Survey
* Your assessment is very important for improving the workof artificial intelligence, which forms the content of this project
* Your assessment is very important for improving the workof artificial intelligence, which forms the content of this project
CLASS NOTES: Repeated Measures of Analysis of Variance (ANOVA) CONCEPT CALCULATION/EXAMPLES APPLICATION Null Hypothesis (Ho): There is no difference b/t treatment effects & any differences that exist occur due to chance. Remember here that for the null hypothesis, it can simply be written as A = B = C = D. They all are the same, so no changes exist. However for the alternative hypothesis, we are comparing more than 2 treatment methods, & there are a number of different alternatives that can exist (A > B, but B = C; A = B = C; A = B, but B < C, etc..). So, we are looking for at least one difference b/t treatments. It is too much to write every possible outcome in the case of the alternative hypothesis. * Researchers usually have some idea of what difference they are looking for in their study. ANOVA for Repeated-Measures Design: This is where the same sample is used multiple times, or that the same study group of individuals participates in all of the different treatment conditions. * The general purpose of a repeated-measures ANOVA is to determine whether the differences that are found between treatment conditions are significantly greater than would be expected by chance. Single factor is used here meaning that only one factor or one variable is manipulated on two or more levels w/ the same set of individuals. Hypotheses for the RepeatedMeasures ANOVA Alternative Hypothesis (H1): At least one treatment mean (µ) is different from another. F-Ratio for the RepeatedMeasures ANOVA F= Variance/differences between Treatments (without individual _________differences________ Variance/differences expected By chance (w/ individual With this formula, you do not have to account for individual differences as you do in independent measures. * For repeated-measured ANOVA, the denominator is called the differences removed) residual variance or the error of variance & measures how much variance is to be expected just by chance after the individual differences have been removed. This is the same for independentmeasures ANOVA The Logic of the RepeatedMeasures ANOVA: There are 2 primary sources for chance differences: 1. Individual differences: There are variances b/t scores for each participant. 2. Experimental Error: There is always potential for some degree of error in general Testing Hypotheses w/ the Repeated-Measures ANOVA There are 2 different variances that we are looking at for this ANOVA: This is the same logic as in independent- measures design. 1) Between-Treatments Variance: The variance b/t treatment conditions 2) Within-Treatment Variance: The variance w/ the sample ANOVA Notation & Formulas Example: S = the number of subjects This is not a complete list of all notations & formulas, but those that are different from earlier notations & also those that are new. There are other notations & formulas used in ANOVA that are not listed here, but that you are already familiar w/. 6 participants = 6 subjects k = Number of treatment conditions 3 different treatments for Alzheimer’s Disease k = 3 (treatment conditions) n = Number of scores in each If there are 5 scores in each Remember that w/ Repeatedmeasures ANOVA, you only have one sample group w/ multiple treatment methods treatment N = Total # of scores in the entire study C = Correction Factor G = The sum of all the scores in the research study (the grand total) SS = Sum of Squares MS = Mean square df = Degrees of Freedom P = Used to represent the total of all the scores for each individual in the study. treatment condition, n = 5 If there are 9 scores altogether (3 treatment conditions, 3 scores in each treatment condition), then N = 9 For Repeated-Measures ANOVA, “N” represents the total number of all scores as opposed to total number of all research participants. C = (∑X1 + ∑X2 + ∑X3…..)2 N Add up all the N scores or add up the treatment totals G = Σ (∑X1 + ∑X2 + ∑X3…..) SS = ∑X2 – (∑X)2 N MS = SS df Remember that these formulas in this column are general. The more specific formulas used in the process are listed below. N–1 (see full example for calculations of each degrees of freedom) ∑X1 + ∑X2 + ∑X3….. “P” can be seen as “Personal totals” or “Participant totals.” ANOVA Formula Step 1: For each subject, calculate.. a) ∑Xm, (∑Xm)2 (symbol P may also be used to represent personal totals) For each treatment condition, calculate… The sum of each subject’s scores & this value squared Review example in text a) ΣX, ΣX2 & (ΣX)2 b) n n = number of participants/ sources in each sample c) M The mean for each treatment condition The subscript ‘m’ represents values associated w/ each subject. Another notation representing personal totals is also P. ‘P’ does not exist in independent measures as these values are not used in calculation. For the full or TOTOAL set of values, calculate…. d) N N = The total number of all scores in the experiment e) M Mean of each treatment mean f) ΣX, ΣX2 & (ΣX)2 For the total set of values (see example for calculations) ANOVA Formula Step 2: By adding the mean values together & dividing by the number of treatment measures ANOVA SS formula: Calculate Sum of Squares for… a) SSTOT total SSTOT = ∑X2 – (∑X)2 N Using your total values. b) SSb Between treatments ∑ The sum of squares between treatment groups. The Sigma (∑) represents the sum. The ‘g’ beneath the Sigma indicates that you should repeat the formula for each treatment. The subscript ‘g’ represents the value associated w/ each g (∑Xg)2 - (∑X)2 ng N c) SSsubj Within Subjects ∑ (∑Xm)2 - (∑X)2 s k N d) SSerror Error SSerror = SSTOT – SSb - SSsubj ANOVA Formula Step 3: Calculate the df for… a) dftot total N–1 b) df between treatments k–1 The ‘s’ beneath the sigma here works the same as the ‘g’ beneath the sigma in the between treatments SS formula. It still means that you are repeated the formula, but for each subject as opposed to each treatment. c) df Between subjects d) df error S–1 (k – 1) (S – 1) ANOVA Formula Step 4: Compute Mean Square for… a) MSb Between Treatments b) MSerror Error SSb dfb SSerror dferror Sum of Square between treatments divided by degrees of freedom between treatments Sum of Square error divided by degrees of freedom error ANOVA Formula Step 5: Compute your F-Ratio So that you can compute your final F-Ratio, which is… F = MSbetween treatments MSerror Mean Square between treatments divided by Mean Square error The Distribution of F-Ratios Since there are 2 df’s, then it is expressed as : df= 2, 12 The first df listed as your between treatments df & your second the within treatments df. Once you have computed your FRatio score… 1) Go to the F distribution table with your dfbetween treatments score (calculated in the numerator portion of your F-Ratio formula) & dfwithin treatments (subjects) score (found in the denominator portion of your F-Ratio formula), which in the case of repeated measures, you will use your dferror as your denominator. 2) Locate these 2 df’s on the table (numerator is listed in the row above whereas the denominator is listed in the column on the right). 3) Connect these 2 scores together in the middle. Regular type scores give you the critical value for alpha level of .05. Bold will give you the critical value for alpha level .01. Example of ANOVA Repeated-Measures Design Visual Detection Scores Participant No Distraction Visual Distraction Auditory Distraction A 47 22 41 B 57 31 52 C 38 18 40 D 45 32 43