Research Methods Final Part 1
incidence
A study was done to determine how many first-year college students develop back pain during their first year in college. Only students who never had back pain prior to starting their first year of college were included in the sample. Students were followed during their first 12 months of college to identify development of back pain. Of the 200 student participants, 10 developed back pain during the first 12 months of college resulting in a rate of 10/200/year or 5% per year. What type of measurement of back pain is this? Group of answer choices Incidence Prevalence
chi-square
A study was done to determine if hearing loss (yes, no) is associated with age group (<50, 50-70, >70 years old) among older adults. What is an appropriate statistical test for this analysis? Group of answer choices Chi-square ANOVA Linear regression
true
A study where the results can be generalized to a population has a high degree of external validity. Group of answer choices True False
true
A study with a very high level of internal validity will very often have limited or narrow generalizability (external validity). Group of answer choices True False
no
Based on the above results, are college athletes at significantly higher risk of ankle ligament trauma as compared to high school athletes? Yes No
linear regression
Identify if the statement most often applies to Linear Regression or (binary) Logistic Regression: Results report the beta coefficient Group of answer choices Linear regression Logistic regression
intention to treat analysis
In a randomized controlled trial, analyzing the data using all subjects randomized, whether or not they finished the intervention, is termed: Group of answer choices Intention to treat analysis Per protocol analysis
paired t-test
Of the statistical tests listed below, which is most appropriate for statistical analysis in the following scenario? A researcher will enroll 100 research participants in an intervention composed of daily guided meditation for 2 months. Stress will be measured on a continuous scale, ranging from 0-100, before (PRE) and after (POST) the 2 month intervention. The researcher will compare PRE vs POST stress values in the sample of 100 participants. Assume stress is a normally distributed variable. Chi-square Independent t-test Pearson correlation Paired t-test
paired t-test
Of the statistical tests listed below, which one is the most appropriate for data analysis in this scenario: A study where the intervention was daily guided meditation among 50 adult participants 2 months. Stress (0-100) was measured before (PRE) and after (POST) the 2-month intervention. Data were analyzed to compare PRE vs POST values within the 50 subjects. Assume stress level is a continuous, normally distributed variable. Group of answer choices Chi-square analysis Linear regression Independent t-test Paired t-test Pearson correlation analysis
ANOVA
Of the statistical tests listed below, which one is the most appropriate for data analysis in this scenario? A study randomized 80 adult participants, 20-60 years old, to an exercise intervention or meditation intervention for two months. The investigators compared the average change in stress (0-100) between the two groups, BUT they hypothesized that there might be an interaction between life stage (young adult, middle age) and the intervention so they included life stage in the analysis. Assume stress level is a continuous, normally distributed variable. Group of answer choices Chi-square analysis Binary logistic regression ANOVA Independent t-test Paired t-test
all of the answers can be used to reduce threats to internal validity
Researchers can reduce threats to internal validity by using _____________. Group of answer choices appropriate instruments and calibrating equipment all of the answers can be used to reduce threats to internal validity a single blind or double-blind study design a control group
randomized controlled trial
Review the PEDro scoring scale. Which study design will usually score a higher PEDro score? Group of answer choices Randomized Controlled Trial Quasi-experimental study Observational study Qualitative study
reliability
The repeatability of measurements is related to: Group of answer choices Validity Reliability
the percent loss for each group was 2/12 = 17%
The research study reviewed in Module 8, Lecture 2 (Chuang, 2006) compared outcomes between cardiac rehab patients who performed treadmill training with virtual reality (VR group) to those who performed treadmill training without virtual reality (non-VR group). 12 subjects were randomized to each group. Both groups received cardiac rehab that consisted of 30 min treadmill training 2x/week for 3 months. Randomization was performed by having each subject draw a ball from a box (Ball A=non-VR, and Ball B=VR). Investigators were not involved in this randomization. During the study, 4 subjects were lost to follow up, resulting in 20 subjects completing the study. Refer to the Tables and Figures in Chuang to answer the question: What was the percent loss to follow up for each group? Group of answer choices The percent loss for each group was 10/12 = 83% The percent loss for each group was 1/12 = 8% The percent loss for each group was 4/12 = 33% The percent loss for each group was 2/12 = 17%.
aquatic versus land-based physical therapy
What are the two treatment groups being compared in Figure 3? Group of answer choices Swimmers versus soccer players Aquatic versus no physical therapy Aquatic versus Land-based Physical Therapy Older patients versus younger patients
is used when you have two continuous, normally distributed variables
Which answer is true regarding Pearson correlation (r) ? Infers "causation" (i.e. cause-effect relationship) Ranges from 0 (no correlation) to 1 (strong correlation) Is used when you have two continuous, normally distributed variables Relates strength but not direction
is valid when you want to examine the relation between two normally distributed continuous variables
Which answer is true regarding the Pearson correlation coefficient (r)? Group of answer choices Ranges from 0 (no correlation) to 1 (strong correlation) Is valid to use when you want to examine the relation between two normally distributed continuous variables Relates strength but not direction Infers causation (i.e. cause-effect relationship)
chi-square analysis
Which of the following statistical tests is most appropriate for examining the relationship between X and Y if both are dichotomous variables? Group of answer choices Pearson correlation Independent t-test Chi-square analysis ANOVA
pearson correlation
Which of the following statistical tests is most appropriate for examining the relationship between X and Y, assuming both X and Y are normally distributed continuous variables? Group of answer choices Chi-square analysis Spearman correlation Independent t-test Pearson correlation
The weight loss difference between the treatment and control groups, although statistically significant, may not be clinically meaningful.
A research study investigating subject satisfaction with weight loss programs suggested a benchmark loss of at least seven pounds is needed for a person to be "satisfied" with their weight loss. Assuming, in the study described in Question 10, weight loss in the experimental group was 3 pounds and weight loss in the control group was zero pounds (p<0.09), which statement below is the most appropriate interpretation of the study results? Group of answer choices The weight loss difference between the treatment and control groups is statistically significant and therefore clinically meaningful. The weight loss difference between the treatment and control groups is neither statistically significant nor clinically meaningful. The weight loss difference between the treatment and control groups, although statistically significant, may not be clinically meaningful.
face validity
A research team created a questionnaire to evaluate physical stress at work. The team read over the questions in the final draft of the questionnaire and decided that the questions generally make sense. What type of validity does this statement represent? Group of answer choices Face validity Concurrent criterion validity Predictive criterion validity Construct validity
it would increase
A researcher is planning a pilot study to determine the effect of a motivational interviewing approach on weight loss. The hypothesis is that the treatment group (motivational interviewing) will decrease their weight more than a control group (handout regarding healthy eating). Based on a priori specification of alpha = 0.10 and beta of 0.20 (i.e., power of 0.80), the sample size calculation indicated that 10 subjects were needed in each group (i.e., 20 subjects total). The researcher randomly assigned 20 people to each group, measured each person's weight when they enrolled in the study, and measured each person's weight again at the end of the study. The average weight loss in the treatment group was significantly greater than the average weight loss in the control group (p=0.09). If the researcher wanted to repeat the study, but with alpha = 0.05, and power of 0.95 (beta=0.05) in a second randomized trial testing the same hypothesis, how would this change the sample size calculation? Group of answer choices It would stay the same It would decrease It would increase
intention to treat analysis
Analyzing the data using all subjects enrolled irrespective of whether they complete the study according to the protocol is: Group of answer choices Intention to treat analysis Per protocol analysis
per protocol analysis
Analyzing the data using only subjects who complete the study according to the protocol is: Group of answer choices Intention to treat analysis Per protocol analysis
yes
Based on the above results, are female college basketball players at significantly higher risk of ankle ligament trauma as compared to female college lacrosse players? Yes No
6
How many individual studies are included in the meta-analysis shown in Figure 3? Group of answer choices 7 6 5 4
Former smokers, as compared to never smoked, have significantly elevated risk of experiencing a myocardial infarction
Myocardial infarction rate ratio by smoking status Smoking Status Myocardial Infarction Rate Ratio (95% CI) Never Smoked 1.0 (reference group) Former Smoker 1.6 (1.1-2.4) Current Smoker 2.4 (1.7-3.4) Based on this results table, which of the statements below is an accurate conclusion regarding former smokers? Group of answer choices Former smokers, as compared to current smokers, have a significantly elevated risk of experiencing a myocardial infarction. Former smokers, as compared to never smoked, have significantly elevated risk of experiencing a myocardial infarction. Former smokers, as compared to never smoked, do not have a significantly elevated risk
predictive criterion validity
The Berg Balance Scale of <45 predicts a future fall among older adults. What type of validity does this statement represent? Group of answer choices Construct validity Concurrent criterion validity Face validity Predictive criterion validity
content validity
A research team developed a new questionnaire to evaluate spasticity in persons with stroke. The team sent the questionnaire to experts in the field for their input. The experts provided systematic input on the representativeness of each construct in the questionnaire. The researchers then revised the questionnaire based on their input. What type of validity does this process represent? Group of answer choices Predictive criterion validity Content validity Face validity Concurrent criterion validity
it would increase
A researcher is planning a study to determine the effect of a motivational interviewing approach on weight loss. The hypothesis is: motivational interviewing will decrease weight more than a handout regarding healthy eating (control group). Based on a priori specification of alpha = 0.05 and beta =0.10 (i.e., power of 0.90), the sample size calculation indicated that 100 subjects were needed in each group (i.e., 200 subjects total). The researcher randomly assigned 100 people to the treatment group and 100 people to the control group, measured each person's weight when they enrolled in the study, and measured each person's weight again at the end of the study. The average weight loss for participants in the treatment group was 5 pounds, and the average weight loss in the control group was 3 pounds (p=0.06). If the researcher wanted power of 0.95 (beta=0.05), what would happen to the required sample size? Group of answer choices It would stay the same It would decrease It would increase
discussion
A study was done to compare the effects of 8 weeks of meditation versus 8 weeks aerobic exercise on reducing stress in college students. 20 students were randomized to either 1) meditation or 2) exercise. The null hypothesis was: there will be no difference in stress between the groups after 8 weeks of treatment. The published article didn't have a sample size calculation in the Methods section. The results showed no difference in stress between the groups after 8 weeks of treatment. The article did not include a sample size calculation in the Methods article. Which section of the article would you most likely find the authors mention the possibility of the study being limited by a small sample size? Group of answer choices Discussion Results Background/Introduction Methods
You are suspicious of a Type II error (accepting the null hypothesis falsely)
A study was done to compare the effects of 8 weeks of meditation versus 8 weeks of aerobic exercise on reducing stress in college students. 20 students were randomized to either 1) meditation or 2) exercise. The null hypothesis was: there will be no difference in stress between the groups after 8 weeks of treatment. The published article didn't have a sample size calculation in the Methods section. The results showed no difference in stress between the groups after 8 weeks of treatment. As a critical reviewer of the scientific literature, what are your thoughts regarding this study? Group of answer choices You believe the results are valid because the study was a randomized control trial and that is the strongest study design for showing a "cause -effect" relationship. You are suspicious of a Type II error (accepting the null hypothesis falsely).
you are suspicious of not enough power due to small sample size
A study was done to compare the effects of 8 weeks of meditation versus 8 weeks of aerobic exercise on stress in college students. 20 students were randomized to either:meditation, Exercise The null hypothesis was: There will be no difference in stress between the groups after 8 weeks of treatment. The published article didn't have a sample size calculation in the Methods section. The results showed no difference in stress between the groups after 8 weeks of treatment. What is your conclusion regarding this study? Group of answer choices You are suspicious of a Type I error (rejecting the null hypothesis) You are suspicious of not enough power due to small sample size You believe the conclusion to be valid because it was a randomized control trial and that is the strongest study design to show "cause -effect" relationship
prevalence
Back pain is common. A study was conducted to determine how much back pain is experienced among first-year college students. Students were asked: 'Do you currently have back pain?' Of the 100 students answering the question, 20 reported currently having back pain (5%). What type of measurement of back pain is this? Group of answer choices Incidence Prevalence
no
Based on the above results, are male college soccer players at significantly higher risk of ankle ligament trauma as compared to male college lacrosse players? Yes No
yes
Based on the results table above, do current smokers have a significantly higher rate of myocardial infarction than never-smokers? No Yes
no
Based on the results table above, do former smokers have a significantly higher rate of myocardial infarction than never-smokers? Yes No
is typically used for randomized controlled trials
Choose the best answer below regarding the PEDro scale. Group of answer choices Is typically used for randomized control trials Total score ranges from 0-10 A score of 0 suggests there are no problems with the study and it appraises well
You are slightly concerned about the percent loss to follow up, because it is a little over 15%.
Chuang et al (2006) compared outcomes between cardiac rehab patients who performed treadmill training with virtual reality (VR group) versus those who performed treadmill training without virtual reality (non-VR group). 24 subjects were randomized to one of two groups: 1) VR group (n=12), or 2) non-VR group (n=12). Study design was a randomized control trial. All subjects received cardiac rehab (including 30 min of treadmill training) 2x/week for 3 months. During the study, 4 subjects were lost to follow up, resulting in 20 subjects completing the study. Each group lost 2 subjects. Choose the best answer below. Group of answer choices You are slightly concerned about the percent loss to follow up, because it is a little over 15%. You conclude that differential loss to follow up is a major limitation of this study.
Although randomization cannot guarantee that baseline characteristics will be similarly distributed between groups, it does increase the probability of similar distribution of baseline characteristics
Chuang et al (2006) compared outcomes between cardiac rehab patients who performed treadmill training with virtual reality (VR group) versus those who performed treadmill training without virtual reality (non-VR group). 24 subjects were randomized to one of two groups: 1) VR group (n=12), or 2) non-VR group (n=12). Study design was a randomized control trial. All subjects received cardiac rehab (including 30 min of treadmill training) 2x/week for 3 months. There were no significant differences between groups in age, body mass index, V02max (oxygen capacity) or months since cardiac surgery. Choose the best answer below. Group of answer choices Randomization will always ensure that baseline characteristics are similarly distributed between groups. Although randomization cannot guarantee that baseline characteristics will be similarly distributed between groups, it does increase the probability of similar distribution of baseline characteristics.
Even though group assignment was randomized, in this type of study it is not possible to truly mask the participant to the type of rehab that they are receiving
Chuang et al (2006) compared outcomes between cardiac rehab patients who performed treadmill training with virtual reality (VR group) versus those who performed treadmill training without virtual reality (non-VR group). 24 subjects were randomized to one of two groups: 1) VR group (n=12), or 2) non-VR group (n=12). Study design was a randomized control trial. Study investigators provided all subjects with cardiac rehab (including 30 min of treadmill training) 2x/week for 3 months. Randomization was performed by "lottery" where each subject (blindly) picked a ball from a box that originally contained 24 balls (12 of ball A=non-VR group, and 12 of ball B=VR group). Investigators were not involved in this randomization. Choose the best answer below. Group of answer choices Even though group assignment was randomized, in this type of study it is not possible to truly mask the participant to the type of rehab that they are receiving. Subjects were masked (blinded) to group assignment. The research coordinators who performed the treadmill training were masked (blinded) to group assignment.
concurrent criterion validity
Comparing a new measure to the current gold standard is used to establish what type of validity? Group of answer choices Content validity Concurrent criterion validity Construct validity
The hypothesized effect size and variability (e.g. standard deviation) of the outcome are both important factors to consider in determining the study sample size.
Having an adequate sample size in a study is important for being able to statistically test a hypothesis. Choose the best answer regarding sample size. Group of answer choices The sample size justification (e.g. calculation) may be found in the Results section. The hypothesized effect size and variability (e.g. standard deviation) of the outcome are both important factors to consider in determining the study sample size. Having a too small sample size may result in a Type I error.
Sample size needed for a study is dependent on several factors including effect size expected and the variability (e.g., standard deviation) of the outcome
Having an adequate sample size in a study is important to be able to statistically test a hypothesis. Choose the best answer regarding sample size. Group of answer choices Sample size needed for a study is dependent on several factors including effect size expected and the variability (e.g., standard deviation) of the outcome The sample size justification (e.g. calculation) is usually found in the Results section Having a small sample size (not enough subjects) may result in a Type I error
234
How many total research subjects are represented in the meta-analysis in Figure 3? Group of answer choices 234 535 115 119
logistic regression
Identify if the statement most often applies to Linear Regression or (binary) Logistic Regression: Results report the odds ratio (OR) Group of answer choices Logistic regression Linear regression
linear regression
Identify if the statement most often applies to Linear Regression or (binary) Logistic Regression: The beta coefficient can range from -infinity to +infinity Group of answer choices Logistic regression Linear regression
logistic regression
Identify if the statement most often applies to Linear Regression or (binary) Logistic Regression: The confidence interval can range from 0.00 to +infinity Group of answer choices Logistic regression Linear regression
logistic regression
Identify if the statement most often applies to Linear Regression or (binary) Logistic Regression: The result is not significant if the odds ratio confidence interval (CI) includes 1.00 Group of answer choices Linear regression Logistic regression
linear regression
Identify if the statement most often applies to Linear Regression or (binary) Logistic Regression: The standardized beta coefficient is expressed in terms of SD units Group of answer choices Linear regression Logistic regression
linear regression
Identify if the statement most often applies to Linear Regression or (binary) Logistic Regression: Used when the dependent variable is a continuous, normally distributed variable Group of answer choices Logistic regression Linear regression
logistic regression
Identify if the statement most often applies to Linear Regression or (binary) Logistic Regression: Used when the dependent variable is dichotomous Group of answer choices Logistic regression Linear regression
is is double-barreled
Identify the primary problem with the following survey question: "How satisfied are you with the variety of food items available in the cafeteria and with the price of the food items?" Group of answer choices It is overly complex It is overly simple It is loaded It is double-barreled
discussion
In the study presented in the previous question (in question 8), the article did not include a sample size calculation in the Methods section. Which section of the article would you most likely find the authors mention the possibility of the study being limited by a small sample size? Group of answer choices Discussion Results Background/Introduction Methods
r = -1.00
Indicate the correlation coefficient (r) that would most accurately describe the data illustrated in the scatterplot below. Group of answer choices r = + 0.10 r = - 0.10 r = -1.00 r = +1.00
basketball
Inversion ankle trauma is disabling, yet little is known regarding the incidence of first-time ankle sprains among athletes and how it is influenced by factors including gender, level of competition, and sport. In this study, 901 athletes who started the academic year without ever having experienced an ankle sprain were enrolled and followed for the academic year. Of the 901 athletes overall, 43 (4.8%) experienced an ankle ligament sprain during the academic year, but the risk of sprain differed by sport. The following table summarizes the risk of ankle ligament sprain by sport for female athletes. Sport Rate Ratio (95% confidence interval) Lacrosse 1.00 (Reference group) Soccer 1.14 (0.35-3.74) Basketball 2.81 (1.02-7.76) Among female athletes, which sport has significantly greater risk of ankle ligament sprain as compared to Lacrosse? Group of answer choices Both Soccer and Basketball are significantly greater risk of ankle injury as compared to Lacrosse Neither Soccer or Basketball are significantly greater risk of ankle injury as compared to Lacrosse Soccer Basketball
incidence
Inversion ankle trauma is disabling, yet little is known regarding the incidence of first-time ankle sprains among athletes and how it is influenced by factors including gender, level of competition, and sport. In this study, 901 athletes who started the academic year without ever having experienced an ankle sprain were enrolled and followed for the academic year. Of the 901, 43 (4.8%) experienced an ankle ligament sprain during the academic year. What type of measurement of ankle sprain does the 4.8% represent? Group of answer choices Incidence Prevalence
time spent outside and fast food meals
Logistic regression analysis was performed to determine the predictors of poor cardiovascular health (yes, no) among high school students. The independent variables examined included: Time engaged with electronic media (< 1 hour/day versus >= 1 hour /day), time spent outside (< 1 hour/day versus >= 1 hour /day), number of fast food meals/week (<=3 versus >=4). The Odds Ratio and 95% Confidence Interval (OR (95% CI)) was used to determine statistical significance. Unadjusted (crude) logistic regression analyses produced the following OR (95% CI): Electronic media use, 1.5 (1.1, 1.8); Time spent outside, 0.7 (0.5, 0.8); Fast food meals, 2.5 (1.5, 4.3). Then, adjusted logistic regression analysis of the same data produced the following adjusted OR (95% CI): Electronic media use, 1.1 (0.9, 2.1); time spent outside, 0.5 (0.4, 0.7); Fast food meals, 2.6 (1.6, 4.4). Which independent variable(s) were associated with significantly increased odds of poor cardiovascular health in ADJUSTED analysis? Group of answer choices Electronic media use and time spent outside Electronic media use Fast Food meals Time spent outside and fast food meals Time spent outside
all three variables were significant in unadjusted analysis
Logistic regression analysis was performed to determine the predictors of poor cardiovascular health (yes, no) among high school students. The independent variables examined included: Time engaged with electronic media (< 1 hour/day versus >= 1 hour /day), time spent outside (< 1 hour/day versus >= 1 hour /day), number of fast food meals/week (<=3 versus >=4). The Odds Ratio and 95% Confidence Interval (OR (95% CI)) was used to determine statistical significance. Unadjusted (crude) logistic regression analyses produced the following OR (95% CI): Electronic media use, 1.5 (1.1, 1.8); Time spent outside, 0.7 (0.5, 0.8); Fast food meals, 2.5 (1.5, 4.3). Then, adjusted logistic regression analysis of the same data produced the following adjusted OR (95% CI): Electronic media use, 1.1 (0.9, 2.1); time spent outside, 0.5 (0.4, 0.7); Fast food meals, 2.6 (1.6, 4.4). Which independent variable(s) were associated with significantly increased odds of poor cardiovascular health in UNADJUSTED analysis? Group of answer choices Fast food meals and electronic media use Fast food meals Electronic media use Time spent outside All three variables were significant in unadjusted analysis
independent t-test
Of the following statistical tests, which is most appropriate for determining if there is a difference in aerobic capacity (a continuous, normally distributed variable) between college athletes versus intramural sport participants? Group of answer choices Independent t-test Pearson correlation Paired-t-test Chi-square analysis
ANOVA
Of the following statistical tests, which is most appropriate for determining if there is a significant difference in the average daily temperature in San Diego, CA, between Fall, Winter, Spring, and Summer seasons? Assume average daily temperature is a normally distributed variable. Group of answer choices Independent t-test ANOVA Chi-square analysis Pearson correlation
ANOVA
Of the following statistical tests, which one is the most appropriate for examining if the average number of minutes/day spent studying is significantly different between freshman, sophomore, junior, and senior students at UC? Assume minutes spent studying is a normally distributed variable. Group of answer choices Pearson correlation Independent t-test Chi-Square ANOVA
linear regression
Of the statistical tests below, which one is the most appropriate for determining the linear model that predicts College GPA from these possible predictors: high school GPA, major in college, and SAT score. Assume GPA is a continuous, normally distributed variable that ranges from 0.00 to 4.00. Group of answer choices Chi-square analysis Binary logistic regression Independent t-test Linear regression
chi-square analysis
Of the statistical tests listed below, which is most appropriate for examining if there is a significant association in young adults between current smoking status (yes, no) and high school graduation status (high school diploma versus did not graduate from high school)? Pearson correlation Chi-square analysis Logistic regression Paired t-test
logistic regression
Of the statistical tests listed below, which is most appropriate for identifying significant predictors of smoking status among college students (formatted as current versus not current smoker), where the potential predictors are gender, history of smoking before college, smoking status of the parents, number of friends who smoke, works at least 10 hours per week, year in college, major in college. [Hint: ask yourself, is the outcome variable dichotomous or continuous] Logistic regression Pearson correlation ANOVA Linear regression
independent t-test
Of the statistical tests listed below, which one is the most appropriate for data analysis in this scenario? A study randomized 80 adult participants, 20-60 years old, to an exercise intervention or meditation intervention for two months. The investigators compared the average change in stress (0-100) between the two groups. Assume stress level is a continuous, normally distributed variable. Group of answer choices Pearson correlation analysis Independent t-test Paired t-test ANOVA Chi-square analysis
pearson correlation analysis
Of the statistical tests listed below, which one is the most appropriate for determining if there is a significant, linear association between minutes per week of exercise and stress level (0-100) in teens? Assume continuous, normally distributed variables. Group of answer choices Independent t-test ANOVA Paired t-test Pearson correlation analysis Binary logistic regression Chi-square analysis
chi-square analysis
Of the statistical tests listed below, which one is the most appropriate for examining if there is a significant association between participation in any kind of exercise class at least once weekly (yes/no) and ability to place 10 lbs on a shelf at shoulder level (yes/no), among older adults? Group of answer choices ANOVA Linear regression Paired t-test Independent t-test Pearson correlation analysis Chi-square analysis
ANOVA
Of the statistical tests listed below, which one is the most appropriate for examining if there is a significant difference in average stress level (0-100) by year in college (1, 2, 3, 4)? Assume stress level is a continuous, normally distributed variable. Group of answer choices Pearson correlation analysis Independent t-test Paired t-test Chi-square analysis Binary logistic regression ANOVA
binary logistic regression
Of the statistical tests listed below, which one is the most appropriate for this scenario: A study is designed to determine the significant predictors of whether a college student uses nicotine in any format (yes, no). Possible predictors you will analyze include: gender, history of smoking before college, at least 1 parent smokes, has a friend that smokes, works at least 10 hours per week, year in college, and major in college. Group of answer choices Paired t-test Chi-square analysis Binary logistic regression Independent t-test Pearson correlation analysis
large
One element of planning the sample size for a study is deciding on a meaningful effect size. In other words, how big should the difference between groups, or before and after treatment, be in order to be a clinically meaningful difference? When examining differences between means, a standardized effect sizes of >0.70 are considered: Group of answer choices Large Minimal/Small Moderate
false
Standardized effect sizes of .90 are considered small True False
concurrent criterion validity
The ankle anterior drawer test has a good validity when compared to MRI in diagnosing ankle ligament sprain. What type of validity does this statement represent? Group of answer choices Concurrent criterion validity Construct validity Face validity Predictive criterion validity
inter-rater reliability
The correlation between measurements from multiple observers observing the same participant at the same event is used to establish ______________. Group of answer choices Inter-rater reliability Internal consistency reliability Test-retest reliability
larger
The greater the variability (e.g. standard deviation) of the study outcome, the _____ sample size required Group of answer choices Larger Smaller
larger
The greater the variability (e.g. standard deviation) of the study outcome, the _____ sample size required. Group of answer choices Smaller Larger
sample
The individuals selected to participate in a research study are called the__________. Group of answer choices Sample Target population Accessible population
true
The lectures present two types of regression analysis - linear regression and logistic regression. Linear regression analysis is used when the outcome is a continuous, normally distributed variable True False
true
The lectures present two types of regression analysis - linear regression and logistic regression. Logistic regression is used when the dependent variable is dichotomous True False
true
The lectures present two types of regression analysis - linear regression and logistic regression. Logistic regression results are considered statistically significant if the odds ratio confidence interval (CI) does not include 1.00 True False
construct validity
The number of push-ups a person can do reflects the operational definition of strength. What type of validity does this statement represent? Group of answer choices Construct validity Predictive criterion validity Concurrent criterion validity Face validity
10
The research study reviewed in Module 8, Lecture 2 (Chuang, 2006) compared outcomes between cardiac rehab patients who performed treadmill training with virtual reality (VR group) to those who performed treadmill training without virtual reality (non-VR group). 12 subjects were randomized to each group. Both groups received cardiac rehab that consisted of 30 min treadmill training 2x/week for 3 months. Randomization was performed by having each subject draw a ball from a box (Ball A=non-VR, and Ball B=VR). Investigators were not involved in this randomization. During the study, 4 subjects were lost to follow up, resulting in 20 subjects completing the study. Refer to the Tables and Figures in Chuang to answer the question, How many subjects in the VR group completed the study? Group of answer choices 10 12 20 24
This is a slight limitation of the study. Lecture 1 discussed loss greater than 15-20% may be a concern because it may cause bias in the results obtained.
The research study reviewed in Module 8, Lecture 2 (Chuang, 2006) compared outcomes between cardiac rehab patients who performed treadmill training with virtual reality (VR group) to those who performed treadmill training without virtual reality (non-VR group). 12 subjects were randomized to each group. Both groups received cardiac rehab that consisted of 30 min treadmill training 2x/week for 3 months. Randomization was performed by having each subject draw a ball from a box (Ball A=non-VR, and Ball B=VR). Investigators were not involved in this randomization. During the study, 4 subjects were lost to follow up, resulting in 20 subjects completing the study. Refer to the Tables and Figures in Chuang to answer the question: Is the percent loss to follow up a possible limitation in drawing conclusions from the study results? Group of answer choices This is not at all a concern, because the loss to follow up was less than 15%. This is a slight limitation of the study. Lecture 1 discussed loss greater than 15-20% may be a concern because it may cause bias in the results obtained. This is a major concern, because the loss to follow up exceeded 20% of participants.
there is not evidence of differential loss, each group lost 2 subjects
The research study reviewed in Module 8, Lecture 2 (Chuang, 2006) compared outcomes between cardiac rehab patients who performed treadmill training with virtual reality (VR group) to those who performed treadmill training without virtual reality (non-VR group). 12 subjects were randomized to each group. Both groups received cardiac rehab that consisted of 30 min treadmill training 2x/week for 3 months. Randomization was performed by having each subject draw a ball from a box (Ball A=non-VR, and Ball B=VR). Investigators were not involved in this randomization. During the study, 4 subjects were lost to follow up, resulting in 20 subjects completing the study. Refer to the Tables and Figures in Chuang to answer the question: Is there evidence of differential loss to follow up? Group of answer choices There is not evidence of differential loss. Each group lost 2 subjects. There is evidence of differential loss to follow up. One group lost more participants than the other.
correlation
The statistic test (type of analysis) you would use to examine the association between two continuous variables is... Anova Chi-square T-test correlation
ANOVA
The statistical test (type of analysis) you would use to analyze data with an independent variable that is categorical with three categories, and a dependent variable that is continuous and normally distributed is... Correlation Chi-square ANOVA T-test
chi-square
The statistical test (type of analysis) you would use to analyze the association between two dichotomous variables is... Correlation T-test Chi-square ANOVA
internal consistency reliability
The type of reliability assessed using the statistic Cronbach's alpha is: Group of answer choices Test-retest reliability Internal consistency reliability Construct reliability Interrater reliability
intention to treat analysis
Using imputation to extrapolate missing data for persons who were lost to follow up is part of: Group of answer choices Intention to treat analysis Per protocol analysis
ANOVA
What is the best statistical test to use to answer the following research question: "does body weight differ between freshman, sophomore, junior, and senior students at UC"? Correlation Chi-square T-test ANOVA
correlation
What is the best statistical test to use to answer the following research question: "is body weight significantly associated with minutes of exercise per day among UC students?" under the assumption that data is normally distributed ANOVA Chi-square T-test Correlation
activities of daily living score
What is the outcome measure analyzed in Figure 3? Group of answer choices Quality of Life Index score Land-based Physical Therapy Activities of Daily Living score Aquatic Physical Therapy
forest plot
What type of figure is shown in Figure 3? Group of answer choices Forest plot Ratio plot Bar chart Scatter plot
internal
When designing an experimental research study, having detailed selection criteria for enrollment and a detailed study protocol to be followed will increase ____________ validity. Group of answer choices External Internal
by choosing a blind or double-blind study design
When designing an experimental/interventional research study, how can you control the placebo effect? Group of answer choices By choosing a naturalistic setting By incentivizing subjects' participation By minimizing subject drop out By choosing a blind or double-blind study design
r will always range between -1.00 and +1.00
Which of the following best describes the possible values for the Spearman or Pearson correlation coefficient, 'r' ? Group of answer choices r can be any value depending upon the data r will always be a positive number greater than zero. r will always range between -1.00 and +1.00
pearson correlation
Which of the following statistical tests is most appropriate for determining the relationship between number of minutes/day spent studying and number of minutes/day spent exercising? Assume data are normally distributed for both variables. Group of answer choices Chi-square analysis ANOVA Independent t-test Pearson correlation
spearman correlation
Which of the following statistical tests is most appropriate for examining the relationship between X and Y, if both variables are ordinal? Group of answer choices Pearson correlation Spearman correlation Independent t-test Chi-square analysis
purposive sampling, stratified random sampling
_____ ______ is an example of non-probability sampling, while ____ ____ _____ is an example of probability sampling. Group of answer choices Purposive sampling, stratified random sampling Cluster random sampling, snowball sampling Homogenous sampling, purposive sampling Stratified random sampling, purposive sampling
reliability, validity
_______________ refers to the reproducibility (consistency) of the measurements method, while ________________ is indicative of your method actually measuring the attribute that you intend to measure. Group of answer choices Reliability, validity Validity, reliability Dependability, stability Stability, dependability
