Investigating the Impact of an Intervention to Promote Mastery Goal Orientation

Investigating the Impact of an Intervention to Promote Mastery Goal Orientation

Experts acknowledge that motivation is amenable to change as a function of contextual factors (Wigfield et al., 2015), highlighting the importance of external influences in motivation. Accordingly, interventions have been designed to alter motivation and, ultimately, improve academic outcomes (Lazowski & Hulleman, 2016). One such intervention is grounded in the achievement goal theory, which proposes that individuals adopt achievement goals or specific purposes for engaging in academic activities (Dweck, 1986), resulting in unique patterns of affect, cognitive processing, and actions toward success or failure in learning situations (Ames & Archer, 1988). Although multiple achievement goal theoretical models exist, most relevant to the study’s purpose is the dichotomous framework, distinguishing between mastery (focus on developing competence) and performance (desire to demonstrate competence compared to others) goal orientations (Dweck, 1986).

A consistent pattern of positive outcomes associated with a mastery goal, as opposed to a performance goal, is well-documented (for a review, see Senko et al., 2011). Particularly, students who adopt a mastery goal report optimal academic experiences, whereas performance goal adoption is linked to both negative and positive academic outcomes. Given that a mastery goal dependably produces positive outcomes, achievement goal interventions usually seek to enhance a mastery-focused goal, utilizing two techniques: adopting pedagogical practices consistent with a mastery goal climate or encouraging individuals to adopt a personal mastery goal and abandon performance goal orientations (Elliot & Hulleman, 2017; Linnenbrink-Garcia et al., 2016). We tested the latter technique.

Researchers have investigated this type of mastery goal intervention across varied contexts, including college-level courses (Elliot & Hulleman, 2017). For instance, Hoyert and O’Dell (2006) taught introductory psychology students about achievement goals and relayed the importance of adopting a mastery goal. Participants then completed exercises related to the lesson. The intervention increased academic performance and mastery goal endorsement and decreased performance goal adoption. Later, Hoyert et al. (2012) found that the same intervention increased course enrollment and graduation rates of introductory psychology students 5 years post-study.

Generally, these studies suggest that mastery goal interventions are robust in promoting mastery goal adoption and enhancing positive academic outcomes. The present study attempted to replicate the findings from Hoyert and O’Dell (2006) and Hoyert et al. (2012) and to extend the research in four ways. First, we aimed to demonstrate the generalizability of the intervention’s impact in an upper-division asynchronous online elective psychology course. Examining the intervention in this context is critical, especially considering that mastery goal adoption tends to decline as students advance through school (Anderman & Midgley, 1997), and also knowing that motivation is critical for persistence in online classes (Bawa, 2016). Second, whereas prior studies predominantly utilized a pretest/posttest experimental design, this study assessed the outcomes four times over a semester. Third, we assessed potentially important outcomes not measured in the prior studies, including fear of failure and self-efficacy. Fourth, given that perceived classroom goal structure could impact personal achievement goals, self-efficacy, achievement, and other outcomes (Federici et al., 2015; Urdan, 2010), it was used as a covariate, a phenomenon unexplored in the prior research.

In sum, we investigated the effect of a mastery goal adoption intervention on upper-division college students’ mastery and performance goal orientations, fear of failure, self-efficacy, and academic performance across multiple time points within a semester, and in comparison to a control group. Given Hoyert and O’Dell (2006) and Hoyert et al. (2012) findings along with prior research linking mastery goal orientation to higher self-efficacy (Sakiz, 2011) and achievement (Grant & Dweck, 2003) and lower fear of failure (Elliott & Church, 1997; Lou & Noels, 2016, 2017; Madjar et al., 2017), we hypothesized that the current intervention would increase self-efficacy, academic performance, and personal mastery goal orientation and decrease fear of failure and performance goal adoption.

Method

Data Analysis

The manipulation check responses were analyzed using descriptive statistics and repeated measures analysis of variance (ANOVA). To address the main research question, we used generalized estimating equations (GEEs) to model the outcome variables accounting for repeated measures (Liang & Zeger, 1986). The generalized estimating equation is typically used for repeated measures but accounts for dependence among observations and considers within-subject correlations (Liang & Zeger, 1986). The quasi-likelihood under independence model criterion (Pan, 2001) was utilized in a stepwise modeling approach designed to minimize the value. The same approach was used to select the correlation structure. In the final step, type III analysis was considered to specifically test the intervention’s impact over time, accounting for other significant variables in the model. This approach was utilized for each reported outcome.

Importantly, the primary function of GEE is to present the model with the best fit. To accomplish this, we analyzed crude GEE models with all potential contributors and determined which were significant based on their p-values. Some classroom goal structure measures did not make a significant impact, and therefore, they were dropped from the final models. Table 1 shows means and standard deviations (SDs) of each outcome variable by group over time. The full GEE results appear in Table 2.

OPEN IN VIEWER

Table 1.

Means and Standard Deviations of the Outcome Variables by Group Over Time.

	Intervention		Control
	M	SD	M	SD
Mastery goal
PreI	6.15	.71	6.25	.72
Immediate PI	6.19	.58	5.89	.70
Delayed PI	6.00	1.00	5.91	1.25
Follow-up	6.04	.79	6.04	1.08
PAP goal
PreI	3.53	1.26	3.40	1.50
Immediate PI	2.87	1.14	3.22	1.46
Delayed PI	2.82	1.32	3.04	1.32
Follow-up	2.60	1.25	2.81	1.34
Fear of failure
PreI	2.99	1.05	3.21	1.13
Immediate PI	2.63	1.15	3.05	1.16
Delayed PI	2.59	1.11	2.80	1.11
Follow-up	4.04	.68	3.70	.80
Self-efficacy
PreI	5.79	.82	5.99	.65
Immediate PI	6.04	.77	5.79	.77
Delayed PI	5.44	.90	5.47	1.44
Follow-up	5.79	.76	5.75	.80
Exam grades
PreI	85.16	12.82	86.24	7.73
Immediate PI	93.45	5.90	92.83	7.18
Delayed PI	83.10	7.96	80.76	10.43
Follow-up	86.97	9.95	81.31	6.95

Note. PAP = performance-approach; PreI = pre-intervention; PI = post-intervention; SD = standard deviation; M = mean. The fear of failure scale ranged from 1 (do not believe at all) to 5 (believe 100% of the time). All other scales ranged from 1 (strongly disagree) to 7 (strongly agree).

OPEN IN VIEWER

Table 2.

Generalized Estimating Equation (GEE) Results for Each of the Five Outcome Variables.

Variable	β	SE	Wald 95% CI	P
Mastery goal
Intervention	.09	.19	[−.28, .46]	.64
Control	0
PreI	0
Immediate PI	−.17	.08	[−.34, −.01]	.03
Delayed PI	−.25	−.14	[−.53, .02]	.07
Follow-up	−.18	.09	[−.35, −.01]	.04
PAP goal
Intervention	−.08	.29	[−.64, .49]	.79
Control	0
PreI	0
Immediate PI	−.36	.13	[−.62, −.11]	.005
Delayed PI	−.51	.16	[−.81, −.20]	<.001
Follow-up	−.74	.18	[−1.10, −.39]	<.001
PAP CGS	.23	.11	[.01, −.44]	.04
Fear of failure
Intervention	−.13	.20	[−.52, .27]	.52
Control	0
PreI	0
Immediate PI	−.25	.11	[−.47, −.02]	.03
Delayed PI	−.40	.11	[−.61, −.18]	<.001
Follow-up	.78	.17	[.45, 1.11]	<.001
Self-efficacy
Intervention	−.03	.17	[−.36, .29]	.85
Control	0
PreI	0
Immediate PI	.03	.09	[−.13, .20]	.69
Delayed PI	−.41	.17	[−.74, −.08]	.02
Follow-up	−.12	.11	[−.34, .09]	.25
MAP CGS	.32	.09	[.13, .50]	<.001
Exam grades
Intervention	.38	1.79	[−3.12, 3.89]	.83
Control	0
PreI	0
Immediate PI	7.44	1.18	[5.12, 9.76]	<.001
Delayed PI	−3.77	1.34	[−6.39, −1.15]	.005
Follow-up	1.44	1.34	[−1.19, 4.07]	.28

Note. PAP = performance-approach; PreI = pre-intervention; PI = post-intervention; CI = confidence interval; SE = standard error; PAP = performance-approach; MAP = mastery-approach; CGS = classroom goal structure.

Results ¹

Discussion

Hoyert and O’Dell (2006) and Hoyert et al. (2012) demonstrated that encouraging undergraduate introductory psychology students to adopt a personal mastery goal enhances academic outcomes. We sought to replicate these findings in an upper-division online elective educational psychology course. The current results, however, did not replicate prior findings as the intervention and control groups yielded similar post-intervention scores over time. Our findings suggest that a mastery goal adoption intervention had no impact on mastery and performance-approach goals, fear of failure, self-efficacy, nor academic performance. Nonetheless, there are plausible explanations for these discrepant results, possibly expanding our understanding of mastery goal adoption interventions.

First, the course under investigation was an elective. Students likely enrolled because of its perceived usefulness in accomplishing goals (e.g., becoming school psychologists, teachers, and academic counselors) or personal interest. Considering that research has demonstrated that endorsing mastery-approach goals is linked to higher perceived utility value and interest (Hulleman et al., 2008), it is reasonable to think that at enrollment, students were committed to understanding course topics. This could help explain the potency of mastery goal interventions among introductory psychology students demonstrated in Hoyert and O’Dell (2006) and Hoyert et al. (2012) studies. Perhaps students enrolled in required lower-level courses begin with lower levels of a mastery goal, desiring merely to “get by” rather than understanding course material. In these circumstances, the intervention would have been more robust in boosting a personal mastery goal. This is an area of research warranting further investigation.

Unlike previous studies, our study was conducted in a fully asynchronous online course. Differences in course delivery format might explain inconsistencies in results. In face-to-face classes, students can directly communicate with and receive more personally scaffolded instruction from instructors and peers. However, online students have limited physical interactions with instructors and peers who can guide them. Instead, they must engage in effective self-regulation for academic success (Sansone et al., 2011). Self-regulation theories postulate that motivation is critical to self-regulation (see Muis, 2007; Winne & Hadwin, 2010), especially when learning online (Winters et al., 2008). Hence, it is possible that students in the current study, recognizing the importance of endorsing, directing, and maintaining optimal motivation for success, developed strategies for managing motivation. Thus, they did not require additional guidance toward adopting a mastery goal orientation. Further research is needed to explore this phenomenon.

Additionally, two competing theoretical perspectives explain the benefits of adopting varying types of achievement goals (i.e., mastery goal perspective and multiple goals perspective; Hulleman et al., 2010). The mastery goal perspective proposes that a mastery goal is more beneficial than performance goals. The multiple goals perspective, however, posits that adopting performance-approach and mastery goals can be potent (Harackiewicz et al., 2002). The current intervention was based on the mastery goal perspective suggesting that students are most academically empowered when they adopt a mastery goal. Yet research has corroborated the multiple goals perspective, revealing that students adopt different combinations of mastery and performance-approach goals, impacting academic outcomes (see Senko et al., 2011). Highlighting the usefulness of both types of goals could have been a more potent intervention in an upper-division course. These classes are often more rigorous than introductory courses, requiring deeper learning strategies to achieve academic success. Prior research has shown that in advanced classes and for challenging tasks, performance goals predict grades (Barron & Harackiewicz, 2003; Darnon et al., 2009; Senko et al., 2013), while mastery goals predict interest (Barron & Harackiewicz, 2003). Evidence is decidedly mixed, given that other researchers have revealed that mastery goals more strongly predict achievement than performance goals (Bereby-Meyer & Kaplan, 2005). Accordingly, we suggest that a combined goal intervention might have had a more broad-ranging impact on our sample’s academic outcomes.

We also acknowledge the study’s limitations. First, studies note that goal endorsements can change because of other contextual factors, including types of tasks (Fryer & Elliott, 2007; Muis & Edwards, 2009) and instructor feedback (Winne et al., 2003). These findings suggest that mastery goal adoption is amenable to change not simply via interventions but by other course elements. Therefore, it is difficult to ascertain whether the current results are related to the intervention or a combination of the intervention and other environmental features.

We also recognize that the present study included a relatively small pool of students enrolled in an upper-division online elective course at one university. Given the sample’s distinct features, we cannot generalize the study’s findings to the larger population of college students. The sizeable proportion of women in the sample is an additional limitation. This issue is particularly pertinent as evidence suggests that women are generally more likely to endorse a mastery goal than men (Anderman & Young, 1994). That means women likely enroll in courses with higher levels of a mastery goal than men, tempering any attempts to encourage their endorsement of a mastery goal. We used the means of the pre-intervention mastery goal scores to examine this possibility in our sample. Women had a mean score of 6.20 and men, 5.85. However, given the male sample size (n = 8), we did not have adequate statistical power to detect accurate gender differences. Nonetheless, a larger and more diverse sample might have curtailed any potential gender effect.