Table of Contents

• Introduction



• Theoretical Framework



• 
  
  
  
  • Value-Expectancy
    Theories
  
  
  
  • Self-Efficacy and
    Attribution
  
  
  
  



• Significance and
  Objectives



• Methodology and Procedures



• 
  
  
  
  • Measurements and Pilot
    studies
  
  
  
  • Participants
  
  
  
  • Data collection
  
  
  
  



• Data Analysis



• Results



• 
  
  
  
  • Research Question 1: Student
    Expectations
  
  
  
  • Goals for the course
  
  
  
  • Expectations and Perceived Goal
    Attainment
  
  
  
  • Speaking Goals
  
  
  
  • Cultural goals
  
  
  
  • Research Question 2: Student
    Attributions
  
  
  
  • Research Question 3: Goal
    Attainment
  
  
  
  



• Discussion and
  Recommendations



• Limitations



• Future Directions



• Notes



• References



• Appendix A: First Survey



• 
  
  
  
  • Part I: Background Information href="#Notes">¹
  
  
  
  • Part II: Reasons for Studying
    Portuguese
  
  
  
  • Part III: Your Goals in this
    Class
  
  
  
  



• Appendix B: Second Survey



• 
  
  
  
  • Part I: Background Information href="#Notes">¹
  
  
  
  • Part II: Goal Attainment this
    Semester
  
  
  
  • Part III: Factors Influencing
    Continuance of Portuguese Study
  
  
  
  • Part IV: Factors Influencing
    Discontinuance of Portuguese Study
  
  
  
  



• Appendix C: First Interview Questions
  (30 Selected Students)



• Appendix D: Second Interview Questions
  (30 Selected Students)



• 
  
  
  
  • 1. Motivation Type
  
  
  
  • 2. Goal
    Modification
  
  
  
  • 3. Additional Questions
    for Specific Students
  
  
  
  • 4. Representative
  
  
  
  

src="../../images/uparrow.gif" alt="Arrow Up" align="right" border="0" height="13"
width="20" />

Introduction

"Life is largely a matter of expectation." Horace (65 BC-8 BC)

Over the past decade, second language acquisition researchers have added greatly
to their understanding of motivation. Their discoveries are often relevant to
educational disciplines outside language learning, but are rarely mentioned in
academic journals or texts for more general education. The present study looks
specifically at language student expectation and motivation to see how student
expectations relate to their motivation for continued language study. This paper
examines several affective aspects of expectancy. These include students’
attributions for success and failure as well as students’ self-efficacy -
defined as "assurance of capabilities" (Bandura, 1994). These aspects were chosen
because they are under the learners' control and can therefore be changed through
interventions (Weinstein, 1994).

The relationship between expectations and motivation is relevant for educators
in disciplines other than language learning, particularly for instructors or
researchers of mathematics. In studying language, much like studying math, students
come to the class with preconceived notions of their abilities to succeed. Just as
students of math "tend to see themselves as either mathematically inclined or
disinclined" (Watson-Acosta, 2003), language students tend to decide early in their
studies whether or not they have the "special gift" of language learning ability
(Horwitz, 1989).

A brief history of motivational theories is offered, followed by the
study’s significance to current literature.

src="../../images/uparrow.gif" alt="Arrow Up" align="right" border="0" height="13"
width="20" />

Theoretical Framework

href="#tableofcontents"> align="right" border="0" height="13" width="20" />

Value-Expectancy Theories

Most relevant to this study is the social-cognitive approach to motivation, a
theory generally used by educational psychologists. It emphasizes the influence
that students' beliefs and interpretations of their experiences have on cognitive
processes (Weiner, 1986 in Pintrich et al, 1993). The decisions involved in goal
setting are influenced by the degree to which individuals expect their goals will
be met and by their beliefs about the importance of effort and abilities. Described
formulaically, the attraction to a certain subject or task equals the value the
person places on it "times the apparent probability it will be attained" (Klinger,
1977, p. 303). Value levels affect both initial and continued motivation.
Learners’ reactions to difficulties faced throughout the goal attainment
process are influenced by whether or not they feel what is gained from their
efforts is worthwhile-- meaning, is the effort put forth moving them in the
direction they want to go (helping them make a certain grade, giving them a sense
of accomplishment, etc.) (Noels, 1999).

The importance a task holds for an individual relates to what goal theorists
term "goal orientation." Orientation involves the reasons affecting students'
initial choice, the degree and direction of effort, and whether or not students
persist in that activity (Nam Yung, 1996). Individuals may be oriented towards
learning goals (also called mastery goals) or performance goals (Covington, 2000).
Students with learning goals demonstrate an incremental belief about ability,
wherein ability level is perceived as changeable, not fixed. They acknowledge the
possibility for growth and focus their attention on mastery instead of just trying
to get by. Those with performance goals, conversely, demonstrate an entity belief
about ability, wherein ability is fixed and not affected by increased effort. They
tend to avoid difficult tasks for fear of failure. Learning orientations have been
shown to affect motivation and student achievement. Mastery goals have been shown
to lead to more active engagement in learning than performance goals (Pintrich,
2000). Students who are less cognitively engaged employ fewer learning strategies
and self-regulatory practices which in turn affect their achievement (Covington,
2000). Schommer (1990) sees a direct link between beliefs and achievement. Her
study found college students who perceived knowledge as "fixed" demonstrated less
appropriate, overly-simplistic writing styles compared to students who saw learning
as more incremental and multi-dimensional (Schommer, 1990 in Mori, 1997). This is
perhaps because those self-regulating strategies that help students try multiple
solutions to challenges are the same strategies used in the complex thought of
writing tasks (Diener & Dweck, 1980).

src="../../images/uparrow.gif" alt="Arrow Up" align="right" border="0" height="13"
width="20" />

Self-Efficacy and Attribution

The connection between student beliefs and volition has received little
attention in language learning motivation research. Further investigation could aid
instructors and program developers in determining how to best meet students’
needs.

Albert Bandura’s work in aptitude beliefs is the cornerstone of
self-efficacy research. He asserts that highly efficacious students see difficult
tasks as challenges to be mastered, not threats to be avoided: "Such an efficacious
outlook fosters intrinsic interest and deep engrossment in activities" (Bandura,
1994, p. 73). Self-efficacious students employ more strategic planning towards
accomplishing their goals (Oxford, 1994). Efficacy levels also affect the type of
linguistic information they choose to pay attention to, which in turn affects
proficiency (Mori, 1997).

Student efficacy often comes from vicarious experiences (Schunk, 1991). For
example, students may assess their language learning ability based on
‘myths’ propagated by classmates or on advertisers' promises of quick
success. Many university students expect to be fluent after one or two years of
study (Horwitz, 1989). They become frustrated when they do not progress that
quickly and often discontinue study of the language when their expectations go
unmet (Horwitz, 1989; Altman, 1985). Bandura (1994) argues that the resulting
frustration lowers self-efficacy and makes students "slacken their efforts and give
up quickly in the face of difficulties" (Bandura, 1994, p. 8).

This frustration is not unique to language study. The negative stereotype about
women and math, for example, affects students’ efficacy. Female students are
so averse to reinforcing the stereotype that they become overly anxious in math
courses, impairing their performance (Oswald & Harvey, 2003). Their lowered
self-efficacy makes them use fewer autonomous learning behaviors necessary for
achievement (Greene et al, 1999) and causes attrition in future math-related
courses (Oswald & Harvey, 2003).

Clearly, however, not all students desist when faced with challenges. To explain
this variability, researchers in educational psychology point to student
attributions, defined as perceived causes for success or failure (Schunk, 1991).
"Locus of control," a generalized control over outcomes, describes how individuals
perceive success or failure as either independent of their own actions and thus
"externally controlled" or dependent on the way they behave and thus "internally
controlled" (Rotter, 1966 in Schunk, 1991). An attribution such as "motivation" or
"effort" would most likely be considered "controllable," whereas "luck" or task
difficulty would be considered uncontrollable (Weiner et al., 1983 in Schunk,
1991). Whether students believe they have control over learning outcomes affects
how much effort they expend in learning and how long they persist in their efforts
(Oxford, 1994). According to Dolinger’s (2000) study of college
students’ locus of control, students who feel they have internal control may
be more successful learners because they are more perceptive of their environments:
"Internals more readily acquire and utilize information that is relevant to their
goal situation" (Dolinger, 2000, p. 1). Other studies show that students with
internal attributions of control demonstrate higher achievement because they are
better at planning how to complete academic tasks (Biggs, 1987 in Hall, 2001).
Conversely, attributions of low ability negatively affect long-term success and
student retention, as students desist in the face of lower self-esteem and a sense
of helplessness (Graham, 1990 in Tse, 2000)

In recent years, these causal attributions have been more frequently mentioned
in interpreting results of foreign language studies (Nam Yung, 1996; Wen, 1997).
Yet few foreign language researchers have included attributional theories in the
design of their studies. The present study operationalizes the concept of
attributions within the context of the foreign language classroom.

src="../../images/uparrow.gif" alt="Arrow Up" align="right" border="0" height="13"
width="20" />

Significance and Objectives

This study will examine university students of Portuguese, a population chosen
for two reasons. Portuguese is the world’s eighth most widely spoken language
(Family Education Network, 2003) and is considered a "critical language" for the
federal government. Survey figures, however, show Portuguese language programs in
U.S. universities are not keeping up with other language programs in terms of
growth (National Security Education Program, 2001; Silva, 2000).

Students of Portuguese were also chosen as participants for examining
expectancies and attributions since changes in beliefs may be most needed for
students of languages that are less commonly taught (LCTLs). Over-efficaciousness
has been shown to be problematic for students of more commonly taught languages
such as French and Spanish (Horwitz, 1989). It can be even more problematic for
students of lesser commonly taught languages because they have generally had less
exposure to the language before its introduction in the classroom. They therefore
have less real world experience on which to base their assumptions. Of the limited
studies on LCTL students, several found that students were not aware of the level
of difficulty of the language; the ensuing over-efficaciousness proved a barrier to
continued motivation (Wen 1997, p. 236). In fact, two years of instruction, the
amount required in most universities, leaves the LCTL student at only the very
beginning stages of the language (McGinnis, 1994). In the case of Portuguese
students, many already know a bit of Spanish and therefore may hold unrealistically
high expectations of reaching advanced levels of Portuguese within a short amount
of time.

This study seeks to answer the following research questions to see how
students’ expectations, attributions, and beliefs about learning affect their
motivation and their decisions to continue or discontinue formal study of the
language:

1. What are students’ expectations for goal attainment?



2. What are students’ attributions for success and failure in meeting
   their language learning goals? What are their beliefs about language
   learning?



3. What role, if any, might students' perceived goal attainment play in
   students’ motivation and their decision to continue studying the
   language?

src="../../images/uparrow.gif" alt="Arrow Up" align="right" border="0" height="13"
width="20" />

Methodology and Procedures

A combination of quantitative and qualitative methods were employed to survey a
large pool of participants while also obtaining more in-depth data from open-ended
questions and interviews. If we consider qualitative and quantitative research as
two ends of a continuum rather than two completely distinct methods, this study
would fall closer to the qualitative end of the spectrum. The study is therefore
more exploratory than confirmatory. It offers a new perspective on this student
group, but avoids the causal "certainties" that purely quantitative data
pursues.

The various types of data in this exploration worked together. Numerical survey
data (Appendices A and B) gave general background information on students and an
overview of their goals, expectations and motivations. It also helped determine
which students to interview (a full description of this decision process is found
in the "Participants" section of this work). The more qualitative open-ended survey
items added details to responses to the more numerical survey questions. They also
pointed out key informants to interview and helped guide interview questions. For
example, several students mentioned an advisor who had told them how easy it was to
learn Portuguese. I therefore interviewed this advisor to find out more about her
interaction with students.

Post-survey interviews asked students to elaborate on their survey responses and
thus obtained a more in-depth picture of students’ beliefs and motivations.
These measures were developed through a series of pilot studies.

src="../../images/uparrow.gif" alt="Arrow Up" align="right" border="0" height="13"
width="20" />

Measurements and Pilot studies

Several pilot studies at the University of Texas at Austin (U.T.), where the
actual study would be performed, elicited information about students' motivations.
This information helped formulate interview questions and modify previously used
survey instruments to fit this specific population. In the first pilot study, the
researcher observed 30 class hours at the University of Texas to become familiar
with the Portuguese classroom setting. Opening and closing interviews were
performed to see how students would articulate concepts like goals and goal
attainment.

The next pilot used three Portuguese classes as its subjects. Students were
given an open-ended questionnaire that asked them to brainstorm their reasons for
taking Portuguese and their goals for their class. They then circled their three
most important reasons and goals. The reasons students gave were coded and used to
modify the motivational survey "Reasons for Studying Spanish" (Ely, 1986), which
after further pilot testing became the survey section "Reasons for Taking
Portuguese" (Appendix A Part II). Similarly, the goals students mentioned were used
to modify a goals survey previously used with students of French and Spanish
(Harlow & Muyskens, 1994). At the close of the semester, students completed
Speiller’s (1988) questionnaire "Factors Influencing Students to Continue or
Discontinue Language Study" and were asked to comment on any confusing or
non-applicable parts of the study. Based on students’ comments, the survey
was broken down into two separate surveys, one for continuers and one for
discontinuers.

The third pilot study administered the revised surveys to students in two other
Portuguese classes. Results indicated the importance of students' expectations to
their perceived goal attainment. Though no specific hypotheses were made based
solely on pilot data, some preliminary assumptions served as a skeleton for
examining patterns, testing the conditions of various relationships, and building
theories.

To gain richer data and ensure that students’ individual responses would
be elicited, the open-ended questions were added to the surveys. To test this
combination of open-ended and Likert scale items, the modified surveys were
piloted. The actual study would begin soon, so this version of the surveys was
given to a group of Spanish students at another Texas university to avoid having
students see the surveys in both the pilot and in the actual study.

This pilot was helpful in ensuring that students would understand the
survey’s breakdown of cultural items. Cultural items were broken down into
four elements, as recommended by the American Council for the Teaching of Foreign
Languages. The council notes that the culture taught in the classroom is not just
one entity. Instead, it consists of both products distinct to different
countries and cultural patterns as well as the actual understanding
(perspectives) of those products and patterns of behavior. In the surveys,
the breakdown was described as follows:

• country's products (e.g. books, art, music, political systems,
  etc.);



• cultural perspectives on those products (attitudes explaining
  why certain products exist and are valued);



• cultural practices (how people use cultural products; patterns
  of behavior such as how people celebrate, dress, etc.); and



• cultural perspectives on those practices (i.e. the attitudes
  and ideas that explain why people behave as they do) (National Standards in
  Foreign Language, 1999).

This breakdown is important to avoid problems faced in previous research on
foreign language goals. A prime example is found in Alalou’s (2001)'s
questionnaire on students' perceived needs in foreign language courses. As Alalou
admits, since the term "culture" was not defined, results were difficult to
interpret:

Although in this study, 'culture' is understood in its broad sense, referring to
both high and popular culture,..we know very little about students' definition of
'culture' because students in this study were not asked to provide a specific
definition of the term (Alalou, 2001, p. 461).

In concluding this description of measure development, it should be noted that
before beginning the actual study, the researcher realized the limitations of using
self-report measures. One could argue, for example, that students may say that in
the next few weeks they plan to register for another Portuguese class and continue
their studies, when in fact they may not actually register for that class when the
time comes. This study examines students’ motivations, however, as opposed to
all the many factors that affect course registration (illness, finances, etc.). As
such, self-report measures appear to reveal the data necessary to better understand
what encourages students to want to learn more.

src="../../images/uparrow.gif" alt="Arrow Up" align="right" border="0" height="13"
width="20" />

Participants

Participants in this study included 101 lower-division Portuguese students at
U.T. Austin, their instructors, and four student advisors. The research was
reviewed and passed by the university’s Human Subjects Review Board. Before
any surveys or interviews were completed, the subjects received a consent form
discussing the purpose of the research, their anonymity as subjects, and that their
participation/non-participation would not be discussed with their instructors or
otherwise affect their standing in the department.

Information on individual student participants was gathered through the
background questions completed by eighty-four students- forty-three females and
forty-one males. Most students were undergraduates (71.3%) with most of those
sophomores or seniors, 26.2% were graduate students, and two students had already
graduated. Nearly half were humanities majors with the remaining students studying
sciences or registered as ‘undeclared.’ Well over half of the students
were Caucasian (63.1%). A large percentage were Hispanic (29.8%). Other students
described their ethnicity as Asian-American, African-American, or "other." Most
spoke English as a first language though over half of all respondents had studied
Spanish formally for over two years. Twenty percent spoke Spanish as a first
language. Just over half were taking the Portuguese to fulfill a requirement.

src="../../images/uparrow.gif" alt="Arrow Up" align="right" border="0" height="13"
width="20" />

Data collection

Data sources included surveys and interviews with students, instructors and
advisors. In the actual study, the first week of class, the researcher administered
the first survey on students' backgrounds, initial motivations, goals, and
expectations (See Appendix A). The second survey, administered the antepenultimate
week of class, examined students' perceived goal attainment and reasons for
(dis)continuing formal study (Appendix B).

Telephone interviews were conducted with the following student groups: seven of
the eight students who dropped the class before the end of the semester; all six
students who were auditing courses; and five ‘extra’ students whose
survey responses merited further inquiry. In addition, thirty students taking the
course for a grade were selected to do both opening and closing interviews (See
Appendices C and D). These thirty subjects were selected using a stratified
purposeful sampling technique (Mertens, 1997) based on students' goal values and
expectations ratings. Groups were formed by coding goal value and expectation
sections from the first surveys (Appendix A Part III C and D). Using EXCEL,
students were divided into four groups: those with a tendency to have low-valued
goals paired with low expectations, those with low-valued goals paired with high
expectations, etc. Participants were separated into level (beginner, intermediate,
etc.). Each class was separated into the four groups and participants names were
then randomly chosen from each group. To follow up on students’ responses,
the researcher interviewed four teachers and four administrators whom students
mentioned as having influenced their decision to study the language.

Data was also collected at Tulane University to compare their students’
motivations for taking Portuguese with University of Texas students. Tulane was
chosen because it is a small private university (12,000 students) compared to the
University of Texas, a large public institution (49,000 students). The purpose of
Tulane’s inclusion was not to show that results from the University of Texas
study are transferable to all other universities. Instead, by comparing
U.T.’s data to a different university, it was useful in giving thicker
description of U.T.’s population and environment. This may aid readers in
deciding the degree of transferability this study has to their own situations.
Survey and interview measures and selection procedures were identical to those used
with the University of Texas sample, except that the second survey was not
administered to the Tulane population. Thirty-five participants from Tulane took
part in the study. Judging from non-participant observation, course descriptions
and syllabi, and instructors' and students' comments, U.T. and Tulane’s first
and second semester courses seemed fairly similar in content. One Tulane course, a
Portuguese literature class with five students, was not offered to U.T.
undergraduates, and was therefore not included in the results.

src="../../images/uparrow.gif" alt="Arrow Up" align="right" border="0" height="13"
width="20" />

Data Analysis

Survey data analysis used the statistical software SPSS and included
frequencies, means, and standard deviations, and factor analysis of Likert scale
responses. These figures and the reliability for each scale are found in the
results section of this work.

Qualitative analysis used the qualitative research program "QSR NUD*IST"
(Qualitative Solutions and Research Pty Ltd.’s Non-numerical Unstructured
Data: Indexing, Searching, and Theorizing). It is a code-based theory-building
program useful in forming and describing categories, making connections between
categories, constructing theories, and validating or rejecting theories about
categorical relationships. Its flexible searching features were helpful in working
with large amounts of text and coding. Its compatibility with SPSS helped link the
quantitative and qualitative data.

As the data were collected, I transcribed interviews and employed a grounded
theory approach for analysis and for analysis of the qualitative survey responses.
I used Strauss and Corbin's (1990) analysis method: a "systematic set of procedures
to develop an inductively derived grounded theory about a phenomenon" (Strauss
& Corbin, p. 24). Grounded theory involves going into a research situation (in
this case, a college classroom) and finding out what is taking place and how the
people relate to that situation.

The first stage of procedures involved the "open-coding" of interview data,
defined as "breaking down, examining, comparing, conceptualizing, and categorizing
data." (Strauss and Corbin, 1990, p. 61). All 60 transcripts from the thirty main
interviewees, along with those from the seven "drop" and the five "extra"
interviews, were read for emerging commonalties and patterns. I used a line-by-line
approach analyzing each sentence and separating data into categories relevant to
students' motivation. Categories were given descriptive names like "Beliefs about
learning culture," and the categories were described in memos. Throughout the
coding process, I reread category names, their descriptions, and their coded
information to ensure new information fit the categories. Often times these
re-readings, coupled with the constant addition of new data, prompted the collapse
of two similar categories into one. Or the addition of new data prompted the
expansion of one category into several, for a richer description of the phenomenon.
For example, at first I had only one category for students’ beliefs. Later,
however, this category was expanded into two categories: one for
"Beliefs-about-language-learning" (with subcategories for beliefs about
"Natural ability," the importance of "Early exposure", etc.) and one for "Beliefs
about learning culture" (with students’ comments about how culture was
learned separately from the language, how teaching culture was laden with bias
etc.). The split was made as it became obvious, in analyzing student responses,
that patterns emerged differently under those two categories.

The next step involved "axial coding" in which connections were formed among the
categories found in open coding. A tree-like structure contained each category,
with "motivation" as the root from which branches (categories) and limbs
(sub-categories) emerged. The tree diagram was modified as I worked and reworked
the connections among categories, confirming relationships with data from other
sources such as teachers' and administrators' interview data or responses to
open-ended survey items. There was a constant interplay between the interview data
and responses to questionnaire items to validate and refine relationships among
categories. The tree sketches were helpful when interpreting and confirming
quantitative findings. For example, I was not surprised when, during factor
analysis, a survey item about students’ desire to "Translate Portuguese"
clustered with less practical items such as "Enjoy myself" and "Improve my accent."
Interview data had similarly shown that many students whose goals fell into a
"Hobby/fun" category also had more "Practical goals" and wanted to put Portuguese
to work-related use.

Data analysis took place both during and after data collection, in line
with Creswell’s (1998) emphasis on a zigzag approach between data gathering
and its analysis. It was essential that the analysis begin during data collection
because the patterns, commonalties and differences that emerged early in the
collection process could then be examined in further detail in later interviews,
follow-up emails or extra interviews.

src="../../images/uparrow.gif" alt="Arrow Up" align="right" border="0" height="13"
width="20" />

Results

src="../../images/uparrow.gif" alt="Arrow Up" align="right" border="0" height="13"
width="20" />

Research Question 1: Student Expectations

Reasons for taking the class: Initial Motivation

To give a better idea of students’ expectations for goal attainment, it is
important to examine why students initially decided to take the class. The "Reasons
for studying Portuguese" section consisted of thirty items, including an "other"
item; each item was ranked in importance ranging from 0-3. Rankings of the top 10
means for responses for the "Reasons for studying Portuguese" section appear in
descending order in Table 1 below.

Table 1: Top 10 Students' Ranking for Reasons for Studying Portuguese

The scale was subjected to an internal consistency reliability analysis. Very
few students responded to the "other" item so it was not included in determining
internal consistency. The Cronbach Alpha was .73 which is somewhat low, suggesting
multidimensionality within the survey items. As such, some clusters of items may
tend to vary together more than others, bound by an underlying commonality (SPSS,
1999). A factor analysis was performed (Tables 2-7) to show the underlying factors
that link certain survey items.

Factor analysis was chosen over principal components analysis since it is
recommended in cases where items are correlated. In fact, items did appear to be
correlated as shown by the .718 Kaiser-Meyer Olkin Measure sampling adequacy.
Maximum Likelihood extraction method with a Promax rotation was used as it is
suggested for use in cases where items are correlated (Rennie, 1997). There were
nine eigenvalues greater than 1.0. I decided to set the number of factors at six
rather than nine since, judging from the scree plots, there appeared to be a
leveling off after six factors, indicating that six factors were sufficient to
account for the variance. Values lower than .3 were suppressed. It is questionable
in a confirmatory factor analysis to include variables whose correlations with the
other variables are below .4 in absolute value (Hedderson, 1993, p. 174). However,
since this was an exploratory factor analysis and since the two variables with
loadings just under .4 appear to fit conceptually, I felt the information gained
from including the two variables outweighed any reservations about relatively low
values.

Six factors accounted for 45.4% of the total variance and the factor correlation
matrix showed a low correlation of .28, indicating that the six factors, or
clusters, were distinct. These groupings of items were helpful in interpreting
survey responses because items clustering together could be considered part of the
same underlying concept. For example, the first factor to emerge was labeled
"Cultural interests" since it housed items related to students' interests in the
target countries' cultures, including their history and their importance among
other countries in the world. Table 2 illustrates the questionnaire items that form
this cluster and the factor loadings. This cluster accounted for 17.81% of the
total variance and the Cronbach alpha for this factor was .87.

Table 2: Factor Loadings for the "Cultural Interests" Cluster (for Tables
2-9, see Appendix A Part II for full survey items)

The second cluster of items that emerged was labeled "Career/academic
advancement." The total variance explained by this cluster was 7.96% and the
Cronbach alpha for this factor was .73. The items and loadings for this cluster are
listed in table below.

Table 3: Factor Loadings for the "Career/academic Advancement"
Cluster

Though it may appear that the "Portuguese-speaking friends" item is illogically
grouped among more practical motivations, interviews with students revealed that
this cluster is not purely pragmatic in nature. The following student quotes show
how work and pleasure cannot be so easily divided.

The future career wasn't at all important at first. I was just interested in
being able to understand what my (Brazilian) girlfriend was thinking. But the
sounds, I always liked..the cultural items came up later. Then after taking the
language, I started moving in the company towards Latin American and then became
interested..in Latin American and Brazil.

(The main reason I am taking Portuguese is that) I wanted to learn another
language and have covered the western hemisphere with English and Spanish.I never
considered the job part. The job part I didn’t think of until later when it
was like, ‘If I keep doing this, I could put it on my resume.’

I'm a Spanish speaker and I inherently love the (Portuguese) language and
there's a trilingual fascination because of business which grew out of my love
for Spanish and Brazil. So it (my reason for studying Portuguese) does
have to do with love of other cultures and Brazilian friends, so I love to sell
Brazil.

The third cluster, "Requirement motivation," had a Cronbach alpha of .37 and
accounted for 7.55% of the total variance. The "travel for pleasure" item loaded
positively for "Requirement motivation" while other variables loaded negatively.
The sign difference indicates that as interest in "travel for pleasure" increased,
interest in "requirement" and "scheduling" decreased and vice versa.

Table 4: Factor Loadings for the "Requirement Motivation " Cluster

The fourth cluster, "Language as hobby," accounted for 4.99% of the total
variance and its Cronbach alpha was .71. Its items were characterized by a love of
language study due to ease, experience, and intrinsic interest.

Table 5: Factor Loadings for the "Language as Hobby" Cluster

The fifth cluster, "Fun," accounted for 3.98% of the total variance and its
Cronbach alpha was .73. Items in this cluster described students' desire to enjoy
learning Portuguese due to its interesting sounds and to the fact that is less
commonly taught than Spanish or French.

Table 6: Factor Loadings for the "Fun" Cluster

The final cluster, "Heritage," accounted for 3.08% of the total variance and its
Cronbach alpha was .62.

Table 7: Factor Loadings for the "Heritage" Cluster

These results indicate that students were taking Portuguese for a variety of
reasons, in particular future career and travel-for-pleasure plans; cultural
reasons were also frequently mentioned.

Tulane University’s Results for Initial Motivation

In comparing Tulane’s results of the Reasons section of the survey,
independent sample t-tests found that only eight of the twenty-nine reasons were
significantly different, and all eight were related either to language
requirements, language-as hobby-items, or cultural items. The table that follows
compares the ranking of means of U.T. and Tulane students' reasons for enrolling,
with the statistically significant reasons highlighted for the university with the
higher mean.

Table 8: Reasons and Tulane Reasons Rankings (SD=Standard Deviation)