Friday, February 22, 2019

Biopsychosocial Analysis of the Effects of Self-Efficacy

 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, self-efficacy is a psychological concept defined as a personal judgement about the capability of the self to innately achieve set goals (Bandura, 1977). Self-efficacy was first approached in 1977 in Bandura’s journal article “self-efficacy” from a psycho-social perspective. Bandura argued that self-efficacy 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). Self-efficacy 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 self-efficacy and repercussions, the biopsychosocial model is an ideal fit.

Knowledge Gap

Since self-efficacy is being analyzed through the biopsychosocial model, it is a good fit to measure the relationship of different biological, psychological, and social variables with self-efficacy. 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 self-efficacy 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 r2 (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 (self-efficacy).

Hypotheses

Gender and Self-Efficacy

H0 μ1 = μ2:  There is no significant relationship between gender and self-efficacy  

H1 μ1 ≠ μ2:  There is a significant relationship between gender and self-efficacy  

Math Grades and Self-Efficacy

H0 μ1 = μ2:  There is no significant relationship between math grades and self-efficacy  

H1 μ1 ≠ μ2:  There is a significant relationship between math grades and self-efficacy  

Science Grades and Self-Efficacy

H0 μ1 = μ2:  There is no significant relationship between science grades and self-efficacy  

H1 μ1 ≠ μ2:  There is a significant relationship between science grades and self-efficacy  

Confidence and Self-Efficacy

H0 μ1 = μ2:  There is no significant relationship between confidence and self-efficacy  

H1 μ1 ≠ μ2:  There is a significant relationship between confidence and self-efficacy  

Usefulness and Self-Efficacy

H0 μ1 = μ2:  There is no significant relationship between usefulness and self-efficacy  

H1 μ1 ≠ μ2:  There is a significant relationship between usefulness and self-efficacy  

Male Dominated Field and Self-Efficacy

H0 μ1 = μ2:  There is no significant relationship between male dominated field and self-efficacy  

H1 μ1 ≠ μ2:  There is a significant relationship between male dominated field and self-efficacy  

Tutor Attitude and Perception and Self-Efficacy

H0 μ1 = μ2:  There is no significant relationship between tutor attitude and perception and self-efficacy  

H1 μ1 ≠ μ2:  There is a significant relationship between tutor attitude and perception and self-efficacy  

Simple Regression Analyses

Age Predictor Variable

 

Variables Entered/Removeda

Model

Variables Entered

Variables Removed

Method

1

Ageb

.

Enter

a. Dependent Variable: Self Efficacy

b. All requested variables entered.

 

 

 

Model Summaryb

Model

R

R Square

Adjusted R Square

Std. Error of the Estimate

1

.115a

.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 self-efficacy.

R = 0.115 indicates weak predictive quality of age on self-efficacy

R2 = 0.013 age explains 1.3% of the variability of self-efficacy

 

ANOVAa

Model

Sum of Squares

df

Mean Square

F

Sig.

1

Regression

1341.447

1

1341.447

1.309

.255b

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.

 

Coefficientsa

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, self-efficacy increases by 2.65 points

β = .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 Diagnosticsa

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 non-indicative.

 

Residuals Statisticsa

 

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/Removeda

Model

Variables Entered

Variables Removed

Method

1

Math Gradeb

.

Enter

a. Dependent Variable: Self Efficacy

b. All requested variables entered.

 

 

Model Summaryb

Model

R

R Square

Adjusted R Square

Std. Error of the Estimate

1

.056a

.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 self-efficacy.

R = 0.056 indicates weak predictive quality of math grades on self-efficacy

R2 = 0.003 math grades explains 0.3% of the variability of self-efficacy

 

 

ANOVAa

Model

Sum of Squares

df

Mean Square

F

Sig.

1

Regression

320.311

1

320.311

.309

.579b

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.

 

 

Coefficientsa

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, self-efficacy increases by 1.639 points

β = .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 Diagnosticsa

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 Statisticsa

 

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/Removeda

Model

Variables Entered

Variables Removed

Method

1

Science Gradeb

.

Enter

a. Dependent Variable: Self Efficacy

b. All requested variables entered.

 

 

Model Summaryb

Model

R

R Square

Adjusted R Square

Std. Error of the Estimate

1

.035a

.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 self-efficacy.

R = 0.035 indicates weak predictive quality of science grades on self-efficacy

R2 = 0.001 science grades explains 0.1% of the variability of self-efficacy

 

 

ANOVAa

Model

Sum of Squares

df

Mean Square

F

Sig.

1

Regression

123.608

1

123.608

.119

.731b

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.

 

 

Coefficientsa

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, self-efficacy decreases by 1.078 points

β = -.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 Diagnosticsa

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 Statisticsa

 

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/Removeda

Model

Variables Entered

Variables Removed

Method

1

Tutor Attitudes, Male Dominated Field, Usefulness, Confidenceb

.

Enter

a. Dependent Variable: Self Efficacy

b. All requested variables entered.

 

 

Model Summaryb

Model

R

R Square

Adjusted R Square

Std. Error of the Estimate

1

.567a

.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 self-efficacy.

R = 0.567 indicates strong predictive quality of math grades on self-efficacy

R2 = 0.322 confidence, usefulness, male dominance, teacher attitude and perception explains 32.2% of the variability of self-efficacy. The low value of R2 does not indicate a goodness of fit for the model.

 

ANOVAa

Model

Sum of Squares

df

Mean Square

F

Sig.

1

Regression

32761.211

4

8190.303

11.279

.000b

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.


 

Coefficientsa

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, self-efficacy decreases by -1.132 points

β = -.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, self-efficacy decreases by -.448 points

β = -.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, self-efficacy decreases by 0.3 points

β = -.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, self-efficacy decreases by .631 points

β = -.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 Diagnosticsa

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 Statisticsa

 

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 self-efficacy. However, the rest of the predictor variables were not able to establish a correlational relationship with self-efficacy, since all of their P values were greater than 0.05, hence lacking in statistical significance (Dahiru, 2008).

This negative correlation between confidence and self-efficacy is in accordance with the literature on self-esteem. Self-esteem and confidence are positively correlated, since both are defined by feelings of self-worth (Brummelman, Thomaes, Nelemans, & Castro, 2017). However, motivational research indicates that self-efficacy is a better predictor of academic performance than self-esteem and confidence, and that the latter can negatively impact performance (Twenge & Campbell, 2017).


 

References

Bandura, A. (1977). Self-efficacy: toward a unifying theory of behavioral change. Psychological Review, 191-215.

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, 502-518.

Brummelman, E., Thomaes, S., Nelemans, S., & Castro, B. O. (2017). When parents praise inflates, childrens' self-esteem deflates. Child Development, 1799-1809.

Dahiru, T. (2008). P – value, a true test of statistical significance? A cautionary note. Ann Ib Postgraduate Med, 21-26.

Engle, G. L. (1981). The clinical application of the biopsychosocial model. Journal of Medicine and Philosophy, 101-124.

Gallo, A. (2015, November 4). A refresher on regression analysis . Retrieved from Harvard Business Review : https://hbr.org/2015/11/a-refresher-on-regression-analysis

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/spss-tutorials/multiple-regression-using-spss-statistics.php

Minitab. (2017). What are type I and type II errors? . Retrieved from Minitab: https://support.minitab.com/en-us/minitab-express/1/help-and-how-to/basic-statistics/inference/supporting-topics/basics/type-i-and-type-ii-error/

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/collinearity-what-it-means-why-its-bad-and-how-does-it-affect-other-models-94e1db984168

SPSS T. (2019). ANOVA - simple introduction. Retrieved from SPSS Tutorials : https://www.spss-tutorials.com/anova-what-is-it/

ST. (2019). Risidual analysis in regression . Retrieved from Stat Trek: https://stattrek.com/regression/residual-analysis.aspx

Twenge, J., & Campbell, K. (2017). Motivation . In Personality Psychology (pp. 179-209). New York : Pearson .

Yi, T., Ji, J., & Yu, F. (2018). The effect of metacognitive knowledge on mathematics performance in self-regulated learning framework—multiple mediation of self-efficacy and motivation. Frontiers in Psychology, https://doi-org.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 locale-specific (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).