PSYCH 210 Lecture 11 Outline of Last Lecture I. Continue Sampling Distributionsa. Constructing a Sampling Distributionb. The Central Limit Theoremi. Estimated Mean Valueii. Standard Errorc. Sample Sizei. ‘Law of Large Numbers’Outline of Current Lecture I. Hypothesis Testinga. Use of probability to answer experimental questionsb. Steps of the processII. Hypothesesa. Nullb. Alternativei. Non-directionalii. DirectionalCurrent LectureI. Continue Supermax Prison examplea. μ=500, σ=100, n=25, M=525b. p = 0.1056i. What does this tell us?1. Difference due to experimental treatment/chance2. Difference due to chance – 0.1056a. Is this a ‘good’ or ‘bad’ probability?II. Hypothesis TestingThese notes represent a detailed interpretation of the professor’s lecture. GradeBuddy is best used as a supplement to your own notes, not as a substitute.a. Def: An inferential procedure that uses sample data to evaluate the credibility of a hypothesis about a populationb. Basic Assumption we needi. Recall how adding/subtracting a constant affects scores in a distribution: Standard Deviation & Variation remain same1. Shape of distributions stay the same when Tx (Treatment) is applieda. Tx is acting like a constantc. Ex) Reproductive inhibition in cotton-top tamarinsi. ‘Cooperative breeders’ii. Test the hypothesis that removal of helper females from the natal group will increase their hormone levelsIII. Basic Overview of Hypothesis Testing Procedurea. POPULATION: State hypothesis about a population of interest (In words for now)b. Removing female CTT helpers from mother will increase LH levelc. Steroid Hormone (LH) Level: μ=26; σ=4i. SAMPLE: choose sample sizeii. n=161. APPLY TREATMENT2. Remove daughters from mother; house individuallya. MEASURE DVb. LH level of 16 monkeys living away from motheri. OBTAIN SAMPLE MEAN (M) 1. Compare to POP. MEAN (μ): Are they different?IV. Step 1: We actually need two hypotheses!a. Null Hypothesis: Ho: Predicts that Tx has NO effecti. H0: μ=26ii. NOTE: When you state hypotheses, they are always in terms of population parameters! (μ)b. Alternative: H1: Tx DOES have an effecti. Nondirectional (two-tailed): H1≠26ii. Directional (one-tailed): H1>26 (or H1<26) 1. How do you choose which alternative hypothesis to use?a. Experimenter gets to chooseb. Choice does affect resultsc. We can never work directly with the alternative hypothesisi. You can NEVER prove a hypothesis correct (only prove them wrong)d. We have to work with the null hypothesis. Why?i. Show this is wrong; go with Alternative by defaulte. Deciding btwn null and alternativei. If there is enough evidence to show null is wrong = REJECT the null1. If large discrepancy btwn M and μ = REJECTii. If there isn’t enough evidence = FAIL TO REJECT null1. If small difference btwn M and μ = FAIL TO REJECTiii. How are we going to figure out what constitutes a large vs. small discrepancy?1. Use a sampling distribution!2. Level of significance/alpha levela. (Tells us where to place lines on distribution)b. The probability value that defines unlikely values for M if the null is truei. Probability that result due to chancec. Ex) α=.05, .01, .001, etc.i. Two unlikely tail areas add up to alpha (α) valueii. Convention has it that .05 is largest α you can
View Full Document
Unlocking...