A New View of Statistics	© 1997 Will G Hopkins
Go to: Previous · Contents · Search · Home

QUIZ

1. A frequency distribution can be shown as

• a statistic
• a histogram
• a scatter plot
• a stem and leaf plot
  answer · Basics

2. Simple statistics are

• for simpletons
• presented in stem and leaf plots
• things like correlations
• things like standard deviations
  answer · Simple Statistics

3. What would you do with a median?

• Use it do show spread.
• Use it for normally distributed data.
• Cross it against oncoming traffic
• Indicate the middle of some data.
  answer · Simple Statistics

4. The following are measures of spread:

• standard deviation
• root mean square errors
• percentile ranges
• polyunsaturated margarine
  answer · the Spread

5. Which arrow indicates the standard error of the estimate?

• A
• B
• C
• D
  answer · SEE

6. A relative risk is

• a risk of matrimony
• an outcome statistic
• a relative of the odds ratio
• a relative frequency
  answer · Relative Frequency

7. Differences between means are best thought about in terms of

• p values
• standard errors of the mean
• percentages of the mean
• standard deviations
  answer · Effect Size · Mean ± SD or Mean ± SEM?

8. Dimension reduction

• describes loss of precision.
• describes factor analysis.
• is an example of ANOVA.
• is a weight-loss program.
  answer · Dimension Reduction

9. Concerning reliability:

• It impacts most on descriptive studies.
• It can be expressed as an ICC.
• It can be expressed as a CV.
• It is quantified by 2-way ANOVA.
  answer · Reliability

10. Concerning validity:

• It impacts most on descriptive studies.
• It is the correlation between true and observed values.
• A valid measure must be reliable.
• A reliable measure must be valid.
  answer · Validity

11. Correct or incorrect expressions?

• height = 175 ± 6 cm
• VO2peak = 67 ± 5.1 ml/min/kg
• ICC = 0.87
• CV = 1.4%
  answer · How Many Digits

12. Confidence intervals...

• are a new form of sprint training.
• are calculated routinely by most stats packages.
• define the likely range of a population value.
• are inferior to p values for indicating magnitude of outcomes.
  answer · Confidence Intervals · What is a P Value?

13. A correlation coefficient and its confidence interval are shown in the figure.

We can conclude that:

• The correlation is significant.
• The true value of the correlation is likely to be 0.45.
• More subjects should be tested.
• A type II error has occurred.
  answer · More on the Lower and Upper Limits · Type II Errors

14. Concerning tests and test statistics:

• One-tailed tests are sometimes justified.
• Test statistics should always be shown.
• Chi-squared is a common test statistic.
• P = 0.06 means there is no effect.
  answer · What is a P Value? · Using P Values

15. Many samples, each of 100 observations, are drawn from a population in which there is a correlation of 0.70 between two variables. How often would you expect to find a statistically significant correlation?

• hardly ever
• about one time in 100
• about one time in 20
• almost always.
  answer · Type I Errors

16. What are appropriate comments about these data, which show mean weekly training durations for three groups of athletes? (Bars are SDs.)

• Differences between all groups are substantial. (See Effect Size.)
• The data should be analyzed by repeated-measures ANOVA.
• Log transformation appears to be necessary before analysis.
• Runners are lazier than cyclists.
  answer · One-Way ANOVA

17. An outcome measured on a five-point scale (not at all to always)...

• is an example of an ordinal variable.
• has a behavior problem when it comes to residuals.
• should be analyzed by logistic regression.
• can be analyzed by ANOVA.
  answer ·Ordinal Dependent Variables

18. Log transform a variable...

• if the values are too big.
• if the residuals (error) get bigger for bigger values of the variable.
• if you don't get statistical significance.
• if non-parametric tests are inappropriate.
  answer · Log Transformation · Non-Parametric Models

19. Non-parametric tests usually...

• are parametric tests in disguise..
• involve rank transformation of the dependent variable.
• work for grossly non-normal data..
• should be attempted if parametric tests give p > 0.05.
  answer · Non-Parametric Models

20. If we studied the effect of gender and body mass on sprint performance time, we would use the following model:

• unpaired t test
• ANCOVA
• ANOVA
• MANOVA
  answer

21. Concerning multiple linear regression:

• Use it to fit curves as well as straight lines.
• Use it to control for the effect of numeric variables.
• It gives misleading results for highly correlated independent variables.
• Use it to fit multiple straight lines with several groups.
  answer · Multiple Linear Regression

22. Repeated-measures models...

• are used in descriptive studies.
• can be analyzed by modeling variances.
• are used when you have to repeat a failed test.
• are straightforward to analyze with stats programs.
  answer ·  Repeated-Measure ANOVA · RM·ANOVA: Three or more trials and no between·subjects effect

23. In a longitudinal study aimed at enhancing sport enjoyment, the following results were obtained

We can conclude that:

• Initial randomization to the two groups was poor.
• There is one between- and one within-subject factor.
• The time effect in the model is substantial.
• The time effect in the model is significant.
  answer · Two trials plus...

24. Concerning sample sizes for a controlled longitudinal study:

• Sample size is proportional to (1 - r), where r = reliability correlation.
• Controlled studies need 4x as many subjects as uncontrolled studies.
• Get sample size "on the fly" by testing until you get an acceptable confidence interval.
• None of the above.
  answer · What Determines Sample Size · Sample Size "On the Fly"

25. The size of a sample needed for a cross-sectional study...

• depends on the size of your research grant.
• is inversely proportional to the square of the validities of your measures.
• is a function of the largest effect you want to detect.
• depends on how many student researchers you have on the project.
  answer · The Right Number of Subjects · What Determines Sample Size

26. When you come home from climbing in the statistical mountains, you will tell the folks, amongst other things, that...

• from now on you will show as few numbers as possible.
• statistical modeling is no substitute for knowing your data.
• it's important to play with stats programs.
• from now on you will test rather that estimate.
  answer · Summary