SPSS Workshop - FHSS Research Support Center

Report
SPSS Workshop
Research Support Center
Chongming Yang
Causal Inference
• If A, then B, under condition C
• If A, 95% Probability B, under condition C
Student T Test
(William S. Gossett’s pen name = student)
• Assumptions
– Small Sample
– Normally Distributed
• t distributions: t = [ x - μ ] / [ s / sqrt( n ) ]
df = degrees of freedom=number of
independent observations
Type of T Tests
• One sample
– test against a specific (population) mean
• Two independent samples
– compare means of two independent samples that
represent two populations
• Paired
– compare means of repeated samples
One Sample T Test
• Conceputally convert sample mean to t score
and examine if t falls within acceptable region
of distribution
x u
t 
s
n
Two Independent Samples
t
x1  x2
(n1  1)s  (n2  1)s 1 1
(  )
n1  n2  2
n1 n2
2
1
2
2
Paired Observation Samples
• d = difference value between first and second
observations
t 
d
Sd
n
Multiple Group Issues
• Groups A B C comparisons
– AB AC BC
– .95 .95 .95
• Joint Probability that one differs from another
– .95*.95*.95 = .91
Analysis of Variance
(ANOVA)
• Completely randomized groups
• Compare group variances to infer group mean
difference
• Sources of Total Variance
– Within Groups
– Between Groups
SSB
df1
F 
• F distribution
SSW
– SSB = between groups sum squares
df 2
– SSW = within groups sum squares
Fisher-Snedecor Distribution
F Test
• Null hypothesis: 1 = 2 = 3 . . . = 
• Given df1 and df2, and F value,
• Determine if corresponding probability is
within acceptable distribution region
Issues of ANOVA
• Indicates some group difference
• Does not reveal which two groups differ
• Needs other tests to identify specific group
difference
– Hypothetical comparisons Contrast
– No Hypothetical comparisons Post Hoc
• ANOVA has been replaced by multiple
regressions, which can also be replaced by
General Linear Modeling (GLM)
Multiple Linear Regression
• Causes  cab be continuous or categorical
• Effect  is continuous measure
y  0  1x1  2 x2  3 x3...k xk  
• Mild causal terms  predictors
• Objective  identify important 
Assumptions of Linear Regression
•
•
•
•
Y and X have linear relations
Y is continuous or interval & unbounded
expected or mean of  = 0
 = normally distributed
not correlated with predictors
• Predictors should not be highly correlated
• No measurement error in all variables
Least Squares Solution
• Choose 0 , 1 , 2 , 3 , . . .  to minimize the
sum of square of difference between observed
 and model estimated/predicted 
ˆ
(
y

y
)
 i i
2
• Through solving many equations
Explained Variance in 
(yi )
2
y 
 ( yi  yˆi )
2
n
R 
2
2 (yi )
yi 
n
2
2
i
Standard Error of 
( yi  yiˆ )
1
SE 
2
2
n  k 1 ( xi  xi ) (1  R )
2
T Test significant of 
• t =  / SE
• If t > a critical value & p <.05
• Then  is significantly different from zero
Confidence Intervals of 
Standardized Coefficient
()
• Make s comparable among variables on the
same scale (standardized scores)
stdx
 eta  
stdy
Interpretation of 
• If x increases one unit, y increases  unit,
given other values of X
Model Comparisons
• Complete Model:
y  0  1 x1  2 x2  3 x3 ...k xk  
• Reduced Model:
y  0  1 x1  2 x2 ... g xg  
• Test F = Msdrop / MSE
– MS = mean square
– MSE = mean square error
Variable Selection
• Select significant from a pool of predictors
• Stepwise undesirable, see
http://en.wikipedia.org/wiki/Stepwise_regression
• Forward
• Backward (preferable)
Dummy-coding of Nominal 
• R = Race(1=white, 2=Black, 3=Hispanic, 4=Others)
R
1
1
2
2
3
3
4
4
d1 d2
1 0
1 0
0 1
0 1
0 0
0 0
0 0
0 0
d3
0
0
0
0
1
1
0
0
• Include all dummy variables in the model, even if
not every one is significant.
Interaction
y  0  1x1  2 x2  3 x3  4 x2 x3...k xk  
• Create a product term X2X3
• Include X2 and X3 even effects are not
significant
• Interpret interaction effect: X2 effect depends
on the level of X3.
Plotting Interaction
• Write out model with main and interaction
effects,
• Use standardized coefficient
• Plug in some plausible numbers of interacting
variables and calculate y
• Use one X for X dimension and Y value for the Y
dimension
• See examples
http://frank.itlab.us/datamodel/node104.html
Diagnostic
• Linear relation of predicted and observed
(plotting
• Collinearity
• Outliers
• Normality of residuals (save residual as new
variable)
Repeated Measures
(MANOVA, GLM)
•
•
•
•
•
Measure(s) repeated over time
Change in individual cases (within)?
Group differences (between, categorical x)?
Covariates effects (continuous x)?
Interaction between within and between
variables?
Assumptions
• Normality
• Sphericity: Variances are equal across groups
so that
• Total sum of squares can be partitioned more
precisely into
– Within subjects
– Between subjects
– Error
Model
yij     i   j   ij   ij
•  = grand mean
•  = constant of individual i
•  = constant of jth treatment
•  = error of i under treatment j
•  = interaction
F Test of Effects
• F = MSbetween / Mswithin
(simple repeated)
• F = Mstreatment / Mserror
(with treatment)
• F = Mswithin / Msinteraction
(with interaction)
Four Types Sum-Squares
•
•
•
•
Type I  balanced design
Type II  adjusting for other effects
Type III  no empty cell unbalanced design
Type VI  empty cells
Exercise
• http://www.ats.ucla.edu/stat/spss/seminars/R
epeated_Measures/default.htm
• Copy data to spss syntax window, select and
run
• Run Repeated measures GLM

similar documents