Week-7-Slides-Mean-Tests-Parametrics-Test-selection-module.ppt

Mean Tests & X2
Parametric vs Nonparametric
Errors
Selection of a Statistical Test
SW242

 4 Levels of Measurement
 Nominal, ordinal, interval, ratio
 Variable Types
 Independent Variables, Dependent Variables
 Moderator variables
 Discrete Variables
 Finite answers, limited by measurement e.g. test scores,
 Continuous variables
 All values possible (GPA not exceed 5.0)
 Dichotomous variable
 Only 2 values, yes or no, male or female
 Binary variable
 Assign a 0 (yes) or 1 (no) to indicate presence or absence of something
 Dummy variable
 Assign a binary value to a dichotomous value for analytic purposes
(e.g. male = ‘0’, female = ‘1’).
Creation & Description of a Data Set

 Variable Types
 Independent Variables, Dependent Variables
 Moderator variables
 Discrete Variables
 Finite answers, limited by measurement e.g. test scores,
 Continuous variables
 All values possible (GPA not exceed 5.0)
 Dichotomous variable
 Only 2 values, yes or no, male or female
 Binary variable
 Assign a ‘0’ or ‘1’ to indicate presence or absence of
something
 Dummy variable
 Assign a binary value to a dichotomous value for analytic
purposes (e.g. male = ‘0’, female = ‘1’).

Number of Variables Analyzed
Univariate analyses
 Examine the distribution of value categories
(nominal/ordinal) or values (interval or ratio)
Bivariate analyses
 Examine relationship between 2 variables
Multivariate analyses
 Simultaneously examine the relationship
among 3 or more variables
Categories of Analysis

Descriptive
 Summaries of population studied
(parameters)
 Preliminary to further analysis
Inferential
 Used with sample from total
population and how well can results be
generalized to total population
Purpose of Analysis

 Parametric Tests require:
 One variable (usually the DV) is at the interval or ratio
level of measurement
 DV is normally distributed in the population;
independent samples should have equal or near equal
variances
 Cases selected independently (random selection or
random assignment)
 Robustness how many and which assumptions above
can be violated without affecting the result (delineated
in advanced texts).
Parametric vs. Non-parametric

 Nonparametric Tests require nominal or ordinal level
data
 Samples complied form different populations and we
want to compare the distribution of a single variable
within each of them
 Variables are nominal or can only be rank ordered
 Very small samples: 6 or 7 are available
 Statistical power is low, increases with sample size (as
with parametric tests)
Parametric vs. Non-parametric

 Frequency Distributions
 Array is an arrangement of data from smallest to highest
 Absolute/simple frequency distribution displays number of times a
value occurs (all levels of measurement)
 Cumulative frequency distribution adds cases together so that it last
number in distribution is the total number of cases observed
 Percentage distribution adds the percent of occurrence in the table
 Cumulative Percentage

Example of an Array
Initial Age Frequency
Cumulative
Frequency %
Cumulative
%
A + T 21 2 2 10 10
B+G+C 26 3 5 15 25
R+W+S 27 3 8 15 40
K+V+R+
D
31 4 12 20 60
Q+F 32 2 14 10 70
S+O+P 37 3 17 10 95
M+A 49 2 19 10 95
B 69 1 20 5 100

Bar Graph/Histogram (bars touch)
Line Graph/Frequency Polygon
Pie Chart
Keep graphs simple.
 Limit to only salient information.
 Collapse categories/distributions when possible.
Graphical Representations

 Typical representation of data, e.g. find a number or
groups of numbers that is most representative of a dataset
 Mode
 Values within a dataset that occur most frequently, if two occur
equally then bimodal distribution, etc.
 Median
 The value in the exact middle of a linear array, mean between 2
values if even number of values.
 Mean: arithmetic mean
 Trimmed mean (outliers removed) minimize effect of extreme
outliers
 Weighted mean: compute an average for values that are not
equally weighted (proportionate / disproportionate sampling)
Measures of Central Tendency

 Variability/Dispersion
 Nominal or Ordinal use a frequency distribution or graph (bar
chart)
 Interval or ratio use range
 Range = maximum value – minimum value +1
 Informs about the number of values that exist between
the ends of the distribution e.g. 31 to 46 there are
potentially 16 values possible. The larger the range, the
greater the variability. However, outliers make the range
misleading. Therefore use median, or mean and standard
deviation whenever possible for interval & ratio data.
Measures of Central Tendency

 Statistical significance of the Chi Square only tells us
that the result is not due to sampling error.
 Nominal variables only
 Small group use
Chi Square Test of Independence

1. Test the relationship of nominal Variables.
 Not a statistic that can infer causality.
 Tells us the difference between Observed and Expected
frequencies.
 Pre-Experimental research designs, limited practice
issues.
- Use for one direction hypothesis testing
(directionality).
- Chi Square only tells us that the result is not due to
sampling error.
Purpose of X2 --- Chi-Square

 Implying cause rather than association
 Overestimating the importance of a finding,
especially with large sample sizes
 Failure to recognize spurious relationships
 Nominal variables only (both IV and DV)
Limitations & Common Problems

 One Sample t-test
 You have the test scores for one population only.
Compare to a known value e.g. average non-clinical
population test score = 30
 Analyze
 Research design for this type of comparison
 Pre-experimental
 Other uses: establish external validity of the sample not
significantly different from the general population
parameter (what is a population parameter?)
Mean Comparison Tests

 Independent Samples t-Test
 2 different samples/populations are tested for
comparison
 not matched, receive same treatment (if matched they
would be paired sample t-test
 Prepare dataset all scores in one column, tx_time = 2.
 Paired T-Test
 Same populations is tested for comparison
Mean Comparison Tests (cont.)

 Squared Pearson product moment correlation
tells the strength of a relationship which
represents the amount of variation explained by
the DV about the IV.
 Used for Comparing Two Interval/Ratio Groups
 Obtain a Correlation Coefficient that Ranges
from -1 (perfect inverse relationship) 0 (no
relationship) or +1 (perfect positive relationship)
Measures of Association

Errors
Our Decision
Real World Reject Null (-1)
Does work =
difference
Accept Null (1)
Doesn’t work =
no difference
Null Hypothesis
False (-1) Does
work =
difference
No Error (+1) Type II (-1)
Null Hypothesis
True (+1) Does
n’t work = no
difference
Type I (-1) No Error (+1)

 The smaller the p value, the less likelihood of committing a
type I, the greater the p value, the greater the chance of a
type II error. p values range from 0 (total significance) to
1.0 (least significance).
 Generally p values less than .05 are considered significant,
while those more than .05 are not.
Errors (continued)

 Sampling Method Used
 How was sample selected?
 What is the size of the sample?
 Were the samples related?
 Was probability sampling used? Disproportionate selection?
 What type of variables were used
 Variable Distribution among Population
 Evenly distributed?
 Judgement call
How to Select a Statistical Test

 Level of Measurement of the Independent & Dependent
Variables
 Inclusionary/ exclusionary criteria (screening mechanisms)
 Variable measurement levels (nominal, ordinal, etc.)
 Measurement precision (best measurement level used) Use
of low level measurement reduces the availability of
stronger statistical techniques.

 Statistical Power (Reduction in Type II error)
 True relationship between variables is strong not weak
 Variability of variables is small rather than large
 A higher p value is used (e.g. .1 vs .05) thereby increases
risk of Type I error
 Directional hypothesis used (one tailed)
 Large sample versus a small sample (power analysis)
 Cost effective sample just right for analysis
 Avoid too small a sample since even if the IV is effective, it
would not yield a statistically significant relationship

Week-7-Slides-Mean-Tests-Parametrics-Test-selection-module.ppt

More Related Content

Similar to Week-7-Slides-Mean-Tests-Parametrics-Test-selection-module.ppt

Recently uploaded

Week-7-Slides-Mean-Tests-Parametrics-Test-selection-module.ppt