WK7: QUANTIATIVE
Null hypothesis: statement of no difference
Alternate hypothesis: definite difference between groups of a subject
Statistical analysis: Organisation and interpretation of data according to well defined, systematic, and mathematical procedures and rules
Levels of statistical analysis: descriptive stats, inferential stats, associational stats, difference
Descriptive statics: data reduction
Inferential statistics: inferences to known population
Associational statistics: causality
Difference: difference between two or more groups of data
Descriptive statistics types: MCT, measures of variability, bivariate descriptive statistics
Measures of central tendency: mode, median, mean
Measures of variability: range, IQR, sum of squares, variance, standard deviation
Bivariate descriptive statistics: contingency tables and correlational analysis
Normal distribution: both halves are identical
Nonsymmetric: either positively (point towards the left) or negatively (point towards the right) skewed
Kurtosis: distribution characterised by its shape
Correlational analysis: determines relationships among variables
Level of significance: statement of the expected degree of accuracy of the findings based on sample size
inferential statistics – processes for drawing conclusions about a population based on data
confidence interval and level: range of values observed in sample that accurately reflects the population
significance level: indicates whether the samples being tests are from the same or different population
one-tailed and two-tailed level of significance: two extreme scores of the bell curve
type I error (alpha): researcher fails to accept null hypothesis
type II error (beta): accepts null hypothesis when it should be rejected
parametric statics: formulas that test hypothesises based on three assumptions
parametric stats assumptions: sample is serviced from population with a normal distribution, variance if homogeneous, data measures at interval level
t-test: Compare the difference of means of one variable
one-way analysis of variance (ANOVA): Test the difference between the means of two or more variables
pearson: measures the strength and direction of a linear relationship between the x and y variables
non-parametric statistics: test hypothesis when the data violates one or more of the assumptions of parametric procedures
chi-squared test: when data is nominal, and computation of mean is not possible
mann-whitney test: test differences between two independent groups
Wilcoxxin signed test: nonparametric alternative to t-test for correlated samples
Kruskal-wallis test: non-parametric alternative to ANOVA
Kappa: determine the degree of agreement between two or more judges
Hawthorne effect: attention given to subjects increases positive outcome
Placebo effect: receiving treatment may produce positive expectations
Honeymoon effect: a short-term effect of a new treatment procedure subjects hope will impact on the disease
Research bias: researcher affecting results through enthusiasm and interest for the treatment to be effective
Test administer bias: person aware which subjects get treatment and which don’t
Sampling error: sample not representative of population
Systematic variance: researcher fails to control extraneous variables such as age, race, gender
Error variance: researcher overlooks unexpected problems in test environment
Null hypothesis: statement of no difference
Alternate hypothesis: definite difference between groups of a subject
Statistical analysis: Organisation and interpretation of data according to well defined, systematic, and mathematical procedures and rules
Levels of statistical analysis: descriptive stats, inferential stats, associational stats, difference
Descriptive statics: data reduction
Inferential statistics: inferences to known population
Associational statistics: causality
Difference: difference between two or more groups of data
Descriptive statistics types: MCT, measures of variability, bivariate descriptive statistics
Measures of central tendency: mode, median, mean
Measures of variability: range, IQR, sum of squares, variance, standard deviation
Bivariate descriptive statistics: contingency tables and correlational analysis
Normal distribution: both halves are identical
Nonsymmetric: either positively (point towards the left) or negatively (point towards the right) skewed
Kurtosis: distribution characterised by its shape
Correlational analysis: determines relationships among variables
Level of significance: statement of the expected degree of accuracy of the findings based on sample size
inferential statistics – processes for drawing conclusions about a population based on data
confidence interval and level: range of values observed in sample that accurately reflects the population
significance level: indicates whether the samples being tests are from the same or different population
one-tailed and two-tailed level of significance: two extreme scores of the bell curve
type I error (alpha): researcher fails to accept null hypothesis
type II error (beta): accepts null hypothesis when it should be rejected
parametric statics: formulas that test hypothesises based on three assumptions
parametric stats assumptions: sample is serviced from population with a normal distribution, variance if homogeneous, data measures at interval level
t-test: Compare the difference of means of one variable
one-way analysis of variance (ANOVA): Test the difference between the means of two or more variables
pearson: measures the strength and direction of a linear relationship between the x and y variables
non-parametric statistics: test hypothesis when the data violates one or more of the assumptions of parametric procedures
chi-squared test: when data is nominal, and computation of mean is not possible
mann-whitney test: test differences between two independent groups
Wilcoxxin signed test: nonparametric alternative to t-test for correlated samples
Kruskal-wallis test: non-parametric alternative to ANOVA
Kappa: determine the degree of agreement between two or more judges
Hawthorne effect: attention given to subjects increases positive outcome
Placebo effect: receiving treatment may produce positive expectations
Honeymoon effect: a short-term effect of a new treatment procedure subjects hope will impact on the disease
Research bias: researcher affecting results through enthusiasm and interest for the treatment to be effective
Test administer bias: person aware which subjects get treatment and which don’t
Sampling error: sample not representative of population
Systematic variance: researcher fails to control extraneous variables such as age, race, gender
Error variance: researcher overlooks unexpected problems in test environment