Flashcards in NUTR Capstone - Exam #1 (Part 2) Deck (52)
What is a Prospective Study?
A group FREE of the disease or outcome of interest is followed over time
What is a Cross-Sectional Study?
A group examined at ONE point in time
What is a Case-Control Study?
Two groups, based on the outcome
What are the types of OBSERVATIONAL Studies?
What are the types of Experimental Studies?
1. Cross-Over Randomized;
2. Randomized blinded trial (short term)
3. Randomized binded trial (long term)
What is a Cross-Over Randomized Trial?
Two groups created by a random process, one group starts placebo and the other in a treatment….washout period….treatments switch groups
What is a Randomized Short, Blinded Trial?
Two groups created by a random process, one group gets a placebo and the other receives treatment — days, weeks, months
What is Randomized Long, Blinded Trial?
Two groups created by a random process, one group gets a placebo and the other receives treatment — several years
What is Statistics?
an objective, mathematical means to interpret a collection of observations
How must a researcher use Statistics?
Researchers must use statistics COMPETENTLY:
- Use the right test(s) (need to understand basic theories)
- Apply them correctly
- Interpret the results appropriately
- Know when to consult a statistician for assistance (best at the beginning)
What about the experiment guides the Statistical analysis?
-Scale of measurement in which the data are (data are always plural) collected
-Relationship among samples
-Number of samples evaluated
-Test’s assumption about the normal distribution
-Whether one- or two-sided test for significance
What are the stats for Comparative Research questions?
-Do differences exist?
-Statistical tests detect differences in means or medians (t-tests or Wilcoxon rank sum test)
What are the stats for Relational Research questions?
-Do correlations or associations exist?
-Statistical analysis assesses correlations or associations (Pearson or Spearman’s rho correlations or chi-squared test)
What is a Comparison study?
-Scientific “intuition” the researcher has about the study outcome;
-Statistics provide the means to evaluate the data and ultimately determine whether to accept or reject the hypothesis… is it true or not?
What is the Null Hypothesis (H0)?:
“there is no difference between population means, no relationship between two variables” (H0 = µ1 - µ2 =0)
What is the Alternative Hypothesis (HA)?
Logical state of reality that MUST exist if the null hypothesis is not true (H0 = µ1 - µ2 ≠0)
What are the errors with Comparative Studies?
Type I and Type II Error
What are Type I Errors?
-REJECTING the null hypothesis when it is TRUE;
-The probability of rejecting a true null hypothesis is equal to the alpha level.
What are Type II Errors?
-Accepting the null hypothesis when it is not true.;
-The probability of accepting the null hypothesis when it is false is the beta value.
What are Variables?
-Discrete or Continuous
What are Discrete Variables?
-Discrete random variables (values for which a few possible values exist);
1. NOMINAL (non-ordered)(with 2 categories = binary);
3. ORDINAL (ordered by categories but the space between the categories is undefined)
What are Continuous Variables?
Continuous random variables (variables with a numerical meaning and the space between the values is defined and can be measured)
How are Discrete variables reported?
Discrete (values for which a few possible values exist) = Reported as frequency/proportions (numbers/percentages)
-Nominal but not binary
Reported as =
1. Relative risk (RR) – cohort or prospective design
2. Odds ratio (OR) – cross-sectional or case-control
How are Continuous variables reported?
-Continuous (variables with a numerical meaning and the space between the values is defined and can be measured)
1. Mean (average)
2. Median (middle score)
3. Mode (most common score)
4. Standard deviation (SD) or range
5. Interquartile range
What are the different types of relationships and samples?
— Pairing and matching
— Serial measurements – Examples???
— Replicate measurements – Examples ????
What are the problem with multiple comparisons or multiplicity?
-As the number of groups increase, the possible pairwise comparisons increase.
-This increases the chance of finding a spurious significant result.
-Have to be careful when making many comparisons. The differences might be seen due to chance.
-Statistician correct by lowering the alpha level indicating significance.
What is the Assumption of Normality?
-The validity of many statistical tests depends upon the assumption that the data is normally distributed and that the variability within groups is similar.;
-Such tests are termed parametric tests;
-Mean, Median and Mode are ALL the same (at 0 on normalized scale)
One Standard Deviations from the Median
60% of data
Two Standard Deviations from the Median
95% of data