7126 Chapter Notes - Chapter MODULE 2A: Linear Regression, Parametric Statistics, Statistical Inference
20 Jun 2018
School
Department
Course
Professor
Descriptives
Getting to know the data
-Remove bad cases to maximise the truth
-Minimise error / noise
-Data checking: one person reads the survey responses aloud to another person who
checks the electronic data file
oFor large studies, check a proportion of the surveys and declare the error-
rate in the report
-Data screening: carefully screening a data file helps to remove errors and maximise
validity
oScreen for out of range values, mis-entered data, missing cases, duplicate
cases, missing data
-Describe the datas main features
oFind a meaningful, accurate way to depict the ‘true story’ of the data
-Test hypotheses
oTo answer research questions
Level of measurement and types of statistics
-Golden rule of data analysis
oA variables level of measurement determines the type of statistics that can
be used, including types of
Descriptive statistics
Graphs
Inferential statistics
-Levels of measurement and non-parametric vs parametric
oCategorical and ordinal data DV
Non parametric
(does not assume a normal distribution)
oInterval and ratio data DV
Parametric
(Assumes a normal distribution)
Non-parametric
find more resources at oneclass.com
find more resources at oneclass.com
(if distribution is non-normal)
-Parametric statistics
oEstimate parameters of a population based on the normal distribution
Univariate
Mean, SD, skewness, kurtosis, t tests, ANOVAS
Bivariate
Correlation, linear regression
Multivariate – many variables together
Multiple linear regression
oMore powerful, more sensitive; able to pick up differences
oMore asusmptions
Pop is normally distributed
Vulnerable to violations of assumptons
-Non parametric statistics
oWhich do not assume sampling from a population which is normally
distributed
There are non parametric alternatives for many parametric statistics
Eg chi square
oLess powerful, less sensitive
oFewer assumptions
oLess vulnerable to assumption violations
Univariate descriptive statistics
-Number of variables
oUnivariate = one variable
Mean, median, mode, histogram, bar chart
oBivariate = two variables
Correlation, t test, scatterplot, clustered bar chart
oMultivariate
More than two variables
Reliability analysis, factor analysis, multiple linear regression
find more resources at oneclass.com
find more resources at oneclass.com
-What do we want to describe?
oThe distributional properties of variables, based on
Central tendency
Eg frequencies, mode, median, mean
Shape
Eg skewness, kurtosis
Any graphing patterns?
Symmetrical or non-symmetrical?
Spread (dispersion)
Min, max, range, IQR, percentiles, variance, SD
-Measures of central tendency
oStatstics which represent the centre of a frequency distribution
Mode: most frequent
Median (50th percentile)
Mean (average)
oWhich ones to use depends on
Type of data (level of measurement)
Shape of distribution – skewness
o
oNominal: cant order data – so only mode is relevant!
oOrdinal: will be no mean because its in order without equal intervals
oInterval: any
find more resources at oneclass.com
find more resources at oneclass.com
Document Summary
Data checking: one person reads the survey responses aloud to another person who checks the electronic data file: for large studies, check a proportion of the surveys and declare the error- rate in the report. Data screening: carefully screening a data file helps to remove errors and maximise validity: screen for out of range values, mis-entered data, missing cases, duplicate cases, missing data. Describe the datas main features: find a meaningful, accurate way to depict the true story" of the data. Golden rule of data analysis: a variables level of measurement determines the type of statistics that can be used, including types of. Levels of measurement and non-parametric vs parametric: categorical and ordinal data dv. Non parametric (does not assume a normal distribution: interval and ratio data dv. Parametric statistics: estimate parameters of a population based on the normal distribution. Mean, sd, skewness, kurtosis, t tests, anovas.
