

Comparing two different measures, ex person iq vs score Or difference score over time and same time. Different scores are less reliable than the scores used to make it up. Test on how a persons scores verbally vs visual task or give same test at two diffrent times. two different to see if there is different construct Problem 2 different scores

What are two kinds of criterion-related validity?

Concurrent and predictive validity Concurrent- we assess the operationalization ability to distinguish between groups that it should theoretically be able to distinguish between. Predictive- assess the operationalization ability to predict something it should theoretically be able to predict.

The question, "What trait is the test really measuring?" is addressed by what kind of validity?

Construct Validity

What kind of validity addresses the extent to which the test may be said to measure a theoretical trait?

Construct Validity

The analysis of developmental changes on an intelligence test is most relevant to what kind of validation?

Construct Validity ex Multimethod matrix Assessing behavior- observe behavior, test them, ask a friend Traits(good listener) vs Method(observe behavior) **Multitrait matrix A=True Positive B=False Positive C=True Negative D=False Positive

.The question, "Are the test items representative of all possible test items?" is addressed by what kind of validity?

Content Validity

What kind of validity essentially involves the systematic examination of the test to determine if it covers a representative sample of the behavioral domain?

Content Validity

What is correlation? How does this concept relate to sums of products and covariance?

Correlation is simply how one variable interacts with another. Notated As "R". Correlations can range from -1 to 1, the closer to -1 or 1 the value is, the stronger the correlation is. Sum of Products- (SSx)(SSy)/N (variance) Covariance-Sum of products= SP/N Covariance is negative Variance is positive

In the Larry P. v. Riles case (which was filed on behalf of African-American children placed in classes for students with Mild Mental Retardation in the San Francisco public schools), greater degrees of error were found when using IQ scores to predict teacher grades for African-American children than for Caucasian children. What kind of bias does this demonstrate?

Criterion related Bias

What is the difference between a norm-referenced and a criterion-referenced test?

Criterion-Referenced test- Has #'s,Set criterion that a person must meet set standards. Norm-Referenced- Comparison, comparing individual to all other individuals in the test.(Ex percentile- the % of score that falls below or above your score.

What is the minimal acceptable level for a validity coefficient?

Depending on population,different levels or hits are optimal, no minimum level.

What is age and grade equivalence?

Dont know SD or variance, scaled in ordinal units not interval, cant add or subtract set up false standard derived by interplation(between diff grades guessing procedure) and extrapolation

What is Section 504 of the Rehabilitation Act of 1973?

Eliminates discrimination against people with disabilities and requires buildings to have ramps, wheelchair accessible entrances, bathrooms. This law Included students with ADD and ADHD!! (ME) included IEP"s

What is the difference between exploratory and confirmatory factor analysis?

Exploratory Factor analysis-The first step in determining how many constructs we are measuring. Use Eigen Values to determine the cutoff, the number of factors before the drop off in eigen values let us know how many factors we use. Confirmatory factor analysis is what's actually being measured, hypothesis orthogonal ex verbal and reading skills.

What kind of validity refers to what the test superficially appears to measure?

Face Validity

Should face validity be considered in test development? Why or why not?

Face validity is considered to be a public acceptance Yes because it can affect how a person responds.

What is factor analysis ?

Factor Analysis a way of determining how many constructs are tested on an exam. For example, there may be two different measures being tested, (a) and (b). We use factor analysis to see how many there are. Including Rotation Factors

The statistical technique of factor analysis is particularly relevant to what kind of validity?

Factor Analysis( Construct Validity)

What is orthogonal rotation?

Factors that are independent of each-other, not related and no correlation. Graph's maintain a 90 degree angle.

How did Wündt contribute to the development of psychological testing?

Grandfather of psychology light-switch experience to see when people noticed when the light was on/off aka Psychophysic test. Experimental Psychology

What is oblique rotation?

Graph's do not maintain a 90 degree angle, the factor dimensions are related to each-other.

If the correlation between predictor and criterion is +1.00 or -1.00, what is the standard error of estimate equal to?

Highest correlation between predictor and criterion is 1 or -1 and the SD=0 The standard error of estimate is equal to 0.

. What are some characteristics of speeded tests?

How fast you do the exam. (timed test) Demand Characteristics- The test taker changes his/her behavior depending on what you think the purpose of the test. EX) Social Desirability- based on what we think is acceptable. Malingering- try to lie about how well your doing. EX) lying on a exam. Expectation bias- Based on person giving the test. Ex teacher expects perfect memo's bc he expects she's smart.

.Error and reliability are indirectly/directly related.

If reliability is decreasing then error increases, low area of observe or obtained score. If error decreases the reliability increases, not going to be close to true score. Classical theory- Xo= Xt + E

If a test is speeded (i.e., timed), what would be the most suitable kind of reliability estimate?

If test is speeded the most suitable kind of reliability estimate would be test retest, equivalent forms. DON'T USE INTERNAL CONSISTENCY WITH SPEEDED TEST. AKA Split Half Reliability- indirect/ directed related to error Relablity increases/error decreases Relability decreases/ error increases

.If a test systematically underpredicts (or overpredicts) the criterion performance for a particular group, what kind of bias does it exhibit?

Intercept Bias

What are some kinds of split-half reliability coefficients?(can be split up any type of way)

Internal Consistency- Given only at one time, one test, no test-retest, no second form of test.Test to see if items measure the same thing or different. Odd-even method- dividing test in half so for example 1,3.5,7(creates one form of the test), 2,4,6,8( creates a 2nd form of the another test) Cronbach's alpha- internal reliability consistency, as if theoretically split test in half in all possible ways and find average for all correlation consistency and has a stable measurement for testing consistency. SPLIT HALF TEST SHOULD NOT BE USED WITH TESTS THAT ARE TIMED,EX) SPEED TEST. The longer the test the more variability, the higher the variability and the higher the correlation, the higher the correlation for reliability is.Test that are longer are more reliable. Spearman-Brown correlation- way to correct split half internal consistency reliability, our test= 0.72, you can tell what the reliability would be based on test length, shorter the interval, the higher the correlation. Kuder richardson-another kind of split half rulon -another kind of split half.

What is the relationship between the standard error of estimate and a criterion-related validity coefficient?

Inverse relationship,higher the error the lower the validity coefficient. r(xy) increases SEM decreases, r(xy) goes down SEM goes up (inversely related) The standard error of estimate is to validity what the standard error of measurement is to reliability When a predictor is used to predict a criterion, there is error in measurement or prediction. the standard error of estimate is the standard deviation of the predictor errors, and its provides a range in which a client's score will fall, given a clients score on a test... Because criterion related validity is a correlation coefficient, when scores do not variability, or there is a restriction in range, the validity coefficient tends to be low. this is usually the case, especially for concurrent validity coefficients, and they tend to underestimate predictive coefficients. since criterion related validity is based on a correlation coefficient, it is affected by nonlinear data.

How do issues such as variability and test length influence correlation Coefficients?

Issues such as variability and test length influence correlation coefficient is that the longer the test the higher the variability. More variability the higher the correlation, longer the test the more variability. Variability- how much difference in test scores.

if a test shows score differences in favor of one of two subgroups, but has the same regression line for both groups in predicting the criterion, what can be said about the test?

It Does not have bias.

What was the outcome of Parents in Action on Special Education (PASE) v. Hannon

Judge was Grady- did not care about correlation, was interested in the questions on test and white and black students abilities to answer them. Based on similar situation to Larry p that occurred in Chicago. Test was ruled not unfair for black children. Bias items were few

Give some examples of psychological tests.

Multiple Choice test Open-ended test-people can answer test questions by saying anything they want in response (how are you)to the questions on the test. Close ended test-require people to answer questions by choosing among alternative answers provided in the test.(yes or no) Ink blot test- personality assessment,operation of a latent trait. TAT(Thematic Aptitude Test)- People draw in all different types of situations. Speed test, projectective test (unstructured)

Explain and give examples of the four scales of measurement.

Nominal Scale-ethnic background, Bar graphs, has identity. Ordinal scale- property of order and property of identity. Naming ex 1st, 2nd,3rd Interval scale-property of identity, property of order, property of quantity. Numbers all equal,number line, zero is a place holder. Ratio scale-property of identity, property of order,property of quantity, property of absolute zero. Can't get below zero

On a selection test for firefighters, average scores for men and women were 85 and 75, respectively. Is this test biased?

Not necessarily bias, Can't say test is bias because the means are different elligartian fallacy look at the regression lines Same lines= no bias Different lines= bias

Explain the following (a) obtained score, (b) true score, and (c) error.

Obtained score-(Xo) observe test score, scores that are observed.(raw score, actual # of corrected answers) True score-(Xt)- What the person actually scored on the test.( Student would get if measurement were accurate and error free) Error-(E)- When marked the wrong answer but knew the right answer or if you guess and got the answer right. (anything that is irrelevant to test content If there is no error in a test, than the obtained score would be your true score. Goal: get tester's true score. Construct test that there is no error.

What patterns of correlations would cause you concern in a multitrait-multimethod matrix?

Patterns of correlations that cause concern are different trait/same method Different trait/ different method ' When c's are high because we're measuring, the method not the trait is being measured and it has no construct validity

What kinds of validity studies would address the accuracy of the Scholastic Aptitude Test(SAT) taken during the senior year of high school for estimating cumulative grade-point average (GPA) in college?

Predictive Validity

The question, "How well do scores on the test predict performance on some other behavior?" is addressed by what kind of validity?

Predictive validity

What kind of validity indicates the effectiveness of a test in predicting an individual's performance in specified activities?

Predictive validity

How do publishers of tests safeguard security and use of tests?

Publishers safeguard test by having certian qualifications for a test in order to buy a test. an example of this is buying an SAT. a teacher can buy them but a student cannot.

What kinds of questions best illustrate the concept of content validity?

Questions that best illustrate concept of content validity are questions that shows what someone learned in the past. Does this reflect what we're hoping that it will reflect? Can those who have a deep understanding of the construct in question successfully approve/evaluate the construct in question?

Explain the concept of regression to the mean and provide an example of this.

Regression to the mean the score u predict is closer to the mean then the score u use to predict it with. High score- closer to mean,lower Low score- closer to mean but higher. Ex) obtainted score to predict a true score regression=line of best fit

Item response theory

Related to parameters Y axis- easy item(1) to Hard item(0) P value X axis Trait- low to high 50% chance and 100% (T/F) to get right Ex #12 is more difficult item, discriminates two levels of theta level of these similar items P values don't just depend on items but also on the group population Theta-measures the trait being measured Items with broder slopes are more broad. Discriminatory of power of the item is based on the steepness of slope Three parameters, difficulty,discriminability parameter guessing parameter- doesn't go lower

How would you decide which test has the highest amount of error?

SEM DIFFERENCE FORMULA , Test with lowest relablity has highest amount of error

How would you calculate and interpret the standard error of Measurement?

SEM-Way in which we determine the error of a test. SEM= SD 1-rtt rtt= reliability Ex if rtt=1 Then the Sem= 0 which means the SEM=0 This states their is no error at all on this test. Worst it can be is rtt=o, rtt=SD of test= Error increases as SD of entry test Ranges from low=0(rtt=perfect) to high which is the rtt= SD of the entire test.

What are the differences between the formula for computing standard error of measurement (SEM) and the formula for computing standard error of estimate (SEest)?

SEest( Standard error of estimate) validity= r2, bc it has not been squared Seest=0 (no error of prediction at all),Y represents degree of error No error of estimate of criterion measures relates to scatter plot error= non bc exactly predict score rxy = 1 no error Strong correlation=1, Positive correlation less than 1 Highest validity correlation= SD of Criertion not test, lowest=0 Seest ranges from 0 to the SD of criterion Doesn't matter which way you write (predicting criterion achievement vs IQ) Standard error estimate can be different bc of SD/Variance/criteion may be different if the SDy = 10 SDx= 15 that has different values Standard error of measurement= Reliability r not squared because it's already been squared, index of reliability

Homoscedity-hot dog scale

Same distrubutions of x, influences correlation coffeiecent stronger than Hetro

In a multitrait-multimethod matrix, what do same trait--same method correlations (i.e., the "a"s) represent?

Same trait-same method refers to reliability. Expect correlation coefficient to be positive and would approach one reliability, high if it has good construct validity.

What was the outcome of Larry P case?

San francisco 70's, special education classes Validity of individually administered test. Blacks were overrepresented in schools Claimed children were not intellectually disabled and the test who placed Blacks were biased. The Judge of this case was Peckham. 28.5% of children were black, 66% were in Mild ID classes( Disproportion of Race) School system tried to prove test was biased. Section 504 Court decided that the correlational was the IQ score on there test with Teacher grades. Court said test were bias against blacks due to the fact that the correlation was very low Criterion validity used.

How would you interpret a scatter plot? What is an outlier?

Scatter plot- if the line lower left upper right= positive correlation(describes linear relationship),slope of points suggest direction of relationship between 2 variables(+/-)Point scatter show stregth or correlation. If line upper left lower right = negative correlation Closer to the line the stronger relationship, further weaker relationship. Points closer to mean= strong Points far from the mean=weaker Outlier- does not fit with the rest of the scores.

How would you calculate and interpret the standard error of Difference

Sem diff=Sem2 +Sem2 ex) IQ= 3, Achivement=4 sqaure root of 3 sqaured plus 4 squared is 5 the amount of error in the different score larger than SEM of either two test Standard error is bigger than both SEM.

What was the outcome of Brown v. Board of Education

Southern Us Law Required schools to desegregate. Psychological test were used in order to get rid of segregated schools.(tried to find loopholes)By only giving blacks the exam and not whites. If blacks did not score high enough they couldn't attend white school. The test was never questioned Legal question: Was this test used properly. Court found it illegal to use this test. If test is used, it must be used with everybody the same way.

What is the Spearman-Brown formula used for?

Spearman-Brown- way to correct split half internal consistency reliability, ex our test=.72, you can tell what what the reliability would be based on test length. Item analysis- Removing items from an exam Relablity for attenunation and increases the accuracy of the relability coffiecent.

What sources of information are needed to determine the standard error of estimate?

Standard deviation and validity coefficient

What is the relationship between standard deviation and variance?

Standard deviation is sample statistics, how far the score is from the mean on the distribution chart. Variance- Variance is SD squared(Sums of squares) Standard Deviation cannot be negative

What is the relationship between standard error of measurement and standard error of estimate?

Standard error of estimate( how much is in criterion predict from test score)= validity Standard error of measurement(how much error is in a test)= reliability

How would you interpret the standard error of estimate for a test?

Terms of degree of error in terms of prediction(SD) The standard error of estimation is interpreted in the same way as the standard deviation. The standard deviation tells us how spread out a distribution of scores is with respect to the distribution mean. The standard error of estimate tells us how spread out scores are with respect to their predicted values. If the error scores (E = Y - YPred) are normally distributed around the prediction line, about 68% of actual scores will fall between ±1 sE of their predicted values.

What factors contribute to error in test scores?

Test itself(worded hard, not understanding wording) Environment(room may be too cold or hot) Person administering the test Test Error- Calculate reliability. Test taker(tired,anxiety)

Criterion-related-predictive validity

The ability of a test to predict a subject's test performance. An example of this is a student's score on the SAT compared to how well they do in college. The important thing for this test is to decide if the results are accurate.

What does the standard error of the estimate indicate?

The degree of accuracy in predicting criterion, function of validity coefficient and SD, Sdy square root 1-rxy 2 The standard error of the estimate is a measure of the accuracy of predictions. Recall that the regression line is the line that minimizes the sum of squared deviations of prediction (also called the sum of squares error). Standard estimate= SDy1-rxy2 It turns out that test is the standard deviation of the errors of prediction

What does a reliability coefficient of rtt = .90 indicate

The reliability coefficient of rtt=.90 indicates that's it's already been squared, correlated these two items and gives use the internal consistency, test to see if items measure same thing or different. 90% true score 10%error

If the correlation between predictor and criterion is zero, what is the standard error of estimate is equal to?

The standard error of estimate is an indicator of the accuracy of prediction. It is equivalent to the standard deviation of the residuals. If there is perfect prediction all of the residuals will be zero and the standard error of estimate will be zero. If there is no prediction (zero correlation), the residuals will be the same as the deviation scores and the standard error of estimate will be the same as the standard deviation of the Y scores (with a slight adjustment for one less degree of freedom). SEest= 0 since there is zero correlation between predictor and criterion.

Against what axes are item characteristic curves plotted?

Theta Horizontal, x is the traits Y is difficulty, vertical probability of being correct

What is non random or systematic error in test scores called?

This is the definition of bias, error is not randomly distributed

Explain how truncated range, preselection,and moderator variables can affect correlation coefficients.

Truncated Range- restricted range of IQ, can't tell if it's positive or negative, higher variability the higher the correlation, longer test have more variability, correlation establish variability. Preselection-goes with truncated range, if make range smaller, preselect out ur making range more restricted. (prechoosing the individuals to take a test) mediator variable- Explains the relationship between 2 variables,ex) poverty and outcome of academic achievement. Moderator variables-any variable that changes the degree of correlation between 2 variables, If z score increases,increase correlation. If z decreases, no correlation can lead to a 3rd variable present. Ex) poverty + academic outcome= negative correlation, 3rd variable that changes relationship between 2 other variables.

Explain and give examples of the following kinds of reliability-Equivalent forms delayed(alternate)

Two test given but wait a period of time to give the second test(everyone takes the same test), sampling across time(temporary/response consistency)

How Do The various kinds of correlation coefficients align with scales of measurement? Examples of each of the following kinds of correlation coefficients-Phi

Two true dichotomy(naturally occuring split passing. failing item on a test)

What is the field of psychometrics concerned with?

Type of information (scores) generated by the use of psychological tests. Reliability of data from psychological tests. Issues concerning the validity of data obtained from psychological tests.

What are some benefits of understanding the field of psychological measurement?

Useful to understand, to be able to select test that are valid in order to get accurate information.

In communicating test results, how can one assure not only accurate comprehension but also acceptance by a recipient?

Using lanugage their familar with, dont use techincal terms ex) mean instead use above or below.

In our class discussion of the multitrait multimethod matrix, I described a study that showed that the different trait--same method correlations (i.e., the "c"s) were higher than expected. What did this suggest?

Value of validity/reliability= sometimes high=same method is counting for variance not the trait That there isn't any construct validity A highest, b high,c low,d lowest

Variablity

Varabilty Increases, correlation increases Variablity decreases, correlation decreases

What was the outcome of Hobson v. Hansen

Washington D.C public schools Blacks were placed in lower tracks of the school(low end). Students were branded as low track students. Or previously known as EMR(Education mental retardation) Levels- EMR, TMR(trainable moderately intellectually disabled) Legal question: If this violated the 14th protection amendment clause. Court ruled it okay to use if test measured natural abilities of the child, if used for something learned or achieved. (aptitude) Ruled tests as inappropriate can be seen as a disapproval validity.but improved educational placement.

Hetroscedity-pear shape

Y scores can differ from the x-scores Depending on the distrubutions of X diffrerent distrubution of for Y Ex ) acedemic achievement- more variablity increases hetro increase correlation

Can a test show mean group differences and still be considered unbiased? Why or why not?

Yes as long as there's a regression line, the mean group differences and still be considered unbiased.

How do T-scores, z-scores and percentiles align in normal distributions?

Z scores- tells us how above or below the score is from the mean. Z=x-u/o -mean of Z score is always zero, SD is always 1. T-scores- mean of 50, SD=10. Always T, converted standard score and z is the individual's original z score. Percentiles-indicates the percentage of scores that are below or above a specific test score.(ordinal scale) 2 to 98 Example if we know that a test taker has scored at the 85th percentile, then we know that the person has a relatively high score. IQ= 70 to 130

Face Validity

a degree in which a measure appears to be related to a specific construct. This validity focuses more on the test taker themselves and how it relates to them specifically than the test creator. In order for this test to be valid the test taker must care about the test

How Do The various kinds of correlation coefficients align with scales of measurement? Examples of each of the following kinds of correlation coefficients-Tetrachoric(rt)

both variables have reduced arbitrary(make a decision yourself) to dichotomy ex) pass/fail a test 50 or 70% fail

How Do The various kinds of correlation coefficients align with scales of measurement? Examples of each of the following kinds of correlation coefficients- Spearman's Rho

correlation coefficient,range -1 to 1, same as pearson but scale of measurement for x and y variables are in ordinal units. Ex) first second third in a race, something else. (both variables)

What is Eigen value ?

determines how much variance is accounted for or the amount. Determines how many factors are in the experiment Screenplot- plot of Eigen values against numbers this data set has 2 numbers in it(2 factors)

Content Validity

extent to measure all the items on the test that is relevant to the topic that is provided on the test. In other words, this type of validity measures what you want to measure.

How did Binet contribute to development of psychological testing?

helped to differentiate children vs children who were less able to benefit from the school system designed to see which children benefit from the school structured system known as a intelligence test. It was later renamed the stanford binet test. Some people called it intelligence scale- which test mental abilities, know known as IQ- ratio. (psychiatric/ clinical perspective)

Construct Validity

how accurate the results of the materials that are provided on a test. It also shows how well they can use these items on the test for grading. By using the multiple intelligence scale, it allows you to focus on a person's traits or skills, and can be determined by what items on the test are related and what's not related. However, construct validity focuses on the trait that is being measured.

crietion and content validity

levels of difficulty of items within the test.

What is discriminant( or divergent) validity

low correlation between unrelated traits,something that is not related to,items are not related to what is being measured on the test

Response Bias

malingering- lying about the how well your actually doing acquiescence(saying yes to everything)

What is the purpose of a multiple-choice test?

measure candidate ability with regard to a specific content area, asses the latent trait.

How Do The various kinds of correlation coefficients align with scales of measurement? Examples of each of the following kinds of correlation coefficients- Pearson

most common, requires you to have 2 sets of data in interval units or ratio units,ranges from -1 to 1, closer to -1 or +1 the stronger the correlation, close to 0 weaker the correlation, Negative correlation=opposite direction, Positive correlation=Same direction. Correlation between two test scores ex) IQ test vs Achievement test requires peason

How Do The various kinds of correlation coefficients align with scales of measurement? Examples of each of the following kinds of correlation coefficients-Point biserial (rpb)

one continuous variable and on that's dictomised. Ex) Binary- Gender=women/men,item on a test). Used most often with how you would use a particular item on a test.Correlates with the total score of test. If negative correlation then it's a bad item or bad test. True dictomiy- naturally occuring Dictomoy- splitting something in half

What is malingering?

people might change their behavior to convey a poor impression to the person doing the measurement. Lie about how well you are doing.

Item difficulty

related to how well a certain population can answer an item on a test. and P value

Explain and give examples of the following kinds of reliability -Test retest

reliability correlation coefficient between obtained scores in test retest problems:Test is given twice same test two different times. shorter the interval, the higher the correlation Example or problems of test retest would be Participants could remember answers, fatigue,same test may be given these 81% of obtained score in variance is a true score. Looks for consistency across time.

What is social desirability?

some people might become apprehensive, others might change their behavior to try to impress the person doing the measurement. Ought to act then how we really act.

How did Seguin development of psychological testing?

taught people with disabilities and formed boards with shapes that are caught out for certain shapes to be placed in the right hole.(history/ clinical)

What is Observer bias?

tendency to see what we expect to see, or what we want to see. When a researcher might have a vested interest in certain intelligence test score outcomes, and as a result, he might allow a bias, perhaps even an unconscious one, to influence the testing procedures.

What is meant by the index of Reliability?

the correlation between the obtained score and true score, decrease of reliability, rxoxt=.90(square this, index of reablity) r2=.81( r coefficient,correlation determation), rtt=.70(already squared) (test retest, reablity coffiecent) Square root of the reliability. IT is a statsic that provides a theortical estimate of the correlation.

Criterion-related- concurrent validity

the opposite of predictive in which the degree of test scores that are correlated with other relevant variables. This test measures the present and not the future.

Explain and give examples of the following kinds of reliability- Equivalent forms- immediate(alternate)

two different test given at the same time, response consistency- testing/ process items(rtt).

Explain and give examples of the following kinds of reliability-Split Half

wants to be corrected for test length. One test given at one time and results are split in half. Aka internal consistency.

What is a latent variable-(Hypothetical construct)

working memory skills, that we are tempting forms to test. ex) ink-plot theoretical psychological characteristics, attributes, processes, or states that cannot be directly observed, and they include things such as learning, intelligence, self-esteem, dreams, attitudes, and feelings.