GoConqr has a new look! Let us know what you think.

Sign up free

Flashcards

Which statistical test to use?

María Jesús Grueso Spanish 40 cards

Cards in this deck (40)

  1. 1Alternative non-parametric test for TWO-SAMPLES UNPAIRED T-TEST (when non-normally distributed data)MANN-WHITNEY-U TEST
  2. 2Alternative non-parametric test for ONE-SAMPLE T-TEST and PAIRED T-TEST (when non-normally distributed data)WILCOXON TEST
  3. 3Alternative non-parametric tests for T-TESTS (when non-normally distributed data)MANN-WHITNEY-U TEST and WILCOXON TEST
  4. 4Comparison of the means of two groups: which workflow must be followed?1) Shapiro test or Kolmogorov-Smirnov test to check if populations are nomally distributed:
    - Non-parametric.
    - Parametric.
    2) In parametric cases, F-test to check if variances are equal.
    - T-test.
    - Welch's correction.
  5. 5Which variants of t-test have we studied?One sample t-test, two-samples t-test, paired t-test, one-tail t-test and two-tailed t-test
  6. 6When to use ONE-SAMPLE T-TEST?One group compared to a standard value and the dependent variable is quantitative
  7. 7Which test would you use for the following example? Measure blood pressure in mice before and after a certain treatment.PAIRED T-TEST (same individuals in both groups; not independent observations)
  8. 8Which test would you use for the following example? Check if a mutation increases Drosophila lifespan (known value)ONE-SAMPLE T-TEST (comparison of one group to a known value; quantitative outcome, lifespan; categorical predictor, mutation)
  9. 9Which test would you use for the following example? Compare the acidity of a medium conditioned by microglia to a neutral pH of 7.ONE-SAMPLE T-TEST (comparison of one group to a known value; quantitative outcome, pH; categorical predictor, condition by microglia)
  10. 10When is SHAPIRO TEST used?- To check if the populations are normally distributed (before going for a t-test; first step)
    - High power for 2 < n < 50
  11. 11When is KOLMOGOROV-SMIRNOV test used?- To check if the populations are normally distributed (before going for a t-test; first step)
    - High n necessary
  12. 12Which correction can be used in multiple comparison t-tests?BONFERRONI CORRECTION (alpha' = alpha/number of hypothesis or comparisons)
  13. 13Which test would you use for the following example? Expression difference of 10,000 genes in control vs. cancerMultiple t-test with bonferroni correction (two groups ; several quantitative outcomes)
  14. 14Alternative non-parametric test for ONE-SAMPLE T-TEST (when dependent variable is NOMINAL/CATEGORICAL)- Chi-square.
    - Binomial test (two possible outcomes)
  15. 15Alternative non-parametric test for TWO-SAMPLES T-TEST (when dependent variable is NOMINAL/CATEGORICAL)- Chi square for large samples.
    - Fisher's test for small samples.
  16. 16Which test would you use for the following example? Check is a mutation (bias) alter the ratio of M/FChi square or binomial test (comparison of one group, "mutation", to a known value for nominal dependent variable, "male" or "female")
  17. 17Which test would you use for the following example? Investigate mRNA expression of 10 ion channels in 50 control mice and 50 epileptic mice.T-test (two groups) with Bonferroni correction (several outcomes want to be compared between the groups)
  18. 18Which test would you use for the following example? You measure EPSP during baseline and EPSP after LTP (you calculate the ratio). You have two groups of slices: treatment and control.Mann-Whitney-U test (two groups; ratios probably not normally distributed)
  19. 19Which test would you use for the following example? Placebo and treated group (n small) and then, you count how many are alive and dead to test if the proportion is the same between the two treatmentsFisher's test (comparison of two groups, categorical nominal outcome)
  20. 20Which test would you use for the following example? Study on a group of blindfold people, ask them to hold the arm at 45% (known value)One-sample t-test (one group compared to a known value)
  21. 21Which test would you use for the following example? Compare mitochondrial acidification in a group treated with drug A and a group treated with drug BTwo-samples t-test
  22. 22Alternative non-parametric test for REPEATED MEASURES ANOVA (when non-normally distributed data)Friedman's test
  23. 23Alternative non-parametric test for ONE-WAY ANOVA or TWO-WAY ANOVAKruskal-Wallis test
  24. 24When do we use ANOVA tests?When comparing the means of three or more groups
  25. 25Which variants of ANOVA have we studied?One-way ANOVA, two-way ANOVA, MANOVA (+ repeated measures ANOVA)
  26. 26When do we use one-way ANOVA?One categorical predictor and one quantitative dependent variable
  27. 27When do we use two-way ANOVA?Two categorical predictors and one quantitative dependent variable
  28. 28When do we use MANOVA?One categorical predictor and several quantitative outcomes
  29. 29Which test would you use for the following example? Groups: treatment A, treatment B and control (no treatment). We measure the expression levels of two different ion channels.MANOVA
  30. 30Which test would you use for the following example? Two different treatments in two different mouse strains, which produces 4 groups and we measure the blood pressure of the miceTWO-WAY ANOVA (treatment and strain are the two categorical predictors)
  31. 31Difference between post-analysis tests: Tukey vs. Studen-Newman-KeulsTukey: more conservative (prone to beta error, miss real differences)
    Student-Neuman-Keuls: more powerful (prone to alpha error, find false differences)
  32. 32Which post-analysis test would you use if you only want to compare different treatment groups to the control?Dunnet's test
  33. 33What should you report after an ANOVA test?Type of test, F value, number of groups, degrees of freedom and p-value
  34. 34Predictor: quantitative. Dependent variable: quantitative. If only one predictor variable, which test would you use to make predictions?SINGLE LINEAR REGRESSION
  35. 35Which test would you use for the following example? Does the amount of time sleeping vary with age in the mutant?SINGLE LINEAR REGRESSION
  36. 36Alternative non-parametric test for SINGLE LINEAR REGRESSIONNon-parametric regression (non-normally distributed data) or simple logistic regression (nominal data)
  37. 37Predictor: quantitative. Dependent variable: quantitative. If more than one predictor variable, which test would you use to make predictions?MULTIPLE LINEAR REGRESSIONS
  38. 38Alternative non-parametric test for MULTIPLE LINEAR REGRESSIONSMULTIPLE LOGISTIC REGRESSION
  39. 39Which test would you use for the following example? Speed of crickets under different temperatures.SINGLE LINEAR REGRESSION
  40. 40Which test would you use for the following example? Measure air temperature, humidity, body mass, leg length, see how they relate to chirping speed in cricketsMULTIPLE LINEAR REGRESSION

Information

Description
Flashcards on Which statistical test to use?, created by María Jesús Grueso on 05/07/2022.
Views
9
Rating

Feedback

What is it about?

What do you think of the new look?

How do you like it?