Literature Review
The biopsychosocial perspective was proposed by the
physician George Engle in 1988 as an alternative to the traditional biomedical
model for the analysis of health behavior. Engle argued that the traditional
biomedical model was too reductionist, and was not capable of accounting for
all the variables that influence health (Engle, 1981).
Engle’s biopsychosocial model is based on the general systems model where
humans are at one level of a multidimensional system, where analysis can be
conducted vertically by widening the scope to include larger ecosystems in the
analysis, or narrowing the analysis into the cellular or atomic level. He also
argued that a horizontal dimension of analysis of health behavior can be
conducted by factoring in time and other influences in the environment of the
same vertical level (Laureate online education, 2016).
Developed by the social psychologist Albert Bandura,
selfefficacy is a psychological concept defined as a personal judgement about
the capability of the self to innately achieve set goals (Bandura, 1977). Selfefficacy was first
approached in 1977 in Bandura’s journal article “selfefficacy” from a
psychosocial perspective. Bandura argued that selfefficacy is attained mainly
through three things: (1) what we attempt to perform, (2) what we observe
others doing, (3) how we are persuaded (Bandura, 1977). Selfefficacy is also related to
motivational psychology, which deals with what drives human behavior (Yi, Ji, & Yu, 2018). Moreover, what
drives human behavior can be biological, for example a state of sexual arousal greatly
influences behavior and impulsivity (Born, Wolvaardt, & Mcintosh, 2015). Since there are
biological, psychological, and social influences on selfefficacy and
repercussions, the biopsychosocial model is an ideal fit.
Knowledge Gap
Since selfefficacy is being analyzed through the
biopsychosocial model, it is a good fit to measure the relationship of
different biological, psychological, and social variables with selfefficacy.
In order to approach the knowledge gap, the predictor variables of confidence,
usefulness, male dominance, and teacher attitude and perception will be
compared statistically with selfefficacy in order to explore correlational
outcomes.
Statistical Requirements
Factorial analysis of variance (ANOVA)
which compares means of two or more independent variables is usually the
utilized statistical tool in experimental design. However, it can also be used
in correlational research (SPSS T., 2019). However since we have more than 3
predictive variables, a factorial ANOVA cannot be used, since it would be
increasing the chance of conducting a Type 1 error (Minitab, 2017).
Regression
analysis is another statistical tool that enables the researcher to explore the
relationships among multiple variables, where they examine the influence of the
predictor variables on the outcome variable. There are several types of
regression analyses, the most basic form is simple regression, it’s when
there’s one predictive variable and it’s called a simple regression, and
represented by a correlation coefficient denoted by r^{2} (Lund, 2018). When analyzing
multiple predictive variables, it is called multiple regression analysis,
however in order to conduct it, the predictor variables should be continuous or
dichotomous (Gallo, 2015).
Since Age, math scores, and science scores are categorical predictor variables,
they will not be included in the multiple regression analysis. Instead, three
independent simple regression analyses will be conducted in order to measure
their respective relationships with the outcome variable (selfefficacy).
Hypotheses
Gender and SelfEfficacy
H0 Î¼1 = Î¼2: There is no
significant relationship between gender and selfefficacy
H1 Î¼1 ≠ Î¼2: There is a
significant relationship between gender and selfefficacy
Math Grades and SelfEfficacy
H0 Î¼1 = Î¼2: There is no
significant relationship between math grades and selfefficacy
H1 Î¼1 ≠ Î¼2: There is a
significant relationship between math grades and selfefficacy
Science Grades and SelfEfficacy
H0 Î¼1 = Î¼2: There is no
significant relationship between science grades and selfefficacy
H1 Î¼1 ≠ Î¼2: There is a
significant relationship between science grades and selfefficacy
Confidence and SelfEfficacy
H0 Î¼1 = Î¼2: There is no
significant relationship between confidence and selfefficacy
H1 Î¼1 ≠ Î¼2: There is a
significant relationship between confidence and selfefficacy
Usefulness and SelfEfficacy
H0 Î¼1 = Î¼2: There is no
significant relationship between usefulness and selfefficacy
H1 Î¼1 ≠ Î¼2: There is a
significant relationship between usefulness and selfefficacy
Male Dominated Field and SelfEfficacy
H0 Î¼1 = Î¼2: There is no
significant relationship between male dominated field and selfefficacy
H1 Î¼1 ≠ Î¼2: There is a
significant relationship between male dominated field and selfefficacy
Tutor Attitude and Perception and SelfEfficacy
H0 Î¼1 = Î¼2: There is no
significant relationship between tutor attitude and perception and selfefficacy
H1 Î¼1 ≠ Î¼2: There is a
significant relationship between tutor attitude and perception and
selfefficacy
Simple Regression Analyses
Age Predictor Variable
Variables
Entered/Removed^{a}

Model

Variables
Entered

Variables
Removed

Method

1

Age^{b}

.

Enter

a. Dependent Variable: Self Efficacy

b. All requested variables entered.

Model Summary^{b}

Model

R

R Square

Adjusted
R Square

Std.
Error of the Estimate

1

.115^{a}

.013

.003

32.00833

a. Predictors: (Constant), Age

b. Dependent Variable: Self Efficacy

A simple linear
regression was conducted to measure the effect of age on selfefficacy.
R^{ }=
0.115 indicates weak predictive quality of age on selfefficacy
R^{2} =
0.013 age explains 1.3% of the variability of selfefficacy
ANOVA^{a}

Model

Sum of Squares

df

Mean
Square

F

Sig.

1

Regression

1341.447

1

1341.447

1.309

.255^{b}

Residual

100404.263

98

1024.533



Total

101745.710

99




a. Dependent Variable: Self Efficacy

b. Predictors: (Constant), Age

F(98) = 1.309 ; P = 0.255 > P =
0.05
The results are not statistically
significant.
Coefficients^{a}

Model

Unstandardized
Coefficients

Standardized
Coefficients

t

Sig.

Collinearity
Statistics

B

Std.
Error

Beta

Tolerance

VIF

1

(Constant)

63.716

46.877


1.359

.177



Age

2.652

2.317

.115

1.144

.255

1.000

1.000

a. Dependent Variable: Self Efficacy

B = 2.652
indicates that for every increase of one year in age, selfefficacy increases
by 2.65 points
Î² = .115
t (98) = 1.114 ; P
= .225 > P=0.05
There
is some difference in mean scores, however the results are not statistically
significant, and thus the null hypothesis cannot be rejected.
Collinearity Diagnostics^{a}

Model

Dimension

Eigenvalue

Condition
Index

Variance
Proportions

(Constant)

Age

1

1

1.998

1.000

.00

.00

2

.002

29.256

1.00

1.00

a. Dependent Variable: Self Efficacy

Collinearity
is measure to account for the variance in the predictor variable that might
affect the significance of the findings (Saslow, 2018). The Eigenvalue = 0.002 which is close
to 0 usually indicates that the predictor variables are intercorrelated.
However, since age is the only predictor variable in this table, the Eigenvalue
is nonindicative.
Residuals Statistics^{a}


Minimum

Maximum

Mean

Std.
Deviation

N

Predicted Value

108.7972

127.3599

117.2300

3.68103

100

Residual

73.40449

61.55095

.00000

31.84626

100

Std. Predicted Value

2.291

2.752

.000

1.000

100

Std. Residual

2.293

1.923

.000

.995

100

a. Dependent Variable: Self Efficacy

The
residual plots are randomly dispersed along the X axis, and thus indicating
that regression analysis is a good model fit for analyzing the data (ST, 2019).
Math Grades Predictor Variable
Variables
Entered/Removed^{a}

Model

Variables
Entered

Variables
Removed

Method

1

Math Grade^{b}

.

Enter

a. Dependent Variable: Self Efficacy

b. All requested variables entered.

Model Summary^{b}

Model

R

R Square

Adjusted
R Square

Std.
Error of the Estimate

1

.056^{a}

.003

.007

32.17069

a. Predictors: (Constant), Math Grade

b. Dependent Variable: Self Efficacy

A simple linear
regression was conducted to measure the effect of math grades on selfefficacy.
R^{ }=
0.056 indicates weak predictive quality of math grades on selfefficacy
R^{2} =
0.003 math grades explains 0.3% of the variability of selfefficacy
ANOVA^{a}

Model

Sum of
Squares

df

Mean
Square

F

Sig.

1

Regression

320.311

1

320.311

.309

.579^{b}

Residual

101425.399

98

1034.953



Total

101745.710

99




a. Dependent Variable: Self Efficacy

b. Predictors: (Constant), Math Grade

F (98) = 0.309 ; P = 0.579 > P =
0.05
The results are not statistically
significant.
Coefficients^{a}

Model

Unstandardized
Coefficients

Standardized
Coefficients

t

Sig.

Collinearity
Statistics

B

Std.
Error

Beta

Tolerance

VIF

1

(Constant)

113.347

7.686


14.747

.000



Math Grade

1.639

2.945

.056

.556

.579

1.000

1.000

a. Dependent Variable: Self Efficacy

B = 1.63 this
indicates that for every increase of one math grade, selfefficacy increases by
1.639 points
Î² = .056
t (98) = .056 ; P
= .556 > P=0.05
There
is negligible difference in mean scores, however the results are not
statistically significant, and thus the null hypothesis cannot be rejected.
Collinearity Diagnostics^{a}

Model

Dimension

Eigenvalue

Condition
Index

Variance
Proportions

(Constant)

Math
Grade

1

1

1.908

1.000

.05

.05

2

.092

4.559

.95

.95

a. Dependent Variable: Self Efficacy

Residuals Statistics^{a}


Minimum

Maximum

Mean

Std.
Deviation

N

Predicted Value

113.3467

119.9008

117.2300

1.79874

100

Residual

70.62376

60.37625

.00000

32.00780

100

Std. Predicted Value

2.159

1.485

.000

1.000

100

Std. Residual

2.195

1.877

.000

.995

100

a. Dependent Variable: Self Efficacy

The
residual plots are randomly dispersed along the X axis, and thus indicating
that regression analysis is a good model fit for analyzing the data.
Science Grades Predictor Variable
Variables
Entered/Removed^{a}

Model

Variables
Entered

Variables
Removed

Method

1

Science Grade^{b}

.

Enter

a. Dependent Variable: Self Efficacy

b. All requested variables entered.

Model Summary^{b}

Model

R

R Square

Adjusted
R Square

Std.
Error of the Estimate

1

.035^{a}

.001

.009

32.20187

a. Predictors: (Constant), Science Grade

b. Dependent Variable: Self Efficacy

A simple linear
regression was conducted to measure the effect of science grades on
selfefficacy.
R^{ }=
0.035 indicates weak predictive quality of science grades on selfefficacy
R^{2} =
0.001 science grades explains 0.1% of the variability of selfefficacy
ANOVA^{a}

Model

Sum of
Squares

df

Mean
Square

F

Sig.

1

Regression

123.608

1

123.608

.119

.731^{b}

Residual

101622.102

98

1036.960



Total

101745.710

99




a. Dependent Variable: Self Efficacy

b. Predictors: (Constant), Science Grade

F (98) = 0.119 ; P = 0.731 > P =
0.05
The results are not statistically
significant.
Coefficients^{a}

Model

Unstandardized
Coefficients

Standardized
Coefficients

t

Sig.

Collinearity
Statistics

B

Std.
Error

Beta

Tolerance

VIF

1

(Constant)

119.839

8.214


14.590

.000



Science Grade

1.078

3.122

.035

.345

.731

1.000

1.000

a. Dependent Variable: Self Efficacy

B = 1.078 this
indicates that for every increase of one science grade, selfefficacy decreases
by 1.078 points
Î² = .035
t (98) = .345 ; P
= .731 > P=0.05
There
is negligible difference in mean scores, however the results are not
statistically significant, and thus the null hypothesis cannot be rejected.
Collinearity Diagnostics^{a}

Model

Dimension

Eigenvalue

Condition
Index

Variance
Proportions

(Constant)

Science
Grade

1

1

1.920

1.000

.04

.04

2

.080

4.897

.96

.96

a. Dependent Variable: Self Efficacy

Residuals Statistics^{a}


Minimum

Maximum

Mean

Std.
Deviation

N

Predicted Value

115.5267

119.8389

117.2300

1.11739

100

Residual

71.68278

60.39526

.00000

32.03882

100

Std. Predicted Value

1.524

2.335

.000

1.000

100

Std. Residual

2.226

1.876

.000

.995

100

a. Dependent Variable: Self Efficacy

The
residual plots are randomly dispersed along the X axis, and thus indicating
that regression analysis is a good model fit for analyzing the data.
Multiple Regression Analyses
Regression
[DataSet1] C:\Users\Roy
Riachi\Desktop\Research\University of Liverpool\03  Data Analysis For
Psychology\Week 8\Trial 1 .sav
Variables
Entered/Removed^{a}

Model

Variables
Entered

Variables
Removed

Method

1

Tutor Attitudes, Male Dominated Field, Usefulness, Confidence^{b}

.

Enter

a. Dependent Variable: Self Efficacy

b. All requested variables entered.

Model Summary^{b}

Model

R

R Square

Adjusted
R Square

Std.
Error of the Estimate

1

.567^{a}

.322

.293

26.94722

a. Predictors: (Constant), Tutor Attitudes, Male Dominated
Field, Usefulness, Confidence

b. Dependent Variable: Self Efficacy

A
multiple linear regression was conducted to measure the effect of confidence,
usefulness, male dominance, teacher attitude and perception on selfefficacy.
R^{ }=
0.567 indicates strong predictive quality of math grades on selfefficacy
R^{2} =
0.322 confidence, usefulness, male dominance, teacher attitude and perception
explains 32.2% of the variability of selfefficacy. The low value of R^{2}
does not indicate a goodness of fit for the model.
ANOVA^{a}

Model

Sum of
Squares

df

Mean
Square

F

Sig.

1

Regression

32761.211

4

8190.303

11.279

.000^{b}

Residual

68984.499

95

726.153



Total

101745.710

99




a. Dependent Variable: Self Efficacy

b. Predictors: (Constant), Tutor Attitudes, Male Dominated
Field, Usefulness, Confidence

F (95) = 11.279; P = 0.000 < P =
0.05
The results are statistically
significant.
Coefficients^{a}

Model

Unstandardized
Coefficients

Standardized
Coefficients

t

Sig.

Collinearity
Statistics

B

Std.
Error

Beta

Tolerance

VIF

1

(Constant)

219.800

40.157


5.474

.000



Confidence

1.132

.469

.346

2.411

.018

.347

2.882

Usefulness

.448

.368

.142

1.217

.227

.525

1.906

Male Dominated Field

.300

.482

.056

.622

.536

.869

1.150

Tutor Attitudes

.631

.610

.124

1.035

.304

.494

2.022

a. Dependent Variable: Self Efficacy

Confidence
B = 1.132 this
indicates that for every increase of one point in confidence score,
selfefficacy decreases by 1.132 points
Î² = .346
t (98) = 2.411 ;
P = .018 < P=0.05
There
is a difference in mean scores, and the results are statistically significant,
and thus the alternative hypothesis can be accepted.
Usefulness
B = .448 this
indicates that for every increase of one point in usefulness score,
selfefficacy decreases by .448 points
Î² = .142
t (98) = 1.217 ;
P = .227 > P=0.05
There
is some difference in mean scores, however the results are not statistically
significant, and thus the null hypothesis cannot be rejected.
Male
Dominated Field
B = .300 this
indicates that for every increase of one point in male dominated field,
selfefficacy decreases by 0.3 points
Î² = .056
t (98) = .662 ; P
= .536 > P=0.05
There
is minor difference in mean scores, however the results are not statistically
significant, and thus the null hypothesis cannot be rejected.
Tutor
Attitudes
B = .631 this
indicates that for every increase of one point in tutor attitude, selfefficacy
decreases by .631 points
Î² = .124
t (98) = 1.035 ;
P = .304 > P=0.05
There
is some difference in mean scores, however the results are not statistically
significant, and thus the null hypothesis cannot be rejected.
Collinearity Diagnostics^{a}

Model

Dimension

Eigenvalue

Condition
Index

Variance
Proportions

(Constant)

Confidence

Usefulness

Male
Dominated Field

Tutor
Attitudes

1

1

4.835

1.000

.00

.00

.00

.00

.00

2

.086

7.518

.01

.05

.34

.10

.00

3

.048

10.088

.01

.03

.05

.80

.01

4

.030

12.637

.01

.58

.61

.06

.00

5

.002

50.133

.97

.33

.00

.04

.99

a. Dependent Variable: Self Efficacy

Collinearity
is measure to account for the variance in the predictor variable that might
affect the significance of the findings (Saslow, 2018). The Eigenvalues of .086 for
confidence, .048 for usefulness, .030 for male dominated field, and .002 for
tutor attitude and perception, which are close to 0 usually indicates that the
predictor variables are intercorrelated.
Residuals Statistics^{a}


Minimum

Maximum

Mean

Std.
Deviation

N

Predicted Value

66.6626

147.4701

117.2300

18.19124

100

Residual

75.09506

58.58260

.00000

26.39722

100

Std. Predicted Value

2.780

1.662

.000

1.000

100

Std. Residual

2.787

2.174

.000

.980

100

a. Dependent Variable: Self Efficacy

Charts
The regression
standardized residual is normally distributed in terms of frequency.
The residual plots
are randomly dispersed along the X axis, and thus indicating that regression
analysis is a good model fit for analyzing the data.
Discussion
After performing simple
regression analysis for the categorical predictor variables, which are age,
math scores, and science scores; we also performed a multiple regression
analysis for continuous predictor variables, which are confidence, usefulness,
male dominated field, and teacher attitude and perception. We were able to
establish a statistically significant negative correlation between confidence
and selfefficacy. However, the rest of the predictor variables were not able
to establish a correlational relationship with selfefficacy, since all of
their P values were greater than 0.05, hence lacking in statistical
significance (Dahiru, 2008).
This negative correlation
between confidence and selfefficacy is in accordance with the literature on
selfesteem. Selfesteem and confidence are positively correlated, since both
are defined by feelings of selfworth (Brummelman, Thomaes, Nelemans, & Castro, 2017). However,
motivational research indicates that selfefficacy is a better predictor of
academic performance than selfesteem and confidence, and that the latter can
negatively impact performance (Twenge & Campbell, 2017).
References
Bandura, A. (1977). Selfefficacy:
toward a unifying theory of behavioral change. Psychological Review, 191215.
Bandura, A. (1977). Social learning theory.
New Jersey: Englewood Cliffs .
Born, K., Wolvaardt, L., & Mcintosh, E.
(2015). Risky sexual behaviour of university students: Perceptions and the
effect of a sex education tool. African Journal For Physical, Health
Education, Recreation & Dance, 502518.
Brummelman, E., Thomaes, S., Nelemans, S., &
Castro, B. O. (2017). When parents praise inflates, childrens' selfesteem
deflates. Child Development, 17991809.
Dahiru, T. (2008). P – value, a true test of
statistical significance? A cautionary note. Ann Ib Postgraduate Med,
2126.
Engle, G. L. (1981). The clinical application of
the biopsychosocial model. Journal of Medicine and Philosophy, 101124.
Gallo, A. (2015, November 4). A refresher on
regression analysis . Retrieved from Harvard Business Review :
https://hbr.org/2015/11/arefresheronregressionanalysis
Laureate online education. (2016, Nevember 10).
Week 7: the biopsychosocial perspective. systems, holism and reductionism. Mind,
Brain and Behavior. Netherlands: Laureate Online Education B.V.
Lund. (2018). Multiple regression analysis
using spss statistics. Retrieved from Lund Research Ltd:
https://statistics.laerd.com/spsstutorials/multipleregressionusingspssstatistics.php
Minitab. (2017). What are type I and type II
errors? . Retrieved from Minitab:
https://support.minitab.com/enus/minitabexpress/1/helpandhowto/basicstatistics/inference/supportingtopics/basics/typeiandtypeiierror/
Saslow, E. (2018, July 11). Collinearity  what
it means, why its bad, and how does it affect other models? Retrieved from
Medium :
https://medium.com/@Saslow/collinearitywhatitmeanswhyitsbadandhowdoesitaffectothermodels94e1db984168
SPSS T. (2019). ANOVA  simple introduction.
Retrieved from SPSS Tutorials :
https://www.spsstutorials.com/anovawhatisit/
ST. (2019). Risidual analysis in regression .
Retrieved from Stat Trek:
https://stattrek.com/regression/residualanalysis.aspx
Twenge, J., & Campbell, K. (2017). Motivation
. In Personality Psychology (pp. 179209). New York : Pearson .
Yi, T., Ji, J., & Yu, F. (2018). The effect of
metacognitive knowledge on mathematics performance in selfregulated learning
framework—multiple mediation of selfefficacy and motivation. Frontiers in
Psychology,
https://doiorg.liverpool.idm.oclc.org/10.3389/fpsyg.2018.02518.
Appendix
Age Regression Syntax
REGRESSION
/MISSING LISTWISE
/STATISTICS COEFF OUTS R ANOVA COLLIN TOL
/CRITERIA=PIN(.05) POUT(.10)
/NOORIGIN
/DEPENDENT SE
/METHOD=ENTER AGE
/PARTIALPLOT ALL
/SCATTERPLOT=(*ZRESID ,*ZPRED)
/RESIDUALS HISTOGRAM(ZRESID) NORMPROB(ZRESID).
Math Grades Regression Syntax
REGRESSION
/MISSING LISTWISE
/STATISTICS COEFF OUTS R ANOVA COLLIN TOL
/CRITERIA=PIN(.05) POUT(.10)
/NOORIGIN
/DEPENDENT SE
/METHOD=ENTER MATH
/PARTIALPLOT ALL
/SCATTERPLOT=(*ZRESID ,*ZPRED)
/RESIDUALS HISTOGRAM(ZRESID) NORMPROB(ZRESID).
Science Grades Syntax
REGRESSION
/MISSING LISTWISE
/STATISTICS COEFF OUTS R ANOVA COLLIN TOL
/CRITERIA=PIN(.05) POUT(.10)
/NOORIGIN
/DEPENDENT SE
/METHOD=ENTER SCIENCE
/PARTIALPLOT ALL
/SCATTERPLOT=(*ZRESID ,*ZPRED)
/RESIDUALS HISTOGRAM(ZRESID) NORMPROB(ZRESID).
Multiple Regression Syntax
GET
FILE='C:\Users\Roy Riachi\Desktop\Research\University of Liverpool\03 
Data Analysis For Psychology\Week 8\UKL1_LPSY_303_Week08_statsconfidence.sav'.
Warning # 5281. Command name: GET FILE
SPSS Statistics is running in
Unicode encoding mode. This file is
encoded in
a localespecific (code page)
encoding. The defined width of any
string
variables are automatically
tripled in order to avoid possible data loss.
You
can use ALTER TYPE to set the
width of string variables to the width of the
longest observed value for each string
variable.
DATASET NAME DataSet1
WINDOW=FRONT.
SAVE OUTFILE='C:\Users\Roy
Riachi\Desktop\Research\University of Liverpool\03  Data Analysis '+
'For Psychology\Week 8\Trial 1 .sav'
/COMPRESSED.
REGRESSION
/MISSING LISTWISE
/STATISTICS COEFF OUTS R ANOVA COLLIN TOL
/CRITERIA=PIN(.05) POUT(.10)
/NOORIGIN
/DEPENDENT SE
/METHOD=ENTER Confidence Usefulness Male.domain
Teacher.attitude.perceptions
/PARTIALPLOT ALL
/SCATTERPLOT=(*ZRESID ,*ZPRED)
/RESIDUALS HISTOGRAM(ZRESID) NORMPROB(ZRESID).