Choosing a test: ... start from thinking whether our variables are continuous or discrete.

Slides:



Advertisements
Similar presentations
Hypothesis Testing making decisions using sample data.
Advertisements

Parametric Inferential Statistics. Types of Inference Estimation: On the basis of information in a sample of scores, we estimate the value of a population.
Statistical Issues in Research Planning and Evaluation
Beyond Null Hypothesis Testing Supplementary Statistical Techniques.
AP Statistics – Chapter 9 Test Review
Statistical Significance What is Statistical Significance? What is Statistical Significance? How Do We Know Whether a Result is Statistically Significant?
HYPOTHESIS TESTING Four Steps Statistical Significance Outcomes Sampling Distributions.
Evaluating Hypotheses Chapter 9. Descriptive vs. Inferential Statistics n Descriptive l quantitative descriptions of characteristics.
Statistical Significance What is Statistical Significance? How Do We Know Whether a Result is Statistically Significant? How Do We Know Whether a Result.
Evaluating Hypotheses Chapter 9 Homework: 1-9. Descriptive vs. Inferential Statistics n Descriptive l quantitative descriptions of characteristics ~
PY 427 Statistics 1Fall 2006 Kin Ching Kong, Ph.D Lecture 6 Chicago School of Professional Psychology.
PSY 307 – Statistics for the Behavioral Sciences
Chapter 12 Inferential Statistics Gay, Mills, and Airasian
Statistical hypothesis testing – Inferential statistics I.
INFERENTIAL STATISTICS – Samples are only estimates of the population – Sample statistics will be slightly off from the true values of its population’s.
© 2008 McGraw-Hill Higher Education The Statistical Imagination Chapter 9. Hypothesis Testing I: The Six Steps of Statistical Inference.
Hypothesis Testing II The Two-Sample Case.
Chapter 8 Introduction to Hypothesis Testing
1 Today Null and alternative hypotheses 1- and 2-tailed tests Regions of rejection Sampling distributions The Central Limit Theorem Standard errors z-tests.
Evidence Based Medicine
Introduction To Biological Research. Step-by-step analysis of biological data The statistical analysis of a biological experiment may be broken down into.
STA Statistical Inference
User Study Evaluation Human-Computer Interaction.
© 2008 McGraw-Hill Higher Education The Statistical Imagination Chapter 10. Hypothesis Testing II: Single-Sample Hypothesis Tests: Establishing the Representativeness.
Psy B07 Chapter 4Slide 1 SAMPLING DISTRIBUTIONS AND HYPOTHESIS TESTING.
Hypothesis Testing Introduction to Statistics Chapter 8 Mar 2-4, 2010 Classes #13-14.
PCB 3043L - General Ecology Data Analysis. OUTLINE Organizing an ecological study Basic sampling terminology Statistical analysis of data –Why use statistics?
Warsaw Summer School 2011, OSU Study Abroad Program Difference Between Means.
Educational Research Chapter 13 Inferential Statistics Gay, Mills, and Airasian 10 th Edition.
Correlation Assume you have two measurements, x and y, on a set of objects, and would like to know if x and y are related. If they are directly related,
Hypothesis Testing An understanding of the method of hypothesis testing is essential for understanding how both the natural and social sciences advance.
Review I A student researcher obtains a random sample of UMD students and finds that 55% report using an illegally obtained stimulant to study in the past.
Chapter 10 The t Test for Two Independent Samples
PCB 3043L - General Ecology Data Analysis.
Tuesday, April 8 n Inferential statistics – Part 2 n Hypothesis testing n Statistical significance n continued….
URBDP 591 I Lecture 4: Research Question Objectives How do we define a research question? What is a testable hypothesis? How do we test an hypothesis?
Hypothesis Testing Introduction to Statistics Chapter 8 Feb 24-26, 2009 Classes #12-13.
Sampling Distribution (a.k.a. “Distribution of Sample Outcomes”) – Based on the laws of probability – “OUTCOMES” = proportions, means, test statistics.
PCB 3043L - General Ecology Data Analysis Organizing an ecological study What is the aim of the study? What is the main question being asked? What are.
Methods of Presenting and Interpreting Information Class 9.
Agenda n Probability n Sampling error n Hypothesis Testing n Significance level.
Lecture #8 Thursday, September 15, 2016 Textbook: Section 4.4
AP Biology Statistics From BSCS: Interaction of experiments and ideas, 2nd Edition. Prentice Hall, 1970 and Statistics for the Utterly Confused by Lloyd.
Logic of Hypothesis Testing
Hypothesis Testing.
Two-Sample Hypothesis Testing
Inference and Tests of Hypotheses
Review Ordering company jackets, different men’s and women’s styles, but HR only has database of employee heights. How to divide people so only 5% of.
Review You run a t-test and get a result of t = 0.5. What is your conclusion? Reject the null hypothesis because t is bigger than expected by chance Reject.
1. Estimation ESTIMATION.
PCB 3043L - General Ecology Data Analysis.
APPROACHES TO QUANTITATIVE DATA ANALYSIS
POSC 202A: Lecture Lecture: Substantive Significance, Relationship between Variables 1.
Inferential statistics,
Inferential Statistics
INTRODUCTORY STATISTICS FOR CRIMINAL JUSTICE Test Review: Ch. 7-9
Difference Between Means Test (“t” statistic)
CHAPTER 26: Inference for Regression
Review: What influences confidence intervals?
Hypothesis Testing.
Introduction to Hypothesis Testing
UNDERSTANDING RESEARCH RESULTS: STATISTICAL INFERENCE
What are their purposes? What kinds?
Inferential Statistics
Chapter 7: Statistical Issues in Research planning and Evaluation
15.1 The Role of Statistics in the Research Process
  Pick a card….
Inferential statistics Study a sample Conclude about the population Two processes: Estimation (Point or Interval) Hypothesis testing.
Type I and Type II Errors
Introduction To Hypothesis Testing
Presentation transcript:

Choosing a test: ... start from thinking whether our variables are continuous or discrete.

The power of a test is the probability that the test finds a statistically significant effect... ....in case the effect of certain strength actually occurs in the population. - can be used in two situations - planning experiments – how large sample to take? - concluding from negative results. A non-significant result as such is not a strong argument. We can never prove that there is no relationship!

But we can show that the strength of the relationship is not larger than .... (some biologically relevant value), we could not show statistical significance, but if the relationship would have been as strong as ..., would it have been very likely (e.g. 84,5% - this is the power!) to get it significant, but as we did not, then probably there still isn’t that strong relationship in the populations. An easier way – confidence limits to parameters.

Hypotheses testing: null hypothesis and research hypothesis: HO: - there is no difference; H1: - there is a difference; Type I error: declare H1 correct, when it actually isn’t; Type II error: remain with HO, although H1 is actually correct. Conservative tests have lower type I error risk but have higher risk of type II error.

Information criteria in model selection, AIC – Akaike Information Criterion, IT-approach; ... simplifying model, which independent variables to include/ drop? ... in situations in which we wish to sort out determinants of our independent variable, especially on the landscape, occurence of a species, especially when we want to predict. Entire models (sets of independent variables) are compared - the models as such get points; - not based on p-values of particular variables;

We can compare the models on the basis of AIC values, AIC score depends on: - model fit (likelyhood); - complexity of the model; the model with lowest AIC is declared the best! Model fit... also R-square but that is always larger for more complex models; in AIC approach, the models get punished for their complexity; did increasing the complexity improve its fit much enough to justify it?

What does the abundance of toads depend on? - abundance of slugs; - abundance of earthworms; - density of pools. Slugs Worms Pools error

Two ways how to make conclusions: the best model (smallest AIC); model averaging: a set of „good“ models is found, they get weights according to AIC values; for particular independent variables, importances can be calculated, according to presence of the variable in good models; ... the variables get more importance points for being present in better models; .... variables can be ranked according to their importance.

Occurrence of Phengaris arion on Saaremaa island, Margus Vilbas et al. 2015

Bayesian statistics, ... we have a sample and a priori information about which value of a variable is how likely to occur; ... in ordinary statistics we have only sample; .... we change our prior understanding based on our sample; ... we have a positive relationship with 99% probability = there is 1 % probability that there is not = the positive relationship in our sample was obtained by chance with 1% probability; .... the ordinary p-value cannot be interpreted this way!

Bayesian statistics in more detail .... a gambler gets „six“ throwing a dice, …. ordinary dice vs. cheating dice; .... for ordinary dice, the probability is 16,7%, this is p. … this is the probability to get „6“ by chance, this is p. … but it is not the probability that „6“ was got by chance (= probability, that the dice is OK); ... 16,7% is not the probability that the dice is OK, we do not know...

But if we know in advance, that in half of the cases, the gambler uses the cheating dice, we can calculate, that the probability of having a cheating dice today is 85,7%, vs 14,3% for the ordinary dice. This 50:50 is prior distribution; 85,7:14,3 is our posterior distribution. Posterior distribution depends 1) on prior distrobution; 2) on our sample.

The gambler gets „six“, the probability that he has a cheating dice depends on - probability to get „6“ using ordinary dice; - probability that cheating dice is used; We have caught 6 female and 10 male bears from the forest, we can calculate the probability of getting this by chance if in the population there is 1:1 ratio; this is p; - we need to know only the sample; .... to answer the question „how likely it is that our population is female biased?“, we must know the probability of the occurrence of such populations in nature. If female-biased populations are very rare, it is much more likely that we just got an odd sample from a male-biased population.

.... the same for a continuous variable:

There is a significant relationship in one group but there was not a significant relationship in another? We study the effect of fertilisation on the growth of birch and aspen, two treatments (fertilised or not), dependent variable is three height. For aspen, an effect of fertilisation was found (p = 0,01), but not for birch (p = 0,08). can we say that there is a (statistically significant) difference in the effect of fertiliser between the tree species?

We have to test the treatment*species interaction! No we cannot. We have to test the treatment*species interaction! zero effect sign effect effect strength aspen birch

If to be absolutely honest, then we should have the hypothesis before looking at the data p value for the hypothesis that last lecture of this course happens on 19th? Null hypothesis and research hypothesis: HO: last statistics lecture happens on whatever date; H1: last statistics lecture happens on 19th day of a month; p = ???

.... p value for the hypothesis that last lecture of this course happens on 19th? Is it then statistically significant, that on 19th day of a month? No, because we set up or hypothesis looking at our data, the hypothesis should be independent of the data.

statistics does not answer the question about causality; do not divide a continuous variable into classes for the analysis, but you can do so in a figure;

„Nothing reveals mathematical illiteracy better than excessive accuracy in calculations “.