-
no statistical difference is detected between treatment groups or p value is >0.05. Which Hypothesis do you accept?
Ho is the null hypothesis – The null hypothesis is accepted when no statistical difference is detected between treatment groups. Typically in clinical studies, researchers will study a medication versus placebo or standard of care. In biostatistics, the null hypothesis is accepted when the p-value is greater than 0.05.
-
A study was done and they found a statistical significance EXISTS between the groups or p-value was <0.05. Which hypothesis will you accept and which one will you reject?
Ha is the alternative hypothesis – This is the opposite of the null hypothesis; when Ho is rejected, Ha is accepted. In drug related studies, generally this means that there is a significant difference between the treatment and placebo (or standard of care) groups.When the p-value is low, the null hypothesis (Ho) must go!
-
Independent and Dependent Variables. While
Dependent variables are variables that will DEPEND upon the independent variables. What does INDEPENDENT variables depend on?
-
Independent
variables are variables that are set by the researcher. These variables are the ones that we (researchers) can control.
- The independent variables are what we set as researchers. The independent variables are Ibuprofen
- use and Naproxen use. The dependent variable is what DEPENDS on the independent variables. We do
- not control the BIMS score, so that will be the dependent variable.
-
Assuming no statistical difference is detected between a new drug and placebo or p value is >0.05. Which Hypothesis do you accept?
-
The null hypothesis (Ho) would be that
there is no difference between the two groups. The alternative hypothesis (Ha) would be that there is a difference in blood pressure.
-
In their study the researchers found that madeuprolol significantly lowers blood pressure more than placebo with a p-value of 0.01. What happens to:
(1) Null Hypothesis?
(2) Alternate Hypothesis?
(3) What type of error stands to be made here?
- So we would reject the null hypothesis in this case. In regards to Type 1 error, 0.01
- would equal a 1% chance that a Type 1 error has been committed. Or stated another way: there is a 1%
- chance that madeuprolol does NOT reduce blood pressure more than placebo based upon the results of
- this study.
- If later on, the results of this study are proven false, a Type 1 error would have been committed.
- By convention, a <5% chance (or a p value of less than 0.05) is considered statistically significant.
-
How do you define a Type 1 error?
Type 1 error is defined as detecting a difference when in reality one doesn’t actually exist.
- In regards to Type 1 error, 0.01
- would equal a 1% chance that a Type 1 error has been committed. Or stated another way: there is a 1%
- chance that madeuprolol does NOT reduce blood pressure more than placebo based upon the results of
- this study.
-
The formular for power is Power = 1 – Beta
Whats the benchmark for Power in studies?
-
In studies, the benchmark Power is generally desired is 80%, which would be equivalent to a beta value
of 0.2.
-
Propbablity of making a TYPE 1 error can be checked with which value?
P-Value
-
Probability of making a TYPE 2 error can be checked with which value?
- Beta value
- If power is 80% than beta is 20%. Thereby the probability of a TYPE 2 error is 20%.
Type 2 error is the percentage or chance of NOT detecting a difference when one actually exists.
-
Confidence interval and Statistical Significance.
Example:
95% confidence interval for the change in hemoglobin is (0.3 - 1.5)
what would make this study statistically significant or not?
- If this range contained the value 0, we
- would deem that the 95% confidence interval did not reveal statistical significance
-
In statistical variables, there are Categorical and Quantitative.
Whats the difference between categorical and quantitative?
What are the two types of categorical variables?
(1) A categorical variable cannot have any value between two points while Continuous variables have a distinct, measurable distance between each value
- (2)CATEGORICAL:
- Nominal-Name, yes or no
- Ordinal-order, pain scale, surveys
-
what is Parametric Data Versus Non-Parametric Data?
- Parametric-Continous data
- Non Parametric-Categorical (ordinal vs nominal)
-
Discrete variables are continous variables but not all continous variables are discreet. Whats the difference between the two?
-
Discrete variables are a countable type of variable, but how this differs from continuous, is that you
- can’t have fractions of a number. A couple examples is number of Accuchecks done per week or CHF
- exacerbations per year. There is a number associated, but you can’t have a fraction (i.e. the data is not
- continuous) like 1.36 CHF exacerbations or Accuchecks. The discrete variable will be a whole number
-
STATISTICAL TESTING FOR NON PARAMETRIC DATA
What type of tests to use for:
Nominal data
Ordinal data
Nominal Data: Chi-squared
Ordinal Data: Mann-Whitney U-Test and Wilcoxon Signed-Ranks Test
-
what are other references associated with continuous data?
- normal distribution
- parametric tests
- discrete variables
- quantitative variables
-
Appropriate test for continuous variable (ratio or interval) is?
-
The appropriate statistical test for
- continuous (ratio or interval) variables will be the t-test or ANOVA(analysis of variance). The t-test (sometimes referred to as student’s t-test) will be used for 1 independent variable with two or less
- groups. The ANOVA will be used for multiple groups (usually 3 or more, but can be used for 2).
-
Differential the different types of tests?
-
WHICH literature is generally peer-reviewed and contains original research done by scientists
and researchers. A clinical trial done by a pharmaceutical company would be a classic example
PRIMARY
-
literature is compiling primary and secondary literature and creating a larger, broader
view of clinical information. A classic example of this would be a textbook on pharmacokinetics
TERTIARY
-
literature consists of interpretations and evaluations of primary literature. A
review article breaking down a bunch of clinical trials relating to a given topic would be an example of
secondary literature.
SECONDARY
|
|