Sei sulla pagina 1di 23

Regression Analysis

Regression analysis procedures have as


their primary purpose the development of
an equation that can be used for predicting
values on some DV for all members of a
population.
A secondary purpose is to use regression
analysis as a means of explaining causal
relationships among variables.

Standard multiple regression


All IVs are entered into the
simultaneously.
The effect of each IV on the DV is
assessed as if it had been entered into the
equation after all the IVs have been
entered.
Each IV is then evaluated in terms of what
it adds to the prediction of the DV, as
specified by the regression equation.

RESEARCH QUESTIONS
How well do the two measures of control i.e.
total perceived control of internal states
(tpcoiss) and total mastery scale (tmast)
predict perceived stress?
How much variance in perceived stress
scores can be explained by scores on these
two measures of control?
Which IVs is the best predictor of perceived
stress?

Descriptive Statistics
Mean
Total perceived stress
Total Mastery
Total PCOISS

Std. Deviation

26.73

5.848

433

21.76

3.970

436

60.63

11.985

430

Correlations
Total
perceived
stress
Pearson Correlation

Sig. (1-tailed)

Total Mastery

Total PCOISS

Total perceived stress

1.000

-.612

-.581

Total Mastery

-.612

1.000

.521

Total PCOISS

-.581

.521

1.000

.000

.000

Total Mastery

.000

.000

Total PCOISS

.000

.000

Total perceived stress

433

433

426

Total Mastery

433

436

429

Total PCOISS

426

429

430

Total perceived stress

Model Summary

Model
1

R
.684(a)

Adjusted R Std. Error of


R Square
Square
the Estimate
.468

.466

a Predictors: (Constant), Total PCOISS, Total Mastery


b Dependent Variable: Total perceived stress

4.274

Model
1

Sum of
Squares

df

Mean Square

Regression

6806.728

3403.364

Residual

7725.756

423

18.264

14532.484

425

Total

a Predictors: (Constant), Total PCOISS, Total Mastery


b Dependent Variable: Total perceived stress

F
186.341

Sig.
.000(a)

Unstandardized
Coefficients

Model
1

Std.
Error

B
(Constan
50.971
t)
Total
-.625
Mastery
Total
PCOISS

-.175

Standardized
Coefficients

Beta

95% Confidence
Interval for B

1.273

Lower
Sig. Bound

Upper
Bound

40.035 .000

48.469

53.474

Collinearity
Statistics

Correlations
Zeroorder

Partial

Part

Tolera
nce

VIF

.061

-.424

.000
10.222

-.745

-.505

-.612

-.445

-.362

.729

1.372

.020

-.360

-8.660 .000

-.215

-.136

-.581

-.388

-.307

.729

1.372

Checking the assumptions - MULTICOLLINEARITY

CORRELATIONS
Check the correlation coefficients of DV with
the IVs preferably .3 and above (Tabachnick
and Fidell, 2007)
r= -.61 for Total Mastery and r=-.58 for Total
Perceived Control of Internal States
Check that the correlations between the IVs is
not too high !! Any thing less than .70 is OK to
retain (Tabachnick and Fidell, 2007)
r= .52 which is less than .70 => retain the
variables

Evaluating the model : step 2


Look in the Model Summary box and check the value
given under the heading R square.
This tells you how much of the variance in the
dependent variable (perceived stress) is explained by
the model (which includes the variables of Total
Mastery and Total PCOISS).
In this case, the value is 0.468.
Expressed as a percentage (multiply by 100,by shifting
the decimal point two places to the right), this means
that our model ( which includes Mastery and PCOISS)
explains 46.8 per cent of the variance in perceived
stress.

This is quite a respectable result (particularly when you


compare it to some of the results that are reported in
the journals).
Adjusted R square value in the output.
When a small sample is involved, the R square value in
the sample tends to be a rather optimistic
overestimation of the true value in the population (see
Tabachnick & Fidell, 2007)
The adjusted R square statistic corrects this value to
provide a better estimate of the true population value.

If you have a small sample, you may wish to consider


reporting this value, rather than the normal R square
value.
To assess the statistical significance of the result, it is
necessary to look in the table labeled ANOVA.
This tests the null hypothesis that multiple R in the
population equals 0.
The model in this example reaches statistical
significance (Sig. = 0.000, this really means p< 0.0005).

Evaluating each of the independent variables STEP3


To determine which of the variables included in the
model contributed to the prediction of the dependent
variable.
We find this information in the output box labeled
Coefficients.
Look in the column labeled Beta under Standardized
Coefficients.
To compare the different variables, it is important that
you look at the standardized coefficient, not the
unstandardized ones.

Standardized means that these values for each of the


different variables have been converted to the same
scale so that you can compare them.
If you were interested in constructing a regression
equation, you would use the unstandardized
coefficients values listed as B.
In this case, we are interested in comparing the
contribution of each independent variable; therefore we
will use the beta values.
Look down the Beta column and find which beta value
is the largest (ignoring any negative signs out the front)
In this case, the largest beta coefficient is 0.42, which
is for Total Mastery.

This means that this variables makes the strongest


unique contribution to explaining the dependent
variable, when the variance explained by all other
variables in the model is controlled for.
The Beta value for Total PCOISS was slightly lower (.36), indicating that it made less of a contribution.
For each of these variables, check the value in the
column marked Sig.
This tells you whether this variables is making a
statistically significant unique contribution to the
equation.

This is very dependent on which variables are included


in the equation and how much overlap there is among
the independent variables.
If the Sig. value is less than 0.05 (.01,.001,etc), the
variable is making a significant unique contribution to
the prediction of the dependent variable.
If greater than .05, you can conclude that variable is not
making a significant unique contribution to the
prediction of your dependent variable.
This may be due to overlap with other independent
variables in the model.

In this case, both Total Mastery and Total PCOISS made


a unique, and statistically significant, contribution to
the prediction of perceived stress scores.
To other potentially useful piece of information in the
coefficients table is the Part correlation coefficients.

The total R square value, however, includes the unique


variance explained by each variables and also that
shared.
In this case, the two independent variables are
reasonably strongly correlated (r = .52 as shown in the
Correlations table); therefore there is a lot of shared
variance that is statistically removed when they are
both included in the model.
The results of the analysis presented above allow us to
answer the two questions posed at the beginning of this
section.

Our model, which includes control of external events


(Mastery) and control of internal states (PCOISS),
explains 46.68 per cent of the variance in perceived
stress.
Of these two variables, mastery makes the largest
unique the contribution (beta = - .42), although PCOISS
also made a statistically significant contribution (beta =
-.37)
The beta values obtained in this analysis can also be
used for other more practical purposes than the
theoretical model testing shown here.

The other useful information in the coefficients table is the Part


correlation coefficients.
If you square this value, you get an indication of the contribution of
that variable to the total R square.
It tell how much of the total variance in the dependent variable is
uniquely explained by that variable and how much R square would
drop if it wasnt included in the model.

OVERALL WRITE-UP
Standard multiple regression was conducted to
determine which independent variables (Total Perceived
Control of External Events and Total Perceived Control
of Internal States) were predictors of perceived stress in
the population.
Regression results indicate an overall model of two
predictors (internal and external control) significantly
predict perceived stress, R2 = .468, R2 adj = .466, F(2,
423) = 186.341, p < .001

OVERALL WRITE-UP
Therefore the model which includes Total Perceived
Control of External Events and Total Perceived Control
of Internal States explains 46.8 per cent of the variance
in perceived stress. Of these two variables, control of
external events makes the largest unique contribution
(beta= -.424), although the control for internal states
also made a statistically significant contribution (beta= .360).

Potrebbero piacerti anche