Download Statistics Assumptions and more Quizzes Statistics in PDF only on Docsity!
we believe something about a population
Regression t-test 1) Linearity 2) Independence 3) Normally distributed QQ plot linearity 4) Equalvariance 2-proportions z-test and z-interval 1) 2 simple random samples 2) Independence 3) Successes and failures aregreater than 10
Choose an answer 1 Research hypothesis 2 Null hypothesis 3 Directional hypothesis 4 Significant hypothesis Don't know?
Terms in this set (58) Hide definitions
One mean t-test and t-interval 1) Normality of QQ plot 2) Representative sample
2 sample t-test 1) 2 random independent samples 2) Large sample sizes, assume normality One-proportion z-test 1) Random sample 2) Expected successes and failures are both at least 15 2 sample t-interval 1) 2 simple random independent sample 2) Normality with large enough samplesizes
Paired t-test 1) simple random sample of paired data 2) Histogram is unimodal and symmetric3) QQ plot points are linear 4) Large sample size, assume normality
Paired t-interval 1) simple random sample of paired data 2) Histogram is unimodal and symmetric3) QQ plot points are linear 4) Large sample size, assume normality
One proportion z-interval 1) Randomization of representative sample 2) At least 15 observed successes and15 observed failures
Categorical (qualitative) variable non-numerical variable with different categories Quantitative variable numerical variable Discrete counting (shoe size), possible outcomes are a set of separate numbers (0,1,2,3...)
Continuous measuring (height, weight), possible outcomes are any value in an interval(includes decimals)
Independence when probabilities are the same Left skewed mean is smaller than median Right skewed mean is larger than the median
Specificity the probability that the test will give a negative result, given that the conditiontested for is not present
z-scores - used to find the number of standard deviations an observation falls above orbelow the mean o Z= observation - mean/standard deviation z-score of 0 indicates that the observation is the mean Normal distribution often used for continuous random variables, and defined by two parameters, themean and standard deviation
- Symmetric and bell-shaped- Standard normal: mean= 0 and standard deviation= 1
Binomial Distributions 1) fixed number of trials 2) Each trial has 2 mutually exclusive outcomes (successand failure) 3) probability of success is the same for each trial 4) Trials are independent CLT Assumptions 1) Randomization Condition 2) Independence assumption 3) 10% condition 4)Sample assumption (at least 15)
Confidence interval point estimate +/- margin of error Margin of error 1/2 width of confidence interval
Width double the margin of error point estimate mean of the upper and lower limits of the confidence interval Null hypothesis we believe something about a population Alternative hypothesis we want to determine if something else is true Simple linear regression 1) Linearity 2) Independence 3) Normality 4) Equal variance mutually exclusive events have a probability of zero Independent events if they equal each other Unusual value when z-score is outside of -3 and 3 and provides evidence against the claim
SRS population standard deviation, normally distributed population or large enoughsample size to overcome deviations from normality
interpret Correlation coefficient, r describes relationship correlation between variables Interpret coefficient of determination, r(2) Approximately % of the variability in y can be explained by x population parameter There is only one value When asked for a probability distribution o Make sure values are 0 to 1o All numbers should equal 1
Explanation for wider CI o Higher confidence level
- With a correlation coefficient of 0 o Variables are not relatedo Variables have a relationship but its not linear
Statistical description useful summaries Lurking variables - don't actually measure in data Confounding variables in data set Response variable example whether or not the bullseye was hit at least once
