Test 90    (statistic) 1. The following are true about confidence intervals: a. the intervals are larger with smaller sample size b. they indicate the presence or otherwise of a statistical      difference between two groups c. a 95% confidence interval means that 95% of all      observed values fall within that interval. d. in an odds ratio, if the 95% confidence interval includes      unity then no significant difference may apply. e. the intervals give a range of values within which the true      value will lie. Answera.T   b.F  c.T  d.T  e.T Confidence intervals describe the range of values  around a mean, an odds ratio, a p value or a standard deviation within which the true value lies.  A confidence interval which is narrow emphasizes  the significance of the result, but is the p value which  describes significance, not the confidence interval around it. 2. True statements about non-parametric tests include: a. they can be sued on small samples b. they can be used to analyse samples that are normally      distributed c. Student's paired t-test is a non-parametric test d. they can be applied to ordinal data e. they can not be used if the nature of the distribution of      the data is unknown.   Answera.T   b.T   c.F  d.T  e.F Non-parametric tests are used for data that is not  normally distributed and data where the distribution is unknown. It can be applied to normally distributed data as well.  Student's paired t-test is a parametric test applied to  two groups and is more powerful than the standard t-test if the same subjects are examined before and after treatment. 3. In statistics: a. null hypothesis describes the probability that a      relationship exists between two samples.  b. analytical statistics are the same as inferential statistics.  c. descriptive statistics produce mean, median and mode     from data. d. the mode is the measurement which lies exactly      between each end of a range of values ranked in order. e. skewed data invalidates further statistical analysis. Answera.F   b.T  c.T   d.F  e.F The null hypothesis sets out the assertion that a relationship between two samples does not exist;  analytical statistical methods set about disproving the null hypothesis ie. demonstrating a relationship. The mode is the most commonly observed value; the median is the measurement which lies exactly  between each end of a range of values ranked in order. . 4. Correlation coefficient: a. is denoted by the symbol "r"  b. is measured on a scale of 0 to 1. c. a positive value implies that a rise in one variable      accompanies a rise in the other. d. describes the degree of association between two      variables e. describes the degree of agreement between two      variables. Answera.T   b.F  c.T  d.T  e.F The values varies from -1 to +1 and is normally represented by the gradient of a best fit line applied to a scatter plot of the two variables in question. Complete correlation of the two variables is expressed by 1. The positive or negative sign denote the relationship (increasing or decreasing) between the two variables.  A degree of correlation is always present between linked variables. 5. Non-parametric tests include a. ANOVA b. Student's t-test c. Chi-squared test d. Wilcoxon signed rank test e. Mann-Whitney U test. Answera.F   b.F  c.T  d.T  e.T Non-parametric tests are applied to data that is not normally distributed, or to data where the distribution is not known.  The commonly used non-parametric tests are the  Mann-Whitney test for ordinal data and Chi-squared test  for nominal data.  The Wilcoxon signed-rank test is a non-parametric alternative to the paired Student's t-test for the case of two related samples or repeated measurements on a single sample. The test is named for Frank Wilcoxon (1892–1965) who proposed this, and the rank-sum test for two independent samples (Wilcoxon, 1945).