Abstract

1
What Do I Do Best As A Father A Two-Step
Analysis Technique For Qualitative Data Norma J.
Perez-Brena, Jeff Cookston
William Fabricious
Scott DeGuzman San Francisco State University
Arizona State
University
San Francisco State University
Abstract

• Measures (cont.)
• 2) Parent Child Interactions (PCI) Scale
• This 5-item scale assesses the frequency in which
  a parent participates in parent-child activities.
  Three of the Items are
• categorized as Traditionally Feminine
  activities (i.e., going to the store) and 2
  items are categorized as Traditionally
• Masculine activities (i.e., playing ball ). The
  scale had good reliability for fathers (a .57).
• 3) Monitoring
• This 6-item scale was created by Stattin and
  Kerr (2000) to assess the frequency that a parent
  monitors his/her children. This
• monitoring scale asks parents How often in the
  past 3 months, did you know what (child) did in
  (his/her) free time. Within our
• fathers data set, a composite was created
  indicating the mean of a fathers and
  adolescents report of a fathers monitoring
• frequency (a .74).
• 4) Time Spent
• This is a two-item scale that was developed
  within a previous study (The DADS project) in
  order to get an estimate of the
• average amount of hours per week that a father
  was spending with his children a composite of the
  fathers and adolescents
• report of time spent was created (a .38)
• 5) Gate-Keeping
• A scale was developed using 9-items from the work
  of Allen and Hawkins (1999) and Beitel and Parke
  (1998). These 9-items
• asked parents if their partner often did things
  that (they) had done around the house or if
  (their) partner liked to be in
• charge with the household decisions. Using a
  composite of both fathers and mothers data, there
  was low reliability within
• the measure (a .56)

In the current study we introduce a two-stage
technique for the analysis of narrative data and
report its efficacy in the extraction of
theoretical domains from narratives obtained from
mothers and fathers. Narratives obtained from
parents in the Parents and Youth Study (PAYS
http//devpsych.sfsu.edu/pays) were coded in two
stages. First, coders were trained to separate
the transcripts into separate prepositional
statements. Next, each prepositional statement
was coded for the presence and level of thematic
dimensions derived from theory and a careful
review of a subset of the transcripts.
Participants were mothers and fathers from 392
PAYS families. Although all of the mothers were
biological parents, 175 (45) of the fathers were
stepfathers and 193 (49) were from Mexican
American families. Working over the course of a
year, coding groups were trained to separate and
code the statements with good consistency (k
.81). Seven thematic dimension of father
behavior were assessed Emotional Quality,
Responsiveness, Investment, Provisioning,
Discipline, Role Modeling, and Team Parenting.
These seven dimensions were found to have good
reliability and be well correlated with other
measures of parenting behavior. Implications for
future research on the mediating role of social
constructions are considered.
Image 1. View of unitizing comparisons between
coder and Goldie
Narratives - Dimensional Coding
Once the scripts were unitized, each unit was
coded for 7 dimensional meaning codes. 1.
Emotional quality (EQ), the emotional aspects of
the parent-child relationship 2. Responsiveness
(RE), the fathers ability to talk to the child
and respond to the childs needs 3.
Investment(IN), the time commitment a father
gives his children 4. Provisioning (PR), the
amount of financial support a father provides his
children 5. Discipline (DI), the level of
control that the fathers attempts to have on his
children 6. Role modeling (RM), the valence of
the types of behavior that a father displays his
children 7. Team parenting (TP), how well the
father works with the mother to build a cohesive
home environment The dimensional codes were
determined using a theoretical and
person-centered approach. As the Dimensions of
Emotional Quality, Responsiveness and Investment
were guided by the work of Lamb, Pleck, Charnov,
Levine (1987), and Provisioning, Discipline,
Role modeling, and Team parenting were developed
after mining the narrative for themes. Each
code was given a score of 1-3 where 1 meant the
father displays a low level or negative form of a
dimension, 2 meant a father showed a dimension at
a medium rate, and 33 meant a father displayed a
dimension at a high rate or in a positive
manner Each unit could be coded for the
presence of many dimensions as fit the statement.
For example, the statement Im a little hard
on them (EQ2)/ especially since I discipline them
a lot (DI3). would be coded as emotional
quality and discipline due to the reference to
the fathers discipline tactics and how
it affects the tone of the parent-child
relationship. Also the score on one unit did not
affect the score of another unit. For example,
in the previous statement the father would have
received a 2 for Emotional Quality and a 3 for
discipline, because he has expressed a semi
negative relationship with his children but high
amounts of discipline. Reliability Similar to
the unitizing reliability protocol, a group of
top coders met to develop gold standard thematic
codes for 20 of the narratives, which where
used to test each coder for reliability
dimensional coding. A Cohen's Kappa was used to
determine there was high reliability for all
codes (EQ, k .64 RE, k .55 IN, k .79 PR,
k .88 DI, k .61 RM, k .66 TP, k
.62). Validity To determine the construct
related evidence of validity, a series of
bivariate correlations were estimated to
understand the dimensions differences between
Good and Change stems (Table 1), in addition
to running correlations with other measures of
parental behavior and parent-child interaction
(Table 2). Finally, to understand whether there
were significant mean level differences between
stepfathers and biological father, a series of
t-test were used (Table 3).
Project Overview
The relationships between parenting and child
internalizing and externalizing behaviors have
been found in many studies (Botta Dumlao, 2002
Cooley, 1998 Kielcolt Acock, 1988), however,
parent-child dynamics are constantly changing
(Feldman, Gowen, Fisher, 1998). It is,
therefore, important to use methods that are
person-centered to gain first hand knowledge of
how parents evaluate their own parenting. The use
of narratives can be useful in this respect
because they allow researchers to gain first-hand
knowledge of important themes and issues that
parents face. Narratives may then open up new
ideas and ways of conceptualizing relationships.
However, the lack of an explicit set of methods
for the analysis of narratives makes their
inclusion in large projects a difficult decision.
Narratives have been used in a variety of ways
they are coded for (a) the presence of themes
(Fiese Sameroff, 1999), (b) level of coherence,
or (c) co-construction among multiple reporters
(Fiese Spagnola, 2005). Also, one must decide
between molar (wholistic) and molecular (unit)
coding (Fiese Spagnola, 2005). Depending on the
needs of the study, one may choose to code the
whole narrative for its general themes molar
or code each meaningful unit within the narrative
molecular. In the current study we report
results of a molecular coding scheme that
explores the level of seven important fathering
dimensions. It is our intention to describe a
two-part technique which consisted of (1)
unitizing father narratives into meaningful units
of analysis, and (2) coding each unit for these
seven fathering dimensions. Finally,
reliabilities estimates for unitizing and
dimensional coding were gathered by comparing
each coders results on a gold-standard set of
narratives.
Narratives - Unitizing
Discussion
For this analysis, fathers were asked two
open-ended questions about the resident
father. 1. Think about yourself as a
(Dad/Step-Dad). Think about the kind of
(father/step-father) you are. Think about the way
you act, the things you say and do, as a parent.
Take a moment to think about yourself and the
kind of (father/step-father) that you are to
(child). What are the 2 or 3 things you do as a
(father/step-father) that you think you do
best? 2. Now think about the things about
yourself as a (father/step-father) that you would
like to change, or you think you need to work on
or improve. Tell me the 2 or 3 things that you
would change about yourself as a
(father/step-father). Unitization First, the
narratives were unitized, or separated, into
unique prepositional statements. Statements were
separated into new units when the parents (1)
gave examples, (2) offered closely related but
distinct statements, (3) stated a focus change,
or (4) made references to a time change. Each
coder unitized the transcripts using an online
computer program created by the research team
through consultation with Dr. John Kim at San
Francisco State University. The online unitizing
program was created so that each coder had cases
assigned each week and completed their coding
work online. Interrater reliability to observe
the beginning of a new unit was calculated based
on comparisons of 20 of cases that were coded
individually by the first two authors with
differences resolves in conference. Estimates of
Cohen's kappa averaged .81 for the coders
ability to evaluate a unit as distinct. The
images below show how the program allowed for
Cohens Kappa analysis (Image 1) and compared a
coder against the gold-standard coder (Image 2).
The use of a two-step process of unitizing the
narratives and coding for the presence of
thematic units offers a new systematic method
for the analysis of data with several reliability
check points along the way. In addition, this
two-step method of analysis has allowed for the
quantification of qualitative data. The
significant differences in responses that were
observed to the what do you do well and what
do you need to change questions suggest the
participants were answering the question
thoughtfully. As illustrated by Table 1, when a
father mentioned a dimension within the Good
narrative, his scores were on average much
higher compared to fathers that mentioned the
dimension in the Change narrative.
Additionally, our thematic codes appear to show
construct-related evidence of validity through
associations with other measures of parenting.
Specifically, higher levels of emotional quality
were related to greater acceptance and less
rejection from father as well as less
interparental conflict and greater perceived
mattering from the adolescent. Additionally,
father investment and responsiveness were related
to time spent together as reported by the
adolescent. Interestingly, fathers from different
family types did not tend to significantly
differ across dimension, suggesting these
fathers evaluate themselves similarly despite
differences in family history. In sum, our
results appear to suggest thematic dimensions of
fatherhood can be reliably extracted from the
narratives and show evidence of validity.
Implications for Future Research In future
studies we plan to explore whether fathers
evaluations of their parenting using these
narrative data serve to add unique predictive
explanation to adolescent adjustment over time
net the contributions of father and mother
behavior. It is possible that social
constructions for father behavior of which our
narrative data are but one example serve to
mediate between actual father behavior and child
adjustment over time. In future studies we plan
to report on such associations in greater detail.
Method-Participants
PAYS Overview The Parents and Youth Study (PAYS)
is an ongoing 5-year-longitudinal multifaceted
research project that seeks to understand the
temporal relations among social constructions of
fathering behavior and adjustment over time.
Besides using a battery of questionnaires, an
narratives were obtained from fathers to asses
evaluations of their own parenting as well as,
conceptualizations of fatherhood, and a
description of a fathering ideal.
Participants Participants were 393 Mexican- and
Anglo-American families in which a target
adolescent, a mother and a resident father were
all interviewed. All adolescents were interviewed
in English, while the mothers and fathers were
interviewed in English or Spanish depending on
their language preference. For the purposes of
this study only data from the English speaking
fathers (n 280) were used. Measures 1) Child
Report of Parenting Behavior Inventory
(CRPBI) This 18-item 3-point scale assessed the
childs perceptions of the father/stepfathers
accepting, rejecting, and discipline behavior.
Examples of CRPBI questions are During the past
3 months, your (dad/step dad) made you feel
better after talking over your worries, and
During the past 3 months, your (dad/step-dad)
sometimes allowed you to do things that he said
were wrong. The reliability for subscales
accepting (a .88), rejecting (a .770, and
discipline (a .71) had good reliability.
Composites for each subscale were created by
taking the mean of mom and adolescent report.
For more information please contact Norma J.
Perez-Brena Jeff Cookston, PH.D. Normajpb_at_sfsu.e
du
Cookston_at_sfsu.edu Or visit http//devpsych.sfsu
.edu/PAYS
Image 2. View of interrater reliability estimates
between coder and Goldie.