Research/Stats Flashcards
This experimental research variable is the one that is manipulated by the researcher; it is sometimes referred to as the "input variable."
Independent
variable
What term is used to denote the values a research variable, most often the independent variable, could take (e.g., placebo only, placebo + treatment, treatment only)?
Levels (of
a variable)
In experimental research, the hypothesis can usually be stated: changes in the \_\_\_\_\_\_\_\_ will be caused by changes in the \_\_\_\_\_\_\_\_.
Dependent
variable;
independent
variable
In what type of research design are all levels of one independent variable combined with all levels of another independent variable?
Factorial design (if one IV has 3 levels and the other has 2, it would be called a 3 x 2 factorial design)
Referred to as the "outcome variable," it is hypothesized to change as a result of particular manipulations of the input variable in experimental research.
Dependent
variable
A study is said to have \_\_\_\_\_\_\_\_ validity when a causal relationship is determined to exist between the independent and dependent variables.
Internal
In correlational research, independent variables are often termed \_\_\_\_\_\_\_\_ to differentiate them from variables that can be manipulated, and dependent variables are referred to as \_\_\_\_\_\_\_\_.
Predictor;
criterion
TRUE or FALSE: A study in which no significant effect of the independent variable is found will always have no internal validity.
FALSE: It will have internal validity if the conclusion is that the IV has no causal effect on the DV
Poor internal validity in a study is the result of numerous possible extraneous variables, typically called \_\_\_\_\_\_\_\_ variables, that might explain the change in the dependent variable.
Confounding
(or
confound)
Changes in the behavior of test subjects (DV) that is more a function of the expectations of the experimenter as opposed to the independent variables is referred to as what?
The
Rosenthal
effect
What is the single best way for a researcher to ensure internal validity when conducting an experiment?
Treat all "levels" of the IV with equivalence (equally) in every respect, except for their IV status
This is likely the best technique for ensuring internal validity, as it distributes all extraneous subject characteristics equally to all research groups, thereby reducing the likelihood of confounded results.
Random assignment (or randomization)
Regarding confounding variables effecting change in the
DV, ________ is a result of biological or psychological
change in the research subjects; ________ is a result of
one’s prior experience with a test; ________ might occur
when there have been changes in the measuring instrument
(e.g., psychologists gets better at diagnosing); ________ is
caused by pre-existing factors in the research subjects
(e.g., higher IQs); and ________ is the result of people who
drop-out of one research group systematically differ from
those who stay in the study.
Maturation; testing; instrumentation; selection; differential mortality
Random \_\_\_\_\_\_\_\_ occurs after study subjects have already been selected to participate and refers to the equal probability of being assigned to a group, whereas random \_\_\_\_\_\_\_\_ occurs prior to starting the study and refers to a method whereby all members of a population have an equal chance of being chosen to participate in a study.
Assignment;
selection (or
sampling)
This term is used to describe the tendency for a test subject who initially produced extreme scores on a measure to score closer to the mean on subsequent testings; it has an effect on internal validity when extreme scorers are used as research subjects.
Statistical
regression
Experimenters are interested in measuring the
effect that different treatment modalities (IV)
have on relapse rates for alcoholics; however,
there is concern that the duration of each
subject’s alcohol use may confound the results.
The researchers decide to split the subjects
based on their duration of use, then randomly
assign them to treatment groups. This is an
example of what technique to increase internal
validity?
Matching
What technique for ensuring internal validity involves essentially making the confound variable another independent variable, allowing the researcher to separate the effects of each variable?
Blocking
When the result of a study performed in the 1920s does not apply to the population in 2010, there is an interaction between \_\_\_\_\_\_\_\_ and \_\_\_\_\_\_\_\_.
History;
treatment
This statistical strategy for increasing internal validity involves adjusting DV scores after the data has been analyzed so that subjects statuses are equalized on 1 or more variables; it is similar to matching, only it is performed post-hoc.
Analysis of
Covariance
(ANCOVA)
An interaction between \_\_\_\_\_\_\_\_ and \_\_\_\_\_\_\_\_ occurs when the results of a study where pretests were used do not generalize to cases in which pretests were not used.
Testing;
treatment
If the results of a study can be generalized to other settings and situations, it would be said to have high \_\_\_\_\_\_\_\_.
External
validity
What is the term used to describe the phenomenon where research participants form an interpretation of the experiment's purpose and, consequently, unconsciously change their behavior or responses accordingly?
Demand
characteristics
When the effects of a given treatment are not generalizable to other members of the target population, it can be said that there is an interaction between \_\_\_\_\_\_\_\_ and \_\_\_\_\_\_\_\_.
Selection;
treatment
A research subject who behaves differently merely due to the fact that she are aware of her participation in an experiment exemplifies what threat to external validity?
The
Hawthorne
Effect
In studies where the same subjects are exposed to more than 1 treatment (called repeated measures design), there is a risk that external validity will be decreased as a result of the influence of the multiple treatments, thereby reducing generalizability. This is referred to as what?
Order effects, carryover effects, or multiple treatment interference
In contrast to naturalistic research, which has high external and low internal validity, \_\_\_\_\_\_\_\_ research tends to have higher internal and lower external validity.
Analogue (research performed in more laboratory-like setting)
A good way of increasing external validity, \_\_\_\_\_\_\_\_ ensures that all members of a population have an equal chance of being selected to participate in the research, and thereby produces results that accurately represent the whole population (it is generalizable).
Random sampling (or selection)
A \_\_\_\_\_\_\_\_ study involves the subjects lacking awareness of the purpose of the experiment and the treatment they have been assigned to; in a \_\_\_\_\_\_\_\_ study, both the subjects and the experimenters lack awareness of the group subjects have been assigned to.
Single-blind;
double-blind
This procedure involves taking a random sample from each of numerous subgroups of the total target population, thus ensuring proportionate representation of the defined population subgroups in the study.
Stratified
random
sampling
One way of controlling for order effects is to administer treatment to groups of subjects in a different order, which is referred to as what?
Counterbalancing
While a threat to internal validity, \_\_\_\_\_\_\_\_ research increases external validity by observing subjects' behavior in a real-life setting.
Naturalistic
What counterbalancing technique orders the administration of treatment so that each appears only once in every position?
Latin
square
design
The defining feature of a \_\_\_\_\_\_\_\_ experiment is that subjects are randomly assigned to different groups; in contrast, \_\_\_\_\_\_\_\_ experiments are performed when it is impossible to randomly assign subjects to groups (e.g., intact groups of subjects).
True;
quasi
In cross-sectional research, what term describes the observed differences between different age groups that are more a function of experience than age?
Cohort (or
intergenerational)
effect
In this type of research, variables of interest are not manipulated, there is no internal validity, results are typically used for the purposes of prediction, and associative rather than causative relationships are discussed.
Correlational
research
Longitudinal research tends to \_\_\_\_\_\_\_\_ true age-related change (due to drop-out of less-abled subjects and practice effects), while cross-sectional tends to \_\_\_\_\_\_\_\_ true age-related change (due to cohort effects).
Underestimate;
overestimate
What type of research is used to address variables as a function of time, such as aging or maturation?
Developmental
research
This type of research design consists of taking multiple measurements over time in order to assess the effects of an independent variable; the series of measurements on the dependent variable is interrupted by administration of the treatment.
Interrupted
time-series
design
The 3 types of developmental research are \_\_\_\_\_\_\_\_ (same people studied over a long period of time), \_\_\_\_\_\_\_\_ (different groups of subjects who are divided by age are assessed at the same time), and \_\_\_\_\_\_\_\_ (representative samples of different age groups are assessed on 2 or more occasions).
Longitudinal; cross-sectional; cross-sequential (which is a combination of the first 2)
In \_\_\_\_\_\_\_\_ designs, which work well for research on behavior modification, one subject serves as his/her own control; behavior (dependent variable) is analyzed numerous times during a baseline phase where there is no treatment, then again during a treatment phase (independent variable).
Single-subject
Of the 3 single-subject designs, a/an ________
involves collecting data from a single baseline
phase and a single treatment phase; in a/an
________, data is collected after the treatment
phase has stopped and, if the behavior returns,
there is a higher chance the change was due to
treatment; and ________ involves applying
treatment sequentially and measuring the DV
after each application, rather than removing
treatment (this is used primarily when removing
treatment would be unethical).
AB design; reversal (withdrawal) design; multiple baseline design
Identify which of the 4 scales of
measurement would be used for each of
the following variables: (1) IQ scores or
scores on most standardized tests, (2)
diagnostic category, (3) the length of time
(in minutes) it takes to complete a task,
and (4) rankings of one’s preferences of
listed therapist characteristics.
(1) Interval, (2)
nominal, (3)
ratio, and (4)
ordinal
What 2 reversal designs involve (1) collecting data at a third non-treatment/baseline phase and (2) re-administering treatment for a second time, collecting data, and comparing the results to the first treatment phase?
(1) ABA
design; (2)
ABAB design
What is the only difference between interval and ratio scales of measurement?
Ratio variables have an absolute zero point (e.g., distance, height, weight, time); interval variables have no absolute zero, so negative values can be used, and ratios are arbitrary- values cannot be multiplied or divided (e.g., IQ scores, temperature)
In \_\_\_\_\_\_\_\_ research, theories are derived from the gathered data rather than developed prior to the experiment; methods for collecting data include surveys (e.g., personal interviews, phone/mail surveys), case studies, and protocol analysis (collecting and analyzing word-for-word reports from subjects).
Qualitative,
or
descriptive
What descriptive statistic would be used if a researcher was interested in simply showing a summarized grouping of data she has collected, organized by the number of cases that fall within a given range or category?
Frequency distribution, which are often displayed on a bar graph (histogram), table graph, or polygon
The two types of statistical methods are \_\_\_\_\_\_\_\_ statistics, which are used to quantitatively describe the main features data collected from a sample, and \_\_\_\_\_\_\_\_ statistics, which are used to make inferences regarding some unknown aspect of a population based on sample data.
Descriptive;
inferential
A bell-curve exemplifies a ________ distribution,
which is symmetrical due to half of the scores
falling above the mean and half falling below the
mean; a ________ distribution occurs when the
majority of scores fall on the high end of the
scale (e.g., test was too easy), leaving a long tail
on the left; a ________ distribution occurs when
most scores fall on the low end of the scale (e.g.,
test was too difficult), leaving a long tail on the
right.
Normal; negatively; positively (hint: "the tail tells the tale," so whatever end the tail is on determines the labeling)
What are the 3
most common
measures of
central tendency?
Mean,
median,
and mode
What are the 3 most common measures of variability (or dispersion)?
Range,
variance, and
standard
deviation
The ________ is determined by dividing the sum
of all values by the number of cases/scores (or
N), providing the most useful measure of central
tendency (MCT) for normal distributions; the
________ is the middle value when data is
ordered from lowest to highest, making it less
sensitive to extreme scores than the prior MCT
and thus more useful with skewed distributions;
the ________ is the most frequent value in a
data set (can have more than one).
Mean;
median;
mode
The \_\_\_\_\_\_\_\_ is determined by simply subtracting the lowest score for the highest score in a data set, which limits its usefulness; the \_\_\_\_\_\_\_\_ is used by many statistical tests in their formulas and it measures a distribution's variability; the \_\_\_\_\_\_\_\_, in simple terms, shows how much variation there is from the mean.
Range;
variance;
standard
deviation
How is the median calculated when there are an even number of cases in the data set?
Take the
mean of the 2
middle values
The \_\_\_\_\_\_\_\_ is the square of the \_\_\_\_\_\_\_\_, which means the \_\_\_\_\_\_\_\_ is the square root of the \_\_\_\_\_\_\_\_.
Variance; standard
deviation; standard
deviation; variance
In a negatively skewed distribution, \_\_\_\_\_\_\_\_ > \_\_\_\_\_\_\_\_ > \_\_\_\_\_\_\_\_; in a positively skewed distribution \_\_\_\_\_\_\_\_ > \_\_\_\_\_\_\_\_ > \_\_\_\_\_\_\_\_.
Mode > median > mean; mean > median > mode (hint: mean is always closer to the tail)
This type of transformed score measures the amount of standard deviations a raw score is from the mean; once all raw scores have been converted, the distribution has a mean of 0 and a standard deviation of 1, permitting comparisons across different measures/tests.
Z-score
TRUE or FALSE: When raw scores are transformed into z-scores, the shape of the distribution stays the same.
TRUE: When score transformation does not change distribution shape, it is called "linear transformation"
Percentile ranks will always have a \_\_\_\_\_\_\_\_ distribution, meaning that within a given range of percentile ranks, there will always be the same number of scores.
Flat (aka
rectangular
or uniform)
What transformed score has a mean of 50 and a standard deviation of 10 (e.g., MMPI-2)?
T-score
TRUE or FALSE: When raw scores are transformed into percentile ranks, the shape of the distribution stays the same.
FALSE: When score transformation changes distribution shape, it is called "nonlinear transformation"
This type of transformed score divides a distribution into 9 equal intervals with a mean of 5 and a standard deviation of 2.
Stanine scores
(shortened form
of STAndard
NINE)
In a normal distribution, what percentage of scores falls within -1 and +1 standard deviations of the mean? Within -2 and +2? Within -3 and +3?
Approximately 68%, 95%, and 99% of scores fall within -1 and +1, -2 and +2, and -3 and +3 standard deviations, respectively
Expressed as the percentage of individuals in a standardized group scoring below an individual's raw score, \_\_\_\_\_\_\_\_ are often helpful when providing feedback from intelligence and psychological tests.
Percentile
ranks
TRUE or FALSE: When raw scores are transformed into z-scores, the shape of the distribution stays the same.
TRUE: When score transformation does not change distribution shape, it is called "linear transformation"
In a normal distribution, a z-score of +1.0 is equal to a percentile rank of \_\_\_\_\_\_\_\_ and is therefore the cutoff point for the top \_\_\_\_\_\_\_\_.
84; 16% (the same is true for the converse- a z-score of -1.0 has a percentile rank of 16 and is therefore the cutoff for the bottom 16%)
The extent to which a sample mean can be expected to deviate from its corresponding population mean is referred to as the \_\_\_\_\_\_\_\_ and is calculated using the formula: \_\_\_\_\_\_\_\_.
Standard Error of the Mean (SEM); standard deviation/square root of N (N = sample size)
In a normal distribution, a z-score of \_\_\_\_\_\_\_\_ is equal to a percentile rank of 98 and is therefore the cutoff point for the top \_\_\_\_\_\_\_\_.
\+2.0; 2% (the same is true for the converse- a z-score of -2.0 has a percentile rank of 2 and is therefore the cutoff for the bottom 2%)
When calculating the standard error of the mean, what occurs as the sample size (N) becomes larger?
Error becomes smaller (there is an inverse relationship between sample size and SEM)
The formula for a \_\_\_\_\_\_\_\_ involves subtracting the sample mean from the obtained raw score, then dividing that value by the standard deviation.
Z-score
The standard error of the mean is \_\_\_\_\_\_\_\_ proportional to the standard deviation and \_\_\_\_\_\_\_\_ proportional to the sample size.
Directly;
inversely
This term refers to the inaccuracies that are a result of observing a sample rather than the whole population when using inferential statistical methods; it is the difference between the statistic (sample value) and the parameter (population value).
Sampling
error (or “error
of the mean”)
The difference between \_\_\_\_\_\_\_\_ hypothesis testing and \_\_\_\_\_\_\_\_ hypothesis testing is that the former states the hypothesis in theoretical terms while the latter states the hypothesis in terms that are quantitatively testable.
Research;
statistical
The hypothesis that the independent variable will have no effect on the dependent variable is called the \_\_\_\_\_\_\_\_ hypothesis, while the \_\_\_\_\_\_\_\_ hypothesis hypothesizes the independent variable will have a significant effect on the dependent variable.
Null (H1:
u1=u2);
alternative (H1:
u1/=u2)
On a graphical representation, when alpha level is changed from .05 to .01, the \_\_\_\_\_\_\_\_ region gets smaller and the \_\_\_\_\_\_\_\_ region gets larger.
Rejection (i.e., level at which null hypothesis is rejected); retention (i.e., level at which null hypothesis is retained)
The hypothesis that the implementation of a certain OCD treatment program will significantly reduce obsessions and compulsive behaviors exemplifies a \_\_\_\_\_\_\_\_ hypothesis, while the hypothesis that a new therapy approach will change the mean test scores of patients with schizophrenia exemplifies a \_\_\_\_\_\_\_\_ hypothesis.
One-tailed (the direction of the change is hypothesized); two-tailed (no direction of change is hypothesized)
A ________ occurs when the null hypothesis is
not rejected when it is false- basically, the
researcher finds no significant difference when in
the population a difference does exist, creating a
“false-negative.” ________ is the probability of
making such an error, while ________ refers to
probability of not making such an error (or, the
probability of rejecting the null when it is false).
Type II error; beta; power (refers to "sensitivity of a statistical test;" equal to 1-beta)
A ________ occurs when the null hypothesis is
rejected when it is true- basically, the researcher
finds a significant difference when in the
population no difference exists, creating a
“false-positive.” The ________ level refers to the
probability of making such an error and is usually
set at the .01 or .05 level, which means a
researcher would reject the null hypothesis if the
test indicates that there is a 1% or less or 5% or
less chance of the null hypothesis being true.
Type I
error; alpha
(p value)
While the power of a statistical test is not known prior to a study and usually remains unknown, what are some ways a researcher can increase power?
Increase sample size, increase the pre-set alpha level, use a one-tailed (directional) test, increase the difference between levels of the independent variable
The probability at which a researcher rejects the null hypothesis as being true is called what?
The significance level, usually expressed "the results were significant (p < .01)"
TRUE or FALSE: As the probability of making a Type I error increases, the probability of making a Type II error decreases, and vice-versa.
TRUE: Hence researchers must consider real-world consequences of making these errors when setting alpha
In research where the hypothesis is
counter-intuitive and contradicts previous
findings/theory, alpha should be set at ________
in order to reduce the chances of making a
________ error (remember “COLD FUSION”);
however, alpha should be set at ________ in
situations where making a ________ error might
have devastating effects (e.g., research on new
medicine to treat a life-threatening illness).
A lower value
(.01); Type I error;
a higher value
(.05); Type II error
What assumption is shared by both parametric and nonparametric tests?
Random
selection of
subjects from
the population
The ANOVA and the t-test are both examples of \_\_\_\_\_\_\_\_ tests, which are used only for data measured on \_\_\_\_\_\_\_\_ or \_\_\_\_\_\_\_\_ scales.
Parametric;
interval;
ratio
The probability of: rejecting a true null hypothesis = \_\_\_\_\_\_\_\_; retaining a false null hypothesis = \_\_\_\_\_\_\_\_; rejecting a false null hypothesis = \_\_\_\_\_\_\_\_.
Alpha;
beta;
power
Parametric tests are based on the assumptions of \_\_\_\_\_\_\_\_ (dependent variable values are normally distributed in the population) and \_\_\_\_\_\_\_\_ (variance of all research groups is equal).
Normal
distribution;
homogeneity of
variance
In determining whether to reject or retain the null hypothesis, the statistical value is compared to a \_\_\_\_\_\_\_\_, which is dependent on 1) the pre-set alpha level and 2) the degrees of freedom for the statistical test used.
Critical
value
Designed to test hypotheses based
on dependent variables measured
on a \_\_\_\_\_\_\_\_ or \_\_\_\_\_\_\_\_ scale,
\_\_\_\_\_\_\_\_ tests make no
assumption of distribution
("distribution free"), which can
decrease power, and include
chi-square and Mann-Whitney U.Nominal;
ordinal;
nonparametric
This statistic is used to test whether there is a significant difference between 2 different means; it cannot be used to compare any more than 2 means.
t-Test (or
Student’s
t-test)
A researcher wants to test the hypothesis that
children in a low-SES school district receive less
direct teacher contact than the national average
for children. The researcher learns from a recent
national survey that children receive an average
of 6 hours per week of direct contact with their
teacher; however, in her survey of children from
a low-SES district, the average was 2.5 hours
per week. What statistic would be used to
determine if significance exists?
One sample t-test: used to compare the mean of a single sample to a known population mean (degrees of freedom = N - 1)
The ANOVA statistic (F-ratio) is based on a comparison of \_\_\_\_\_\_\_\_ variance and \_\_\_\_\_\_\_\_ variance, wherein the greater the former is over the latter, the greater the effect of the treatment (independent variable).
Between-group
(or treatment);
within-group (or
error)
A researcher interested in studying the effects of
a novel therapeutic approach to improve social
function selects 100 people diagnosed with an
anxiety disorder. Subjects are randomly
assigned to either a group receiving the new
treatment or a group receiving CBT, referred to
as ________ samples. Data from a measure of
social function is collected following 12 weeks of
treatment. The researcher utilizes the ________
to test whether there is a significant difference
between the groups.
Independent; t-test
for independent
samples (degrees
of freedom = N - 2)
The first step in determining the F-ratio of a one-way ANOVA involves calculating the between-group and within-group \_\_\_\_\_\_\_\_, which is a measure of variability of a data set.
Sum of
Squares
This test is used to compare the means of samples that are related to each other, such as in a pretest-posttest design where the same group of subjects is measured before and after treatment, then the means compared.
t-Test for correlated samples (degrees of freedom = N - 1, where N = pairs of scores)
For a one-way ANOVA, degrees of freedom-between = k - 1 where k equals \_\_\_\_\_\_\_\_; degrees of freedom-within = N - k where N equals \_\_\_\_\_\_\_\_.
Number of
groups; total
number of
subjects
A researcher interested in determining whether
significant differences exist between 3 or more
independent groups on a single independent
variable would utilize a/an ________, yielding
a/an ________ statistic that would indicate
whether there was a difference in group means
without telling where the differences lie; should
the test yield significance, ________ tests can be
used to specify the differences.
One-way ANOVA (can be used with 2 groups, but usually t-test is simpler); F-ratio; post-hoc
Between-group variance is estimated by \_\_\_\_\_\_\_\_, which is calculated by dividing Sum of Squares-between by degrees of freedom-between; similarly, within-group variance is estimated by \_\_\_\_\_\_\_\_ , which is calculated by dividing Sum of Squares-within by degrees of freedom-within; the F-ratio is then calculated by dividing \_\_\_\_\_\_\_\_ by \_\_\_\_\_\_\_\_.
Mean Squares Between (MSB); Mean Squares Within (MSW); MSB; MSW
Should an F-ratio be higher than the critical value at a significance level of .05, how will the significance be indicated? How about at the .01 level?
A single asterisk indicates significance at the .05 level (p* < .05), while 2 asterisks indicate significance at the .01 level (p** < .01)
A/an \_\_\_\_\_\_\_\_ is used when a study involves more than 1 independent variable and 3 or more groups.
Factorial
ANOVA
Assuming a one-way ANOVA produces significance, \_\_\_\_\_\_\_\_ comparisons can help determine where the differences lie; however, when multiple comparisons are conducted, the probability that a/an \_\_\_\_\_\_\_\_ will occur increases, which is referred to as \_\_\_\_\_\_\_\_.
Post-hoc; Type I
error;
experiment-wise
error rate
The benefit of using a factorial ANOVA over separate one-way ANOVAs is that a factorial ANOVA produces both \_\_\_\_\_\_\_\_ (effect of 1 independent variable by itself) and \_\_\_\_\_\_\_\_ (effects of an independent variable at the different levels of the other independent variables).
Main effects;
interaction
effects
The most conservative post-hoc test is the \_\_\_\_\_\_\_\_ (protects against Type I error, which increases chance of Type II error); the \_\_\_\_\_\_\_\_ protects against Type I error when only pairwise comparisons are made.
Scheffe;
Tukey
In a factorial ANOVA, the differences between the \_\_\_\_\_\_\_\_ are examined to determine the main effects, while \_\_\_\_\_\_\_\_ are examined to determine interaction effects.
Marginal
means; cell
means
In a study where all of the subjects receive all of the levels of the independent variable (so they are not split into separate groups for comparison), or the study involves more than 2 matched groups, what statistic is used?
One way
ANOVA for
repeated
measures
In graphical form,
how is one able to
tell if there is an
interaction effect?
Interaction
effects are
indicated when
lines cross
What parametric statistical procedure is used when a study involves 1 or more independent variables and 2 or more dependent variables?
Multivariate
Analysis of
Variance
(MANOVA)
TRUE or FALSE: In order to use chi-square, it is necessary that all observations be independent of each other.
TRUE: "Before-and-after" studies violate this rule and thus cannot use chi-square
What test is used to assess whether observed frequencies differ significantly from what is expected under the null hypothesis when data is categorical (i.e., nominal)?
Chi-square
Two tests used when data is rank-ordered (i.e.,
ordinal) include the ________, used when a
study involves 2 independent groups, and the
________, used when correlated groups are
compared; both can also be used when interval
or ratio data does not meet the assumptions of
parametric tests (i.e., normal distribution,
homogeneity of variance), in which case data
would be converted to ranks.
Mann-Whitney
U; Wilcoxon
Matched-Pairs
Test
A chi-square that involves collecting categorical data from only 1 sample of subjects is called \_\_\_\_\_\_\_\_ chi-square test, while a \_\_\_\_\_\_\_\_ chi-square test involves adding another variable in addition to the one that gives rise to the classification categories.
Single-sample;
multiple-sample
What test is used when data is rank-ordered and more than 2 groups are to be compared?
Kruskal-Wallis
Test
How are expected frequencies (fe) calculated for both a single-sample and a multiple-sample chi-square?
single-sample fe = N (# of subjects)/number of cells; multiple-sample fe = (column total x row total)/total N
The nonparametric alternative for a: one-way ANOVA is \_\_\_\_\_\_\_\_; t-test for independent samples is \_\_\_\_\_\_\_\_; t-test for correlated samples is \_\_\_\_\_\_\_\_.
Kruskal-Wallis Test;
Mann-Whitney U;
Wilcoxon
Matched-Pairs Test
Typically ranging from -1.0 to +1.0, the \_\_\_\_\_\_\_\_ indicates the \_\_\_\_\_\_\_\_ (indicated by the absolute value) and \_\_\_\_\_\_\_\_ (indicated by the sign) of a relationship.
Correlation
coefficient;
magnitude;
direction
This term refers to the squared correlation coefficient that indicates the percentage of variability in one measure that is accounted for by variability in the other measure; it is one way of interpreting the Pearson r.
Coefficient of
determination
This correlation coefficient is used to calculate the relationship between 2 variables measured on an interval or ratio scale (e.g., relationship between IQ and scores on the SAT); it is the most commonly used correlation coefficient.
Pearson Product
Moment
Coefficient
(Pearson r)
A variable that can have an infinite number of values between 2 given points is called \_\_\_\_\_\_\_\_ (e.g., income), while a \_\_\_\_\_\_\_\_ variable is a variable that categorizes data into 2 distinct groups (e.g., high/low, true/false, male/female).
Continuous;
dichotomous
A correlation is referred to as \_\_\_\_\_\_\_\_ when Y is impacted differently by different levels of X (e.g., children's anxiety level [X] and learning capacity [Y]: children learn best with moderate anxiety, and worse with low or high anxiety), in which case Pearson r cannot be used and the \_\_\_\_\_\_\_\_ coefficient would be used.
Curvilinear;
eta
TRUE or FALSE: It is possible to transform continuous variables into dichotomous variables for use in a correlation coefficient.
TRUE: The variable is then referred to as "artificially dichotomous," though it is not recommended to do this
An assumption of the use of Pearson r is \_\_\_\_\_\_\_\_, which is when the dispersion of scores on a scattergram is equal; the term \_\_\_\_\_\_\_\_ refers to there being more dispersion at some parts of the scattergram than others, which lowers the Pearson r coefficient.
Homoscedasticity;
heteroscedasticity
Identify the correlation coefficient to use: X and Y
= continuous (income x IQ); X = true dichotomy
and Y = continuous (gender x income); X =
artificial dichotomy and Y = continuous (income
as “upper” or “lower” x IQ); X and Y = true
dichotomy (a MMPI-2 question x gender); X and
Y = artificial dichotomy (income as “upper” or
“lower” x IQ as “high” or “low”); X and Y =
ordinally ranked (relationship between scores
given by 2 judges of a competition).
Pearson r; Point
Biserial; Biserial;
phi; Tetrachoric;
Spearman’s rho
An equation used to estimate the value of one variable based on the value of the other, when they are correlated, is referred to as what?
Regression
equation
The predictive power of the multiple regression
equation is highest when predictor variables
have ________ correlations with the criterion
variable and ________ correlations with other,
as this implies that each predictor variable is
providing new information about the criterion;
when predictors are highly correlated with each
other, referred to as ________, combining them
produces no substantially new information.
High; low;
multicollinearity
A researcher is interested in predicting the GRE scores of current undergraduate students based on their college GPA, which had previously been shown to be correlated. In this situation, GRE scores are the \_\_\_\_\_\_\_\_ variable while the \_\_\_\_\_\_\_\_ variable is GPA.
Criterion
(dependent);
predictor
(independent)
TRUE or FALSE:
Multiple R can be
both positive and
negative.
FALSE: Multiple R can never be negative, as the calculations do not allow for negative outcome
A line that passes through as many dots as possible on a scattergram is called a regression line, or \_\_\_\_\_\_\_\_, which implies a linear relationship between 2 variables and is an assumption underlying regression.
Line of
best fit
Is it ever possible for a multiple correlation coefficient (multiple R) to be lower than the highest simple correlation between an individual predictor and the criterion?
NO: Adding predictors will never decrease multiple R, though it may not increase it either
A \_\_\_\_\_\_\_\_ is used to assess the relationship between 2 or more predictor variables and a single criterion variable; the use of scores on more than 1 predictor to estimate scores on a criterion is referred to as \_\_\_\_\_\_\_\_.
Multiple correlation
(Multiple R);
multiple regression
A researcher who has a large number of potential predictor variables wants to maximize predictive power by retaining the predictors that, combined, have a high multiple correlation with the criterion. What technique should be used?
Stepwise
multiple
regression
While multiple correlation indicates the magnitude of the relationship between multiple predictors and a single criterion variables, \_\_\_\_\_\_\_\_ is used when there are multiple predictor and multiple criterion variables.
Canonical
correlation
What technique involves identifying different scores on a series of predictors that a person must score at or above to be predicted as successful on the criterion?
Multiple
cutoff
What coefficient is used to assess 2 nominally scaled variables, each having more than 2 categories (e.g., political party x DSM diagnostic category)?
Contingency
coefficient
Research results reveal that there is a relatively high correlation between 2 variables, though the researcher suspects there may be an additional variable contributing to this relationship. What technique might be used to statistically control for the effects of the other variable?
Partial
correlation
What correlational technique involves predicting a categorical criterion/dependent variable by one or more continuous predictor/independent variables (e.g., membership in "high achievement" or "low achievement" group based on WAIS, Bender-Gestalt, and SAT scores)?
Discriminant function analysis, which is the statistical opposite of MANOVA
What technique is used for testing and estimating causal relationships using a combination of statistical data and qualitative causal assumptions?
Structural
equation
modeling
(SEM)
When would logistic regression, which is used to predict the criterion group one belongs to, be used in place of discriminant function analysis?
When the assumptions of normal distribution and homogeneity of variance are not met and/or when the predictor variables are not continuous (DFA predictors must be continuous)
________ is used to verify directed (one-way)
causal relationships (dependencies) between
variables, while ________ can be used to verify
both one-way and two-way causal flows; further,
________ can be used only in models that
include observed variables (i.e., measurable),
while ________ can be used in models that
specify observed and latent variables (inferred as
being measured).
Path analysis;
LISREL; path
analysis;
LISREL
A researcher is not so much interested in determining the magnitude of relationship between quantitative variables (interval or ratio) as determining the course of change in the dependent variable over time. What correlational technique would be used in this case?
Trend
analysis
According to the \_\_\_\_\_\_\_\_: (1) the shape of the sampling distribution of means approaches a normal shape as sample size increases, regardless of whether the population of distribution scores is normally distributed or not; and (2) the mean of the sampling distribution of means equals the population mean.
Central
Limit
Theorem
The \_\_\_\_\_\_\_\_ describes the principle underlying the construction of a regression line, stating that a regression line is places at a location on the scattergram that results in the lowest possible sum of squared deviations of points from the line.
Least
squares
criterion
The standard deviation of the sampling distribution of means is equivalent to what?
The standard
error of the
mean
A \_\_\_\_\_\_\_\_ distribution is the entire set of cases a researcher is interested in, a \_\_\_\_\_\_\_\_ distribution is a set of scores obtained from the sample chosen from a population, and a \_\_\_\_\_\_\_\_ distribution (of any statistic) is the distribution of the statistic for all possible samples of a given size.
Population;
sample;
sampling
A statistical test is considered "\_\_\_\_\_\_\_\_" when the Type I error rate is not significantly increased when assumptions of parametric tests are violated; basically, if they are violated, the results may still be accurate.
Robust
By choosing numerous samples of size N from a population, recording the value of the mean from each sample, then repeating this procedure thousands of times, what would be constructed?
Sampling
distribution
of the mean
In a time-series analysis, the correlation between values obtained at different points in time is referred to as what?
Autocorrelation
This theorem shows how a conditional probability depends on its inverse (e.g., mammograms are 95% accurate, though a positive mammogram is likely inaccurate based on the relative rarity of breast cancer).
Bayes’ Theorem
(conditional
probabilities and
base rates)
