• Home
  • Swinburne University of Technology
  • Foundations of statistics
  • Statistical Analysis and Interpretation

Statistical Analysis and Interpretation

StuDocu.com STA10003 Cheat Sheet Foundations of statistics (Swinburne University of Technology) StuDocu is not sponsored or endorsed by any college or university Downloaded by Akshata Katarey (akshatakatarey@gmail.com) Significant: P < 0.05 Correlation (Scatterplot): Direction Form Strength Outliers Not Significant: P > 0.05 Y= Constant + (SlopeX) Levene Significant Not Equal Variance Pearson's r: >0.5 Strong, 0.3-0.49 Mode, 0.10-0.29 Weak, <0.10 Ext Weak DV: Y Axis IV: X Axis p= The strength of the linear r/p r2= The amount of variability in the DV explained by the r/p of the linear variables Nuisance Variable: Unwanted factor on independent variable Confounding Variable: A factor that effects both DV and IV Levene Not Sig: Equal Variance Positive skewed tails to right - negative tails to left, Median better for skewed Z-score (Normal Dist) = (x-u )= o where x = value u= mean o= StdDev Cross-tabulations (IV: Column DV: Rows) IQR= Q3 - Q1 Outliers= 1.5 x IQR df= n-1 n= Sample size x mean (not normal dist) u=mean (normal dist) s= std dev (not normal dist) o= std dev (normal dist) Relative Risk: (x/xz) = (y/yv) Expressed as more or less likely where 1=No risk x d = mean difference Sd = Std Dev of the difference Odds Ration: (x/z)=(y/v) Expresses as higher or lower where 1= even odds When describing: Shape (Normal, Symmetric, skewed), Centre (Mean or Median), Spread(Std Dev) Outliers (1.5xIQR or >2s) Calc Stand Value z = x - 11 x x = Sample mean Jx = mean of sampling distribution Yes No Z Total(XZ) 0 0 x = StdDev of the sampling distribution How many StdDev is the sampling mean away from the sampling distribution (multiple samples) Use std error when assessing distribution of multiple sample means Cat A X Cat B Y V Total (YV) Total (XY) Total (ZV) Independent sampling E.G - Ours is bigger than yours- 1 Metric & 1 Categorical Variable - Box Plot, Graph of Means-Independent Groups no neg t-score A researcher hypothesised that the blood cholesterol levels of people who do not exercise regularly would be higher than those of regular exercisers. In a random sample of 78 adults, the average blood cholesterol level of the non-exercisers (x = 247.65mg/dl, s = 33.44mg/dL, n = 43) was higher than the average blood cholesterol of the people who exercise regularly ( x = 189.06mg/dl, s =35.96mg/dL, n= 35), and an independent samples t-test shows that the difference in the mean blood cholesterol levels is significant, t(76)= 7.44. p < . 001. The 95% Confidence interval shows that the mean blood cholesterol level for people who do not exercise regularly is between 42.9 and 74.3mg/dL higher than for people who do. As expected, the blood cholesterol levels of non-exercisers are higher than for those people who exercise regularly. Paired Sampling E.G - Our this is bigger than our that - 1 Metric & 1 Categorical Variable - Box Plot, Graph of Means-Repeated Measures It was hypothesised that the electricity bills for households when using