From her final PPT questions in card format Flashcards by Alexander Lee

This should be in the _______ section of a manuscript

Introduction

How well did you know this?

Not at all

Perfectly

What is a possible statistical test that was used in this study?

Let’s say the relationship was between BMI values and body satisfaction score (I/R data).

Pearson Correlation

How well did you know this?

Not at all

Perfectly

What are the three levels of IRB applications?

Exempt, expedited, full

How well did you know this?

Not at all

Perfectly

What is the level of evidence on the CEBM pyramid?

How well did you know this?

Not at all

Perfectly

If you investigate a clinical research question and can only find evidence from fellow clinicians and continuing education courses. What “grade” of evidence should be assigned?

How well did you know this?

Not at all

Perfectly

PEDro can only be used to evaluate a ____________ article.

RCT

How well did you know this?

Not at all

Perfectly

Who was the primary author of this article?

Whitman (listed first)

Last author - senior author/experienced investigator
Second = major contribution
Middle = everyone else

How well did you know this?

Not at all

Perfectly

Was there a significant difference between the two groups?

What is the chance these authors committed a Type 1 error?

Yes, p=.0015

.15%

How well did you know this?

Not at all

Perfectly

This text should be in the __________ section of a manuscript.

Introduction

How well did you know this?

Not at all

Perfectly

What do the words “a priori” mean?

A power analysis was conducted before the study and was set at .80

How well did you know this?

Not at all

Perfectly

According to this manuscript, there is a ____% chance they will find a significant difference….if there is one to be found.

80%

How well did you know this?

Not at all

Perfectly

Is it easy or harder to get a significant result with a 1-tail test?

Easier

How well did you know this?

Not at all

Perfectly

If you are looking for a difference in the proportion of people in groups, which statistical test should you run? T-test, chi-square, or Mann-Whitney

chi-square

How well did you know this?

Not at all

Perfectly

According to these authors, were there any significant differences between the groups?

How well did you know this?

Not at all

Perfectly

Do you think these were independent or repeated t-tests?

Independent

How well did you know this?

Not at all

Perfectly

Is HOV an assumption for an independent t-test?

Yes

How well did you know this?

Not at all

Perfectly

What is the name of the statistical test that checks for HOV in SPSS?

Levene’s

How well did you know this?

Not at all

Perfectly

Based on these graphs alone, what is a major concern?
A) the large amount of difference in the means
B) the small amount of difference in the means
C) the large amount of variability
D) the small amount of variability.

C) the large amount of variability

How well did you know this?

Not at all

Perfectly

You are not likely to find significance if the confidence interval spans 0 in which column?
A, B, or C

C – since this would be that the mean difference could be 0 or no difference.

How well did you know this?

Not at all

Perfectly

Would this be considered a primary or secondary source of literature?

secondary

How well did you know this?

Not at all

Perfectly

You are a researcher at University of Arizona. You decide to compare 3 groups of individuals with different types of PTSD on quality of life following a 1:1 aerobic and strengthening program with a physical therapist. You code the data into SPSS as 1, 2, and 3 for the groups.

What is the level of data?

nominal

How well did you know this?

Not at all

Perfectly

Is this graph showing a 1-tail or 2-tail test?

2-tail

How well did you know this?

Not at all

Perfectly

Are the numbers on the X axis standardized? Like a z, t, or F.

Yes. Since the mean is a 0.

How well did you know this?

Not at all

Perfectly

Which type of validity has been violated if you don’t recruit or test enough participants, so your power is too low?

Internal validity
Construct validity
External validity
Statistical conclusion validity

Statistical conclusion validity

How well did you know this?

Not at all

Perfectly

Which type of validity has been violated if you over-generalize your results to a larger population than was included in your study. * Internal validity * Construct validity * External validity * Statistical conclusion validity

External validity

History Maturation Attrition/Mortality Repeated Testing Instrumentation Regression to the Mean Experimenter Bias Selection All of these are possible threats to ___________ validity.

internal

History Maturation Attrition/Mortality Repeated Testing Instrumentation Regression to the Mean Experimenter Bias Selection Which of the above is a big problem when treating patients who naturally start to recover?

Maturation

If you need 20 randomly selected students from TWU, and you ask the registrar to simply pull 20 random students from the records. What type of sampling have you just done?

Simple random

True 𝛼=1−(1−𝛼)number of tests What is this formula used to calculate?

Calculating the true alpha level with inflation.

You are testing two groups of randomly assigned participants and plan to compare ROM after an intervention? What is the research design?

Posttest-only randomized group

You are testing two groups of patients, one receiving the standard of care and one receiving a new treatment. Both groups will be tested over three time periods. What is the research design?

Factorial – mixed or split-plot (can be randomized or non-randomized)

Which type of validity are you testing if you are comparing your new survey tool results with a gold standard tool?

Criterion-based validity

If you want to make your new outcome tool more sensitive to change…. should you add more items or remove some items?

Add items

If a new test battery is better at ruling in a sacroiliac disorder as compared to ruling it out. What number would be higher….sensitivity or specificity?

specificity

If a patient receives a positive test result for an autoimmune disease, but truly doesn’t have the disease. This is called a _________________. * False negative * False positive * True negative * True positive

False Positive

Your patient is suspected of having tuberculosis. The patient’s diagnostic lab test is positive for tuberculosis, but you know that only 83% of those with a positive test actually have tuberculosis. Does this mean that the lab test has a +PV of 83% or a –PV of 83%?

+PV of 83%

If a diagnostic test has a –LR of .11, that means there is a decreased probability of disease if the patient has a _________ (negative or positive) test result.

Negative

Based on this ROC curve, a cut-off value of 11 is the best place to maximize which two values?

Sensitivity and specificity

What is this diagram called? What type of skew is present? How do you test for skew on SPSS?

* Histogram * Negative * Histograms, skewness/kurtosis values should be between -2 and 2

To maximize the effect size using Cohen’s d…..you want the difference between the two means to be _______________. (large or small) To maximize the effect size using Cohen’s d…you want the variability to be ___________. (large or small)

Large Small

What BMI value would have a Z-score of +1?

36.6 + 5.1 = 41.7

What does intention-to-treat mean?

* Missing data was included in the analysis * How? Last value was kept in the dataset….or a different method.

What does it mean if you read that an analysis was completed using a 2-tailed test?

* It is possible that the outcome may increase or decrease and still be significant. * Also, it is a more correct way to analyze data since it splits the .05 into two tails. It is easier to reach significance with a 1-tail test.

The chance of a Type 1 error is determined by the researcher when he or she sets the __________ level.

alpha

Is the t value=1.66 a critical t value or a calculated t value?

Calculated you can't know the critical t-value

Will the mean for this distribution be to the left or the right of the median?

To the left

What would be your estimated r value for this association?

+.6 (just a guess)

48. What does inter rater mean?

between 2+ raters

49. What does intra-rater mean?

Within 1 rater

50. For the ANOVA, it would be 1x__

For the ICC, if the raters were the only ones of interest, and each rater measured two times. It has an ICC(__,__)

ICC(3,2)

Here are the results from the 1x3 ANOVAs. Is this an independent one-way ANOVA or R-M one-way ANOVA?

RM (time)

Here are the results from the 1x3 ANOVAs. Do these global ANOVAs need post-hoc follow-up testing?

Yes

Here are the results from the 1x3 ANOVAs. Why are follow-up tests needed?

We need to know where the exact differences are located.

Here are the results from the 1x3 ANOVAs. There are 13 tests. What is the FamilyWise (overall) error rate? Larger or smaller than .05?

1-(.95)13 or 48.7%

Some of these tests are reported as Greenhouse-Geisser tests….why?

Issue with sphericity.

Is the main question here….is the rater reliable or is the participant reliable?

participant test-retest reliability of the modified four step square test (mFSST) in patients

The SEMeasurement was .80. 95% of the time your patient should have a score between +/- ___ points.

The true score is +/- .80 seconds (68% of the time), or 1.60 (95% of the time).

When do you use this test?

Non-parametric independent t-test

When do you use this test?

regular t-test

When do you use this test?

Hint: It is the 2nd row of output for a t-test. t-test results when you don’t have HOV

When do you use this test?

non-parametric Pearson correlation

What does it mean if the r-value is negative?

As Berg goes up, TUG goes down. Or vice versa.

In this study, to reduce a risk of type l error, a Bonferroni-adjusted significance level of 0.00455 was calculated by dividing the p value of 0.05 by the number of statistical analyses performed. Are you happy or sad to read this in an article?

Happy

What ANOVA is going to be run in SPSS? You can assume that the group variable consists of 3 independent groups.

3x3 split-plot ANOVA

A two-way ANOVA was conducted that examined the effect of sex and education level on interest in politics. The interaction between the effects of sex and education level on interest in politics was not significant at F(2,54) = 4.643, p = .614. What does two-way mean?

Two IVs

Sex and education level

Interest in politics

Related to the # of IVs

Related to the # of participants

We don’t know. We could look it up in a table.

4.643

A two-way ANOVA was conducted that examined the effect of sex and education level on interest in politics. The interaction between the effects of sex (2 levels) and education level (3 levels) on interest in politics was not significant at F(2, 54) = 4.643, p = .614. What analysis should be done next?

Post-hocs on the main effect of education level if it was significant.

Would you expect a significant interaction with this plot?

Here are some post-hoc comparisons for our education by sex ANOVA. Are these simple effects or post-hocs on the main effects?

Main effects

How many were significant?

3 pair-wise

What alpha level are you using for comparison (if the alpha hasn’t been adjusted)?

.05/3 or .0167

What statistical test can be used to determine if the outlier is an influential outlier?

Cook's

If this participant had an outlier test value of .88, would you keep him/her in the study?

Yes (OK is less than 1)

This data shows a clear violation of which assumption?

Homoscedasticity

You would like to know if type post-hemispherectomy surgery in epileptic pediatric clients can predict LOS? What type of regression do you run?

Linear Regression

What is the predictor?

Type of surgery

What is the outcome?

LOS

Would you say there is a strong relationship between the two variables?

Yes, .873

How much of a patient’s LOS can be predicted by their type of surgery?

76.2%

You would like to know if the type post-hemispherectomy surgery in epileptic pediatric clients can predict LOS? Is the regression model significant?

Yes, p≤.0005

You would like to know if the type post-hemispherectomy surgery in epileptic pediatric clients can predict LOS? What would be the regression equation to predict LOS?

LOS = 8286.78+.564(type of surgery) Note….if two surgeries are compared…enter a 0 or 1

You would like to know if type post-hemispherectomy surgery in epileptic pediatric clients can predict LOS? Now, let’s say we also added in type of epilepsy as a predictor: 0 or 1. And we entered it into the model with type of surgery. In other words…the two predictors were entered together. Within this new model, is type of surgery a significant predictor?

yes

If the association between two variables (hours of productive studying and GPA) is r=.722. What is the effect size of this relationship?

.722

What is the threat to internal validity? You have just completed your pre-test assessments of spasticity and 3 participants in your control group undergo Botox injections.

History

What is the threat to internal validity? You give the same balance test to 24 participants three times as a warm-up and three times for the study.

Repeated Testing

If you have a significant interaction with a 2x4 split plot ANOVA. What is the TOTAL number of possible pairwise comparisons that can be run?

If you had 3 DVs in this study and you wanted to see their effect when combined ….you should run a 2 x 4 ____.

you should run a 2 x 4 split plot MANOVA.

What are the three omnibus tests for this ANOVA?

Main effect of group Main effect of time Interaction of group and time

The omnibus ANOVA tests are typically run at an alpha level of _________.

.05

If you don’t have a significant interaction with a 2x4 split plot ANOVA. Does the main effect of group need follow-up testing?

If you don’t have a significant interaction with a 2x4 split plot ANOVA. Does the main effect of time need follow-up testing?

Yes, and remember to divide your alpha level by 6.

Based on this 3x4 figure, do you expect a significant interaction?

yes

Based on this 3x4 figure, do you expect a significant main effect of group?

Yes. Regardless of time…Group C is probably significantly different than A and B.

Based on this 3x4 figure, do you expect a significant main effect of time?

Yes. Regardless of group….there is probably a significant difference in time….mostly likely between time 1 and 4.

Why do these researchers report adjusted means?

ANCOVAs were run on adjusted means. They are adjusted based to the covariate.

You work for SW Airlines. You are reviewing the results from a statistical analysis of the survey results. What statistical technique was used?

Factor Analysis

You work for SW Airlines. You are reviewing the results from a statistical analysis of the survey results. How many factors were revealed?

You work for SW Airlines. You are reviewing the results from a statistical analysis of the survey results. What would you name these factors?

Factor 1: Personal touch Factor 2: Cost Factor 3: Comfort it totally doesn't matter this is just what Katy chose

Participant numbers are determined prior to data collection. Quantitative or Qualitative Research?

QUANT

Data collection instrument is the researcher. Quantitative or Qualitative Research?

QUAL

Themes and theories emerge through data collection. Quantitative or Qualitative Research?

QUAL

Hypotheses are tested. Quantitative or Qualitative Research?

QUANT

Often includes quotes from the participants in the results section. Quantitative or Qualitative Research?

QUAL

Concerned more with the process and less with the outcomes. Quantitative or Qualitative Research?

QUAL

1. Chamberlain Mitchell SAF, Garrod R, Clark L, et al. Physiotherapy, and speech and language therapy intervention for patients with refractory chronic cough: a multicentre randomised control trial. Thorax. 2017;72(2):129-136. doi:10.1136/thoraxjnl-2016-208843. Is this APA or AMA format?

AMA

If you expect two groups will differ by 12 points after an intervention with a standard deviation of 24. What is the Cohen’s D effect size?

12/24 = .5

If you find an r value of .2, how much variance is shared between the two variables?

If you find an r value of .2, is it likely to be significant if you have 400 participants?

Yes. Correlations are often significant if you have a large sample size.

The central limit theorem states that if you have at least ____ participants, the assumption of normality will be met.