Professional Documents
Culture Documents
Final Notes
Final Notes
10. In an intervention study, for testing whether the mean change (or difference score) equals zero, the sample mean of the difference score is divided by the standard error of
the mean of the difference score. What is this statistic called? t statistic
11. To obtain the chi-square statistic, differences of observed and expected cell counts are added over all cells of the contingency table. FALSE
12. The T statistic for the Wilcoxon rank sum test is the sum of ranks of the smaller group. TRUE
13. Wilcoxon signed rank test for paired data compares the distribution of ranks in two groups. How are these groups created? Positive differences form one group and
negative differences form another group
14. What is the main idea behind the F test of one-way analysis of variance that allows it to extend beyond a comparison of two groups to a simultaneous comparison of all group
means? Comparison of the variance arising from group means to the variance within the groups.
15. The ratio of between-group mean squares to the within-group mean squares is identified as being the: F statistic (is designed to analyze variation between and within groups,
without having to restrict the group variable to have only two levels. When the proper degrees of freedom are considered, we can compute the mean squares between groups and
mean squares within groups. The ratio of these quantities is the F statistic used in ANOVA)1
16. Select all applicable statements about the degrees of freedom associated with the F statistic resulting from a study that has 4 groups, each of size 25.. Total df (numerator
+denominator) = 99, Numerator df = 3, Denominator df = 96 (The numerator has df that are one less than the number of groups, i.e. 4-1, or 3.The denominator has df
equal to the total sample size minus the number of groups, i.e. (25 x 4) - 4, or 100 -4, or 96)
17. The F test of one-way ANOVA is reliable only if variances are equal in the groups. TRUE
18. What is the advantage of conducting post-hoc tests, such as Tukey's HSD, following analysis of variance rather than doing multiple t tests to compare each pair of means in 3
or more groups? Post-hoc tests control the overall Type I error rate, whereas multiple t tests lead to increase of Type I errors, or false positive results.
19. Match correct test:
t test for independent samples survery to measure mothers attitudes toward labor given to 2 groups, group 1 attend classes, group 2 did not. Test hypothesis at alpha level 0.05.
t test for
paried comparison if height varies during course of 10 days, randomly selected volunteers ht measured every morning and evening. 2 nd example: Researchers would like to know whether salivary cortisol
levels vary during the day in a group of 27 patients who are receiving stress management therapy.
Chi-square test large randomized controlled trial to test improvements in learning outcomes due to new nurse education program compared to existing program.
Wilcoxon rank sum test durations to pregnancy were measured after couple fertility treatment. Test whether mean duration is different between two genotypes of mothers (same size is small)
20. The following output is from a comparison of the mean weight (kg) for smokers and non-smokers using the t test for independent samples with an alpha level = .05. Based on
the Test for Equality of Variances, which t test results are appropriate to use? equal variances assumed What is the p value for the appropriate t test for Equality of Means?
0.207 What is the statistical decision based on the results of the appropriate t test for Equality of Means? Do not reject the null hypothesis that (Smokers) = (NonSmokers) Which value that is included in the 95% confidence interval confirms the statistical decision? 0 Can we conclude that the population mean weights for smokers and
non-smokers are different? - NO
21. The following output is from a comparison of the mean Memory scores for patients with three different types of dementia. Statistical analysis was conducted using a one-way
analysis of variance (ANOVA) with an alpha level = .05.
The F statistic (4.333) is calculated by which ratio? Between Groups Mean Square / Within Groups Mean
Square
What is the statistical decision based on the results of the one-way ANOVA test?
Reject the null hypothesis that the population means for all
3 types are equal For which comparison can we conclude that the population mean Memory scores are different? Type 2 vs. Type 3
22. The table below shows the results of a double blinded RCT examining whether patients with chronic fatigue syndrome (CFS) improved 6 weeks after treatment with
intramuscular magnesium. The group who received the magnesium were compared to a group who received a placebo. The outcome was patient assessment of feeling better.
What is null hypothesis The proportion of those who felt better is identical in treatment and placebo groups
dependent variable frequencies of pts who felt better and who did
not feel better
Independent variable treatment group (magnesium, placebo)
most appropriate statistical test for this analysis? Chi-square test of proportions
23. The table shows the first two cases and summary of data on 24 hour total energy expenditure (Cal/day) of groups of lean and obese women. The aim of this study was to
compare total energy expenditure between the lean and obese women.
What is the null hypothesis? Mean daily enerergy expenditure is equal in lean and obese women
dependent variable? Energy expenditure
independent variable? Weight groups(lean/obese)
most appropriate statistical test for this analysis? Independent
sample t test
24. Nonparametric alternatives to parametric hypothesis tests generally convert continuous data values to: rank values
25.. Students in the TEL group performed better than the controls, both in the final cumulative test and in a graded examination. TEL participants experienced better final
cumulative test results than students not tested (M TEL = 23.11, M Re-study = 20.47, t = -2.57, df = 109.86, p < 0.05). The authors compared the means of the final cumulative
test scores in two groups and found a significantly higher mean score in the TEL group. The value of the t statistic was -2.57 and the degrees-of-freedom were 109.86. The
difference of group means 23.11 - 20.47 = 2.64. The significance of the result implies that the 95% confidence interval of the difference of group means does not include the
value 0. If the authors had found that test scores were not normally distributed, they could have compared the mean ranks of test scores using the Wilcoxon rank sum test or
Mann-Whitney U test.
1. The Pearson correlation coefficient r is a number that lies between (ends included): -1 and 1
2. If the correlation between two variables is estimated to be zero, and assumptions of correlation are satisfied, it suggests that: increase in one variable has no impact on the other variable
3. The correlation between two variables was observed to be -0.7. According to Cohen's guidelines, this is large negative correlation.
4. A positive correlation between variables FEV1 (forced expiratory volume in 1 second, a measure of lung capacity) and height in children provides evidence that: height and
FEV1 are associated; increase in height tends to be accompanied by increase in FEV1 ( Correlation should not be confused with causation. If variables xand y are
correlated, it could be that x is a cause of y or that y is a cause of x or that some other factor z is a cause of both x and y.The positive value of the correlation indicates that an
increase in one variable tends to be associated with an increase in the other variable.)
5. Which of the following assumptions must be satisfied for the proper interpretation of the Pearson correlation coefficient?
variables are normally distributed
variables must be linearly related
Observations are independent between cases.
6. In linear regression the aim is to find the best-fit line by the method of least squares and predict one variable given the value of another. (Note that analysis of variance or t-tests are used for probing
differences between groups. It is not possible to enforce a linear relationship between variables using linear regression if such a relationship does not naturally exist.)
7. Simple linear regression involves one continuous DV and one continuous IV. (Although categorical variables may be added to a general (advanced) regression model, simple
linear regression does not have any categorical variables.)
8. The equation of a straight line is given by: y = b0 + b1 x
9. The intercept of the line, y = b0 + b1 x, represents: the point at which the line intersects the y-axis
10. The relationship between correlation and simple linear regression is that the Pearson correlation coefficient is equal to the: standardized slope (beta). (The Pearson correlation
coefficient equals the slope when the variables are standardized, y = x)
11. Which of the following statements provides a correct interpretation of the summary measure of goodness of fit provided by R2 ? R2 is the proportion of variance of the DV
that is explained by the linear model.
12. Residuals are the differences between observed and estimated valuesof the dependent variable (DV). (Residuals are the error terms or differences between observed values
and estimated values of the DV that lie on a straight line)
13. Which of the following are assumptions of linear regression? The DV and IV have a linear relationship.
Residuals satisfy homoscedasticity, i.e. variance of
residuals is the same for any predicted value.
Residuals are normally distributed.
Observations are independent.
14. When investigating a sample of 556 men working in a cadmium pigment plant, it was found that the cadmium in a man's urine was linearly related to the number of years worked in the plant. The estimated slope of the
line obtained by regressing the amount of cadmium against years worked was 0.9 g/l per year, with 95% confidence interval extending from 0.6 g/l to 1.2 g/l. The intercept was not significantly different from zero. A man
who had worked in the plant for 10 years would, on average, have increased the amount of cadmium in his urine by 9 g/l. For 10 years of work urine cadmium level = 0.9 x 10 = 9 g/l)
15. For a hypothetical data set, a linear regression model was used to predict number of visits to health care providers from health care funding levels in 50 cities. (See tables and
plot below.) Which of the following conclusions is FALSE? (graphs from SSPS to review) The t test for the intercept is significant (p < 0.05), therefore the null hypothesis
may be rejected. The intercept is not equal to zero.
16. In an analysis of the DCCT glucose profile data (Diabetes Care 25:275-278, 2002), mean HbA1c and average glucose (AG) were calculated for each study subject (n= 1439),
and analyzed with a simple linear regression model. Results showed that model R2 was 0.67 for a linear relationship between HbA1c and AG (mg/dL) that was estimated by the
following regression equation: AG = 35.6 x HbA1c - 77.3
Dependent variable in regression model AG Independent variable in regression model- HbA1c For each unit increase in HbA1c the value of AG increased by 35.6mg/DL Linear regression model explain % of AG
variance 67%
17. evaluate the effect of depression on sleep score in nursing home residents. Ninety residents of a nursing home facility were divided into three groups based on their level of
depression: mild, moderate, or severe. The residents were then monitored overnight to obtain their sleep score. Test the hypothesis at the alpha level of .05 that the population
mean sleep scores for the three groups are all equal. Use one way ANOVA to test hypothesis, DV= sleep score IV = level of depression group, if hypothesis test
was p .05 but difference mean sleep score between 3 groups were to small to be clinical importance = statistically significant but no clinically significant
18. Test the hypothesis at the alpha level of .05 that there is no association between spousal cholesterol levels. method for the study hypothesis. simple linear regression or
correlate
19. Intervention program will reduce pressure ulcers in residents of nursing facilities. A sample of 200 nursing facility residents without pressure ulcers were randomized to either
the intervention program or to a control (standard care) program., the researchers determined if the residents had acquired a pressure ulcer (results shown below). Test hypothesis
at alpha level of .05 no association between the intervention program and pressure ulcers in the population. Chi-square test, IV = program (intervention or control)
20. A study was conducted to compare the cognitive scores of a group of adult patients before and after an intervention. A questionnaire designed to assess cognitive score was
administered to 15 adult patients in a treatment facility before and after the intervention. Although the cognitive score was measured on a continuous scale, the change in score
data from this group of patients did not follow a normal distribution. Test the hypothesis at the alpha level of .05 that the population median change in cognitive score after
the intervention = 0.
Wilcoxon signed rank test, DV = change in cognitive score