The Data http coreecuedupsycwuenschkSPSSSPSSDatahtm CorrRegr See Correlation and Regression Analysis SPSS Masters Thesis Mike Sage 2015 Cyberloafing Age Conscientiousness ID: 246662
Download Presentation The PPT/PDF document "Correlation & Regression" is the property of its rightful owner. Permission is granted to download and print the materials on this web site for personal, non-commercial use only, and to display it on your personal computer provided you do not modify the materials and that you retain all copyright notices contained in the materials. By downloading content from our website, you accept the terms of this agreement.
Slide1
Correlation & RegressionSlide2
The Data
http://
core.ecu.edu/psyc/wuenschk/SPSS/SPSS-Data.htm
Cyberloafing
See
Correlation and Regression Analysis:
SPSS
Master’s Thesis, Mike Sage, 2015
Cyberloafing
= Age
, ConscientiousnessSlide3
Analyze, Correlate, BivariateSlide4
Pearson Correlations
Cyberloafing
Age
Conscientiousness
Cyberloafing
Pearson Correlation
1
-.462**-.563**Sig. (2-tailed) .001.000N515151AgePearson Correlation-.462**1.143Sig. (2-tailed).001 .317N515151ConscientiousnessPearson Correlation-.563**.1431Sig. (2-tailed).000.317 N515151**. Correlation is significant at the 0.01 level (2-tailed).Slide5
Spearman Correlations
Cyberloafing
Age
Conscientiousness
Spearman's rho
Cyberloafing
Correlation Coefficient
1.000-.431**-.551**Sig. (2-tailed)..002.000N515151AgeCorrelation Coefficient-.431**1.000.110Sig. (2-tailed).002..442N515151ConscientiousnessCorrelation Coefficient-.551**.1101.000Sig. (2-tailed).000.442.N5151
51
**. Correlation is significant at the 0.01 level (2-tailed).Slide6
Analyze, Regression, LinearSlide7
StatisticsSlide8
PlotsSlide9
r
= .1 is small, .3 medium, .5 large
Model
Summary
b
Model
R
R Square
Adjusted R SquareStd. Error of the Estimate1.563a.317.3037.677a. Predictors: (Constant), Conscientiousnessb. Dependent Variable: CyberloafingSlide10
ANOVA
a
Model
Sum of Squares
df
Mean Square
F
Sig.
1Regression1339.80111339.80122.736.000bResidual2887.5324958.929 Total4227.333
50
a. Dependent Variable:
Cyberloafing
b. Predictors: (Constant), ConscientiousnessSlide11
Coefficients
a
Model
Unstandardized Coefficients
Standardized Coefficients
t
Sig.
BStd. ErrorBeta 1(Constant)57.0397.288 7.826.000 Conscientiousness-.864.181-.563-4.768.000 a. Dependent Variable: CyberloafingCyberloafing = 57.039 - .864(Conscientiousness) + errortConsc. = .864/.181 = 4.77 = SQRT(22.736) = SQRT(F)Slide12
Residuals HistogramSlide13
Graphs, Scatter, Simple, DefineSlide14
Chart Editor, Elements, Fit Line at Total, Method = Linear, CloseSlide15Slide16Slide17Slide18
Construct a Confidence Interval for
the calculator at VassarSlide19
Trivariate
AnalysisSlide20
StatisticsSlide21
PlotsSlide22
R2
Adding Age increased
R
2
from .317 to .466.
Model
R
R Square
Adjusted R Square1.682a.466.443Slide23
ANOVA
ANOVA
a
Model
Sum of Squares
df
Mean Square
F
Sig.1Regression1968.0292984.01520.906.000bResidual2259.3044847.069 Total4227.33350 Slide24
Coefficients
Model
Unstandardized Coefficients
B
Std. Error
1
(Constant)
64.066
6.792Conscientiousness-.779.164Age-.276.075Slide25
Unstandardized Coefficients
Cyberloaf = 64.07 -.78
Consc
- .28 Age
When
Consc
and Age = 0, Cyber = 64.07
Holding Age constant, each one point increase in
Consc produces a .78 point decrease in Cyberloafing.Holding Consc constant, each one point increase in Age produces a .28 point decrease in Cyberloafing.Slide26
How Large are these Effects?
Is a .78 drop in Cyberloafing
a big drop or a small drop?
When the units of measurement are arbitrary and not very familiar to others, best to standardize the coefficients to mean 0, standard deviation 1.
Z
Cyber
= 0 +
1Consc + 2Age Slide27
More Coefficients
t
Sig.
Correlations
Beta
Zero-orderPartialPartConstant 9.433.000 Conscie-.507-4.759.000-.563-.566-.502Age-.389-3.653.001-.462-.466-.386Slide28
Beta Weights
ZCyber
= 0
-.51
Consc - .39Age
Holding Age constant, each one
SD
increase in Conscientiousness produces a .51 SD decrease in CyberloafingHolding Conscientiousness constant, each one SD increase in Age produces a .39 SD decrease in Cyberloafing.Slide29
Semi-Partial Correlations
The correlation between all of Cyberloafing
and that part of Conscientiousness that is not related to Age = -.50.
The
correlation
between all
of
Cyberloafing
and that part of Age that is not related to Conscientiousness = -.39.Slide30
Partial Correlations
The correlation between that part of Cyberloafing
that is not related to Age and that part of Conscientiousness that is not related to Age = -.57.
The correlation between that part of
Cyberloafing
that is not related to Conscientiousness
and that part of
Age that is not related to Conscientiousness= -.47.Slide31
Multicollinearity
The R
2
between
any one
predictor and the remaining predictors is very
high.
Makes the solution unstable.
Were you to repeatedly get samples from the same population, the regression coefficients would vary greatly among samplesSlide32
Collinearity Diagnostics
Tolerance
, which is simply 1 minus
the
R
2
between one predictor and the remaining predictors. Low (.1) is troublesome.
VIF, the Variance Inflation Factor, is the reciprocal of tolerance. High (10) is troublesome.Slide33
Coefficients
a
Model
Collinearity Statistics
Tolerance
VIF
1
Age
.9801.021Conscientiousness.9801.021Slide34
Residuals
Residuals
Statistics
a
Minimum
Maximum
Mean
Std. DeviationNPredicted Value10.2235.4122.676.27451Residual-17.34415.153.0006.72251Std. Predicted Value-1.9832.032.0001.00051Std. Residual-2.5282.209.000.98051No standardized residuals beyond 3 SD.Slide35
Residuals HistogramSlide36
Residuals PlotSlide37
Put a CI on R
2
http://
core.ecu.edu/psyc/wuenschk/SPSS/SPSS-Programs.htm
CI-R2-SPSS.zip
-- Construct Confidence Interval for
R
2
from regression analysisUsing SPSS to Obtain a Confidence Interval for R2 From Regression -- instructionsNoncF.sav -- necessary data fileF2R2.sps -- see Smithson's WorkshopNoncF3.sps -- syntax fileSlide38
Open NoncF.sav
Enter the observed value of
F
and degrees of freedom.Slide39
Open and Run the SyntaxSlide40
Look Back at .sav
FileSlide41
Why You Need Inspect Scatterplots
Data are at http://
core.ecu.edu/psyc/wuenschk/SPSS/Corr_Regr.sav
Four sets of bivariate data.
Bring into SPSS and Split File by “set.”Slide42
Predict Y from X in Four Different Data SetsSlide43Slide44Slide45Slide46Slide47Slide48Slide49Slide50