Q&A Hero

Q: The ANOVA Summary Table for a two-factor, mixed design is divided into two parts; the between subjects section and the __________ section. A. interaction B. within-subjects C. eta-squared D. correlation

Q: When reporting the results of a complex design experiment, which of the following should not be included? A. summary statistics for cells in the design in the text, a table, or a figure B. results for omnibus F-test, with exact probabilities C. verbal description of any statistically significant interaction effects D. each subject's score for each dependent variable

Q: Null hypothesis significance testing is used to compare two means in an independent groups design. (a) What is the null hypothesis? (b) What is required to "reject" the null hypothesis?

Q: Explain the difference between experimental sensitivity and statistical power and identify factors that influence each.

Q: (p. 383, 386) (a) What is a Type I error and what is a Type II error when using NHST? (b) Which is more common in psychological research? (a) Decisions about the outcome of an experiment are based on probabilities. A Type I error occurs when researchers rejecting a true null hypothesis (i.e., there is no difference between population means but a difference is claimed). A Type II error occurs when researchers fail to reject a false null hypothesis (i.e., a true difference between population means is missed). (b) Type II errors are more common in psychological research because many studies have low statistical power.

Q: What does NHST tell us when a "statistically significant" finding is obtained?

Q: When a result is "statistically significant," why should we not immediately claim that our results are "important" (either scientifically or practically)?

Q: Briefly describe the logic of the analysis of variance or F-test for a single-factor random groups design. There are two sources of variation in any random groups experiment. First, variation within each group can be expected because of individual differences among subjects who have been randomly assigned to a group. The second source of variation in the random groups design is variation between the groups. If the null hypothesis is true (no differences among groups), any observed differences among the means of the groups can be attributed to error variation (e.g., the different characteristics of the participants in the groups). Thus, the variation among the different group means, when the null hypothesis is assumed to be true, provides a second estimate of error variation in an experiment. If the null hypothesis is true, this estimate of error variation between groups should be similar to the estimate of error variation within groups. Thus, the random groups design provides two independent estimates of error variation, one within the groups and one between the groups. The ratio of these two estimates, as represented in the F-test, should be 1.0. If the null hypothesis is false, that is, the independent variable has had an effect, there will be systematic differences in the means across the different groups of the experiment. This systematic variation will be added to the differences in the group means that are already present due to error variation. By creating a ratio (F-ratio) of variation between groups and variation due to individual differences, evidence can be obtained regarding the likelihood of systematic variation in the experiment.

Q: In general, a repeated measures design is likely to be ________ than a random groups design. A. less systematic B. less powerful C. more sensitive D. more accurate

Q: The results of an omnibus F-test for a complex design experiment allow the researcher to A. know whether the interaction effects and main effects are statistically significant. B. interpret whether the pattern of means supports the hypotheses. C. know whether the effects in the experiment are linear. D. know which means in the experiment differ significantly from one another.

Q: If the omnibus analysis of variance for a complex design reveals a statistically significant interaction effect, the source of the interaction effect may be identified using A. null comparisons. B. simple main effects analyses. C. correlational analyses. D. complex comparisons.

Q: If the omnibus analyses of variance for a complex design reveals a statistically significant interaction effect, the source of the interaction effect may be identified using simple main effects analysis and, when there are more than two levels of an independent variable, also A. complex comparisons. B. null comparisons. C. simple correlations between variables. D. comparisons between two means.

Q: In order to conclude whether the results of an inferential test support the research hypothesis, a researcher must A. re-state the null hypothesis. B. examine the descriptive (summary) statistics. C. determine whether the mean square error is greater than zero. D. reject probability values greater than .05.

Q: Two measures of effect size for an independent groups design with more than two means are A. the t-test and the F-test. B. Cohen's d and the F-test. C. eta-squared and Cohen's f. D. F squared and rho.

Q: Following an omnibus F-test, a researcher may learn about specific sources of systematic variation in a single-factor independent groups experiment by performing A. comparisons of two means. B. repeated measures t-tests. C. correlation coefficients. D. simple main effects.

Q: The primary way that analysis of variance for repeated measures differs from that for an independent groups design is in the estimation of A. between-group variation. B. variability between means. C. systematic variation. D. error variation.

Q: As the degree of linear relationship between two measures increases, A. the absolute value of the correlation coefficient approaches 1.00. B. the scatterplot resembles a straight line. C. our ability to predict for these variables increases. D. all of these

Q: The direction (positive or negative) of a correlation is indicated by A. the sign of the coefficient. B. numbers between 0.0 and 1.00. C. the absolute value of the correlation. D. our ability to make predictions for the variables.

Q: Suppose that research on intelligence and mental illness suggests that at any level of intelligence, the likelihood of being mentally ill or not is 50/50. That is, the correlation between measures of intelligence and mental illness is close to A. -50. B. 0.00. C. +50. D. +1.00.

Q: When two variables are correlated, our ability to make ___________ increases. A. causal inferences for these variables B. predictions about a third variable C. predictions for these variables D. causal inferences about a third variable

Q: One reason we may not make causal conclusions based only on correlational evidence is that a correlation between two variables A. does not tell us about the possible direction of causality. B. may be negative. C. never involves causally related variables. D. may be greater than 1.00.

Q: A researcher finds a correlation of +.90 between two variables. Assuming the correlation coefficient was calculated correctly, which of the following is definitely not true? A. Ability to make predictions for these variables is good. B. As values on one measure increase, values on the other measure decrease. C. A third variable may be present that explains the correlation. D. The variables may be causally related.

Q: The relationship between a sample's correlation coefficient and the population correlation (r, "rho") is A. analogous to the relationship between a sample mean and a population mean. B. such that the correlation coefficient provides an estimate of rho. C. such that the more sampling error there is for the correlation coefficient, the wider the confidence interval for rho. D. all of these

Q: Assume that the results of an F-test for a single-factor experiment are reported as: F(2, 42) = 7.30, p = .01. On the basis of this information we may conclude that there were _______ independent groups in this experiment. A. two B. three C. four D. five

Q: Assume that the results of an F-test are reported as: F(2, 42) = 7.30, p = .01. On the basis of this information we may conclude that A. there were two independent groups in the experiment. B. there were 42 total subjects in the experiment. C. the results are statistically significant. D. all of these

Q: The value of the F-statistic in a single-factor experiment is determined by dividing the Between-Group Mean Square by the A. Within-Group Mean Square. B. degrees of freedom. C. Between-Group Sum of Squares. D. Total Sum of Squares.

Q: When our results are "statistically significant" we know A. the independent variable likely produced an effect on the dependent variable. B. the direction of the effect of the independent variable. C. the effect size between the independent and dependent variables is large. D. all of these

Q: Which of the following indicates the strongest correlation? A. -.80 B. 0.00 C. +.25 D. +.75

Q: When interpreting confidence intervals when there are three or more means, if the intervals overlap such that the sample mean of one group lies within the interval of another group, we may conclude that A. the sample means are the same. B. the population means do not differ. C. the population means for the groups differ. D. the population mean for another group will be different.

Q: If a scatterplot shows that most of the points fall on a straight line, we can be confident that the correlation between the two measures A. should not be computed. B. is near zero. C. is weakly positive. D. is strong.

Q: One important reason for displaying correlational data in a scatterplot prior to computing a correlation coefficient is to make sure A. the relationship between the variables is linear. B. the means for the two variables are the same. C. the x- and y-axis for each variable has appropriate endpoints. D. the causal variable is graphed on the x-axis.

Q: A negative correlation indicates that as values for one measure ___________, the values for the other measure ____________. A. decrease; remain constant B. increase; remain constant C. increase; decrease D. increase; increase

Q: Which of the following is certainly not positively correlated? A. SAT scores and college grades B. smoking and lung disease C. amount of snowfall and driving speed D. high school grades and college grades

Q: A correlation coefficient is a ________ summary of the degree of relationship between two sets of scores. A. subjective B. qualitative C. quantitative D. graphical

Q: Values for a correlation coefficient range from A. -1.00 to 0.0. B. -1.00 to +1.00. C. 0.0 to -1.00. D. 0.0 to +1.00.

Q: The 95% confidence interval for a population mean is calculated as follows: A. sample mean (t critical) (estimated standard error) B. sample mean (.95) (estimated standard error) C. population mean (.95) (estimated standard error) D. population mean (t critical) (estimated standard error)

Q: The "margin of error" typically reported for survey results estimates the difference between the sample results and the population values due to A. computer calculations. B. differences among survey methods. C. chance or random factors. D. survey results.

Q: The "margin of error" for a mean value provides a range of values that are likely to contain the A. sample mean. B. population mean. C. effect size. D. sample standard deviation.

Q: A confidence interval is basically the same as A. the difference between two means. B. the range. C. the standard deviation. D. a margin of error.

Q: The inferential statistic that is used in the calculation of a confidence interval is A. the t statistic. B. the degrees of freedom. C. the effect size. D. analysis of variance.

Q: Having calculated a 95% confidence interval for a single population mean we may state that the odds are 95/100 that the A. population mean equals 95. B. sample mean is the same as the population mean. C. obtained interval contains the population mean. D. population mean falls in the interval.

Q: To construct a confidence interval for a comparison between two independent group means we substitute the ____________ for a single sample mean. A. difference between two sample means B. difference between the sample mean and the population mean C. difference between two population means D. population mean

Q: Having calculated a 95% confidence interval for a difference between two means, we may conclude that the odds are 95/100 that the A. true population mean difference falls in the interval. B. difference between the sample mean is the same as the difference between the population means. C. population mean difference is less than 95. D. obtained interval contains the true population mean difference.

Q: To find the value of t critical for a repeated measures design, we calculate degrees of freedom based on the A. number of scores obtained. B. number of scores obtained minus one. C. number of pairs of scores minus one. D. size of the group.

Q: A conceptual definition of effect size for an independent variable with two conditions is the difference between the two sample means divided by A. the sample size, N. B. the variability within the groups. C. Cohen's d. D. the population difference between two means.

Q: A student conducts a research project to test the effect of an independent variable with two conditions. The value for Cohen's d for her data is 0.25. Based on this, she concludes that the independent variable had _________ effect on the dependent variable. A. zero B. a small C. a medium D. a large

Q: A major approach to "confirming what the data reveal" is the calculation of A. standard deviations. B. stem-and-leaf displays. C. measures of central tendency. D. confidence intervals for a population parameter.

Q: The most commonly used measure of central tendency is the A. mean. B. median. C. mode. D. range.

Q: Which of the following is not a measure of dispersion or variability? A. range B. variance C. standard deviation D. confidence interval

Q: The standard deviation is equal to the square root of A. the variance. B. the sum of squared deviations from the mean. C. N - 1. D. the standard error of the mean.

Q: The mean of a random sample of scores is a point estimate of A. the center of a distribution. B. standard error of the mean. C. the population mean. D. population variability.

Q: The estimated standard error of the mean is equal to the sample ______ divided by _______. A. variance; N B. standard deviation; N - 1 C. standard deviation; N D. standard deviation; N

Q: In general, the estimated standard error of the mean provides information about how well the sample mean estimates A. the population variance. B. the population mean. C. the true standard error of the mean. D. population variability.

Q: The _________ the value of the estimated standard error of the mean, the _________ our estimate of the population mean. A. larger; better B. larger; higher C. smaller; better D. smaller; lower

Q: A researcher tests a sample of first-grade children for a school district to help the district identify areas in which the curriculum should be changed. The measure is a valid and reliable measure of intelligence. The researcher estimates the population mean for the intelligence scores and observes a large value for the standard error of the mean. The best thing the researcher could do to improve the estimate is A. test children in a different, more heterogeneous school district. B. compute a confidence interval. C. change the testing procedures to increase the variability of the test scores. D. increase the sample size of children tested.

Q: When calculating Cohen's d as a measure of effect size, the difference between two means is A. divided by the population standard deviation. B. divided by N - 1. C. divided by N. D. placed in the denominator.

Q: An advantage of a stem-and-leaf display is that A. it is a graph of the mean, median, and mode. B. it reveals whether the relationship between two variables is linear. C. it reveals the shape of a distribution and the presence of any outliers. D. the intervals tells us how well our sample mean estimates the population mean.

Q: Distributions of scores for two groups in an experiment can be compared by A. arranging the "stems" for both groups in one stem-and-leaf display and the "leaves" for both groups in a second stem-and-leaf display. B. preparing one stem-and-leaf display using the scores for both groups. C. arranging a stem-and-leaf display for each group side-by-side. D. none of these

Q: The value that splits a frequency distribution into two halves, each half with the same number of values, is the A. mode. B. mean. C. median. D. standard deviation.

Q: The ____________ is the best measure of central tendency when there are extreme values in the distribution. A. mode B. median C. mean D. variance

Q: Computer-assisted data analysis requires that the researcher have a good knowledge of A. null hypothesis testing. B. research design and statistics. C. computer hardware. D. computer software.

Q: Which of the following procedures would not be expected during the first stage of data analysis? A. performing tests using inferential statistics B. looking for outliers C. visualizing the distribution with graphical displays D. checking for errors

Q: A score that is extreme and does not "go with" other scores in the distribution is called an A. error. B. impossible value. C. significant value. D. outlier.

Q: If a stem-and-leaf display were created for the following set of numbers, 23, 24, 24, 21, 20, 27, 29, 28, we would expect the digit 2 to be a A. leaf. B. stem. C. tree. D. trunk.

Q: The relationship between basic research and applied research can best be described as A. an independent relationship. B. a reciprocal (or circular) relationship. C. a hierarchical relationship with basic research taking precedence over applied research. D. a top-down relationship with applied research taking precedence over basic research.

Q: Social policy research (such as the research done by Ellen Langer on mindfulness and the effects of living in a nursing home) serves the purpose(s) of A. informing policy decisions only. B. advancing scientific theory and knowledge only. C. advancing scientific theory and knowledge and informing policy decisions. D. advancing scientific theory and knowledge or informing policy decisions but not both.

Q: Perhaps the greatest difference between basic research and program evaluation lies in the differences in A. the political and social realities surrounding each of them. B. statistical procedures. C. the importance of the research. D. methodology.

Q: Which of the following is likely to be the most important benefit from the use of program evaluation based on sound experimental methodology? A. advancing the development of basic research in the social sciences by focusing on social policy B. advancing the influence and prestige of the social sciences in decision making among social policy makers C. providing information to policy makers that can be helpful in making more informed choices among possible treatments for social problems issues D. providing definitive information to policy makers such that program evaluation will eventually become the primary basis of social policy decisions

Q: Even when controlled, scientific research studies established the effectiveness of a home healthcare program for Medicare recipients, decision makers eliminated the program. The best explanation for this decision lies in A. politicians' unwillingness to fund scientific research. B. the political and social realities of complex problems such as healthcare. C. the public's greater acceptance of basic research over applied research. D. none of these

Q: Describe the advantages and disadvantages of the three measures of central tendency: mean, median, mode.

Q: Why is a confidence interval also called a "margin of error"?

Q: What does the width of a confidence interval tell us, and how does the size of a sample affect the width of a confidence interval for a population mean?

Q: What does the "95" in a 95% confidence interval refer to?

Q: Explain the phrase, "Correlation does not imply causation."

Q: Which of the following is not a major stage of data analysis? A. getting to know the data B. confirming what the data reveal C. transforming the data D. summarizing the data

Q: A coherent "story" of data analysis for a study will include A. an explanation of the findings. B. counter arguments for opposing interpretations of the findings. C. justifications for conclusions. D. all of these

Q: Program evaluation represents perhaps the extreme case of A. basic research. B. applied research. C. theory development. D. experimental control.

Q: The effect of a treatment in a simple interrupted time-series design is best indicated by A. an increasing trend in the dependent variable that is present both before and after the treatment. B. a decreasing trend in the dependent variable that is present both before and after the treatment. C. a clear discontinuity (abrupt increase or decrease) in the dependent variable at the point the treatment is administered. D. a gradual change in the dependent variable that begins just as the treatment is implemented.

Q: The major threat to internal validity in the simple interrupted time series design is A. history. B. maturation. C. selection. D. additive effect of selection and history.

Q: A study using a simple interrupted time-series design examined people's responses to the September 11, 2001 terrorist attacks using a questionnaire measure of religion and faith. Archival data were available for the population four months before and after the attacks. The researchers observed that people's religious feelings and faith increased in the months following the attacks. One threat to internal validity that should be considered in this study is A. the September 11, 2001 attacks. B. regression toward the mean level of religious feeling. C. that religious feeling increased, but other measures such as honesty and curiosity did not. D. seasonal variation in people's feelings of religion and faith.