Chapter 4: Findings
The purpose of this quantitative
descriptive and correlational study was to assess the key variables posed in the
TPB model as possible determinants of parent intention for school involvement
behavior (i.e., parental attitudes and beliefs, subjective norms, and perceived
behavioral controls) and to ascertain whether they are significantly related to
and can predict the reported intentions of involvement of immigrant and refugee
parents in their children’s early childhood education programs. The goal was to
examine the constructs that are pivotal to TPB and a test of this theory within
the context of immigrant and refugee families with children enrolled in Head
Start or Early Head Start. The predictor variables of attitudes and beliefs,
subjective norms, and perceived behavioral controls were examined in hopes of
helping devise interventions in the immigrant and refugee communities to
improve their intention for involvement in children’s education at an early age
as these variables are hypothesized to contribute to parental intentions to be
involved in their children’s education.
The Parent Involvement Project
(PIP) survey was the instrument used for data collection. The questionnaire
contains 57 items. All items were measured on a six-point Likert scale with subscales
for attitudes and beliefs, subjective norms, perceived behavioral control, and
parental intentions for involvement. Data obtained from the participants were
analyzed through the Statistical Package for the Social Science (SPSS) program.
The parent participants were Arabic, Hispanic, Vietnamese, and English
speaking. While some were English language learners, some spoke only their
native languages. A total of 122 parents participated in the study.
This chapter presents the findings
with regard to the research questions and hypotheses. First, Cronbach’s alphas
were used to compute and determine the internal consístency reliability of
each construct and subscale. The data was examined to identify missing data and
outliers and check all assumptions required for the inferential statistics to
insure they were met prior to running the required analyses. The results of
these first two steps are presented in the reliability and validity of data
section. Next, the results section begins with descriptive statistics and
demographics of the participants, including the calculations of means, standard
deviations, and ranges for attitudes and beliefs, subjective norms, perceived
behavior controls, and parental intentions. Then, the findings for each
research question are presented. Finally, the evaluation of the findings is
discussed followed by a brief summary.
Reliability
and Validity of the Data
Cronbach’s α provides a measure of
the overall reliability of a set of items creating a subscale; values 0.70 or
greater are considered acceptable (Field, 2013). The previous or original
reliability coefficients obtained from the developers of the subscales were (a)
attitudes and beliefs (0.77), (b) subjective norms (0.88), (c) perceived
behavioral controls (0.83), and (d) parental intentions (0.78) (Hoover-Dempsey,
Sandler, & Walker, 2005). The subscale reliabilities for the present study
are in Table 1. All four subscales had acceptable internal consistency: (a)
attitudes and beliefs (α = 0.84), (b) subjective norms (α = 0.71), (c)
perceived behavioral controls (α = 0.91), and (d) parental intentions (α =
0.79). The current findings align well with previous estimates.
Table
1
PIP
Subscale Reliabilities
Composite
Reliability for the Averaged Constructs Subscale Scores.
|
Subscale |
|
Cronbach's Alpha |
|
||
|
Hoover - Dempsey et al. |
Current Study |
|
Number of Items |
||
|
Attitudes and Beliefs Subjective Norms Perceived Behavioral Controls Parental Intentions |
.770 .880 .830 .780 |
.844 .711 .907 .793 |
24 6 17 10 |
||
The statistical assumptions of
multiple regressions were tested before conducting the analysis and
interpretation of the findings. The normality of the continuous predictors
(Attitudes and Beliefs, Subjective Norms, and Perceived Behavioral Control) and
outcome (Parental Intentions) variables were tested with skewness and kurtosis
statistics. Correlations were run between the predictor variables to assess
multicollinearity, as well as Variance Inflation Factor (VIF) and Tolerance
statistics being computed. Scatterplots of the predictor variables against the
outcome variable were used to assess the assumption of linearity. Durbin-Watson
statistics were used to check for the assumption of autocorrelation. Normality
of residuals was assessed using a histogram, and homoscedasticity was tested
using a P-P plot of the standardized residuals. When assumptions were
met, the predictor variables were entered into the regression model in a
simultaneous fashion. The F test was used to check for the increase in
shared variance (
) associated with entering the predictor variables
into the model. Unstandardized beta coefficients with standard errors, as well
as standardized beta coefficients were reported and interpreted for the
regression model. All analyses were conducted using SPSS version 22 and
statistical significance was assumed at an alpha value of 0.05.
A Shapiro-Wilk’s test (p > .05)
and a visual inspection of histogram, normal P-P, Q-Q, and
box-plots showed that the value scores were normally distributed for the samples.
Univariate normality was met for the three predictor variables and the outcome
variable. Correlations between the predictor variables were acceptable.
Scatterplots shows linear relationships with the outcome. The Durbin-Watson
statistic was 2.024, meaning that an autocorrelation was not likely. VIF and
Tolerance statistics were in an acceptable range. Interpretation of them was
therefore undertaken. All of the subscales were normally distributed as per the
skewness and kurtosis statistics being below an absolute value of 2.0 (Field,
2013).
Table 2
The Shapiro-Wilk Test of Normality
|
Tests of Normality |
|
|||||||
|
|
Kolmogorov-Smirnova |
Shapiro-Wilk |
|||||
|
Statistic |
Df |
Sig. |
Statistic |
df |
Sig. |
||
|
Attitudes and Beliefs (AB) |
.219 |
24 |
.200* |
.925 |
24 |
.542 |
|
|
Subjective Norms (SN) |
.199 |
6 |
.200* |
.916 |
6 |
.475 |
|
|
Perceived Behavioral Control
(PBC) |
.223 |
17 |
.200* |
.931 |
17 |
.586 |
|
|
Parental Intentions (PI) |
.200 |
10 |
.200* |
.926 |
10 |
.546 |
|
|
*. This is a lower bound of
the true significance. |
|||||||
|
a. Lilliefors Significance
Correction |
|||||||
The results of Shapiro-Wilks test for normality in Table 2
indicate that all variables had probability values greater than 0.005 (p
> .05) (Field, 2013); thus, the measured variables in this sample were not
significantly different from a normal distribution with attitudes and beliefs
scores, AB (24) =
0.219, p =.542; subjective norm, SN (6) = 0.199, p = .475;
perceived behavioral control, PBC (17) = 0.223, p = .586; and the
parental intentions, PI (10)=0.200, p = .546.
Table 3
Descriptive
Statistics of Predictors and Dependent Variable
|
|
N |
Mean |
Std. Deviation |
Skewness |
Kurtosis |
||
|
Statistic |
Statistic |
Statistic |
Statistic |
Std. Error |
Statistic |
Std. Error |
|
|
Attitudes
& Beliefs |
122 |
107.0164 |
14.04066 |
.046 |
.219 |
.572 |
.435 |
|
Subjective
Norms |
122 |
30.9754 |
3.51023 |
-.902 |
.219 |
1.825 |
.435 |
|
Perceived
Behavioral Control |
122 |
88.3361 |
9.12956 |
-.676 |
.219 |
.783 |
.435 |
|
Parental
Intentions |
122 |
50.9426 |
5.54366 |
-.371 |
.219 |
.080 |
.435 |
|
Valid
N (listwise) |
122 |
|
|
|
|
|
|
Table 3 shows all the subscales
were normally distributed as per the skewness (AB=.046, SN=-.902, PBC=-.676,
PI=-.371) and kurtosis (AB=.572, SN=1.825, PBC=.783, PI=.080) statistics being
below an absolute value of 2.0 (Field, 2013). All the subscales were normally
distributed as per the skewness and kurtosis statistics being below an absolute
value of 2.0.
Figure 1. Parental Intentions: Normality of Residuals
The data was examined for normality of residuals. Figure 1 shows
the residuals are normally distributed.
Figure 2. Parental Intentions: P-P Plot of Residuals
Figure 2 is the P-P plot
constructed for the residual of the variable of Parental Intentions. The P-P
plots were generated to test the assumption of normality (Field, 2013).
Assumption of normality is met when points on the plot fall closely to the
diagonal line. As observed, the P-P plot met the assumption of
normality; therefore, an adequate level of normality was assumed for the sample
taken for this study.
Q-Q plots were also generated for
checking the assumption of normality. Figure 1, 2, 3, and 4 in Appendix L
provide the output. The assumption of normality is met when points on the plot
fall closely to the diagonal line of the variables. An adequate level of
normality was assumed for this study’s sample.
Figure 3. Boxplots of AB, SN, PBC,
and PI
Figure
3 displays boxplots of the four variables (Attitudes and Beliefs, Subjective
Norms, Perceived Behavioral Control, and Parental Intentions). There was one
univariate outlier (circular point) in perceived behavioral control; however,
this univariate outlier is not considered as troublesome and might be ignored
as assessed by inspection (Field, 2013). There were no other univariate
outliers in the data as assessed by inspection of a boxplot for values greater
than 2.5 box-lengths from the edge of the box. Thus, this assumption was met.
For the planned analyses, the data
must meet a number of assumptions (Field, 2013). By design, the independent and
dependent variables were measured on a continuous interval scale of
measurement. To test the independence of residuals assumption, the
Durbin-Watson statistic was computed and assessed (Field, 2013).
Table 4
Model
Summary: Predictors and Dependent Variable
|
Model Summary |
||||||||||
|
Model |
|
|
Adjusted |
Std. Error of the Estimate |
Change Statistics |
Durbin-Watson |
||||
|
|
F Change |
df1 |
df2 |
Sig. F Change |
||||||
|
1 |
|
|
.495 |
.027 |
.507 |
40.529 |
3 |
118 |
.000 |
2.024 |
|
b. Dependent Variable:
Parental Intentions |
||||||||||
In Table 4, the entry of the predictor variables accounted for a
significant increased in , F (3, 118) = 40.53, p < 0.05. The Durbin-Watson
statistic shows there was no autocorrelation. D = 2.024. Field (2013) suggested
that the values less than 1 or greater than 3 is a cause for concern. A value of
2 indicates the assumption has certainly been met.
The
next assumption tested was for linear relationships between the dependent
variable and each of the independent variables; scatterplots were generated for
examination.
Figure
4. Scatterplot for Combinations of Variables Model
The
12 scatterplots in Figure 4 indicates symmetrical distribution of data points
around a diagonal line, thus confirming the assumption of linearity (Field,
2013).
Figure
5. Partial Regression Attitudes and Beliefs Model
Figure
5 shows a linear relationship exists between the parental intentions and
attitudes and beliefs. There were no outliers observed in the plot that could
invalidate the assumption of linearity (Field, 2013); thus, the assumption of
linearity was met.
Figure
6. Partial Regression Subjective Norms Model
Figure
6 shows a linear relationship exists between the parental intentions and
subjective norms. There were no outliers observed in the plot that could
invalidate the assumption of linearity (Field, 2013); thus, the assumption of linearity
was met.
Figure
7. Partial Regression Perceived Behavioral Control Model
Figure
7 shows a linear relationship exists between the parental intentions and
perceived behavioral control. There were no outliers observed in the plot that
could invalidate the assumption of linearity (Field, 2013); thus, the
assumption of linearity was met. Figure 8 shows homoscedasticity, as assessed by visual inspection of a
plot of standardized residuals versus standardized predicted value (Field,
2013).
Figure
8. Testing for Homoscedasticity
To
check for the assumption that the data must not show multicollinearity,
diagnostics were examined. Multicollinearity
occurs when the sample has two or more independent variables that are
correlated with each other (Field, 2013). In Table 6 as below, all the
Tolerance values were greater than 0.10 (the lowest is 0.585) and the VIF
values were all below 10, so this assumption was met (Field, 2013).
Table
5
Checking
for Multicollinearity
|
Model 1 |
Unstandardized Coefficients |
Standardized Coefficients |
Sig. |
95.0% Confidence Interval for B |
Correlations |
Collinearity Statistics |
||||||
|
B |
Std. Error |
Beta |
Lower Bound |
Upper Bound |
Zero-order |
Partial |
Part |
Tolerance |
VIF |
|||
|
|
(Constant) |
9.765 |
3.864 |
|
.013 |
-18.060 |
25.168 |
|
|
|
|
|
|
AB |
.120 |
.029 |
.304 |
.000 |
-1.526 |
.729 |
-.415 |
-.732 |
-.659 |
.756 |
1.323 |
|
|
SN |
.037 |
.133 |
.023 |
.037 |
-1.895 |
4.019 |
.425 |
.738 |
.670 |
.585 |
1.711 |
|
|
PBC |
.308 |
.051 |
.507 |
.000 |
-4.183 |
3.349 |
-.011 |
-.319 |
-.207 |
.585 |
1.708 |
|
Results
Upon approval from Northcentral
University’s IRB and the AKA Head Start centers, parents were invited to
participate through the distribution of the survey packets at each center’s
gate where parents/caregivers dropped-off their children. Attempts were made to
recruit parent participants during the pick up and drop off times in order to
reach the goal of a maximum number of parents who met the criteria for
inclusion in the study. The AKA Head Start executive director and center
directors collaborated on this project by allowing access to distribute the
survey packets. Twelve sites gave permission for parents to participate. Participants
received a survey package with (1) a PIP survey, (2) a consent form, (3) a
recruitment letter, and (4) a self-stamped envelope for them to return the
survey and consent form to the researcher’s address. All surveys were completed
and collected within an eight-week time-frame.
Sample demographics. A total of 500 surveys was
distributed, and 122 parents who met the criteria for inclusion successfully
completed and returned the surveys and consent forms within the timeframe as requested.
Table 6 provides the domographics of
Table
6
Demographic
Characteristics
|
Variable n Percentage |
Variable n Percentage |
Age of Participants Schooling Parents
Completed
18 – 65 122 100% Some High School
17 13.9%
Early Head Start
22 18.0% High
School Diploma 31 25.4%
Head Start 98 80.3% Some College 31 25.4%
Gender of Participants College Degree 25 20.5%
Male 19 15.6%
Graduate Degree 8 6.6%
Female 103 84.4%
Vocational School 6 4.9%
Gender of Children None 3
3.3%
Boys 56 45.9% Marital
Status of Parents
Girls
66 54.1%
Married 64
52.5%
Parents Described as Divorced/Separated 13 10.7%
Immigrant 92 75.4% Never Married 39 32.0%
Refugee
17 13.9%
Missing 6 4.9%
Missing:
13 10.7% Relationship
to the Child
Time of Living in the USA Mother 96 78.7%
Two years 4 3.3% Father
18 14.8%
Four years 3 2.5% Grandmother 6 4.9%
Five years 6 4.9% Grandfather 1 0.8%
More than five years 101 82.8% Other
1 0.8%
Missing 8 6.6%
Home Language of
Participants
Arabic 15 12.3%
English 50 41.0%
Hispanic 50 41.0%
Vietnamese 7 5.7%
parents
who completed the surveys. All parents were between 18 and 65 years old. There
were 22 parents (18.0%) who had a child enrolled in Early Head Start and 98
parents (80.3%) with a child in Head Start. Nineteen participants were male
(15.6%) and 103 were female (84.4%). The genders of their children were 56 boys
(45.9%) and 66 girls (54.1%). Ninety-two parents described themselves as
immigrants (75.4%) and 17 as refugee (13.9%). Out of 122 parents, four parents
(3.3%) had lived in the USA for two years; three parents (2.5%) for four years;
six parents (4.9%) for five years; 101 parents (82.8%) for more than five
years; and eight parents (6.6%) who did not respond. Among those who participated
in the study, 17 parents (13.9%) completed some high school; 31 parents (25.4%)
had a high school diploma; 25 parents (20.5%) obtained a college degree; six
parents (4.9%) attended vocational school; 31 parents (25.4%) attended some
college; eight parents (6.6%) held a graduate degree; and three parents (3.3%)
did not report their education status. With regard to marital status, 64
parents (52.5%) were married; 13 parents (10.7%) were divorced or separated; 39
parents (32.0%) were never married; and six parents (4.9%) did not report. The
participants’ relationship to a child included 96 mothers (78.7%), 18 fathers
(14.8%), six grandmothers (4.9%), one grandfather (0.8%), and one other (0.8%).
The overwhelming majority of participants spoke English (41.0%), Spanish
(41.0%), 15 speaking Arabic (12.3%), and 7 Vietnamese (5.7%).
Research questions 1 through 4. The first four research questions
posed were descriptive questions and asked the levels of (1) attitudes and
beliefs (AB), (2) subjective norms (SN), (3) perceived behavioral controls
(PBC), and (4) parental intentions (PI) of immigrant/refugee parents/caregivers
regarding parent involvement in AKA Head Start and Early Head Start programs.
Table 7
Descriptive
Statistics of Predictors and Dependent Variable
|
|
N |
Mean |
Std. Deviation |
Skewness |
Kurtosis |
||
|
Statistic |
Statistic |
Statistic |
Statistic |
Std. Error |
Statistic |
Std. Error |
|
|
Attitudes
& Beliefs |
122 |
107.0164 |
14.04066 |
.046 |
.219 |
.572 |
.435 |
|
Subjective
Norms |
122 |
30.9754 |
3.51023 |
-.902 |
.219 |
1.825 |
.435 |
|
Perceived
Behavioral Control |
122 |
88.3361 |
9.12956 |
-.676 |
.219 |
.783 |
.435 |
|
Parental
Intentions |
122 |
50.9426 |
5.54366 |
-.371 |
.219 |
.080 |
.435 |
|
Valid
N (listwise) |
122 |
|
|
|
|
|
|
Table
7 presents descriptive statistics for all variables across all participants.
The mean scores derived for independent variables were attitudes and beliefs (M=107.02,
SD=14.04), subjective norms (M=30.98, SD=3.51), perceived
behavioral control (M=88.34, SD=9.13) and parental intentions in
their actual behaviors (M=50.94, SD=5.54). Given the possible
range for attitudes and beliefs from 24
to 144; the mean score is considered moderate. Given the possible range
for subjective norm is from 6 to 36; the mean score is considered moderately high.
Given the possible range for perceived behavioral control is from 17 to 102;
the mean score is considered moderate. Given the possible range for parental
intentions from 10 to 60; the mean score is considered moderately high (Field,
2013).
Research question 5. The final research question posed was: What is the relationship of measures of
attitudes and beliefs, subjective norms, and perceived behavior control, and
how well, if at all, do they predict parental intentions for parental
involvement by parents in AKA Head Start and Early Head Start programs?
Table
8
Correlation
Coefficients
|
Correlation Matrix |
|||||
|
|
Attitudes and Beliefs |
Subjective Norms |
Perceived Behavioral Controls |
Parental Intentions |
|
|
Spearman’s rho |
Attitudes and Beliefs |
|
|
|
|
|
Subjective Norms |
.982** (.000) |
|
|
|
|
|
Perceived Behavioral
Controls |
.861** (.000) |
.948** (.002) |
|
|
|
|
Parental Intentions |
.837** (.001) |
.980** (.000) |
.950** (.002) |
|
|
**. Correlation is significant
at the 0.01 level (1-tailed).
Table 8 provides the results of the
correlational analysis for the variables. The subscale of attitudes and beliefs
was significantly and positively correlated with subjective norms (r=0.982,
p=0.000), perceived behavioral control (r=0.861, p=0.000),
and parental intentions (r=0.837, p=0.001). Subjective norms and
perceived behavioral controls were significantly and positively correlated (r=0.948,
p=0.002) as were subjective norms and parental intentions (r=0.980, p=0.000).
Perceived behavioral control was positively and significant correlated with
parental intentions (r=0.950, p=0.002). Given the significant
correlations, the null hypothesis was rejected.
Results
of the Multiple Regression Analyses Predicting Parental Intentions
|
Variables t p β Intentions A&B 4.318
.000 .304 SN 9.832
.037 .023 PBC 6.006 .000 .507
|
F
df p adj. |
|
|
|
||||||
|
40.529 |
3,
118 |
0.013 |
0.495 |
|
||||||
|
|
|
|
|
|||||||
|
|
|
|
||||||||
Note: A & B = Attitudes and Beliefs, SN =
Subjective Norms, PBC = Perceived Behavioral Control
The results of the multiple
regression analysis in Table 9 demonstrate significant predictive capacity F(3,
118) = 40.529, p < 0.013 with the predictor variables accounting for 50 % of
the variance in parental intentions. All three variables contributed
significantly to the prediction model: attitudes and beliefs (t=4.318, p=0.000,
b=0.304), subjective norms (t=9.832,
p=0.037, b=0.023), and perceived behavioral
control (t=6.006, p=0.000, b=0.507). The null hypothesis was rejected.
Based on the results of
correlational and regression analyses, the null hypothesis for research
question five was rejected because significant positive correlations were found
among the variables and all three variables were statistically significant in
predicting variation in parental intentions for involvement. It is important to
note that the measure for parental intentions is moderately high for parent
involvement and there were significant and high correlation with parental
intention. This finding is not surprising because it is consistent with the
finding revealed by Hoover-Dempsey et al. (2005) who pointed to the importance
of social norms in their study of parental involvement. Perry and Langley
(2013) also found the attitudes and beliefs and subjective norms were the two
strongest predictors of parental intentions. Additionally, results from this
study are consistent to that of Kiriakidis (2015) who reported that perceived
behavioral control is significant and contribution to the prediction of
intentions above of attitudes and subjective norms. Kiriakidis (2013) also
reported the theory of planned behavior model is superior to the theory of
planned action model in predicting and understanding parent behavior after
testing the two theories in ten different behaviors. This study fits with the
extensive work they have done. The findings are also aligned with prior
research by Bracke and Corts (2012) who yielded several outcomes: (a) affirmed
parents’ positive attitudes and beliefs for their children’s education, (b)
offered support for a long-term, collaborative relationship between Early Head
Start and Head Start programs and the local community.
The findings from this study revealed
that immigrant and refugee parents’ involvement in EHS/HS programs could be
predicted by their reported attitudes and beliefs, subjective norms, and
perceived behavioral controls with regard to their involvement in their
children’s education. All three variables should be considered important
aspects in which teachers and administrators may affect parents’ intentions and
desires to be involved in their child’s schooling. The findings support that
the theory of planned behavior model can be applied to explaining parent
involvement of immigrant and refugee parents whose children are enrolled in
Head Start and Early Head Start programs, thereby expanding the theory to
encompass people with culturally and linguistically diverse backgrounds.
This
study was guided by the theory of planned behavior that stipulates the more
favorable parents’ intentions to engage in their child’s education, based on
their attitudes and beliefs, subjective norms, and perceived behavioral
control, the more likely they will actually engage in the intended behaviors
(Ajzen, 2011). This finding is particularly encouraging because it suggests
that the relatively high parental intentions are likely to have a strong
effect, influencing not only their child’s education but also leading to
parents’ level of involvement in EHS/HS programs.
Summary
A quantitative, descriptive and
correlational study was employed to assess key variables posed in the TPB based
model as possible determinants of parent intention for school involvement
behavior, i.e. parental attitudes and beliefs, subjective norms, and perceived
behavioral controls, within the immigrant and refugee population. They were
significantly related to each other and predicted the reported intentions of
involvement of immigrant and refugee parents in their children’s early
childhood education programs. By assessing the constructs that were pivotal to
TPB, a test of this theory within the context of immigrant and refugee families
with children enrolled in Head Start or Early Head Start was accomplished. In
terms of age group, parents were between 18 to 65 years old. Twenty-two parents
had a child in Early Head Start and 98 parents had a child in Head Start; 19 were
males and 103 were female parents. The genders of children were 56 boys and 66
girls. The parents described themselves as immigrants (92 parents) or refugees
(17 parents).
A correlation
analysis was used to measure the degree of association among the variables, and
a multivariate regression analysis was used to assess the magnitude and
intentions of a relationship of the predictor variables to the criterion
variables of parental intentions for involvement. The measures of attitudes and
beliefs, subjective norms, perceived behavioral control perceptions, and
intentions of immigrant and refugee parents were significantly related to each
other and predicted parents’ intentions for involvement in AKA Head Start and
Early Head Start programs resulting in rejection of the null hypothesis. The findings are consistent with the
theory of planned behavior model and can be applied to explaining parents’
involvement for those with culturally and linguistically diverse backgrounds.
No comments:
Post a Comment