Psychology Research Methods - Chapter 14
A one-tailed test is conducted when one's _____. research lacks variable data research results are externally valid research hypotheses are directional research includes voluntary participants
research hypotheses are directional
Orthogonal comparisons are comparisons between pairs of means in an experiment that yield new information. Any set of means has _____ orthogonal comparisons, where k is the number of treatments. (k - 3) (k - 4) (k - 2) (k - 1)
(k - 1)
To facilitate comparison across variables and experiments, effect size is usually reported as _____. a ratio of between-groups variability to within-groups variability in a statistical test a proportion of the variation in scores within the treatments under comparison a ratio between the variance caused by an independent variable plus experimental error and the variance caused by experimental error alone. a proportion of dependent variables and independent variables of a statistical test
a proportion of the variation in scores within the treatments under comparison
If one wanted to know whether a research hypothesis was either better or worse than the standard method, then _____ is conducted. a one-tailed test a t test a z test a two-tailed test
a two-tailed test
Two types of error that must be considered when making many comparisons are per-comparison error and familywise error. Per-comparison error is the _____. probability of making at least one Type I error with the increase in the number of comparisons alpha for each comparison between means probability of making at least one Type II error with the increase in the number of comparisons obtained p for each comparison
alpha for each comparison between means
In the context of determining statistical significance, the probability of committing a Type I error depends on the criterion one uses to accept or reject the null hypothesis. This criterion is known as the _____. alpha level observed value beta level critical value
alpha level
Unplanned comparisons are often "fishing expeditions" in which one is simply looking for _____. a way to calculate the sample mean the occurrence of Type I error any differences that might emerge alternate explanations
any differences that might emerge
A true statement about unplanned comparisons is that these _____. can be used in lieu of an overall ANOVA are also known as a priori comparisons are used when one does not have a specific preexperimental hypotheses are made using information from one's overall ANOVA
are used when one does not have a specific preexperimental hypotheses
In the context of using the two-factor between-subjects ANOVA, the presence of an interaction in an experiment shows that neither of the independent variables has a simple, independent effect. Therefore, one should _____. always interpret main effects when an interaction is present avoid interpreting main effects when an interaction is present identify how main effects are influenced by the interactions identify the nature of the interactions and its effects on the dependent variable
avoid interpreting main effects when an interaction is present
Inferential statistics let one infer the characteristics of a population from the _____. calculated measure of spread from the population of scores dependent and independent scores in one's data characteristics of the samples comprising one's data calculated mean from the population of scores
characteristics of the samples comprising one's data
When one's dependent variable is a dichotomous decision or a frequency count, the statistic of choice should be _____. the t test the Mann-Whitney U test chi-square ANOVA
chi-square
When using a two-factor within-subjects ANOVA, researchers must _____. only consider the effect that the dependent variable has on the independent variable consider the power of the statistical test when evaluating Type I errors consider the interaction between each of the independent variables and the subjects factor only consider the interaction between dependent variables
consider the interaction between each of the independent variables and the subjects factor
In a distribution of scores with a known mean, the number of scores that are free to vary around the mean are known as the ____ ____ ____ (df).
degrees of freedom
The degree to which the manipulation of an independent variable changes the value of a dependent variable is termed the ____ ____.
effect size
The process of using sample data to estimate population parameters is called ____.
estimation
When the correlation between samples is 0, the t values _____. given by the correlated samples and independent samples t tests (unpooled version) are identical given by the correlated samples t test are more than those of independent samples t test (pooled version) given by the correlated samples t test are less than those of independent samples t test (pooled version) given by the correlated samples and independent samples t tests (pooled version) are identical
given by the correlated samples and independent samples t tests (pooled version) are identical
The degrees of freedom (df) for a single sample are _____. n - 2 n - 4 n - 3 n - 1
n - 1
Statistical theory reveals that sampling distribution of the mean closely approximates the _____, even when the population of scores from which the samples are drawn is not normal in shape. bimodal distribution. positively skewed distribution normal distribution negatively skewed distribution
normal distribution
Statistical theory reveals that sampling distribution of the mean closely approximates the _____, even when the population of scores from which the samples are drawn is not normal in shape. positively skewed distribution normal distribution bimodal distribution negatively skewed distribution
normal distribution
In a weighted means analysis, each group mean is weighted according to the _____. number of dependent variables in the group treatments used in the group number of subjects in the group number of levels of independent variables in the group
number of subjects in the group
One uses the t test for independent samples when _____. one has data from two groups of participants who were assigned at random to the groups one has data from more than two groups of participants who were assigned at random to the two groups the data at one's disposal has do not meet the assumptions of a parametric test the two means being compared come from samples that are not independent of one another
one has data from two groups of participants who were assigned at random to the groups
A special ANOVA is used when _____. all subjects in a within-subjects design with two factors are exposed to every possible combination of levels of one's two independent variables one has used a design mixing between-subjects and within-subjects components one has included a continuous correlational variable in one's experiment one includes two independent variables and randomly assign different subjects to each condition
one has included a continuous correlational variable in one's experiment
If researchers used a multilevel within-subjects design in their experiment, the statistical test to use is the _____. two-factor within-subjects ANOVA two-factor between-subjects ANOVA one-factor between-subjects ANOVA one-factor within-subjects ANOVA
one-factor within-subjects ANOVA
Statistics characterizing a population are called _____. sample statistics the distribution of population the sampling distribution of the mean population parameters
population parameters
If a difference of the treatment means of one's experiment is not statistically significant, then it can have no _____, no matter how big the difference is. practical significance valid significance scientific significance analytical significance
practical significance
The distribution of every possible sample of n scores from one's population is known as _____. sampling distribution of the mean sampling distribution of the variance sampling distribution of the range sampling distribution of the mode
sampling distribution of the mean
For ANOVA, the variation in scores is divided, or partitioned, according to the factors assumed to be responsible for producing that variation. These factors are referred to as _____. coefficients of variance ratios of variance sources of variance measures of variance
sources of variance
The standard error of the mean is used to estimate the _____ of the sampling distribution of the mean. degrees of freedom obtained p variance standard deviation
standard deviation
If one's measure of effect size is the difference between the condition means, divided by the pooled standard deviation, then one can think of it as a _____. chi-square statistic t-score standard z-score f statistic
standard z-score
If the difference between means yields an observed value of a statistic that meets or exceeds the critical value of one's inferential statistic, the difference can be declared as _____. statistically significant practically significant irrelevant for one's study crucial for one's study
statistically significant
If the obtained p is less than or equal to alpha, one's comparison is _____. statistically significant scientifically proved variable in nature statistically invalid
statistically significant
A true statement about parametric tests is that _____. parametric statistics are useful when one's data do not meet the assumptions of nonparametric statistics they usually provide more sensitive test of the null hypothesis than does equivalent nonparametric statistics parametric statistics can also be used when one's dependent variable is scaled on a nominal or ordinal scale appropriate versions of parametric statistics are not always available for complex designs
they usually provide more sensitive test of the null hypothesis than does equivalent nonparametric statistics
An unweighted means analysis gives each group in one's design equal weight in the analysis, despite _____. unequal distribution of data different experiment techniques unequal group sizes different guidelines and procedures
unequal group sizes
A significant F ratio tells one that at least some of the differences among one's means are probably not caused by chance but rather by _____. different subjects' characteristics variation in a dependent variable experimental errors variation in one's independent variable
variation in one's independent variable
If the inequality in sample sizes was planned or reflects actual differences in the population, one should use a(n) _____. weighted variance analysis unweighted means analysis weighted means analysis unweighted variance analysis
weighted means analysis
The _____ for the difference between two proportions is used to determine the significant difference between the samples proportion and the population proportion. Wilcoxon signed ranks test z test Mann-Whitney U test t test
z test
Two types of error that must be considered when making many comparisons are per-comparison error and familywise error. Familywise error can be calculated with the formula _____. αFW = k - 2, where k is the number of groups αFW = 1 - (1 - α)^c, where α is an alpha level αFW = k - 1, where k is the number of treatments αFW = k (n - 1), where k is the number of groups and n is the number of subjects in each group
αFW = 1 - (1 - α)^c, where α is an alpha level
Which of the following have special versions of the t test? (Check all that apply.) Designs involving correlated samples Designs involving uncorrelated samples Designs involving independent samples Designs involving dependent samples
Designs involving correlated samples Designs involving independent samples
Identify the problem with the F ratio obtained in the one-factor between-subjects ANOVA. It fails to tell that the differences among means are caused by chance or by variation in one's independent variable. One cannot find the critical value of F if the table does not list the exact degrees of freedom for the denominator. It fails to determine statistical significance accurately. It fails to tell one where among the possible comparisons the reliable differences actually occur.
It fails to tell one where among the possible comparisons the reliable differences actually occur.
Identify a true statement about the effect of an increased sample size on the power of a statistical test. It makes it difficult to provide stable estimates of population parameters. It adds to the difficulty of rejecting the null hypothesis when it is false. It makes it easier to accept the null hypothesis when it is false. It helps easily detect small differences in population means.
It helps easily detect small differences in population means.
In the context of the uses of data transformations, which of the following is true of the square root? It is useful if data show a moderate positive skew. It is used when basic observations have a binomial distribution. It is useful if data show a negative skew. It normalizes data with severe positive skew.
It is useful if data show a moderate positive skew.
Identify the consequence of serious violation of one or more assumptions for a parametric statistic. It will strengthen the value of the statistic as a guide to decision making. It may lead one to commit a Type II error. One will need to measure the data on a ratio scale. It may bias the statistical test.
It may bias the statistical test.
What will happen if one reduces alpha level from .05 to .01? It will make it more difficult to reject the alternate hypothesis. It will make it more difficult to accept the null hypothesis. It will reduce the probability of making a Type I error. It will increase the probability of making a Type I error.
It will reduce the probability of making a Type I error.
What are used to counterbalance the order in which subjects receive treatments in within-subjects experiments? Cross-Sectional Designs Latin square designs Mixed designs Time Series Designs
Latin square designs
The _____ is used to evaluate the significance of a difference between two independent groups when one's dependent variable is scaled on at least an ordinal scale. Anderson-Darling test Mann-Whitney U test Shapiro-Wilk test Kolmogorov-Smirnov KS test
Mann-Whitney U test
Why do the scores obtained in a between-subjects experiment vary from one another even when all subjects are exposed to the same treatment conditions? (Check all that apply.) The frequency of the experiment varies. All the participants are similar. Measurement error fluctuates. Subjects differ from one another.
Measurement error fluctuates. Subjects differ from one another.
_____ make assumptions about the distribution from which the scores were sampled. Parametric statistics Basic statistics Nonparametric statistics Descriptive statistics
Parametric statistics
Identify the test that can be used as an alternative to chi-square when one has small expected frequencies and a 2 X 2 contingency table. The Wald-Wolfowitz runs test The Mann-Whitney U test The Fisher exact probability test The Wilcoxon signed ranks test
The Fisher exact probability test
Which of the following tests is a good alternative to the t test when one's data do not meet the assumptions of the t test? The Shapiro-Wilk test The Anderson-Darling test The Kolmogorov-Smirnov KS test The Mann-Whitney U test
The Mann-Whitney U test
Which nonparametric test should be used for a single-factor experiment using a within-subjects or matched-pairs design? The Fisher exact probability test The Kolmogorov-Smirnov test The Wilcoxon signed ranks test The Mann-Whitney U test
The Wilcoxon signed ranks test
Which of the following are true of the z test for the difference between two proportions? (Check all that apply.) The logic behind this test is essentially the same as for the t tests. The difference between the two proportions is evaluated against an estimate of population range. The logic behind this test is different from that of t tests. The difference between the two proportions is evaluated against an estimate of error variance.
The logic behind this test is essentially the same as for the t tests. The difference between the two proportions is evaluated against an estimate of error variance.
Which of the following is the probability that a statistical test will reject the null hypothesis when the null hypothesis is false and is one minus β? The effect size of a statistical test The power of a statistical test The sample size of a statistical test The alpha level of a statistical test
The power of a statistical test
Which of the following is true of the sample size in the context of statistical tests? The probability of making a Type I error increases with the size of the sample. Small samples provide stable estimates of population parameters. Large samples make it difficult to detect small differences in population means. The power of a statistical test increases with the size of the sample.
The power of a statistical test increases with the size of the sample.
Which of the following is true of a two-factor within-subjects ANOVA design? The researcher includes two independent variables and randomly assigns different subjects to each condition. The researcher's experiment includes only dependent variables that are scaled on an ordinal scale. The researcher treats subjects as a factor along with the manipulated independent variables. The researcher's experiment includes only one factor and has different subjects in each experimental condition.
The researcher treats subjects as a factor along with the manipulated independent variables.
Identify the assumptions underlying a parametric inferential test. (Check all that apply.) The within-groups variances are heterogeneous. The scores have been sampled randomly from the population. The sampling distribution of the mean is normal. The dependent variables are scaled on a ratio scale.
The scores have been sampled randomly from the population. The sampling distribution of the mean is normal.
Identify the factors that affect the power of a statistical test. (Check all that apply.) The chosen delta level The size of the effect produced by the independent variable The size of the sample The size of the effect produced by dependent variables The use of a one-tailed or two-tailed test
The size of the effect produced by the independent variable The size of the sample The use of a one-tailed or two-tailed test
Which statistical test should one use when one's experiment includes only two levels of the independent variable? The chi-square test The t test The Wilcoxon signed ranks test The Mann-Whitney U test
The t test
Identify the causes of between-groups variability. (Check all that apply.) The variation in an independent variable Individual differences among the different subjects in groups Experimental error Using less than two groups in an experiment The variation in dependent variables
The variation in an independent variable Individual differences among the different subjects in groups Experimental error
In the context of null hypothesis significance testing (NHST), which of the following is true of a two-tailed test? Using the two-tailed test means giving up any information about the reliability of a difference in one of the two directions. Researchers typically use this test when their research hypotheses are directional. The use of this test should always be avoided unless there are compelling a priori reasons not to. The z scores required to reach statistical significance must be more extreme for a two-tailed test than is the case for a one-tailed test.
The z scores required to reach statistical significance must be more extreme for a two-tailed test than is the case for a one-tailed test.
True or false: If a difference between sample means is not statistically significant, then it can have practical significance.
This is false. If a difference is not statistically significant, then it can have no practical significance, no matter how big the difference is. It is not reliable; therefore, even a large difference likely reflects nothing more than the effect of extraneous variables on one's measure.
True or false: In Type II error, one rejects a null hypothesis when it is true, whereas in Type I error, one does not reject a null hypothesis when it is false.
This is false. In Type I error, one rejects a null hypothesis when it is true, whereas in Type II error, one does not reject a null hypothesis when it is false.
True or false: It is easier to reject a null hypothesis with a the two-tailed test than with a one-tailed test.
This is false. It is easier to reject the null hypothesis with a one-tailed test than with a two-tailed test.
True or false: The probability estimates given by the pooled version of the t test for independent samples are always accurate, even if there are large differences in sample sizes and standard errors.
This is false. The probability estimates given by the pooled version may be misleading if there are large differences in sample sizes and standard errors. The pooled version should be avoided if there are large differences in sample sizes and standard errors.
True or false: There is an inverse relation between the alpha level and the power of one's statistical test.
This is false. There is a direct relation between alpha level and power in an experiment. As one reduces one's alpha level, it will reduce the power of one's statistical test.
True or false: Using computerized statistical packages, one can compare the obtained p with one's selected alpha level by using the relevant table of critical values of one's test statistic.
This is false. These days, most statistical analyses are conducted using computerized statistical packages that usually provide the exact probability value p along with the obtained value of the test statistic. One can directly compare this obtained p with one's selected alpha level and avoid having to use the relevant table of critical values of one's test statistic.
True or false: When one's dependent variable is a dichotomous decision or a frequency count, the statistic of choice is ANOVA.
This is false. When one's dependent variable is a dichotomous decision or a frequency count, the statistic of choice is chi-square (x^2).
In the context of hypothesis testing, _____ allows one to compute what researchers actually want—the probability of the null hypothesis, given the data. Fisher's null hypothesis testing a loglinear analysis a Bayesian analysis null hypothesis significance testing
a Bayesian analysis
If one wants to guard more strongly against Type I errors, one can adopt _____. a more stringent alpha level, such as the .01 level an element of replication in every study a more flexible alpha level, such as the .05 level an estimate of the effect size for each statistical test
a more stringent alpha level, such as the .01 level
One would conduct _____ if one is interested only in whether the obtained value of the statistic falls in one tail of the sampling distribution for that statistic. a z test a t test a two-tailed test a one-tailed test
a one-tailed test
By convention, the minimum acceptable alpha has been set at _____. .10 .25 .20 .05
.05
Identify a true statement about the F ratio with reference to the one-factor within-subjects ANOVA. A significant overall F ratio tells one where the significant differences among means occur. A significant overall F ratio indicates that significant differences exist among one's means. A significant overall F ratio counterbalances the order in which subjects receive treatments. A significant overall F ratio randomly assigns different subjects to each condition.
A significant overall F ratio indicates that significant differences exist among one's means.
Match the following types of statistical error (in the left column) to the accurate conclusions about independent variable (in the right column) in an experiment. A. Type I error B. Type II error 1. Error of assuming independent variable having an effect on dependent variable when in fact it does not 2. Error of assuming independent variable not having an effect on dependent variable when in fact it does
A. 1., B. 2.
Match the types of data transformations (in the left column) with their descriptions (in the right column). A. Linear transformations B. Nonlinear transformations 1. They simply change the magnitude of the numbers representing the data, but they do not change the scale of measurement. 2. They tend to correct the problems with the data by changing a skewed distribution of scores into a normal one or by removing heterogeneity of variance.
A. 1., B. 2.
Match the concepts related to the process of estimation (in the left column) with their descriptions (in the right column). A. A point estimate B. The margin of error (MoE) C. The confidence interval (CI) 1. It refers to the best single-number guess of the population percentage. 2. It estimates the range of values within which the true population value lies. 3. It is a range of values within which the true population percentage is likely to lie with a specified probability.
A. 1., B. 2., C. 3.
Which of the following allows one to examine the relationship between experimentally manipulated variables while controlling another variable that may be correlated with them? z test t test Chi-square ANCOVA
ANCOVA
An unweighted means analysis involves a minor correction to the _____. t test ANOVA F ratio one-tailed test
ANOVA
When one's experiment includes more than two groups, the statistical test of choice is _____. the t test ANOVA chi-square the z test
ANOVA
Identify the factors that determine the value of any particular score obtained in a between-subjects experiment. (Check all that apply.) Scale used to measure experimental data Characteristics of the subject at the time the score was measured The value of the dependent variable The value of the independent variable Measurement or recording errors
Characteristics of the subject at the time the score was measured The value of the independent variable Measurement or recording errors
In the context of the limitations of chi-square, what should researchers do if the expected cell frequencies for the chi-square test are less than five? (Check all that apply.) Consider a different test Include more subjects to increase the sample size Reduce the alpha level Combine cells Review independent variables of the experiment
Consider a different test Include more subjects to increase the sample size Combine cells
Given that one's results are statistically significant, what can help one to gauge the practical significance of one's finding? The obtained p The F ratio Effect size Alpha level
Effect size
The carryover effects contained in the Latin square design tend to inflate the error term used to calculate one's _____. p value α level t statistic F ratio
F ratio
Identify a true statement about the one-factor within-subjects ANOVA. Individual differences of subjects contribute to the between-treatments sum of squares. The between-treatments sum of squares are unaffected by experimental error. The between-treatments sum of squares cannot be affected by the level of the independent variable. Individual differences do not contribute to the between-treatments sum of squares.
Individual differences do not contribute to the between-treatments sum of squares.
Which of the following are true of using an alpha level lower than .05 for a study? (Check all that apply.) It makes it easier for one to accept the null hypothesis when the null hypothesis is false. It allows one to have greater confidence in one's decision about the results. It allows one to meet the assumptions of a parametric test. It reduces the possibility of one making a Type I error.
It allows one to have greater confidence in one's decision about the results. It reduces the possibility of one making a Type I error.
The F ratio is simply the ratio of _____. between-groups variability to total variability single-subject within-groups variability to multiple-subjects between-groups variability within-groups variability to total variability between-groups variability to within-groups variability
between-groups variability to within-groups variability
The within-groups variability may be attributed to _____. the variation in dependent variables individual differences between subjects treated alike within groups similarities among the different subjects in groups the variation in an independent variable
individual differences between subjects treated alike within groups
Statistics that assess the reliability of one's findings are called _____. inferential statistics elementary statistics descriptive statistics population statistics
inferential statistics
The two-factor between-subjects ANOVA is more complicated than a one-factor ANOVA because _____. it must determine the scientific applicability of each main effect and its interactions it must prove that an inverse relationship exists between each main effect it must determine the statistical significance of each main effect and of the interaction as well it must determine the influence of each main effect and reduce the chances of Type II errors
it must determine the statistical significance of each main effect and of the interaction as well
A two-tailed test is _____ than a one-tailed test. scientifically significant stronger easier to calculate less powerful
less powerful
The particular level of alpha one adopts is called the _____. level of significance obtained p value critical value two-tailed test
level of significance
In the context of determining statistical significance, the likelihood of making a Type 1 error can be reduced by _____. increasing statistical power making the alpha level larger decreasing statistical power making the alpha level smaller
making the alpha level smaller
In the context of determining statistical significance, the likelihood of making a Type 1 error can be reduced by _____. making the alpha level smaller increasing statistical power making the alpha level larger decreasing statistical power
making the alpha level smaller
The ANOVA computations for _____ designs involve sums of squares for the between factor and for the within factor of an experiment. mixed nested within-subjects between-subjects
mixed
A significant overall F ratio does not _____. show any statistical inferences regarding the experiment tell where significant differences occur indicate significant differences among one's means assign values to calculate the frequency variation
tell where significant differences occur
The statistic used in ANOVA to determine statistical significance is _____. chi-squared statistics the t statistics sample mean the F ratio
the F ratio
The standard error of the mean may be calculated from the given formula: sM = s/√n. In the formula, n refers to _____. the number of scores in the sample the mean of the sample the sample standard deviation the number of scores in the population from which the sample was drawn
the number of scores in the sample
Chi-square for contingency tables compares _____. the sample means with the observed cell frequencies the expected cell frequencies with the expected median the values of the interquartile range with the variations in the study the observed cell frequencies with the expected cell frequencies
the observed cell frequencies with the expected cell frequencies
Chi-square for contingency tables is designed for frequency data in which _____. the relationship between two factors is to be determined the relationship between two variables is to be determined the independent variable is scaled on an interval scale the dependent variable is scaled on an ordinal scale
the relationship between two variables is to be determined
The t test for correlated samples produces a larger t value than the t test for independent samples when applied to the same data if _____. the scores from the two samples are randomly assigned the variance of the two samples is the same the mean of the two samples is different the scores from the two samples are at least moderately correlated
the scores from the two samples are at least moderately correlated
The most complex part of the mixed design ANOVA is _____. the selection of an error term to calculate the F ratios choosing the alpha level for Type I errors determining the sample size deciding between one-tailed and two-tailed tests
the selection of an error term to calculate the F ratios
A true statement about the planned comparisons is that _____. these may be performed in experiments with many levels of an independent variable these are used when one has specific preexperimental hypotheses these are also known as post hoc comparisons these are often "fishing expeditions" in which one is simply looking for any differences that might emerge
these are used when one has specific preexperimental hypotheses
