Lectura 1 resumida

Scholarship of Teaching and Learning in Psychology: Statistics Is Puzzling

Citation

• Author: Miranda M. McIntyre

• Institution: California State University, San Bernardino

• Publication Date: October 19, 2020

• DOI: http://dx.doi.org/10.1037/stl0000204

Introduction

• Statistics courses are perceived as intimidating by many psychology students.

• Students often demonstrate low interest and high anxiety regarding statistics (Druggeri et al., 2008; Rajecki et al., 2005; Zeidner, 1991).

• Low mathematical self-efficacy negatively impacts performance in quantitative courses (Finney & Schraw, 2003; Walker & Brakke, 2017; Williams & Williams, 2010).

• Challenges for instructors arise from students’ lack of interest and negative beliefs regarding statistics.

Objective of the Study

• Introduce and test an active learning strategy using puzzles to engage students and enhance learning.

• Build on evidence-based strategies like collaborative and game-based learning.

• Implement an escape-room-inspired activity aimed at improving student content knowledge and self-efficacy.

Pedagogical Framework

Evidence for Collaborative Learning

• In collaborative learning, students work together with peers, leveraging shared knowledge (Johnson et al., 2007; Li, 2013; Prince, 2004).

• Advantages include:

  • Improved performance via explanation and error correction (Krause et al., 2009; Nokes-Malach et al., 2015; Schoor et al., 2015).

  • Larger learning gains compared to individual learning (Johnson et al., 2007; Lewis, 2011; Miller & James, 2019).

• Particularly beneficial for underrepresented groups (Drane et al., 2014; Lewis, 2011).

Evidence for Game-Based Learning

• Game-based learning incorporates games into curricula, engaging students through playful and immersive experiences.

• Many forms exist including video games, virtual reality, and board games.

• Benefits include:

  • Improved assessment performance (Cardozo et al., 2016; Luchi et al., 2017).

  • Enhanced motivation and positive attitudes towards learning (Banfield & Wilkerson, 2014; Connolly et al., 2012; Plass et al., 2013; Sung & Hwang, 2013).

• Educational escape rooms are a trending method that foster problem solving in groups (Nicholson, 2018).

• Initial evidence suggests these activities do increase student learning (e.g., LaPaglia, 2020).

Current Study

Research Questions

• Main goals to assess:

  • Hypothesis 1 (H1): Increased content knowledge post-puzzle activity versus pre-activity.

  • Hypothesis 2 (H2): Increased self-efficacy post-puzzle activity versus pre-activity.

  • Hypothesis 3 (H3): Better exam performance for participants in puzzle activities versus non-participants.

Method

Design

• A 2×2 mixed ANOVA with:

  • Repeated within-subjects factor (pre-activity, post-activity).

  • Between-subjects factor: laboratory section.

• Primary dependent variables: content knowledge and self-efficacy.

Participants

• 48 students from an introductory psychological statistics course participated (N = 48).

  • Composition: 62.5% sophomores, 31.3% juniors, 68.8% psychology majors.

• Consent obtained from those present in sessions, total sample for analyses was 30 students.

Procedure

• Two lab sections (24 students each) attended joint lectures.

• Review sessions occurred 5 weeks apart, covering material via collaborative puzzles in groups of 3-5 students.

• Topics included:

  • Session 1: levels of measurement, frequency distributions, statistical notation, z-scores, hypothesis statements.

  • Session 2: significance testing, independent and dependent t-tests, degrees of freedom, critical values.

• Pre-activity and post-activity self-efficacy assessed via three question measures ranked from 1 to 7.

Measures

• Content Knowledge Assessment:

  • 12-item multiple-choice quizzes administered pre- and post-activity.

  • Example questions:

  • “A variable has values with a specific order and a true zero starting point. What level of measurement is this variable?”

  • “For an independent t-test, how are degrees of freedom calculated?”

• Self-Efficacy Assessment:

  • 3-item measure regarding comfort, confidence, and preparedness for upcoming exams.

  • Strong internal reliability (Cronbach’s alpha = .933 for Session 1; .901 for Session 2).

Results

Session 1

• Content Knowledge:

  • No significant difference, t(33) = 1.52, p = .067.

  • Pre-activity: average 74.6% (SD = 12.8%); Post-activity: average 81.1% (SD = 12.4%).

• Self-Efficacy:

  • Increased significantly post activity, t(33) = 3.55, p < .001.

  • Pre-activity: average 4.07 (SD = 1.11); Post-activity: average 5.27 (SD = 0.80).

• Exam Performance:

  • No significant difference, t(33) = 1.13, p = .13.

  • Average score: Pre-activity 73.7% (SD = 13.5%); Post-activity 68.6% (SD = 12.5%).

Session 2

• Content Knowledge:

  • Significant increase post activity, t(37) = 4.13, p < .001.

  • Pre-activity: average 48.7% (SD = 21.2%); Post-activity: average 75.0% (SD = 18.5%).

• Self-Efficacy:

  • No significant difference, t(37) = 0.83, p = .207.

  • Pre-activity: average 4.53 (SD = 1.01); Post-activity: average 4.78 (SD = 0.93).

• Exam Performance:

  • Slight increase post activity, t(37) = 2.01, p = .026.

  • Average score: Post-activity 80.0% (SD = 10.0%); Pre-activity 72.4% (SD = 12.0%).

Omnibus Analysis

• Main effects of condition:

  • Content Knowledge: Significant, F(1, 28) = 20.17, p < .001, partial η² = .42.

  • Self-Efficacy: Significant, F(1, 28) = 9.03, p = .006, partial η² = .24.

  • Exam Performance: Not significant, F(1, 28) = 0.16, p = .69.

Exploratory Analyses

• Review attendees showed lower decreases in exam scores compared to non-attendees.

• Compared to prior terms and traditional reviews, the puzzle format proposed possible performance improvement.

Discussion

• Results support all three hypotheses:

  • Higher content knowledge and self-efficacy post-puzzle activity.

  • Review attendees performed better than non-attendees.

• Limitations include small sample size and lack of direct comparison to traditional methods.

• Future investigations should focus on the specific components of the activity that foster learning effects and how engagement strategies can mitigate negative attitudes towards statistics.

References

• Comprehensive list of cited studies and publications.

Supplemental Materials

• Available at: http://dx.doi.org/10.1037/stl0000204.supp

The results of the study show a distinction between psychological confidence and academic performance across the two sessions. In Session 1, the puzzle activity was successful in significantly boosting student self-efficacy (confidence), but it did not lead to a statistically significant increase in content knowledge or exam performance at that time. In Session 2, however, the results flipped: students showed a massive and significant gain in content knowledge (increasing from $48.7\%$ to $75.0\%$) and a significant improvement in exam scores, although their confidence levels did not see the same boost as in the first session. What clearly "worked" overall was the intervention's ability to mitigate the typical decline in grades seen as statistics courses get harder; while exam scores in Session 1 appeared lower numerically, exploratory analysis revealed that attendees still performed better and experienced less of a grade drop than those who did not participate. This suggests the puzzles are effective at building early confidence and eventually translating that engagement into better mastery of complex statistical concepts.

Based on the study, all three hypotheses (H1, H2, and H3) were eventually supported, but the 'success' looked different in each session:

1. Did the Hypotheses come true?
   Yes. According to the Discussion section, the results supported all three main goals:

• H1 (Content Knowledge): Supported by the massive increase in Session 2 ($48.7\%$ to $75.0\%$).

• H2 (Self-Efficacy): Supported by the significant boost in confidence during Session 1 (p < .001).

• H3 (Exam Performance): Supported because attendees performed better than non-attendees and showed smaller grade decreases as the course difficulty increased.

1. What was more successful?
   It depends on how you define success:

• Session 1 was more successful psychologically: It was highly effective at reducing the 'intimidation' factor of statistics, significantly raising student confidence (Self-Efficacy), even though their test scores didn't jump immediately.

• Session 2 was more successful academically: It led to a statistically significant improvement in actual exam scores ($p = .026$) and a major leap in understanding the material (Content Knowledge).

In summary, the puzzle activity was successful because it provided a 'confidence cushion' early on, which then translated into better mastery of the more difficult statistical concepts later in the term.