Research & Statistics-final exam
CHAPTER 8
ADVANCED CORRELATIONAL STRATEGIES
CHAPTER 12
ANALYZING COMPLEX EXPERIMENTAL DESIGNS
CHAPTER 11
ANALYZING EXPERIMENTAL DATA
CHAPTER 10
EXPERIMENTAL DESIGN
CHAPTER 14
SINGLE-CASE RESEARCH
which type of error do researcher usually regard as more serious? why?
The Type I error is more serious, because you have wrongly rejected the null hypothesis
Psychobiography
a biographical case study of an individual with a focus on explaining the course of the persons life using psychological constructs and theories
what is a case study are case studies an example of descriptive, correlational, experimental or quasi-exeprimental research
a case study is a detailed study of a single individual, group or event. they are descriptive
factor analysis
a class of multivariate statistical techniques that identifies the underlying dimensions that account for the observed relationships among a set of measured variables
Narrative description
a descriptive summary of an individuals behavior often with interpretations and explantations such as is geared in a case study
what is a factorial design? why are factorial designs used more frequently than one way designs
a factorial design is one in which two or more independent variables are manipulated the IV are to as factors they are used more often because they can show you the effects of more than one IV
split plot factorial design
a factorial design that combines one or more between subjects facts with one or more within subject factors
statistical significance
a finding that is very unluckily to be due to error variance
bonferroni adjustment
a means of preventing inflation of type 1 error when more than one statistical test is conducted
simultaneous multiple regression
a multiple regression analysis in which all of the predictors are entered into the regression equation in a single step
standard multiple regression
a multiple regression analysis in which all of the predictors are entered into the regression equation in a single step
stepwise multiple regression
a multiple regression analysis in which predictors enter the regression equation in order of their ability to predict unique variance in the outcome variable
hierarchical multiple regression
a multiple regression analysis in which the researcher specifies the order that the predictor variables will be entered into the regression equation
ABACA design
a multiple-I single case experimental design in eh baseline data are obtained (A) one level of the independent variable is introduced (B) this level of the independent veritable is withdrawn (A) a second level of the independent variable is introduced (C) and this level of the independent variable is withdrawn (A)
ABC design
a multiple-I single case experimental design that contains a baseline period (A) followed by the introduction of the independent variable (B) followed by the introduction of another level of the independent variable (C)
subject variable
a personal characteristic of research participants such as age, gender, self esteem or extraversion
participant variable
a personal characteristic of research participants such as age, gender, self esteem or extraversion; also called subject variable
nondirectional hypothesis
a prediction that does no express the direction of a hypothesized effect
directional hyposthesis
a prediction that explicitly states the direction of a hypothesized effect
nested design
a research design in which participants are drawn form various groups such as students being recruited from classroom in a nest design the repossess of participant who come form a single group are not independent of one another which raised special analysis issues
cross-lagged panel correlation design
a research design in which two variables are measured at two point in time and correlation between the variables are examined across time
Multiple-I design
a single case experimental design in which levels of an independent variable are introduced one at a time
Reversal design
a single case experimental design in which the independent variable is introduced and then withdrawn
Multiple baseline design
a single case experimental design in which two or more behaviors are studies simultaneously
ABA design
a single case experimental design in who baseline data are obtained (A) the independent variable is introduced and a behavior is measured again (b) then the independent viable is withdrawn and behavior is observed a third time (A)
power analysis
a statistic that converts the power or sensitivity of a study; power analysis is often used to determine the number of participants needed to achieve a particular level of power
Prep statsitic
a statistic that estimates the probability of replicating an effect obtained in an experiment
one titled test
a statistic used to test a directional hypothesis
structural equations modeling
a statistical analysis that tests the viability of alternative causal explanations of variables that correlate with one another
standard error of the difference between tow means
a statistical estimate of how much two condition means would be expected to differ if their difference is due only to error variance and the independent variable has no effect
what is multiple regression analysis
a statistical procedure by which an equation is derived that can predict one variable (criterion) from a set of other variables (predictor)
multiple regression analysis
a statistical procedure by which an equation is derived that can predict one variable from a set of other variables
regression analysis
a statistical procedure by which an equation is developed to predict scores on now variable based on scores from another burble
multivariate analysis of variance MANOVA
a statistical procedure that simultaneously tests differences among the means of two or more groups on two or more dependent variables
two-tailed test
a statistical test for a nondirectional hypothesis
what does it mean if the difference between two means is statistically significant?
a statistically significant finding is one that has a low probability of occurring as a result of error variance alone
a regression equation is actually the equation for a straight line. what line is described by the regression equation calculated for a set of data?
a straight line
paired t-test
a t-test performed on repeated measures two group design
factor matrix
a table that shows the results of a factor analysis in this matrix the rows are viable and the columns are factors
moderator variable
a variable that qualifies or moderates the effects of another variable on behavior
(1) what would it mean if the proportion of variance effect size in an experiment was .25? .4? .00? what (1) would it means if the means difference effect size was .25? .4? .00?
(1) 25% 40% and 0% of the variance in the DV is caused by the IV (2) 25% 40% and 0% of the variance in the DV is caused by the IV or that the condition means differ by .25 .4 or .0 standard deviations
mixed factorial design
(1) an experimental design that includes one or more between subjects farces and one or more within subjects factors (2) also refers to an experimental design that includes both manipulated independent variables and measured participant variables
what are four primary reasons that behavioral researcher use case studies
(1) as a source of insights and ideas (2) to describe rare phenomena (3) psychobiography (4) illustrative anecdotes- illustrate general principles to other researchers and students
under what two circumstances would you use a multivariate analysis of variance
(1) conceptually related DVs- when a researcher has measured several dependent variables, all of which tap into the same general construct (2) inflation of type 1 error
factor
(1) in experimental designs an independent variable (2) in factor analysis the underlying dimension that is assumed to account for observed relationships among variables
factor
(1) in experimental designs an independent variable (2) in factor analysis the underlying dimension that tis assumed to account for observed relationships among variables
in an experiment with two independent variables an ANOVA partitions the total variance into four components what are they
(1) the error variance (2) the main effect of A (3) the main effect of B (4) the A x B interaction
(1) how many independent variables are involves in 3 x 3 factorial design? (2) how many levels are there of each variable? (3) how many experimental conditions are there?
(1) two IVs (2) three levels (3) nine
once researchers calculate a values for t they compare that calculated value to a critical value of t. what two pieces of information must be known in order to find the critical value of t in a table of critical values?
(1) we need to calculate the degrees of freedom for the t-test (2) we need to specify the alpha level for the test
(9) a case study is detailed descriptive stud of a single individual gourd or event the case is described in detail and conclusions solutions or recommendations are offered
(10) case studies rarely allow a high degree of confidence in the researchers interpretations of the data because extraneous variables are never controlled and the biases of the researcher may influence his or her observations and interpretations. Even so case studies are useful in generating new ideas studying rate phenomena doing psychological studies of famous people and serving as illustrative anecdotes
(9) expericorr factorial designs combine manipulated independent variables and measured participant variables such designees are often used to study participant variables that qualify or moderate the effects of the independent variables
(10) researchers using an expericorr design sometimes classify participants into groups using a median split or extreme groups procedure but others use analyses that allow then to maintain the continuity of the measured participant variable in either case causal inferences amy be drawn only about the variables in the design that were experimentally manipulated
(9) hypothesis about the outcome of two-group experiments may be directional or nondirectional. whether the hypothesis is directional or nondirectional has implications for whether the circle value of t used in the t-test is one tailed or two tailed
(10) the paired t-test is used when the experiment involves a matched subjects or within subjects design
SUMMARY (1) when research designs involve more than two conditions researchers analyze their data using ANOVA rather than t-tests because conducting many t-tests increases the chances that they will make a type 1 error
(2) ANOVA partitions the total variability in participant responses into between groups variability and within groups variance. than an F-test is conducted to determine whether the between-gorups variance exceeds what we would expect based on the amount of within-groups variance in the data. if it does we reject the null hypothesis and conclude that at least one of the means differs form the others
(1) the data from experiments are analyzed by determining whether the means of the experimental conditions differ. however because error variance can cause condition means to differ even then the independent variables has no effect we must compete the difference between the condition means to how much we would expect the means o differ if the difference is due solely to error variance
(2) researchers use inferential statistics to determine whether the observed difference between the measure greater than would be expected o the basis of error variance alone
SUMMARY: (1) a one way experimental design is an experiment in which a single independent variable is manipulated. the simplest possible experiment is the two group experimental design
(2) researchers use three general versions of the one way design-the randomized groups design (in which participants are assigned randomly to two or more groups) the matched subjects design (in which participants are first matched into blocks and then randomly assigned to conditions) ad the reported measures or within-subjects design (in which each participant serves in all experimental conditions)
SUMMARY (1) the principles and empirical findings of behavioral science are probabilistic in nature describing the reactions of most individuals but recognizing that not everyone will fit the general pattern
(2) single case research comes in two basic varieties, single case experimental designs and case studies both of which can be reared to the earliest days of behavioral science
SUMMARY (1) regression analysis is used to develop a regression equation that describes how variables are related and allows researchers to predict peoples scores on one variable (the outcome or criterion variable) based on their scores on other variables (the predictor variables) a regression equation provides a regression constant (equivalent to the y-intercept) as well as a regression coefficient for each predictor variable
(2) when constructing regression equations a researcher may enter all of the predictor variables at once (simultaneous or standard regression) allow predictor variables to tenth the equation based o their ability to account for unique variance in the criterion variable (stepwise regression) or enter the variables in a manner that allows I'm or her to test particular hypothesis (hierarchical regression)
(3) multiple correlation expresses the strength of the relationship between one variable and a set of other variables among other things it provides information about how well a set of predictor variables can predict scores on a criterion variable in a regression equation
(4) cross lagged panel correlation designs and structural equations modeling are used to test the plausibility of causal relationships among a set of correlated variables. both analyses can provide evidence for or against causal hypotheses but our conclusions are necessarily tentative beach the data are correlational
(3) each of these designs may involve as single measurement of the dependent variable after the manipulation of the independent variable or a pretest and a posttest
(4) factorial designs are experiments that include two or more independent variables (independent variables are sometimes called factors a term not to be confused with its meaning in factor analysis)
(3) in a one way design a single F-test is conducted to test the effects of the lone independent variable. in factorial design an f-test is conducted to test each main effect and interactions
(4) for each effect being tested the calculated value of F is compared to critical value of F.if the calculated value of F exceeds the critical value we know that at least one condition mean differs form the others if the calculated value is less than the critical value we fail to reject the null hypothesis and conclude that the condition means do not differ
(3) single case experiments investigate the effects of independent variables on individual reseal participants. Unlike gourd designs in which are are averages across participants for analysis each participants responses are analyzed separately and they are from individual participants are not combined
(4) the most common single participant designs variations of the ABA design involve a baseline period followed by a period in which the independent variable is introduced then the independent variable is withdrawn more complex designs may involve several periods in which the independent variable is successively reintroduced and then withdrawn
(3) if the condition means differ more than expected based on the amount of error variance in the data researchers reject the null hypotheses and conclude the independent variable affect the dependent variable if the means do not differ by more than error variance would predict researchers fair to reject the null hypothesis and conclude that the indent variable does not have an effect
(4) when deciding to reject or fail to reject the null hypothesis researchers may make one of two kinds of errors. a type 1 error occurs when the researcher rejects the null hypothesis when it is true and a type 2 error occurs when the researcher fails reject the null hypothesis when nit is false
(5) multilevel modeling is used to analyze the relationships among variables that are measured at different levels of analysis. for example when several preexisting groups of participants are studied multilevel modeling allows researchers to examine processes that are occurring at the level of the groups and at the level of the individuals
(6) factor analysis refers to a set of procedures for identifying the dimensions or factors that account for the observed relationships among a set of variables. a factor matrix shows the factor landings for each underlying factor which are the correlations between each variable and the factor from this matrix researchers can identify the basic factors in the data
(5) in multiple-I designs several levels of the independent variable are administered in succession often with a baseline period between each administration
(6) multiple baseline designs allow researchers to document that the effects of the independent variable are specific to particular behaviors such designs involves the simultaneous stiffly of two or more behaviors only one of which is hypothesized to be affected the independent variable
(5) if the F-test show that the main effects or interactions are statistically significant follow-up tests are often needed to determine the precise effect of the independent variable. main effects of independent variables that involves more than two levels require post hoc tests whereas interactions are decomposed using simple effects tests
(6) special varieties of ANOVA and MANOVA are used when data from within-subjects designs or mixed designs are being analyzed
(5) the size and structure of factorial designs are described by specifying the number of levels of each independent variable. for example a 3 x 2 factorial design has two independent variables one with three levels and one with two levels
(6) there are four general types of factorial designed-the randomized gourds, matches subjects, repeated measures and mixed factorial designs
(5) researchers can never know for certain whether a type 1 or 2 error has occurred, but they can specify the probability that they have made each kind of error (type 1-alpha) (type 2-beta)
(6) to minimize the probability of making a type 2 error researchers try to design powerful studies. power refers to the portability that a study will correct reject the null hypothesis to ensure they have sufficient power researchers often conduct a power analysis that tells them the optimal number of participants for their study
(7) factorial designs provide information about the effects of each independent variable by itself as well as the combined effects of the independent variables
(8) an interaction between two or more independent variables is present if the effect of one independent variable is different under one level of another independent variable than it is under another level of that independent variable
(7) multivariate analysis of variance MANOVA is used to test the differences among the means of two or more conditions on a set of dependent variables. MANOVA is used in two general cases: when the dependent variables all measure aspects of the same general construct and when the researcher is concerned that performing separate analysis on several dependent variables will increase the possibility of making a type 1 error
(8) in either vase MANOVA creates a new composite variable=a canonical variable-from the original dependent variable and then determines whether participants scores on this canonical variable differ across conditions
(7) because averages are not used the data on single participant experiments cannot be analyzed using inferential statistics rather effect of the independent variables on behavior are detected through graphic analysis
(8) single case experiments are used most frequently to study the effects of basic learning processes and to study the effectiveness of behavior modification in treating behavioral and emotional problems
(7) effect size indicates the strength of the independent variables effect on the dependent variable. it is expressed as either th proportion of the total variability in the dependent variable that is accounted for by the independent variable or as the size of the difference between two means expressed in standard deviation units
(8) the t-test is used to analyze the difference between two means. a value for t is calculated by dividing the difference between the means by an estimate of how much the means would be expected to differ on the basis of error variance alone. this is calculate value of t is then compared to the critical value of t. if the calculated value exceeds the critic lvalue the null hypothesis is rejected
discuss the advantages and disadvantages of single case experimental designs relative to group designs
(A) provide convincing evidence regarding the causal effects of IV on behavior; greater generalizability (D) do not possess greater external validity, generalizability, not well suited for studying interactions, ethical issues
an ANOVA for a one way design partitions the total variance in a set of data into two components what are they?
1. systematic variance 2. error variance total variance=systematic variance+error variance
how many main effects can be tested in a 2x2 design? in a 3x3 design? in a 2x2x3 design
2x2=2 3x3=2 2x2x3=3
how many interactions can be tested in a 2x2 design? a 3x3 design? in a 2x2x3 repeated measures factorial design
2x2=4 3x3=9 2x2x3=12
how many levels of the independent variable are there in an ABACADA design?
3: b, c, d
multilevel modeling
an approach to analyzing data that have nested structure in which variables are measured at different levels of analysis
regression equation
an equation from which one can predict scores on one variable from one or more other variables
mean square between groups
an estimate of between group variance calculated by dividing the sum of squares between groups by the between group degrees of freedom
pretest posttest design
an experiment in which participants repossess are measured twice once before and once after the introduction of the independent variable
posttest only design
an experiment in which participants responses are measured only once-after introduction of the independent variable
two-group experimental design
an experiment with two conditions; the simplest possible experiment
repeated measures design
an experimental design in which each participant serves in more than one condition of the experiment
randomized groups design
an experimental design in which each participant serves in only one condition of the experiment
matched subjects design
an experimental design in which participants are matched into homogeneous block and participants in each block are randomly adding to the experimental conditions
Group design
an experimental design in which several participants serve in each condition of the design and the data are analyzed by examining the average responses of participants in these conditions
Single case experimental design
an experimental design in which the unit of analysis is the individual participant rater than the experimental group
factorial design
an experimental design in which two or more independent variables are manipulated
randomized groups factorial design
an experimental design involving toe or more independent variables in which each participant serves in only one condition of the experiment
repeated measures factorial design
an experimental design involving two or more independent variables in which each participants serves in all conditions of the experiment
matched subjects factorial design
an experimental design involving two or more independent variables in which participants are first matched into homogeneous blocks and then within each block are randomly assigned to the experimental conditions
between-within design
an experimental design that combines one or more between subjects factors with one or more within subjects factors; also called missed factorial or split-plot design
expericorr factorial design
an experimental design that includes one or more manipulated independent variables and one or more preexisting participant variables that are measured rater than manipulated; also called mixed factorial design
one way design
an experimental design with a single independent variable
t-test
an inferential statistic that tests the difference between two means
analysis of variance (ANOVA)
an inferential statistical procedure used to test difference between means
F-test
an inferential statistical procedure used to test for differences among condition means
Case study
an intensive descriptive stud of a particular individual group or event
what is an interaction
an interaction is present when the effect of one IV differs across the levels of other IV; if one IV has a different effect at one level of another IV than it has at another level of that IV we say that the IVs interact and that an interaction between the IVs is present
what statistical analysis is used to analyze data from experimental designs that involve more than two conditions
analysis of variance
median split procedure
assigning participants to two groups depending on whether their scores on a particular variable fall above or below the median of that variable
why do researchers use ANOVA rater than t-tests to analyze data from experiments that have more than two groups
because ANOVA is designed to analyze differences between all condition means in an experiment simultaneously rather than testing the difference between each pair of means as a t-test does ANOVA determines whether any of a set of means differs from another using a single statistical test that hold the alpha level at .05 regardless of how many group means are involved in the test
when analyzing data why do researcher test the null hypothesis rater then the experimental?
because you can show that something did not happen but you cannot prove that something did happen; we cannot determine whether an IV truly confirmed the something but we can determine that the IV probably has an effect on the DV
rejecting the null hypothesis
concluding on the basis of statistical evidence that the null hypothesis is false
failing to reject the null hypothesis
concluding on the basis of statistical evidence that the null hypothesis is true- that the independent variable does not have an effect
what is so bad about making a type 1 error
conducting multiple tests inflates type one error. although one could use multiple t-tests to analyze these data such an analysis creates a serious problem. one test runs a risk of 5% for type one error but if you run them over and over it will increase by 5% every time making it much more likely that you will make a type one error
extreme groups procedure
creating two groups of participants that have unusually low or unusually high scores on a particular variable
when would you calculate a multiple corrections coefficient? what do you learn if yo square it?
describes the degree of relationship between the criterion variable and the set of predictor variables; can be squared to show the percentage of variance in the criterion variable that can be accounted for by the set of predictor variables
power refers to the degree to which a study is likely to
detect the effects of the IV
type 2 error
erroneously failing to reject the null hypothesis when it is false; concluding that the independent variable did not have an effect when in fact it did
type 1 error
erroneously rejecting the null hypothesis when nit is true; concluding that an independent variable has an effect when in fact it did not
if the IV in an experiment had absolutely no effect on participants responses, all of the variance in the data is
error variance
what is the name of the statistic used in an ANOVA
f-test
describe a mixed, or split-plot, factorial design, this design is a hybrid of what two other designs
factorial designs involve more than one IV so they can combine features of both randomized group designs and repeated measures in a single experiment; some IV in a factorial experiment amy involve random assignment whereas other variables involve a repeated measure
if the calculated value of t is less than the critical value do you reject or fail to reject the null hypothesis? explain
fail to reject the null
if you want to have 20 participants in each experimental conditions, how many participants will you need for a 2x3x3 completely randomized factorial design? how many participants will you need for a 2x3x3 repeated measures factorial design
for the randomized design you would need 240 participants but for a repeated measures design you would only need that 20
describe how structural equations modeling works
given the pattern of correlation among a set of variables, certain causal explanations of the relationships among the variables are more logical or likely than others. given the pattern of correlation among the variables, certain causal relationship may be virtually impossible whereas others are plausible ex: we are trying to understand the relationship between x, y, z if we predict x causes y and then y causes z then we should find not only that x and y are correlated but also that the relationship between x and y is stronger than the correlation between x and z
in stepwise regression why might a predictor variable that correlates highly with the criterion variable not enter into the regression equation
if the predictor variable everted into the equation in step 1 is highly correlated with other predictors it may already account for the variance that they could account for in the criterion variable if so that other predictors may not be needed
why do researchers use inferential statistics
if we estimate how much the means of the conditions would be expected to differ even if the IV has no effect than we can determine whether the difference we observe between the means exceeds the estimate
canonical variable
in MANOVA a composite variable that is calculated by summing two or more dependent variables that have been weighted according toothier ability to differentiate among groups of participants
predictor variable
in a regression analysis a variable used to predict scores o the criterion or dependent variable
factor loading
in factor analysis the correlation between a variable and a factor
Interaparticipant replication
in single case experimental research documenting the generalizabiltiy of an experimental effect by demonstrating the effect on other participants
Interparticipant replication
in single case experimental research the attempt to repeatedly demonstrate an experimental effect on a single participant by alternatively introducing and withdrawing the independent variable
Graphic analysis
in single case experimental research the visual inspection of graphs of the data to determine whether the independent variable affected the participants behavior
distinguish among simultaneous or standard, stepwise and hierarchical regression
in standard multiple regression all of the predictor variables are entered into the regression analysis at the same time, stepwise multiple regression analysis builds the regression equation by entering the predictor variables one at a time, hierarchical multiple regression the predictor variables are entered into the equation in an order that is predetermined by the researcher based on hypotheses that he or she wants to test
fit index
in structural equations modeling a statistic that individuate how well a hypothesized model fits the data
how does a cross lagged panel correction design provide evidence to support a casual link between two variables
in this design the correlation between two variables x and y is calculated at two different points in time, then correlations are calculated between measurements of the two variables across time; the relationship between a cause (x) and its effect (y) should be stronger if the causal variable is measured before rather than after its effect
post hoc tests
inferential statistics that are used after a significant F-test to determine which means differ
multiple comparisons
inferential statistics that are used after a significant F-test to determine which means differ from which
follow-up tests
inferential statistics that are used after a significant F-tet to determine which means differ from which
What is the difference between interparticipant and intraparticiapnt variance? Which of these types of variance is more closely related to error variance in group experimental designs? Which type is of primary interest to researchers who conduct single case experiments
interparticipant- not the kind of variability that behavioral researchers are usually trying to understand and explain- individual differences; error variance intraparticipant variability in an individuals behavior when the or she is in the same situation on different occasions; the interest of single case
what is psyhcobiography
involves applying concept and theories from psychology in an effort to understand the lives of famous people; they are trying to explain their life or specific aspects of the individuals behavior are studied
why do researcher use factor analysis
it is used to analyze the interrelationship among a large number of variables; its purpose is to identify the underlying dimensions or factors that account for the relationships that are observed among variables
inferential statistics
mathematical analysis that allows researchers to draw conclusions regarding the reliability and generalizability of their data
MSwg appears in denominator of every F-test why?
mean square within groups because it represents the error variance
why do researcher use multilevel modeling
multilevel modeling allows us to tease apart the various influences on the variable by analyzing variables operating at all levels of the nested structure simultaneously; allows us to capitalize on the opportunity to examine relationships among variables across the levels of the design
if the independent variable has absolutely no effect on the dependent variable will the means of the experimental conditions be the same why or why not?
no because there is still error variance and personal differences that will affect the scores no matter how closely the groups match there will still be differences
Distinguish between the nomothetic and idiographic approaches to behavioral science
nomothetic- seeking the establish general principles and broad generalizations ideographic- seeks to describe, analyze,and compare the behavior of individual participants
distinguish between one tailed and two tailed t-tests
one tailed is directional and two tailed is not
distinguish between an independent variable and a participant variable
participant variables are things like gender or self esteem that can classify them into groups then they are randomly assigned to levels of the IV
distinguish between latent variable modeling and path analysis as types of structural equations modeling
path- when single measures of each construct are used, researchers sometimes call structural equations analysis path analysis latent-each construct in the model is assessed by two or more measures using multiple measures of each construct not only provides a better, more accurate measure of the underlying or latent variable than any single measure can but also allows us to account for measurement error in our modeling
what are the relative advantages and disadvantages of posttest-only v. pretest-poshest experimental designs
posttest only-in each instances the dependent variable is measured only after the experimental manipulation has occurred pretest-posttest-the DV is measured twice-once before and once after; pretest scores on the DV the researcher can verify that participants in the various experimental conditions did not differ with respect to the dependent variable at the beginning of the experiment; can see exactly how much the IV changed participant behavior; pretest provides a baseline for judgment; more powerful but they lead to sensitization or cue the participants in on the topic or purpose of the experiment
do powerful studies minimize alpha or beta? explain
power is the probability that the study will obtain a significant result if the researcher's experimental hypothesis is in fact true; they state their alpha and beta levels and their desired power and then figure out how many participants they need to reach it
describe how participants are assigned to conditions in randomized groups, matched-subjects, and repeated measures experimental designs
randomized groups-a between subjects design in which participants are randomly assigned to one of two or more conditions; sometimes used to increase the similarity of the experimental groups prior to the manipulation of the IV matched-subjects-participants are matched into blocks on the basis of a variable that researcher believes relevant to the experiment; they are then randomly assigned to one of the experimental or control conditions repeated measures-each participant serves in all experimental conditions
distinguished among randomized groups, matched subjects and repeated measures factorial designs
randomized groups-participants are assigned randomly to one of the possible combinations of the independent variables matched subjects-involves first matching participants into blocks on the basis of some variable that correlated with the DV then the participants in each block are randomly assigned to one of the six experimental conditions repeated measures-requires participants to participate in every experimental condition; they can become unwieldy with larger factorial designs
the bonferroni adjustment is used to
reduce the likelihood of a type one error
when do researchers use regression analysis
regression analyses are often used to extend the findings of correlational research; once we know that certain variables are correlated with a particular physiological response or trait regression analysis allows us to develop equations that describe precisely how those variables relate to that response
explain the difference between rejecting and failed to reject the null hypothesis. in which case does a researcher conclude that the independent variable has an effect on the dependent variable?
rejecting the null-means that the researcher concludes that the IV DID have an effect failing to reject the null-mean that the researcher concludes that the independent variable has no effect
Researchers who use group designs replicate their findings by repeating the same experiment on other samples of participants. How do single case researchers replicate their findings
replication is rare. some introduce an IV then remove it then reintroduce it; sometimes they will have a handful of people doing the experiment
Idiographic approach
research that describes analyzes and attempts to understand the behavior of individual participants; often contrasted with the nomothetic approach
Nomothetic approach
research that seeks to establish vernal principles and broad generalizations; often contrasted with the ideographic approach
why do researchers use expericorr designs?
researcher use this for two reasons (1) to investigate the generality of an independent variables effect, participants who possess different characteristics often respond to the same situation in quite different ways (2) used in an attempt to understand how certain personal characteristics related to behavior under varying conditions
what information do researchers obtain form conducting a power analysis
researchers conduct a power analysis to determine the number of participants that is needed in order to detect the effect of particular IV; once they set their alpha level and specify the power they desire researchers can calculate the number of participants needed to detect an effect of a particular size
how does the bonferroni adjustment control for type 1 error when many statistical tests are conducted
researchers sometimes use this in which they divide their desired alpha by the number of tests they plan to conduct that way when we conduct all of them it is at the level that would be acceptable and not really high
when are tests of simple main effects used and what do researchers learn form them
simple main effect is the effect of one IV at a particular level of another independent variable; shows us precisely which condition means within the interaction differ form each other
What criticisms do proponents of single case experimental designs level against groups designs
some researcher insist that research involving the study of individuals is essential for the advancement of behavioral science whereas other researchers see such approaches as having limited usefulness group designs have problems with (1) error variance (2) generalizability (3) reliability
how do mixed/expericorr designs differ form other experimental design?
such designs combine features of an experimental design in which IVs are manipulated and features of correlational designs in which participant variables are measured
write the formula for a t-test
t=(xbar1 - xbar2)/squroot(1/n1 + 1/n2)
if you reject the null hypothesis what is the likelihood that you made a type 1 error? type 2?
the alpha level
mean square within groups
the average variance within experimental conditions the sum of squares within groups divided by the degrees of freedom within group
what kind of variance does the mean square within groups reflect
the average variance within experimental conditions; the sum of squares within groups divided by the degrees of freedom within groups
why is it essential for researchers to establish a stable baseline of behavior during the initial A phase of an ABA design?
the baseline has to be reliable because they are trying to show that IV is what is causing the change and when the IV is removed they will return to the initial baseline if it is not accurate they wont be able to tell the cause of the changes
interaction
the combined effect of two or more independent variables such that the effect of one indecent variable differs across the levels of the bother independent variable
multiple correlation coefficient
the correlation between one variable and a setoff other variables often used in multiple regression to express the strength of the relationship between the outcome variable and the set of predictor variables
power
the degree to which a reseach design is sensitive to the effects of the independent variable; powerful designs are able to detect effects of the independent variable more easily than less powerful designs
main effect
the effect of a particular independent variable ignoring the effects of other independent variables in the experiment
what is a main effect?
the effect of a single IV in a factorial design is the main effect; reflects the effect of a particular IV while ignoring the effects of the other IVs
simple main effect
the effect of one independent variable at a particular level of another independent variable
if the calculated value of F is found to be significant for the main effect of an independent variable with more than two levels what tests does the researcher conduct? why are such tests not necessary if the independent variable has only two levels
the first step in interpreting the results of any experiment is to calculate the means for the significant effects; if an ANOVA reveals a significant effect for an independent variable that has only to levels no further statistical tests are necessary; if a significant main effect is found further tests are needed you need to do post hoc tests and testing for interactions
the sum of squares between groups reflects the degrees to which the condition means vary around the ______________
the grand mean
experimental hypothesis
the hypothesis that the independent variable will have an effect on the depend variable; equivalently the hypothesis that the means of the various experimental conditions will differ form one another
null hypothesis
the hypothesis that the independent variable will not have an effect
alpha level
the maximum probability that a researcher is willing to make a type 1 error; typically set at .05
explain what it means if a researcher sets the alpha level for a statistical test at .05
the maximum probability that a researcher is willing to make a type 1 error; typically set at .05; there is a 5% chance of type 1 error
grand mean
the mean of all the condition means in a n experiment
assuming that all confounds were eliminated the means of the conditions in an experiment may differ from one another for two reasons. what are they?
the means may differ because of error variance; error variance reflects the random influences of variables that remain unidentified in the study, such as individual differences or because the IV did have an effect on their results
critical value
the minimum value of a statistic at which the results would be considered statistically significant
distinguish between the null and the experimental hypothesis
the null hypothesis states that the independent variable did not have an effect on the dependent variable. the experimental hypothesis states that the IV did have na effect
if the independent variable has no effect the calculated value of F will be around 1 why?
the numerator and denominator of the F-ratio are estimates of the same thing and the value of F will be 1 (mean square between groups/mean square within groups)
why must a difference t-test be used for matched subject and within subject designs that for randomized groups designs
the paired t-test takes advantage of the correlation between pairs to reduce the estimate of error variance used to calculate ; we can account for the source of some of the error variance in the data
how do researchers analyze the data from single case experiments?
the preferred method of presenting the data from a single participant design is with graphs that show the results individually for each participant rather than testing the significance of the experimental effect they employ graphic analysis
beta
the probability of committing a type 2 error
in analyzing the data from an experiment why is it not sufficient simply to examine the condition means to see whether they differ?
the problem with merely looking at the differences between the means of the experimental conditions is that means may differ even if the independent variable does not have an effect; if something other than the independent variable differs in a systematic fashion between the means may be due to this confounding variable rather than the independent variable
what is the rationale behind the ABA design?
the researcher who uses these designs attempts to demonstrate that an IV affects behavior, first by showing that the variable causes a target behavior to occur and then by showing that removal of the variable causes the behavior to cease
dependent variable
the response measured in a study typically a measure of participants thoughts feelings behavior or physiological reactions
Why do single case researchers believe that the data from individual participants should not be combines aa when we compute a group mean
the results of an averaged group data do not reflect the behavior of any participant. group data can be misleading whereas single-participant designs show the true pattern
pretest sensitization
the situation that occurs when completing a pretest affect participants repossess on the posttest
regression coefficient
the slope of a regression line
what special problems do nested designs create for researchers? what special opportunities do they offer for understanding how variables relate to one another
the special problem with nested designs in that the responses of the participants within any particular group are not independent of one another but the special opportunity offered is that ability for researchers to study variables that operate at different levels of analysis
effect size
the strength if the relationship between two or more variables usually expressed as the proportion of variance in one variable that van be accounted for by another variable
in what areas have single case experiments been primarily used?
the study of operant conditioning; schedules of reinforcement/punishment; the effects of behavior modification; demonstrations purposes simply to show that a particular behavioral effect can be obtained
sum of squares
the sum of the squared deviations between individual participants scores and the means
sum of squares within groups
the sum of the variances of the scores within particular experimental conditions
total sum of squares
the total variability in a set of data
outcome variable
the variable being predicted in a multiple regression analysis
criterion variable
the variable being predicted in a regression analysis; the dependent or outcome variable
sum of squares between groups
the variance in a set of scores that tis associated with the independent variable the sum f the squared difference between each condition mean and the grand mean
why are behavioral scientists reluctant to trust case studies as a means of testing hypotheses
there is a failure to control extraneous variables, and there is observer biases
describe a 2x2x3 factorial design. how many independent variables are involved and how many levels are there of each variable? how many experimental conditions are in a 2x2x3 factorial designs?
three IVs, levels two two and three, 12 conditions
explain the rational behind the t-test
to analyze data from a two-group randomized groups experiment
what is a multiple baseline design and when are such designs typically used
two or more behaviors are studied simultaneously. after obtaining baseline data on all behaviors, an IV is introduced that is hypothesized to affect only one of the behaviors
how many conditions are there in the simplest possible experiment?
two; there are only two levels of the independent variables; a minimum of two conditions is needed so that we can compare participants responses in one experimental condition with those in another conditions
distinguish between a type 1 and type 2 error
type 1=occurs when a researcher erroneously concludes that the null hypothesis is false and this rejects it type 2= occurs when a researcher concludes that the independent variable did not have an effect when in fact it did
a difference between means that is statistically significant is
unlikely to be due to error variance
explain how you would use regression analysis to see whether variable z mediates the effect of variable x on variable y
use of hierarchical multiple regression is to test meditational hypothesis; many hypotheses specify that the effects of a predictor variable on a criterion variable are mediated by one or more other variables. medication effects occur when the effect of x on y occurs because of an intervening variable z
Interaparticipant variance
variability among the responses of the participants in a particular experimental condition
Interparticipant variance
variable among the responses of a participant when tested more than once in a particular experimental condition
under what circumstances is an ABA design relatively useless as a way of testing the effects of an independent variable?
when the IV produces permanent changes in participants behavior changes that do not reverse when the IV is removed
individual differences among participants contribute to
within groups variance
write the general form of a regression equation that has a single predictor variable. identify the criterion variable the predictor variable the regression constant and the regression coefficient
y= mx +b x=predictor variable b= regression constant m=regression coefficient y=criterion variable
regression constant
yhr y-intercept in a regression equation
imagine that the equation for predicting y from x is y=1.12-.47x. how would you use this equation to predict a particular individual score
you plug the individual score into x and solve