In this current study, “attitudes toward statistics” is defined as a combination of the students‟ “attitudes towards the statistics course” and the students‟ attitudes toward the use of statistics the field of study (Cashin & Elmore, 2005; Wise, 1985). However, according to Gal and Ginsburg (1994), they report that students frequently come in statistical courses with adverse opinions, or far along mature adverse attitudes towards statistics. “Attitudes toward statistics” include four components (“affect”, “cognitive competence”, “value”, and “difficulty”) – while “attitudes toward the field of study”
comprise only two components (“interest” and “effort”).
However, conflictingly, Wise (1985) observed that the students‟ negative SATS score achieve low performance in statistics courses. For some students, statistics is considered a barrier to obtaining their degree (Galusha, 1998; Rose, 2005; Fleiss, Levin & Paik, 2013). This current study addresses the change observed in attitudes toward statistics of the students, given that the Survey Attitudes Toward Statistics (SATS) data are linked to student performance measures and other student characteristics. In addition, a possible investigation of the relation between attitudes and performances, as well as, academic and demographic predictors of current attitudes and changes in attitudes, according to the desired expectations (goals), is also conducted. Additional well-meaning attention is given to how attitudes towards statistics may influence students‟
self-efficacy in learning statistics, in their academic environment.
“Attitudes towards statistics” have been explored by several studies, using undergraduate students (Evans, 2007; Chiesi & Primi, 2009; Zimprich, 2012; Hagen, Awosoga, Kellett & Dei, 2013; Swanson et al., 2014; Kiekkas et al., 2015). Other studies focussed only on graduate students (DeVaney, 2010; Zhang et al., 2012;
Hannigan et al., 2014). The last group concentrated on both undergraduate and postgraduate students (Coetzee & Van der Merwe, 2010). With regard to the context, many of these studies were conducted in the USA (Evans, 2007; DeVaney, 2010;
Griffith et al., 2012; Swanson et al., 2014); in Europe (Chiesi & Primi, 2009; Zimprich, 2012; Kiekkas et al., 2015), in Canada (Hagen et al., 2013); in China (Zhang et al., 2012); and in South Africa (Coetzee & Van der Merwe, 2010; Mji, 2009). Considering the research design, numerous studies applied the quantitative method (Chiesi & Primi, 2009; Mji, 2009; Coetzee & Van der Merwe, 2010; Zimprich, 2012; Hagen et al., 2013;
Hannigan et al., 2014; Swanson et al., 2014; Kiekkas et al., 2015), while only three studies used mixed methods (Evans, 2007; DeVaney, 2010; Zhang et al., 2012), and one research applied a qualitative method (Griffith et al., 2012).
The Survey Attitudes Towards Statistics (SATS) was applied as the instrument to measure statistical anxiety. The first category completes the full set of factors of SATS (Coetzee & Van der Merwe, 2010; Hannigan et al., 2014; Swanson et al., 2014;
Kiekkas et al., 2015), while some scholars (DeVaney 2010; Zhang et al., 2012;
Zimprich, 2012) used only four factors of the SATS instrument, namely, Affect, Cognitive competence, Value, and Difficulty. Chiesi and Primi (2009), as well as Mji (2009), however, used the Attitudes Towards Statistics (ATS) scale, while Evans (2007) assessed the general attitudes towards statistics, by using eleven mnemonics, with four questions about each.
The students, who were randomly selected from three departments at the United States University, regardless of having general positive attitudes, still pronounced selected negative “attitudes toward statistics” (Evans, 2007). Certain students usually considered statistics to be a well-meaning part of study, as well as a substance in which they believed they could complete adequately. Nevertheless, additional students considered statistics to be a subject they would rather not be studying at the time, because they
deemed it unnecessary in their individual professions. The last group of students believed that statistics was not a convenient subject for them, but, precisely for others.
In addition, Evans (2007) interviewed five instructors at the end of the semester, using open-ended questions to explore the strategies that the instructors were applying to progress the attitudes, and remove the misconceptions, of their students. One mentor established the connection among statistics and upcoming courses the students might learn. Additionally the mentor used humour, confidence, and eagerness for statistics to produce extra attention. In addition, one instructor asked students to collect data, using a survey on the opinions of a future governmental appointment. This presented as a real-world illustration to support students to acquire a good understanding of statistical conceptions, and their real-life applications. This additional facet of data assemblage permitted students to practice their individual records for statistical examination.
A second study, conducted at the University of Florence, Italy, focussed on undergraduate psychology students‟ measures and attitudes in the Italian educational context (Chiesi & Primi, 2009). In this study, 313 psychology students at the university completed mutually pre-course versions of the SATS and ATS, while only 263 completed jointly post-course versions of the SATS and ATS. A multivariate analysis established the unchanged time of the SATS instrument. Chiesi & Primi (2009) revealed that the pre-SATS and the post-SATS results in their study have different structural weights. The pre-SATS result showed that, as the original SATS scores upgraded, the closing grade augmented with a small magnitude. However, the post-SATS result designated that as the last post-SATS improved, the final grade also increased with a big magnitude. The hidden difference in means among pre- and post-administrations indicate that the changes for all components were positive values and statistically meaningful. Given that the achievement is measured through written and verbal examinations, they establish that there is a correlation, among college students, between a positive “attitude toward statistics” and achievement in statistics examinations. Therefore, the approaches at the close of the subject were a better prognosticator of attainment than the behaviours at the opening. Once they attended the classes, the students were inclined to be extra self-assured in their academic awareness and expertise, when applied to statistics, while the attendance of the courses moderately condensed the apparent struggle of the statistics subject.
Mji (2009) reports no statistically significant gender differences for the attitudes towards statistics courses, while there are statistically significant differences for the area of study programmes. Bonferroni-adjusted subscale contrasts among averages for attitudes components revealed that the variances were mostly among students studying Taxation and Cost Management Accounting. In addition, a substantial difference was indicated on the subject attitude component concerning the similar clusters. He argued that Taxation students had more adverse approaches towards statistics courses, compared to those in the other two areas (Marketing and Accounting). Students in Taxation recorded negative behaviours about statistics, especially within their area of study. This may be related to their previous experiences with statistics, given that they did not have previous knowledge of mathematics. Self-reports from respondents during data collection, reliance on oral materials present at meetings, or still unpublished researches, restrict occasions for more wide evaluations, and limits several forms of reliability, as well as validity of the instrument.
Ciani, Easter, Summers and Posada (2009) investigated how the autonomic arousal and statistics self-efficacy of undergraduate students could positively affect their scores in final examinations at a Midwest University in the USA. Important Pearson connections were observed among physiological arousal and efficacy, as well as among self-efficacy and positive “affect”. Using artificial cut points (for example, average differences), Ciani et al. (2009) examined interaction effects and revealed a substantial positive key result of self-efficacy on positive behaviour. Once incoming the collaboration term, the leading result of self-efficacy was still important, and the communication term was similarly important for effect. These researchers, therefore, argue that students with high confidence have higher positive behaviuors, while students with low confidence have lower positive behaviours, as autonomic arousal increases, when they are aware of the imminent final exam. Students with low self-assurance may tend to misinterpret their autonomic stimulation as negative; and the adverse behaviour that follows may damage their aptitude to concentrate on the assignment at finger. The results indicate that increased physiological arousal of students is correlated to increase self-efficacy beliefs. In addition, at moderate-to-high levels of self-efficacy, strong reports of arousal correspond with strong reports of
positive behaviours, while at low levels of self-efficacy; there is an opposite association between the strong of physiological arousal and positive behaviour.
Regarding the results of SATS for American students enrolled statistics courses, achieved a substantial difference for the “affect” and “difficulty” by DeVaney (2010).
The on-campus students have more favorable SATS while online students increase considerably on Affect from the pre-test to the post-test. An “independent sample t-test”
reveal substantial modifications among the on-campus and online students for two SATS components “affect” and “difficulty”. The choice and assignment criteria would have affectedly lead deterioration to the mean.
In a study conducted by Coetzee and Van der Merwe (2010), a sample of UNISA undergraduate and graduate psychology students were selected, using a convenient sampling method. They applied a cross-sectional survey design to examine the reliability and validity of the “Survey of Attitudes Toward Statistics” (SATS-36). In addition, they investigated the variances regarding students‟ SATS scores, in terms of their previous mathematics knowledge, levels of statistics courses, and biographical variables (Coetzee & Van der Merwe, 2010). All the latent components, excluding the
“difficulty” factor, attained suitable levels of internal reliability, which implied that the
“difficulty” factor did not describe the data sufficiently. More students likely disagree with the item, “Statistics procedures are easy to comprehend”, or less students agree with the item, “Statistics is a complicated subject”. The Mann-U Whitney test revealed no important change between the levels of statistics courses and the students‟ SATS scores.
Doctoral and master‟s students, in various colleges of education, at 250 universities in the USA, participated in the study lead by Perepiczka et al. (2011). The main objective was to evaluate how these students in counseling and education, from different backgrounds, react to statistics courses, as well as the implications for their educators.
The authors investigated the connection among self-efficacy to learn statistics SELS) and statistics anxiety (STARS), attitude towards statistics (SATS), and social support (MSPSS) of graduate students. The study employed a quantitative method. STARS and SATS were extremely interrelated, demonstrating multi-collinearity. The results from the various instruments revealed negative correlations concerning SELS and STARS, as
well as positive correlations between SELS and SATS. The findings of multiple regression analysis indicated a statistically important association among SELS and STARS, SATS, and MPSS, with a modest influence magnitude at 52.8% of the variance accounted for in the model, R2 = .528. Additionally, MSPSS did not impact the analysis, removing it in the model did not change the result. STARS and SATS are meaningful predictors of SELS.
Griffith et al. (2012) investigated an illustration of undergraduate students from two universities in the USA indicated that student attitudes toward statistics are either positive or negative, and the reasons for their attitudes were given in the written format.
These respondents were students with criminal justice, business, and psychology field marshal. The business major students were the most positive, compared to criminal justice and psychology major students. For positive attitude toward statistics, the business major students tended to have confidence in that indicators were useful in their upcoming profession. However, an interconnection existed among study‟s field and behavior components. The business major students presented more frequent responses compared to criminal justice and psychology major students. For graduate school, more psychology major students have recurrent responses, other major students. Business major students seemed to ensure a better appreciative of the prominence of statistics in their coming profession, whereas psychology major students were inclined to have conviction that statistics was significant for graduate school.
The respondents with negative attitudes of were grouped in five categories across the majors. Business students achieved with fewer responses while their countermates were more pronounced. In addition, criminal justice and psychology students, considered careers that were more in line with their mindsets, and there was no connection of statistics in the field they intended to follow, while business major students required more of a considerate of statistics requests in businesses. Useful information may have been acquired if the students were assessed in a pre/post-tests manner, by discipline.
Additionally, the respondents were from six classes, with six different lecturers, across two universities, implying that the lecturer‟s background and mode of teaching could also influence the attitudes of students towards statistics. Therefore, future research should explore the influence of the lecturer‟s background and mode of teaching on the statistics behaviours
Zhang et al. (2012) directed postgraduate medical students to complete the both pre-test and posttest of the SATS-28 in a single institution in China. They explore the respondents‟ feelings towards statistics and its effect on students‟ achievement, as well as differences across departments. The findings revealed that students with greater levels of statistical instruction and research experience were inclined to have extra positive SATS. Students with a improved mathematics foundation were also extra positive compared to those with a deprived foundation. However, students with clinical academic specialties were more likely to have negative attitudes, compared to their counterparts. Therefore, students with more positive SATS had a tendency to achieve well in the examination. The “Affect” and the “Cognitive Competence” components were strongly and positively connected one to another.
Regarding the qualitative question about the basis of the overall SATS, most of the students‟ behaviours derived from their prior experiences in statistical or mathematical courses. Some students presumed that statistics was a portion of mathematics, and, therefore their behaviours toward mathematics were just shifted to statistics. Additional causes of influence included classmates‟. In addition, the students revealed other diverse causes, such as out-of-school lives. Ultimately, more real involvements are required to assist students overcome their fear and anxiety of statistics.
Undergraduate psychology students from the University of Zurich in Germany completed the Survey Attitudes Towards Statistics (SATS) scale, German version (Schau, Stevens, Dauphinee & Del Vecchio, 1995), and wrote a statistical test (Zimprich, 2012). These resulted in the determination of the factorial structure, predictors and outcomes of the SATS. To compare the relative fit of models, the χ² difference test was applied, and completed by 90% “root mean square error of approximation” (RMSEA) of confidence intervals. The factorial structure of the SATS revealed that Model A achieved an acceptable fit according to the RMSEA, but not according to the χ² difference test. The Comparative Fit Index (CFI) indicated that the model could be improved. Model A did not adequately describe the associations between Items 10 and 19 (referring to the usefulness of statistics in professional life) or by contrast, Items 14 and 21 (referring to relatively strong negative emotions caused by statistics). In Model B, the two residual co-variances are assessed (Items 10 and 19 and Items 14 and 21). Model B completed significantly better than Model A; therefore, the
CFI was suitable. Model B was an adequate description of the data. Items 10 and 19 were both elements of the “value” factor, while Items 14 and 21 were elements of the
“affect” factor. The two residual co-variances did not change the structure of factors.
All factors were significantly associated, with the strongest correlation between
“affect” and “cognitive competence”, and a moderate correlation between “value” and
“difficulty”. Zimprich (2012) revealed that students with more positive emotions toward statistics, as well as those feeling more experienced in statistics indicated higher accomplishment.
In the Model D, there was a small impact for the difficulty of statistics, revealing that female students considered statistics to be more difficult, compared to their male counterparts. The attitude toward statistics was positively associated with mathematics grade, indicating a good performance in mathematics grade was associated with a more positive SATS. The statistics achievement was included as a latent variable; the regression was statistically significant and positive with a large effect size. Students with more positive SATS succeeded in statistics. In Model E, the four factors of attitudes toward statistics represented 30% of the variance in statistics achievement, with the strongest predictor being “Affect”, followed by “Cognitive Competence”, and
“Value”. Notably, the effect of “Difficulty” on statistics achievement was negative and strong, as well, indicating the presence of suppression, as they were bivariate positively correlated. Compared to other research domains addressing subjective and objective perspectives on performance (Mascherek & Zimprich, 2011), the link between SATS and “statistics achievement” was relatively strong. An investigation into the relationship between SATS and “statistics success” should produce more interesting insight, considering that SATS and “statistics achievement” change over time. Therefore, the changes observed in the two variables were strongly correlated. These hypotheses should be examined using latent change models.
In a study conducted by Hagen et al. (2013), nursing students at a university in Western Canada completed pre- and post-surveys, in order to determine their attitudes towards statistics courses, as well as complete fear and anxiety. These students also define their preferred “learning and teaching styles”, and the perceived utility and value of taking a statistics course. Using a pre-experimental research design, the authors assumed that the data set denoted effective couples of the similar group of students to avoid bias, and the
data set was approximately normally distributed. The nursing students described modest levels of emotions towards statistics subjects, reasonable ability in applying mathematics, and fair self-assurance in using computers for statistics (Hagen et al., 2013). Regarding the preferred learning styles factor, the students seemed convinced that what they did not achieve at the beginning, would enable them to learn during the course. According to the preferred teaching styles factor, the students had favourable opinions towards their instructors, who took into consideration perfect descriptions, practicality, persistence, in-depth knowledge of statistics, thorough monitoring of the course, and appropriate feedback to students, as well as clear learning expectations.
Hagen et al. (2013) achieved a surprising result, when the students‟ fear and anxiety scores were decreased. However, the post-test results revealed the intense variation, compared to the pre-test. The students achieved some significant changes in favourites around “learning styles”, among the pre-test and the post-test of the course. The majority of them also registered a positive experience with the “team-based” knowledge method used in the course. The preferences for “teaching styles” had the lowest improvement of all the student behaviours by the termination of the course. While most behaviours remained unchanged, the “instructor teaching styles”, which appeared very significant to the students at the preliminary of the course, decreased rather in significance by the conclusion of the course. This study was conducted in a particular area, with a single programme (nursing), using a few respondents, and as a result, generalization might not be possible. The changes between the pre-test and post-test were difficult to explain, because of elements in the research design, such as the lack of randomization and the use of a control group, which made the study susceptible to various threats, such as internal validity, and many other details for the variations observed over time. Some changes observed among the variables in the pre-test and post-test were statistically significant, though these variations were negligible, which could make it meaningless to nurse educators in the teaching of statistics to students.
Hannigan et al. (2014) explored SATS among students from various backgrounds, at a medical school in Ireland. The respondents completed the survey including the demographic information, as well as the prior learning experiences section, at the
Hannigan et al. (2014) explored SATS among students from various backgrounds, at a medical school in Ireland. The respondents completed the survey including the demographic information, as well as the prior learning experiences section, at the