Psychologically Controlling Teaching:Examining Outcomes, Antecedents, and Mediators

Bart SoenensGhent University

Eline SierensCatholic University of Leuven

Maarten VansteenkisteGhent University

Filip Dochy and Luc GoossensCatholic University of Leuven

Psychologically controlling teaching (PCT) refers to the use of intrusive behaviors that pressure studentsto act, think, and feel in particular ways. The goal of the present research was to examine the dynamicsinvolved in PCT. Study 1 examined self-regulated learning and achievement outcomes of PCT, whereasStudy 2 examined antecedents (i.e., individual and environmental pressures). In Study 1, we found thatPCT was related negatively to students use of self-regulation strategies, which, in turn, was positivelyrelated to academic achievement. Students relative autonomy for studying played an intervening role inthese associations. Findings of Study 2 revealed that both pressure from above (i.e., a pressuring schoolenvironment) and pressure from within (i.e., teachers low relative autonomy for teaching) were relatedto PCT, whereas pressure from below (i.e., students low relative autonomy for studying) was not. Theseassociations could be accounted for by depersonalization, one component of teacher burnout. Thediscussion focuses on how PCT represents one aspect of the dark side of teaching, which isunderstudied in educational psychology.

Keywords: psychological control, teaching style, motivation, self-determination

Most of you scored poorly on the last test, and in the last fewdays, you have been behaving like little children. I am disap-pointed in this class. Such a statement is indicative for teach-ers use of psychological control. Psychologically controllingteaching (PCT) refers to the use of intrusive and sometimessubtle behaviors that pressure students to act, think, and feel inparticular ways. Common to these intrusive behaviors (includ-ing guilt induction, shaming, and expression of disappointment)is that they convey a conditionally approving attitude fromteachers toward students. In this article, we introduce PCT inthe teaching literature, drawing on Barbers work in the par-enting literature (Barber, 1996; Barber & Harmon, 2002) andself-determination theory (SDT; Deci & Ryan, 2000; Vansteen-kiste, Niemiec, & Soenens, 2010).

Psychological Control: From Parenting to Teaching

Historically, the concept of parental psychological control wasdiscovered in a rather inductive, bottom-up fashion. Throughlarge-scale factor analyses, early scholars in the domain of parent-ing identified a number of highly correlated parenting strategiessuch as guilt induction, shaming, and love withdrawal and referredto this set of strategies as psychological control (Schaefer, 1965).During the past two decades, the concept of parental psychologicalcontrol has been examined intensively (Barber & Harmon, 2002;Soenens & Vansteenkiste, 2010). Also, scholars have attempted toprovide a more precise and encompassing definition of psycho-logical control. Parental psychological control has been defined byBarber (1996, p. 3297) as

. . . a rather insidious type of control that potentially inhibits or in-trudes upon psychological development through manipulation andexploitation of the parentchild bond (e.g., love-withdrawal and guiltinduction), negative, affect-laden expressions and criticisms (e.g.,disappointment and shame), and excessive personal control (e.g.,possessiveness, protectiveness).

Parental psychological control essentially involves a condition-ally approving attitude toward children, where parents withdrawlove and appeal to feelings of guilt and shame when children failto succeed at parentally valued tasks (Assor, Roth, & Deci, 2004;Barber & Xia, in press). Numerous studies have shown that psy-chologically controlling parenting is related to maladaptive devel-opmental outcomes in children and adolescents and to internaliz-ing problems in particular (Soenens, Luyckx, Vansteenkiste,Duriez, & Goossens, 2008). In the academic domain, psycholog-

This article was published Online First October 3, 2011.Bart Soenens, Department of Developmental, Personality, and Social

Psychology, Ghent University, Ghent, Belgium; Eline Sierens and LucGoossens, Center for School Psychology and Child & Adolescent Devel-opment, Catholic University of Leuven, Leuven, Belgium; Maarten Van-steenkiste, Department of Developmental, Personality, and Social Psychol-ogy, Ghent University; Filip Dochy, Center for Research on Teaching andTraining, Catholic University of Leuven.

Eline Sierens is now at Artevelde Hogeschool Ghent, Belgium.Correspondence concerning this article should be addressed to Bart

Soenens, Department of Developmental, Social, and Personality Psychol-ogy, Henri Dunantlaan 2, Ghent B-9000, Belgium. E-mail: Bart.Soenens@ugent.be

Journal of Educational Psychology 2011 American Psychological Association2012, Vol. 104, No. 1, 108120 0022-0663/11/$12.00 DOI: 10.1037/a0025742

108

ical control has been found to relate to maladaptive learningstrategies (Vansteenkiste, Zhou, Lens, & Soenens, 2005) and tolow academic achievement (Aunola & Nurmi, 2004).

It has been argued that the use of psychological control is notunique to parents and that other socialization figures, includingteachers, might also engage in psychological control (Barber &Olsen, 1997). Specifically, psychologically controlling teaching(PCT) would be apparent when teachers use their own opinion andvalues as an exclusive frame of reference and ignore their studentsperspective. Using intrusive tactics such as guilt induction, sham-ing, and expression of disappointment, teachers would try tomotivate learners to comply with their frame of reference. Similarto psychologically controlling parents display of conditional re-gard, psychologically controlling teachers would show reducedconcern and involvement when students fail to reach teacher-imposed standards. Although teachers may sometimes use theseintrusive behaviors in a deliberate and intentional fashion, such aswhen they want to prompt attention in the class, at other times theymay engage in these behaviors impulsively, for instance, whenthey experience distress or pressure themselves. The overall aim ofthe current research is to examine how PCT relates to features ofstudents and teachers functioning. To better understand how PCTrelates to other teaching dimensions and how it relates to motiva-tional dynamics in both students and teachers, we draw fromself-determination theory.

Self-Determination Theory

Basic Psychological Needs

Self-determination theory (SDT; Deci & Ryan, 2000) is a broadtheory on motivation and personality development that has beenapplied intensively in educational psychology (Niemiec & Ryan,2009; Reeve, 2009; Vansteenkiste, Lens & Deci, 2006). At theheart of SDT is the formulation of three basic psychological needs,that is, the needs for autonomy (i.e., experiencing a sense ofvolitional and psychological freedom), competence (i.e., experi-encing personal effectiveness), and relatedness (i.e., experiencingcloseness and mutuality in interpersonal relationships). When sat-isfied, these needs would provide energy and foster well-being,adjustment, and high-quality motivation. In contrast, thwartedneeds would deplete individuals energy and vitality and wouldundermine their well-being and motivation.

Teachers Interpersonal StyleOne important factor involved in processes of need satisfaction

and need frustration is the interpersonal style used by socializationfigures, including teachers. Parallel to the distinction between thethree needs, SDT distinguishes between three dimensions of teach-ing style that would affect students need satisfaction and subse-quent motivation and adjustment (Grolnick, Deci, & Ryan, 1997;Reeve, 2009). First, involvement pertains to teachers engagementin a warm and supportive style of communication and to teacherscapacity to respond adequately to students distress. Involvementis contrasted with a cold, distant, and aloof teacher orientation.Second, structure refers to the provision of clear instructions in theclassroom and the communication of realistic goals and expecta-tions for behavior and learning. Structure also involves giving

positive feedback and providing adequate help during the learningprocess. Structure is contrasted with a chaotic class climate. Third,and essential to the topic of this research, autonomy support refersto teachers promotion of volitional functioning and teachersfostering of a sense of initiative and interest in students. Keyautonomy-supportive practices include taking the students frameof reference, demonstrating intrinsic interest in the learning mate-rial, providing choices and opportunities for initiative, and com-municating a meaningful rationale when choice is limited (Assor& Kaplan, 2001; Deci, Eghrari, Patrick, & Leone, 1994).

In SDT, autonomy support is contrasted with a controllinginterpersonal style, that is, a style where teachers ignore thestudents perspective and pressure the students to think, act, or feelin particular ways (Deci et al., 1994; Grolnick, 2003; Reeve,2009). According to SDT, a controlling style can be expressed inat least two different ways, that is, as internally or externallycontrolling (Ryan, 1982; Soenens & Vansteenkiste, 2010). Inter-nally controlling teaching refers to triggering internally pressuringforces in learners by appealing to students feelings of guilt,shame, anxiety, and self-worth. Activating those internal pressuresoften happens in a covert and subtle way. Externally controllingteaching refers to activating a sense of external obligation instudents by using rather overtly controlling strategies, such aspunishments, pressuring rewards, and explicitly controlling lan-guage, like you must (Vansteenkiste, Simons, Lens, Soenens, &Matos, 2005).

When framed from the SDT perspective, PCT clearly representsan autonomy-suppressing and controlling teaching strategy. It isrelatively less clear, however, whether PCT is fully equivalent tothe concept of an internally controlling style in particular (Soenens& Vansteenkiste, 2010). At the definitional level, PCT seems to bea somewhat broader construct than internally controlling teaching.According to Barbers (1996) definition, psychological controlinvolves an intrusive and manipulative parental orientation. Intheory, such an orientation might be expressed in both internallycontrolling practices (such as guilt induction) and in relativelymore externally controlling practices (such as explicit and harshcriticism). At the level of assessment, however, psychologicalcontrol is typically operationalized in terms of mainly internallycontrolling practices such as guilt induction, shaming, and lovewithdrawal. As such, the measure of PCT largely fits SDTsconcept of internally controlling teaching.

In sum, from the perspective of SDT, PCT is largely incompat-ible with teacher autonomy support and with need-supportiveteaching in general (Grolnick, 2003; Vansteenkiste, Zhou, et al.,2005). It is particularly interesting to examine psychological con-trol in the teaching context because SDT-based research on teach-ing style has tended to focus on the adaptive side of each of theteaching dimensions and their beneficial effects on students learn-ing (but see, for example, Deci, Schwartz, Sheinman, & Ryan,1981). An examination of the concept of PCT may provide anopportunity to address the relatively neglected dark side ofteaching.

Consequences of Teaching Style

According to SDT, students perceptions of their teachers asneed-supportive or need-frustrating (including perceptions ofteachers as being psychologically controlling) have important ram-

109PSYCHOLOGICALLY CONTROLLING TEACHING

ifications for their motivational orientation and subsequent learn-ing strategies and achievement (Niemiec & Ryan, 2009; Vansteen-kiste, Zhou, et al., 2005). In SDT, the quality of motivation isconceptualized in terms of the degree to which motivation isautonomous (vs. relatively more controlled) in nature. Autono-mously motivated students learn in a self-endorsed or volitionalfashion because of the perceived value (i.e., identified regulation)or because of the inherent satisfaction (i.e., intrinsic motivation) ofthe learning activity (Ryan & Connell, 1989). In contrast, studentswith a controlled motivation learn to meet external (e.g., rewards,external regulation) or internal (e.g., feelings of guilt, introjectedregulation) pressures (Ryan, Connell, & Grolnick, 1992). Severalstudies demonstrated that an autonomous, relative to a controlled,study motivation is associated positively with students use ofeffective and thorough strategies of self-regulated learning such astime management and deep-level cognitive processing (Vansteen-kiste, Sierens, Soenens, Luyckx, & Lens, 2009) and grades (Black& Deci, 2000).

Whereas need-supportive teaching would foster a relativelymore autonomous motivation in students, need-frustrating teach-ing is thought to relate negatively to relative autonomy for study-ing (Niemiec & Ryan, 2009). Consistent with this hypothesis,research has indeed shown positive associations between teacherautonomy support and autonomous (relative to controlled) motivesfor studying (Vansteenkiste, Zhou, et al., 2005). Few studies,however, have examined how student perceptions of autonomy-suppressing teacher practices relate to students quality of moti-vation. Herein, we propose that perceived PCT, as it represents afeature of need-frustrating teaching, would relate negatively toautonomous (relative to controlled) motivation for studying and tosubsequent self-regulated learning strategies and grades.

Antecedents of Teaching Style

In addition to providing an analysis of the consequences ofneed-thwarting teaching (including PCT), SDT offers a theoreticalperspective on the antecedents of teaching style. Specifically withregard to controlling teaching, it is argued that pressuring featuresin teachers own functioning and interpersonal environment in-crease the likelihood of engagement in a controlling teaching style.Such need-frustrating features and pressures might deplete teach-ers energy. As a consequence, they may fail to engage in avitalizing and autonomy-supportive style and instead look forshortcuts to make the students comply with the teachers agenda.This would be reflected in the use of controlling practices andpsychological control in particular.

We distinguish three sources of pressure similar to the sourcesdiscerned in the parenting and SDT literature: pressure fromabove, from within, and from below (Grolnick, 2003). Within theteaching context, pressure from above refers to pressure from theenvironment, in particular from parents, colleagues, and principals.Such pressure can be characterized, for example, by frequentevaluations and forced conformity to colleagues teaching methods(Pelletier, Seguin-Levesque, & Legault, 2002). Pressure fromwithin and pressure from below refer to stressful conditions in thefunctioning of teachers themselves and of their students, respec-tively. One important feature of pressure from within is teachersown motivational orientation. Much like students motivation forschool can be more or less autonomous (vs. controlled) in nature,

teachers differ in how autonomously (vs. controlled) they aremotivated to do their work. Similarly, teachers may perceive theirstudents as high or low on quality of motivation, and a perceptionof low-quality motivation in students may represent another sourceof pressure for teachers.

Various studies have examined associations between these pres-sures and teaching style, thereby typically focusing on one partic-ular source of pressure. For example, Roth, Assor, Kanat-Maymon, and Kaplan (2007) demonstrated, using multilevelmodeling, that teachers self-reported autonomous, relative to theircontrolled, reasons for teaching was positively associated withchild-reported perceptions of autonomy-supportive teaching. Fewstudies, however, examined the combined effects of the threeforms of pressure, with the exception of Taylor, Ntoumanis, andStandage (2008). These authors found that perceived job pressure(pressure from above), teachers autonomous causality orientation(low pressure from within), and perceptions of students relativeautonomy (low pressure from below) were all related to teachersuse of adaptive motivational strategies. In this research, we furtherexamined whether the three types of perceived pressure wouldyield a significant relation with the maladaptive teaching dimen-sion of PCT. It was expected that the exposure to any of thesepressuring sources would prompt teachers to act in a psychologi-cally controlling way.

The Present Set of Studies

We present the results of two studies intended to address twobroad aims. First, we aimed at developing a self-report measure ofperceived PCT and demonstrating its reliability and validity. Sec-ond, we wanted to study the dynamics of perceived PCT byexamining both outcomes (Study 1) and antecedents (Study 2) ofperceived PCT. Specifically, we hypothesized that perceived PCTwill be negatively associated with students autonomous motiva-tion and self-regulated learning (Study 1, outcomes) and will bepositively related to rigid and controlling forces in teachers ownfunctioning and working climate (Study 2, antecedents). The hy-pothesis that controlling dynamics would be involved in PCT fitswith SDT, which maintains that both teachers and students willfunction in a more controlled and dysfunctional fashion whenencountering pressuring events or persons (Deci & Ryan, 2000).

Study 1: Psychologically Controlling Teaching andStudents Motivation and Learning

The first goal of Study 1 was to develop and validate a measureof perceived PCT. To externally validate our scale, we examinedassociations between perceived PCT and a well-validated measureof dimensions of need-supportive teaching, that is, involvement,structure, and autonomy support (Skinner & Belmont, 1993). Thesecond goal of Study 1 was to examine associations betweenperceived PCT and students motivation, learning, and perfor-mance. On the basis of SDT, we hypothesized that PCT wouldrelate negatively to autonomous (vs. controlled) motives for study-ing as well as to strategies of self-regulated learning and grades.

Self-regulated learning (SRL) has been defined as a set ofproactive and self-focused processes in which students construc-tively monitor their learning toward the completion of academictasks (Zimmerman, 2008). In our study, as in many other studies

110 SOENENS ET AL.

(e.g., Case, Harris, & Graham, 1992), we concentrated on thecognitive and meta-cognitive components of SRL. Cognitive strat-egies comprise deep-level cognitive strategies such as criticalthinking and summarizing. Students who use deep-level process-ing avoid simply memorizing the learning material by repetition,but want to master the learning material thoroughly (Rozendaal,Minnaert, & Boekaerts, 2005). Meta-cognitive strategies refer toplanning, organizing, and self-monitoring ones learning process(Winne, 1995). We hypothesized that perceived PCT would benegatively related to both types of self-regulatory capacities. Be-cause a lack of self-regulatory capacities represents a risk factorfor low academic achievement (Pintrich & De Groot, 1990), wealso examined whether perceived PCT is related to lower schoolachievement through its negative association with self-regulationstrategies. Finally, we hypothesized that associations between per-ceived PCT and SRL (and subsequent grades) would be mediatedby students quality of motivation (i.e., autonomous vs. con-trolled).

MethodParticipants and procedure. The sample comprised 533

adolescents (11th and 12th grade; 41.2% boys) from nine second-ary schools from the academic track in Flanders (Belgium). Par-ticipants mean age was 16.9 years (SD 0.7 years). The ques-tionnaires were administered during a class period of 50 min,with the first author being present to answer questions. One or twoteachers were also present during the administration of the surveys.Students permission to participate in the study and for researchersto scrutinize their exam scores was obtained through a procedureof passive informed consent. Specifically, the researcher orallyexplained the purpose of the study and the necessity to use stu-dents exam scores. Confidential treatment of the data was guar-anteed. All students were told they could refuse participation anddeny permission for their exam scores to be used in the study byfilling out a form. None of the students chose to do so. A total of 511exam scores were provided by the school board 1 month later.

Measures. All questionnaires in Study 1 and Study 2 weretranslated from English to Dutch, the participants mother tongue,using the guidelines of the International Test Commission(Hambleton, 1994). All scales used a 5-point Likert-type scale,ranging from 1 (strongly disagree) to 5 (strongly agree), unless

otherwise indicated. Scale scores were computed as the means ofitems. Means and standard deviations for all study variables areprovided in Table 1.

Psychologically controlling teaching. A group of scholarsexperienced with research on psychological control (i.e., the firstthree authors) selected seven items from two frequently used andvalidated scales, that is, the Psychological Control ScaleYouthSelf-Report (PCSYSR; Barber, 1996) and the Psychological Con-trol Scale of the Childs Report of Parental Behavior Inventory(CRPBI; Schaefer, 1965). Five items were almost literally takenover, but the socialization figure referred to was changed fromparents to teachers (e.g., My teachers often interrupt me). Twoitems were slightly adjusted to tap PCT (e.g., My teachers clearlyshow that I have hurt their feelings when I have failed to live upto their expectations). The resulting scale and its descriptivestatistics can be found in Table 2. Information about the psycho-metric characteristics of this scale is provided in the Resultssection.

Autonomy support, structure, and involvement. Theseteaching dimensions were measured using the short form of theTeacher as Social Context Questionnaire (TASC; Belmont, Skin-ner, Wellborn, & Connell, 1988). The TASC comprises subscalesfor Autonomy Support (eight items; e.g., My teachers give me alot of choices about how I do my schoolwork; .71), Structure(eight items; e.g., If I cant solve a problem, my teachers show medifferent ways to try to; .70), and Involvement (eight items;e.g., My teachers really care about me; .83).

Relative autonomy for studying. To assess students aca-demic motivation, we used an adapted, Dutch version (16 items;Vansteenkiste et al., 2009) of the Self-Regulation QuestionnaireAcademic initially introduced by Ryan and Connell (1989). Themeasure consists of four subscales, representing four differenttypes of motivation for studying, that is, external regulation (fouritems, e.g., Studying is what Im supposed to do; .82),introjected regulation (four items, e.g., I will feel bad aboutmyself if I do not study; .72), identified regulation (fouritems, e.g., It is personally important to me; .79), andintrinsic motivation (four items, e.g., I enjoy studying; .90).Similar to the procedure used in previous research (Soenens &Vansteenkiste, 2005), after assigning a weight to the four types ofmotivation according to their degree of autonomythat is, exter-

Table 1Correlations Among and Means, Standard Deviations, and Gender Differences for Study Variables in Study 1 (N 503)

Variable 2 3 4 5 6 7 8Total

M SDMale

M SDFemale

M SD Fa

1. PCT .22 .05 .08 .08 .15 .27 .14 2.28 0.71 2.50 0.73 2.13 0.66 37.252. Relative autonomy .26 .06 .24 .30 .39 .22 1.80 3.36 2.26 3.16 1.48 3.47 6.903. Elaboration .52 .53 .50 .28 .07 3.18 0.60 3.12 0.62 3.22 0.58 3.774. Organization .16 .33 .20 .04 3.08 0.84 2.90 0.87 3.20 0.80 16.515. Critical thinking .35 .13 .01 2.78 0.74 2.94 0.78 2.66 0.69 18.856. Meta-cognitive self-regulation .48 .23 3.19 0.49 3.14 0.49 3.23 0.49 4.687. Time and study environment .26 3.10 0.66 3.00 0.65 3.17 0.65 8.568. Academic achievement 66.47 7.16 65.15 7.13 67.40 7.06 12.55

Note. PCT psychologically controlling teaching.a df (1, 524) for the SRL variables; df (1, 526) for PCT; df (1, 528) for relative autonomy for studying; df (1, 509) for academic achievement. p .05. p .01.

111PSYCHOLOGICALLY CONTROLLING TEACHING

nal regulation 2; introjected regulation 1; identified regula-tion 1; and intrinsic motivation 2we summed theseweighted scores to create an index of relative autonomy for study-ing ( .78; see, e.g., Niemiec et al., 2006, for this procedure).

Deep-level cognitive strategy use. Participants were admin-istered three scales from the Motivated Strategies for LearningQuestionnaire (MSLQ; Pintrich, Smith, Garcia, & McKeachie,1991), namely, Elaboration (six items, e.g., When reading forclasses, I try to relate the material to what I already know; .56), Organization (four items, e.g., I make simple charts, dia-grams, or tables to help me organize course material; .67),and Critical Thinking (5 items, e.g., I treat the course material asa starting point and try to develop my own ideas about it; .75). Rather than tapping into participants use of deep-level cog-nitive learning strategies for a specific course, as is commonlydone, we assessed participants deep-level learning in general.

Meta-cognitive strategy use. We measured students use ofmeta-cognitive strategies using the subscales for Meta-CognitiveSelf-Regulation (12 items; e.g., Before I study new course mate-rial thoroughly, I often skim it to see how it is organized) andTime and Study Environment (eight items; e.g., I usually study ina place where I can concentrate on my course work) of theMSLQ. These items also pertained to participants use of meta-cognitive strategy use in general rather than with respect to aspecific course. The subscales had a Cronbachs alpha of .65(Meta-Cognitive Self-Regulation) and .70 (Time and Study Envi-ronment).

Academic performance. Academic performance was mea-sured by students overall grades at the end of the first semester(December), which were obtained through the school board.Grades ranged between 41% and 88%, with a mean of 66.47%(SD 7.17).

Plan of analysis. We examined the proposed model in whichPCT relates to motivation, SRL, and achievement using structuralequation modeling with latent variables (LISREL 8.7; Jreskog &Srbom, 1996). In line with Holmbecks (1997) recommendations,we tested the following models: (a) direct-effects models, (b)full-mediation models, and (c) partial-mediation models. Full me-diation is demonstrated when the addition of a direct path in thethird model does not improve fit compared with the second model.This three-step approach was used to test different portions of ourhypothesized model (see Figure 1). Assessment of model fit was

based on multiple criteria: the root-mean-square error of approx-imation (RMSEA), the comparative fit index (CFI), and the stan-dardized root-mean-square residual (SRMR). A RMSEA of .08 orbelow (Hu & Bentler, 1999), a CFI value .90 (Byrne, 1994), anda small SRMR value (.06; Quintana & Maxwell, 1999) indicateacceptable fit to the data.

Results

Preliminary analyses. Initial analyses focused on the psy-chometric characteristics of the newly developed instrument andassociations with related constructs.

Reliability and validity of the PCT ScaleStudent Report.To ensure the adequacy of the PCT Scale as a measure of per-ceived teacher psychological control, we investigated (a) the fac-torial validity and the internal consistency and (b) the externalvalidity of the scale by calculating zero-order correlations withmore adaptive teaching dimensions, that is, perceived autonomysupport, structure, and involvement.

To examine the factorial validity of the PCT Scale, we subjecteditems to a principal component analysis. This analysis revealed oneclear factor, accounting for 46% of the variance, with all itemloadings above 0.58 (eigenvalue of 3.20). Cronbachs alpha was.80. PCT was significantly negatively related to autonomy support(r .44, p .01), structure (r .38, p .01), and involve-ment (r .29, p .01). PCT and autonomy support were morestrongly negatively related than PCT and involvement (Fisher z 3.92, p .001) and PCT and autonomy support were somewhatmore strongly related than PCT and structure (Fisher z 1.67,p .09).

Correlational analyses and descriptive statistics. Correla-tion coefficients among the study variables are presented in Table1. PCT was significantly negatively related to students relativeautonomy for studying, to the two meta-cognitive strategy usesubscales, and to academic achievement. Unexpectedly, PCT wasunrelated to the use of deep-level cognitive strategies. Further,students relative autonomy for studying was significantly posi-tively related to all SRL strategies, except for elaboration, and toacademic achievement. Finally, the two meta-cognitive strategyuse subscales, but not the deep-level cognitive strategy use sub-scales, were significantly positively related to students academicperformance.

Table 2Descriptive Statistics of the Psychologically Controlling Teaching Items: Student Report

Item: My teachers . . .

Student responses (Study 1) Teacher responses (Study 2)

M Skewness Kurtosis M Skewness Kurtosis

1. Are always trying to change me. 2.09 0.89 0.34 2.08 0.59 0.352. Clearly show that I have hurt their feelings when I have failed

to live up to their expectations.2.03 0.76 0.03 2.35 0.34 0.94

3. Are less friendly with me, if I do not see things their way. 2.78 0.24 0.85 1.77 1.15 1.154. Are strict with me if I have disappointed them. 2.54 0.37 0.41 2.67 0.01 0.825. Make me feel guilty when I dissatisfied them. 2.42 0.42 0.67 1.98 0.66 0.586. Avoid talking with me when I have disappointed them. 1.78 1.33 1.61 1.23 3.40 13.287. Often interrupt me. 2.30 0.60 0.37 1.74 0.87 0.29

Note. The teacher report of the Psychologically Controlling Teaching Questionnaire is analogous to the student report. For example, one item from theteacher report version is I always try to change my students.

112 SOENENS ET AL.

Next, we examined gender differences because boys and girlshave been found to differ in various study variables, such asmotivation (Vallerand & Bissonnette, 1992) and meta-cognitivestrategy use (Vrugt & Oort, 2008). The multivariate analysis ofvariance (MANOVA) with gender as the between-subjects vari-able and the self-regulated learning scales as dependent variableswas significant, Wilks .89; F(5, 519) 12.52; p .001,2 .11. Follow-up univariate analyses (see Table 1) revealedthat female, relative to male, students scored higher on organiza-tion, elaboration, time and study environment, and meta-cognitiveself-regulation and lower on critical thinking. Univariate ANOVAson the remaining variables indicated that girls scored lower onperceived PCT and higher on relative autonomy for studying andacademic achievement. Because of these gender differences, weincluded gender as a control variable in all subsequent analyses.

Primary analyses. The primary analyses checked whetherthe constructs in the model were measured adequately and whetherthe expected associations held among these constructs.

Measurement model. We estimated a measurement model toassess whether the indicators represented the latent constructsproperly. We used parcels as indicators of the latent constructsPCT (three parcels) and relative autonomy for studying (fourparcels) by randomly assigning their respective items to one ofthe parcel groupings (Little, Cunningham, Shahar, & Widaman,2002). The subscales critical thinking, elaboration, and organiza-tion were used as indicators of the latent construct deep-levelcognitive strategy use, whereas meta-cognitive self-regulation andtime and study environment were used as indicators of the latentconstruct meta-cognitive strategy use. In addition, gender andacademic performance were each represented as a latent variablewith a single indicator, the error variance of which was set to 0.The measurement model with six latent variables represented by14 indicators approached the criteria for acceptable fit, 2(64) 318.34; RMSEA .09; CFI .93; SRMR .07. After a cross-

loading of meta-cognitive self-regulation on deep-level cognitivestrategy use was added, the fit of the revised model, 2(63) 291.17; RMSEA .08; CFI .94; SRMR .04, improvedconsiderably, 2(1) 27.17, p .001. Adding this cross-loadingseemed logical because meta-cognitive self-regulation and deep-level cognitive strategy use share a cognitive focus. Moreover, themodification did not substantially change the correlations amongthe latent factors, as indicated by the nonsignificant chi-square test,

2(15) 20.31, p .05. Factor loadings ranged from .47 to .99(mean lambda .74) and were all significant (p .001).

Structural model. To examine the association between PCTand academic achievement, we tested a direct effects model in-cluding only PCT as a direct predictor of achievement, 2(4) 3.80; RMSEA .00; CFI 1.00; SRMR .02. The direct pathfrom PCT to achievement was significantly negative, .12,p .05. In a next model, we added deep-level cognitive andmeta-cognitive strategy use as mediators in the relation betweenPCT and achievement. In this model, PCT was only indirectlyrelated to achievement through the SRL strategies, 2(27) 140.13; RMSEA .09; CFI .93; SRMR .07. Results showedthat PCT was negatively related to meta-cognitive strategy use, .38, p .001, but unrelated to deep-level cognitive strat-egies, .03, p .05. Further, whereas meta-cognition waspositively related to achievement, .31, p .001, deep-levelcognitive strategies was unrelated to achievement, .05, p .05. To examine whether the relation between PCT and achieve-ment is fully mediated by meta-cognitive strategy use, we added adirect path from PCT to achievement, thus testing a partial medi-ation model, 2(26) 139.54; RMSEA .09; CFI .93;SRMR .07. The fit of this model was not significantly betterthan the fit of the full mediation model, 2(1) 0.59, p .05,and the direct path from PCT to achievement was no longersignificant, .02, p .05, suggesting that the initial relationbetween PCT and achievement was fully mediated by students

Students Relative Autonomy

R = .11

Grades

Deep-Level Cognitive Strategies

R = .07Parcel 1 Parcel 2 Parcel 3

.84 .62 .79

Perceived Psychologically

Controlling Teaching

Parcel 2 Parcel 3 Parcel 4Parcel 1

Academic Performance

R = .11

Meta-Cognitive Strategies R = .34

.68 .93 .90 .67

Organi-zation

Critical Thinking

Self-Regulation

Elabora-tion

Time & Study

Self-Regulation

.99 .52 .54 .35

.47 .81

1.00

-.28***

.25***

.45***

.36***

-.07

.36***

-.26***

Figure 1. Final model of outcomes of psychologically controlling teaching. For reasons of clarity, the effectsof gender are not shown. p .001.

113PSYCHOLOGICALLY CONTROLLING TEACHING

use of meta-cognitive strategies. The Sobel test (Sobel, 1982)indicated that the indirect effect of PCT to achievement (throughmeta-cognitive strategy use) was significant, z 4.01, p .001.

Next, we examined whether students relative autonomy forstudying would mediate the negative path from PCT to SRLstrategies and to meta-cognitive strategy use in particular. We firsttested a model in which PCT was only indirectly related to SRLthrough relative autonomy for studying (i.e., a full mediationmodel). Next the fit of the full mediation model, 2(67) 309.61;RMSEA .08; CFI .93; SRMR .07, was compared with thefit of a partial mediation model, in which a direct path was addedfrom PCT to meta-cognitive strategy use. There was no need toexamine whether there was a remaining direct pathway to deep-level cognitive strategy use because PCT was unrelated to deep-level cognitive strategy use in the direct effects model. The fit ofthe partial model, 2(66) 293.20; RMSEA .08; CFI .94;SRMR .06, was significantly better than the fit of the fullmediation model, 2(1) 16.41, p .001, and the path fromPCT to meta-cognitive strategy use remained significantly nega-tive, .26, p .001. Accordingly, the association betweenPCT and meta-cognitive strategy use was partially mediated bystudents relative autonomy for studying, and the partial mediationmodel was retained as the final best fitting model. This model isdisplayed graphically in Figure 1. For the sake of clarity, pathsfrom gender to each of the latent constructs were not included.Gender (dummy coded with 0 male and 1 female) wassignificantly related to PCT, .36, p .001, and academicachievement, .14, p .01, with female participants scoringlower on PCT and higher on achievement.

As shown in Figure 1, perceived PCT was related negatively tostudents relative autonomy for studying, which in turn was relatedpositively to deep-level cognitive strategy use and meta-cognitivestrategy use. Only meta-cognitive strategy use was, in turn, relatedpositively to achievement. In addition to its indirect associationthrough relative autonomy for studying, PCT also had a direct andnegative association with meta-cognitive strategy use. The Sobeltesting indicated that the indirect effects of PCT over relativeautonomy for studying to deep-level cognitive and meta-cognitivestrategy use were both significant, z 3.47, p .001, and z 4.28, p .001, respectively.

Summary of Study 1

Study 1 yielded three important findings. First, the PCT scaleproved to be a reliable and valid instrument. Principal componentanalysis on the PCT items yielded one clear factor, Cronbachsalpha of the PCT scale was high, and the scale was significantlynegatively related to dimensions of need supportive to teaching,and in particular to autonomy support. Second, as expected, higherPCT was associated with lower SRL and achievement outcomes.Third, students relative autonomy was found to mediate some ofthese associations. PCT was negatively related to relative auton-omy for studying, which in turn was positively related to studentsuse of deep-level cognitive and meta-cognitive strategies. Meta-cognitive strategy use, in turn, was significantly positively relatedto academic performance. In addition, PCT continued to be sig-nificantly and negatively related to meta-cognitive strategy useabove and beyond its indirect relation through students relativeautonomy for studying.

Study 2: Sources of Psychologically ControllingTeaching

Because of the maladaptive pattern of correlates of PCT, itbecomes imperative to explore its antecedents. Therefore, in Study2, we aimed to examine the role of environmental and dispositionalsources of pressure in the prediction of PCT and the potentialexplanatory role of teacher burnout in these associations. Specif-ically, we examined the role of three sources of pressure, that is,perceived pressure from people in the school environment (such asparents, colleagues, and principals), teachers own quality of mo-tivation for teaching, and teachers perceptions of students qualityof motivation. It was expected that each of these pressuringsources would relate to PCT and that these associations would bemediated by teacher burn-out.

Burnout is defined as a state characterized by frustration as aresult of a perceived discrepancy between what teachers are doingand their expected effects on their students (Friedman, 1995). Thekey aspects of burnout represent feelings of emotional exhaustion,which refers to being overextended, and depersonalization, whichrefers to the adoption of a cynical and deindividuating attitudetoward others (Maslach, Schaufeli, & Leiter, 2001). Herein, wehypothesized that the encounter of pressuring sources would drainteachers energy and, as a result, would be associated with emo-tional exhaustion. Additionally, teachers who are exposed to pres-suring sources might adopt a depersonalizing attitude toward theircolleagues, students, and students parents to cope with the pres-suring atmosphere of their working situation. These predictions areconsistent with previous research showing that organizationalpressure (pressure from above; Levesque, Zuehlke, Stanek, &Ryan, 2004); teacher characteristics, such as maladaptive perfec-tionism (pressure from within; Stoeber & Rennert, 2008) andteachers relative autonomy for teaching (Roth et al., 2007); andnegative student behavior, such as disrespect (Friedman, 1995), areassociated with features of burnout.

In sum, when teachers feel emotionally drained, they wouldhave less energy available to stay attuned to their students wishes.Moreover, the adoption of a depersonalizing attitude would lowerteachers threshold to act in a controlling way vis-a`-vis theirstudents. Therefore, we hypothesized that greater emotional ex-haustion and depersonalization would be related to higher levels ofPCT and would mediate the relationships between perceived pres-sures and PCT.

MethodParticipants and procedure. Teachers voluntarily partici-

pated in this study during a faculty meeting or in the staff roomsof six secondary schools offering the academic track. They wereinformed that participation involved filling out a survey on theirjob conditions and interactions with students. The teachers wereasked to send this questionnaire to the principal researcher of thisstudy by regular mail. Three weeks after the initial invitation toparticipation, a reminder together with the survey were mailed tothe nonrespondents. Out of the sample consisting of 488 Dutch-speaking Belgian teachers, 317 (65%) completed the question-naire. Their age ranged from 21 to 61 years, with a mean of 40years (SD 10.4 years); 62.8% were women. The mean numberof years of teaching experience was 16.8 (SD 10.6 years); 40.4%

114 SOENENS ET AL.

of the teachers had obtained a college degree (not at a university),whereas 59.6% had a university degree.

Measures. Self-report measures were used to assess psycho-logically controlling teaching, its presumed sources, and burnout.

Psychologically controlling teaching. Teachers assessedtheir own perceptions of their PCT with the PCT ScaleTeacherSelf-Report. The seven items used in Study 1 were slightly re-worded to assess PCT from teachers own perspective (e.g., Iavoid looking at my students if they have disappointed me).Cronbachs alpha was .74. A principal component analysis indi-cated that all items loaded on a single factor with a minimumloading of .46, explaining 40.1% of the variance.

Pressure from above. We used an abbreviated version of theConstraints at Work Scale (Pelletier et al., 2002) to measurepressure exerted by colleagues, parents, and school principals.Specifically, we selected the items that most directly tapped intopressure versus sense of choice. The resulting subscale consistedof seven items. Cronbachs alpha was .62. A sample item reads,In this school, I have to conform to my colleagues teachingmethods. Ratings were made on a 7-point Likert scale, rangingfrom 1 (not at all true) to 7 (completely true).

Pressure from withinRelative autonomy for teaching.The same items that were used in Study 1 to assess studentsmotivation were used to assess teachers motivation for teaching.However, the stem of this scale was reworded to: I am motivatedto teach well because . . . The reliabilities of the different moti-vational subtypes (i.e., external, introjected, identified, and intrin-sic motivation) were satisfactory, ranging between .70 and .90. Anindex of relative autonomy for teaching was constructed in thesame way as in Study 1. Cronbachs alpha was .83.

Pressure from belowRelative autonomy for studying.To tap into teachers perceived motivation for studying of theirstudents, we used the same items as in Study 1, although the stemof this scale was reworded to read, My students are motivated tolearn because . . . The reliabilities of the motivation subtypeswere satisfactory, ranging between .73 and .88. The teacher-perceived relative autonomy index for studying had an alpha of.77.

Burnout. Participants completed the subscales for EmotionalExhaustion and Depersonalization of the widely used and vali-dated Maslach Burnout InventoryEducators Survey (MBIES;Kokkinos, 2006; Maslach & Jackson, 1986). Emotional exhaustionassesses participants feelings of tiredness (nine items; e.g., I feel

emotionally drained from my work; .91), whereas Deper-sonalization assesses the development of an impersonal attitudetowards the teaching job and students (five items; e.g., I feel Itreat some students as if they were impersonal objects; .66).

Results

Preliminary analyses. Correlations among the study vari-ables can be found in Table 3. We included years of teachingexperience in the analyses because past research has shown dif-ferences related to teaching experience in our study variables, suchas relationships with students (Pianta, Hamre, & Stuhlman, 2003).Constraints at work and teachers relative autonomous motivationwere significantly negatively related. No other correlations amongthe hypothesized antecedent variables were significant. Constraintsat work and teachers (but not students) relative autonomousmotivation were significantly related to PCT. The three variablesreflecting hypothesized antecedents were significantly related toboth emotional exhaustion and depersonalization, while both ofthem were positively interrelated and positively associated withPCT. Years of teaching experience was negatively correlated withconstraints at work, but positively related to teachers perceivedrelative autonomous motivation in students.

Because past research has shown gender differences for severalof the study variables, such as burnout (Grayson & Alvarez, 2008),we examined mean differences. We inspected gender differencesusing a MANOVA analysis treating gender as between-subjectsvariable and the sources of pressure as outcome variables. Anoverall significant effect was found, Wilks .96; F(3, 305) 3.97; p .01, 2 .04. Follow-up univariate analyses revealedthat female teachers scored higher on perceived constraints at workthan male teachers. A MANOVA with the burnout components asdependent variables also yielded an overall significant effect,Wilks .97; F(2, 314) 3.97; p .01, 2 .04. Follow-upunivariate analyses revealed that female teachers scored lower ondepersonalization than male teachers. A univariate ANOVA onPCT indicated that female teachers scored lower than male teach-ers. The results are shown in Table 3. Because of the effects ofgender and teaching experience, we included them as controlvariables in the primary analyses.

Primary analyses. The actual analyses checked whether theconstructs in the model were measured adequately and whether theexpected associations held among these constructs.

Table 3Correlations, Means, Standard Deviations, and Gender Differences for Study Variables in Study 2 (N 317)

Variable 2 3 4 5 6Total

M SDMale

M SDFemale

M SD Fa

1. Constraints at work (pressure from above) .22 .10 .24 .24 .21 3.06 0.87 2.90 0.93 3.16 0.83 6.642. Relative autonomy teacher (low pressure from within) .08 .37 .33 .28 6.98 3.12 6.62 3.19 7.15 3.10 2.103. Relative autonomy students (low pressure from below) .14 .13 .09 1.76 2.76 1.57 2.67 1.91 2.80 1.144. Emotional exhaustion .38 .28 1.92 0.81 1.91 0.86 1.93 0.79 0.095. Depersonalization .39 1.41 0.50 1.51 0.57 1.34 0.45 8.846. Psychologically controlling teaching 1.97 0.58 2.06 0.59 1.92 0.57 4.21

a df (1, 308) for constraints at work, relative autonomy teacher, and relative autonomy students; df (1, 316) for emotional exhaustion anddepersonalization; df (1, 314) for psychologically controlling teaching. p .05. p .01.

115PSYCHOLOGICALLY CONTROLLING TEACHING

Measurement model. We evaluated a measurement modelincluding six latent constructs, that is, constraints at work, teach-ers relative autonomous motivation, teacher-perceived relativeautonomous study motivation in students, emotional exhaustion,depersonalization, and PCT. Three to four parcels were used asindicators of each of these constructs. Gender and years of teach-ing experience were each represented as a latent variable with asingle indicator. The measurement model provided an excellent fitto the data, 2(183) 274.05; RMSEA .04; CFI .98;SRMR .05. Examination of factor loadings, ranging from .39 to.92 (mean .82), indicated that they were all significant.

Structural model. To verify the relative contribution of eachhypothesized antecedent of PCT, we tested a direct effects model,2(91) 152.81; RMSEA .05; CFI .98; SRMR .05.Whereas constraints at work was related positively to PCT, .27, p .001, teachers relative autonomous motivation wasrelated negatively to PCT, .28, p .001. In line with thecorrelations, the path from teacher-perceived relative autonomousstudy motivation in students to PCT was not significant, .10,p .05.

Next, the burnout components were included in the model asintervening variables in the relations between perceived pressuresand PCT. We first tested a full mediation model, in which theantecedents were only indirectly related to PCT through the burn-out components, 2(186) 282.82; RMSEA .04; CFI .98;SRMR .05. Next, we compared this model with a partialmediation model, in which we allowed direct paths from bothconstraints at work and teachers relative autonomous motivationto PCT. Note that in this model, we did not allow a direct pathfrom teacher-perceived relative autonomous study motivation instudents to PCT because this path was not significant in the directeffects model. Although the fit of the partial mediation model,2(184) 275.98; RMSEA .04; CFI .98; SRMR .05, wasslightly better than the fit of the full mediation model, 2(2) 6.84, p .05, the direct associations of constraints at work and

teachers relative autonomous motivation with PCT were reducedto nonsignificance, s .17 and .11, respectively, both ps .05. This indicates that associations between the antecedents andPCT, if any, were full mediated by the dimensions of burnout.Further testifying to the intervening role of the burnout dimen-sions, Sobel tests indicated that the indirect effects of constraints atwork and teachers relative autonomous motivation over deper-sonalization to PCT were significant, z 3.14, p .01 and z 4.39, p .001, respectively.

Figure 2 graphically displays all structural paths in the fullmediation model. For the sake of clarity, paths from gender andyears of teaching experience to each of the latent constructs werenot included. Gender (dummy coded with 0 male and 1 female) was significantly related to constraints at work, .18,p .05, and depersonalization, .28, p .01, with womenreporting more constraints at work and less depersonalization.Years of teaching experience was related negatively to constraintsat work, .20, p .01. As shown in Figure 2, both con-straints at work and low teacher relative autonomy were related todepersonalization, which in turn was related positively to PCT.Low teacher relative autonomy was also related to emotionalexhaustion but emotional exhaustion was not related to PCT.

Summary of Study 2

The relations between perceived pressure from above (as in-dexed by constraints at work) and pressure from within (as indexedby low relative autonomy for teaching) and PCT were mediated bydepersonalization. Pressure from below, as indexed by studentslow relative autonomy for learning, was not related to PCT.Further, pressure from within was the only source of pressure thatwas significantly related to emotional exhaustion. Emotional ex-haustion was not related to PCT, indicating that it did not functionas an intervening variable in the relationships between differentsources of perceived pressures and PCT.

Psychologically Controlling Teaching R = .35

Parcel 1

Parcel 2

Parcel 3

Constraints at Work

.84

.57

.40

Relative Autonomy for

Teaching

Parcel 1

Parcel 2

Parcel 3

Parcel 4

.74

.92

.88

.81

Students Relative Autonomy for

Studying

Parcel 1

Parcel 2

Parcel 3

Parcel 4

.69

.89

.89

.81

Emotional Exhaustion

R = .19

Depersonalization R = .33

Parcel 1 Parcel 2 Parcel 3

.93 .82 .91

Parcel 1 Parcel 2 Parcel 3

.57 .73 .65

Parcel 1

Parcel 2

Parcel 3

.69

.71

.70

-.22***

.06

.14

-.36***

-.36***

-.11

.30***

-.08

.10

.51***

.49***

Figure 2. Final model of antecedents of psychologically controlling teaching. For reasons of clarity, effects ofgender and years of teaching experience are not shown. p .05. p .01. p .001.

116 SOENENS ET AL.

General Discussion

The present studies applied the construct of psychological con-trol, which involves the communication of a conditionally approv-ing attitude through intrusive and manipulative practices (includ-ing shaming and expression of disappointment) to the teachingcontext. Specifically, we examined associations between PCT anda number of hypothesized antecedents, mediators, and outcomes.Validity and reliability analyses show that PCT can be assessed asa reliable and unidimensional factor that correlates negatively withadaptive teaching dimensions, that is, autonomy support, structure,and involvement. The remaining part of the discussion is organizedaround two important sets of findings, that is, (a) the hypothesizedoutcomes and (b) the hypothesized antecedents of PCT. Across thetwo studies, we aimed to test the SDT-based notion that controllingteaching, and PCT in particular, relates to a controlled orientationin the functioning of both teachers and students.

Outcomes of Psychologically Controlling Teaching

Through structural equation modeling, our first study demon-strated negative associations between PCT and (a) important com-ponents of SRL, that is, deep-level cognitive and meta-cognitivestrategy use and (b) academic achievement. Our findings furtherindicated that the association between PCT and meta-cognitivestrategy use was partially mediated by students low relativeautonomy for studying. Low use of meta-cognitive strategies was,in turn, related to lower academic performance. Together, thesefindings are consistent with the idea that PCT undermines stu-dents autonomous motivation to learn and instead fosters a con-trolled regulation of study activities (Vansteenkiste, Simons, et al.,2005). This low-quality type of motivation would in turn depletestudents energy and vitality that are needed to adequately self-regulate their learning process (De Bilde, Vansteenkiste, & Lens,2011). The direct relationship between PCT and meta-cognitivestrategy use possibly indicates that PCT is a strong and directsource of stress, reducing the energy needed for meta-cognitivestrategy use.

With regard to the self-regulatory learning strategies, it is inter-esting to note that meta-cognitive strategy use was significantlyrelated to academic achievement, whereas deep-level cognitivestrategy use was not. These findings are in line with the results ofprevious research showing stronger relations between meta-cognitive strategy use and performance than between cognitivestrategy use and performance (Wolters & Pintrich, 1998). Further-more, the positive association found between meta-cognitive strat-egy use and academic performance is consistent with investiga-tions showing that the promotion of self-regulated learningincreases academic performance (Lane et al., 2008).

Antecedents of Psychologically Controlling Teaching

Another aim was to identify factors associated with teachersPCT. In doing so, we examined the role of three sources ofpressure in relation to PCT. Results indicated that perceived pres-sure from above, as indexed by teacher-perceived constraints atwork, and perceived pressure from within, as measured by teach-ers controlled relative to their autonomous motivation for teach-ing, but not perceived pressure from below, as measured by

students controlled relative to their autonomous motivation forstudying, were associated with PCT. These findings are consistentwith research reporting relations between perceived pressure fromauthorities and teachers own use of pressure (Deci, Spiegel, Ryan,Koestner, & Kauffman, 1982) and between personal characteris-tics and motivation in teachers and their behavior toward students(Boggiano, Barrett, Weiher, McClelland, & Lusk, 1987; Roth etal., 2007).

Furthermore, we found that depersonalization played a mediat-ing role in the relations between pressures from above and withinand PCT. So, it seems that pressure from above and pressure fromwithin distract teachers from students personal wants and desires,presumably because they foster an objectifying stance towardsstudents (Moller & Deci, 2010; Vansteenkiste, Mouratidis, &Lens, 2010). This attitude of depersonalization might lower teach-ers threshold for adopting a controlling attitude toward students asthey are reduced to objects that can be pushed around. Althoughemotional exhaustion was positively correlated with PCT, it didnot play a unique mediating role in the relation between thesources of pressures and PCT. Possibly, emotional exhaustion isrelatively more strongly related to intrapsychic problems such asinternal distress and somatic complaints (Jackson, Schwab, &Schuler, 1986; Leung & Lee, 2006). In contrast, depersonalizationmay yield relatively more relational costs, as expressed, for in-stance, in a controlling interpersonal style (Moller & Deci, 2010).

Some results were not in line with our predictions. First, pres-sure from below, defined as teachers beliefs about studentsquality of motivation, was not uniquely related to the use of PCT,nor was it related to burnout. These findings were not in line withsome previous studies on pressure from below (Taylor & Ntou-manis, 2007). A possible explanation for this lack of associationmay be that teachers do not really suffer from a controlled moti-vation in their students. Students with a controlled orientation maystill be compliant and obedient, such that the teachers authority isnot immediately threatened. We speculate that students who dis-play rebellious reactions, show apathy, or achieve poorly do in-duce PCT because these behaviors involve a stronger and moreimmediate threat to teachers capacities. Second, there was nosignificant path from pressure from above to emotional exhaustion,whereas there was a significantly positive path from pressure fromwithin to emotional exhaustion. It is possible that pressure fromabove, which refers to an interpersonal experience, is especiallyimportant in developing interpersonal attitudes, like depersonal-ization, whereas experiences of pressure in ones own functioningis detrimental to both intrapersonal and interpersonal functioning.

Limitations and Suggestions for Future Research

The current research has some limitations. First, our focus islimited by primarily relying on self-report data. To avoid possibleresponse bias and thereby to validate our results, multiple infor-mants, such as external observers, and behavioral measures areneeded. In developing an observational coding scheme, futureresearch could rely on Barber (1996) who assessed parental psy-chological control in videotaped family problem-solving tasks. Inour view, the examination of associations between objectiveteacher use of psychological control and subjective perceptions ofpsychological control is not only important to demonstrate con-struct validity. Such research also could contribute to a number of

117PSYCHOLOGICALLY CONTROLLING TEACHING

important and substantive research questions. For instance, dostudents experience objective expressions of psychological controlsimilarly? Or are some students more sensitive to expressions ofpsychological control than others? It could be hypothesized, forinstance, that students motivational orientation might moderatethese associations such that students with a controlled orientationare more likely to subjectively experience even mild expressionsof psychological control as strongly pressuring and intrusive. Assuch, an important direction for future research is to examinepersonal (and perhaps also sociocultural variables) that may mod-erate associations between objective teacher behavior and subjec-tive perceptions of psychological control.

Second, the cross-sectional nature of our studies does not allowus to examine reciprocal relations between the concepts we stud-ied. For that purpose, longitudinal research is recommended. Fur-ther, for causality statements, experimental research is recom-mended (see Vansteenkiste, Simons, et al., 2005, for an example).Third, we would like to raise three issues related to the general-izability of our findings. First, students frame of reference inStudy 1 for responding to the measures was devoid of context.Therefore, to internally validate our findings, investigators con-ducting future research should ask students to respond to theinstruments with reference to a particular course. Consideringdifferent courses (e.g., math, chemistry, and language learning),one could examine whether the structural relationships among thestudy variables are similar across courses or rather course specific.Second, although the return rate of 65% in Study 2 was rather high,it leaves open the question as to how representative of secondaryschool teachers the sample is. Third, the involvement of onlysecondary school students and teachers from the academic tracklimits the generalizability of our findings to the educational systemas a whole. Given the obvious work and organizational differencesbetween educational levels, it may be important to replicate thepresent findings with kindergarten, elementary school, or collegestudents and teachers. Moreover, we encourage future researchersto replicate the findings in countries outside of Europe, such as theUnited States. Such external validation seems necessary due to thecontextual differences between countries, such as the system ofhigh-stakes testing in the United States, which clearly representspressure from above and is absent in the European context.

Finally, we propose two important additional avenues of study.First, due to the partial mediating role of students relative auton-omy for studying (Study 1), future researchers could focus on otherpossible mediators, such as need satisfaction in students. Second,our study is a first step toward elucidating the dynamics involvedin PCT. It would be instructive for future research to more thor-oughly explore the relation between PCT and other teaching di-mensions, such as autonomy support, structure, and external con-trol, and to look for unique effects of these teaching dimensions onstudents learning, well-being, and performance. Of particularinterest is an examination of associations between PCT and SDT-based measures of internally controlling teaching. Roth, Assor,Niemiec, Ryan, and Deci (2009) have recently argued that inter-nally controlling socialization may involve both the use of nega-tive conditional regard (i.e., withdrawing love when children donot meet standards) and positive conditional regard (i.e., showingmore love than usual when children do meet standards). Althoughthe concept of PCT entails features of negative conditional regard,it does not capture the use of positive conditional regard. As such,

future researchers would do well to compare the educationaloutcomes and antecedents of both psychological control and pos-itive conditional regard.

Practical ImplicationsGiven the harmful correlates of PCT, from an applied perspec-

tive, it is important for teachers to refrain from PCT. To modifypsychologically controlling teacher behavior, teachers can be pro-vided with information about what behaviors constitute PCT andtheir effects on adolescent learning and achievement. To the extentthat teachers wish to positively influence their students learning,they can be advised to teach in an autonomy-supportive fashion,for instance, by explaining the relevance of learning strategies(Reeve, 2009). It is equally important that pressure on and controlof teachers is reduced, as indicated by the results of Study 2. Toavoid the development of an objectifying attitude toward students,which seems to catalyze the use of PCT, it is desirable that theentire educational community and the general public recognizethe complexity, responsibilities, and stresses that are inherent inthe teaching profession so that the pressure from above on teachersis reduced. To achieve this aim, principals could create anautonomy-supportive climate in which teachers have a say invarious decisions, a rationale is provided when staff involvementin decisions is limited, and teachers are offered opportunities forself-direction and self-initiative. At the intraindividual level, in-creasing teachers pleasure and importance of their teaching mightbe important because teachers motivation relates significantly tothe use of a psychologically controlling stance. Perhaps, it is betterfor teachers who primarily teach for controlled reasons to reorientthemselves in their professional career and to choose a job that liesmore in line with their values and interests.

ConclusionBy introducing psychological control as a dimension of teaching

style, our aim was to contribute to both research on parentalpsychological control and SDT-based research in educational psy-chology. Developmental research on parental psychological con-trol has convincingly shown its negative ramifications for chil-drens and adolescents general well-being and adjustment. Fewstudies, however, have examined the implications of parentalpsychological control for educational and learning outcomes (seeAunola & Nurmi, 2004; Vansteenkiste, Simons, et al., 2005;Vansteenkiste, Zhou, et al., 2005 for exceptions). This researchcontributes to the literature by showing that perceived teacherpsychological control is related to a nomological network of mal-adaptive motivational and learning outcomes in adolescents. At thesame time, this research adds to SDT-based research on teachingstyle by highlighting one important feature of need-frustratingteaching. There is increasing recognition that experiences of needfrustration do not simply reflect a lack of need satisfaction and thatit is important to identify the specific dynamics involved in need-frustrating interpersonal styles, including psychological control(Bartholomew, Ntoumanis, Ryan, Bosch, & Thogerson-Ntoumani,in press).

ReferencesAssor, A., & Kaplan, H. (2001). Mapping the domain of autonomy support:

Five important ways to undermine students experience of autonomy in

118 SOENENS ET AL.

learning. In A. Efklides, J. Kuhl, & R. M. Sorrentino (Eds.), Trends andprospects in motivation research (pp. 101120). Dordrecht, the Nether-lands: Kluwer.

Assor, A., Roth, G., & Deci, E. L. (2004). The emotional costs of parentsconditional regard: A self-determination theory analysis. Journal ofPersonality, 72, 4788. doi:10.1111/j.0022-3506.2004.00256.x

Aunola, K., & Nurmi, J.-E. (2004). Maternal affection moderates theimpact of psychological control on a childs mathematical performance.Developmental Psychology, 40, 965978. doi:10.1037/0012-1649.40.6.965

Barber, B. K. (1996). Parental psychological control: Revisiting a ne-glected construct. Child Development, 67, 32963319. doi:10.2307/1131780

Barber, B. K., & Harmon, E. L. (2002). Violating the self: Parentalpsychological control of children and adolescents. In B. K. Barber (Ed.),Intrusive parenting: How psychological control affects children andadolescents (pp. 1552). Washington, DC: American PsychologicalAssociation. doi:10.1037/10422-002

Barber, B. K., & Olsen, J. A. (1997). Socialization in context: Connection,regulation, and autonomy in the family, school, and neighborhood, andwith peers. Journal of Adolescent Research, 12, 287315.

Barber, B. K., & Xia, M. (in press). The centrality of control to parentingand its effects. In A. S. Morris, R. E. Larzelere, & A. W. Harrist (Eds.),New directions for authoritative parenting. Washington, DC: AmericanPsychological Association.

Bartholomew, K., Ntoumanis, N., Ryan, R. M., Bosch, J., & Thogersen-Ntoumani, C. (in press). Self-determination theory and diminished func-tioning: The role of interpersonal control and psychological need thwart-ing. Personality and Social Psychology Bulletin.

Belmont, M., Skinner, E., Wellborn, J., & Connell, J. (1988). Teacher associal context: A measure of student perceptions of teacher provision ofinvolvement, structure, and autonomy support (Tech. Rep. No. 102).Rochester, NY: University of Rochester.

Black, A. E., & Deci, E. L. (2000). The effects of instructors autonomysupport and students autonomous motivation on learning organic chem-istry: A self-determination theory perspective. Science Education, 84,740 756. doi:10.1002/1098-237X(200011)84:6740::AID-SCE43.0.CO;2-3

Boggiano, A. K., Barrett, M., Weiher, A. W., McClelland, G. H., & Lusk,C. M. (1987). Use of maximal-operant principle to motivate childrensintrinsic interest. Journal of Personality and Social Psychology, 53,866879. doi:10.1037/0022-3514.53.5.866

Byrne, B. M. (1994). Burnout: Testing for the validity, replication, andinvariance of causal structure across elementary, intermediate, and sec-ondary teachers. American Educational Research Journal, 31, 645673.

Case, L. P., Harris, K. R., & Graham, S. (1992). Improving the mathemat-ical problem-solving skills of students with learning disabilities: Self-regulated strategy development. The Journal of Special Education, 26,119. doi:10.1177/002246699202600101

De Bilde, J., Vansteenkiste, M., & Lens, W. (2011). Understanding theassociation between future time perspective and self-regulated learningthrough the lens of self-determination theory. Learning and Instruction,21, 332344. doi:10.1016/j.learninstruc.2010.03.002

Deci, E. L., Eghrari, H., Patrick, B. C., & Leone, D. (1994). Facilitatinginternalization: The self-determination theory perspective. Journal ofPersonality, 62, 119142. doi:10.1111/j.1467-6494.1994.tb00797.x

Deci, E. L., & Ryan, R. M. (2000). The what and why of goal pursuits:Human needs and the self-determination of behavior. PsychologicalInquiry, 11, 227268. doi:10.1207/S15327965PLI1104_01

Deci, E. L., Schwartz, A. J., Sheinman, L., & Ryan, R. M. (1981). Aninstrument to assess adults orientations toward control versus autonomywith children: Reflections on intrinsic motivation and perceived com-petence. Journal of Educational Psychology, 73, 642650. doi:10.1037/0022-0663.73.5.642

Deci, E. L., Spiegel, N. H., Ryan, R. M., Koestner, R., & Kauffman, M.(1982). Effects of performance standards on teaching styles: Behavior ofcontrolling teachers. Journal of Educational Psychology, 74, 852859.doi:10.1037/0022-0663.74.6.852

Friedman, I. A. (1995). Student behavior patterns contributing to teacherburnout. Journal of Educational Research, 88, 281289. doi:10.1080/00220671.1995.9941312

Grayson, J. L., & Alvarez, H. K. (2008). School climate factors relating toteacher burnout: A mediator model. Teaching and Teacher Education,24, 13491363. doi:10.1016/j.tate.2007.06.005

Grolnick, W. S. (2003). The psychology of parental control: How well-meant parenting backfires. Mahwah, NJ: Erlbaum.

Grolnick, W. S., Deci, E. L., & Ryan, R. M. (1997). Internalization withinthe family: The self-determination theory perspective. In J. E. Grusec &L. Kuczynski (Eds.), Parenting and childrens internalization of values:A handbook of contemporary theory (pp. 7899). London, England:Wiley.

Hambleton, R. K. (1994). Guidelines for adapting educational and psycho-logical tests: A progress report. European Journal of PsychologicalAssessment, 10, 229244.

Holmbeck, G. N. (1997). Toward terminological, conceptual, and statisti-cal clarity in the study of mediators and moderators: Examples from thechild-clinical and pediatric psychology literatures. Journal of Consultingand Clinical Psychology, 65, 599 610. doi:10.1037/0022-006X.65.4.599

Hu, L., & Bentler, P. M. (1999). Cutoff criteria for fit indexes in covariancestructure analysis: Conventional criteria versus new alternatives. Struc-tural Equation Modeling, 6, 155. doi:10.1080/10705519909540118

Jackson, S. E., Schwab, R. L., & Schuler, R. S. (1986). Toward anunderstanding of the burnout phenomenon. Journal of Applied Psychol-ogy, 71, 630640. doi:10.1037/0021-9010.71.4.630

Jreskog, K., & Srbom, D. (1996). LISREL VIII: Users reference guide.Chicago, IL: Scientific Software International.

Kokkinos, C. M. (2006). Factor structure and psychometric properties ofthe Maslach Burnout InventoryEducators survey among elementaryand secondary school teachers in Cyprus. Stress and Health, 22, 2533.doi:10.1002/smi.1079

Lane, K. L., Harris, K. R., Graham, S., Weisenbach, J. L., Brindle, M., &Morphy, P. (2008). The effects of self-regulated strategy developmenton the writing performance of second-grade students with behavioraland writing difficulties. The Journal of Special Education, 41, 234253.doi:10.1177/0022466907310370

Leung, D. Y. P., & Lee, W. W. S. (2006). Predicting intention to quitamong Chinese teachers: Differential predictability of the components ofburnout. Anxiety, Stress, and Coping, 19, 129 141. doi:10.1080/10615800600565476

Levesque, C., Zuehlke, A. N., Stanek, L. R., & Ryan, R. M. (2004).Autonomy and competence in German and American students: A com-parative study based on self-determination theory. Journal of Educa-tional Psychology, 96, 6884. doi:10.1037/0022-0663.96.1.68

Little, T. D., Cunningham, W. A., Shahar, G., & Widaman, K. F.(2002). To parcel or not to parcel: Exploring the question, weighingthe merits. Structural Equation Modeling, 9, 151173. doi:10.1207/S15328007SEM0902_1

Maslach, C., & Jackson, S. E. (1986). Maslach Burnout Inventory manual(2nd ed.). Palo Alto, CA: Consulting Psychologists Press.

Maslach, C., Schaufeli, W., & Leiter, M. (2001). Job burnout. AnnualReview of Psychology, 52, 397 422. doi:10.1146/annurev.psych.52.1.397

Moller, A. C., & Deci, E. L. (2010). Interpersonal control, dehumanization,and violence: A self-determination theory perspective. Group Processes& Intergroup Relations, 13, 4153. doi:10.1177/1368430209350318

Niemiec, C. P., Lynch, M. F., Vansteenkiste, M., Bernstein, J., Deci, E. L.,& Ryan, R. M. (2006). The antecedents and consequences of autono-

119PSYCHOLOGICALLY CONTROLLING TEACHING

mous self-regulation for college: A self-determination theory perspec-tive on socialization. Journal of Adolescence, 29, 761775. doi:10.1016/j.adolescence.2005.11.009

Niemiec, C. P., & Ryan, R. M. (2009). Autonomy, competence, andrelatedness in the classroom: Applying self-determination theory toeducational practice. Theory and Research in Education, 7, 133144.doi:10.1177/1477878509104318

Pelletier, L. G., Seguin-Levesque, C., & Legault, L. (2002). Pressure from aboveand pressure from below as determinants of teachers motivation and teachingbehavior. Journal of Educational Psychology, 94, 186196.

Pianta, R. C., Hamre, B., & Stuhlman, M. (2003). Relationships betweenteachers and children. In W. M. Reynolds, G. E. Miller, & I. B. Weiner(Eds.), Handbook of psychology: Vol. 7. Educational psychology (pp.199234). Hoboken, NJ: Wiley.

Pintrich, P., Smith, D., Garcia, T., & McKeachie, W. (1991). A manual forthe use of the Motivated Strategies for Learning Questionnaire (Tech-nical Report 91-B-004). Ann Arbor: University of Michigan.

Pintrich, P. R., & De Groot, E. V. (1990). Motivational and self-regulatedlearning components of classroom academic performance. Journal ofEducational Psychology, 82, 3340.

Quintana, S. M., & Maxwell, S. E. (1999). Implications of recent devel-opments in structural equation modeling for counseling psychology.Counseling Psychologist, 27, 485527.

Reeve, J. (2009). Why teachers adopt a controlling motivating style towardstudents and how they can become more autonomy supportive. Educa-tional Psychologist, 44, 159175.

Roth, G., Assor, A., Kanat-Maymon, Y., & Kaplan, H. (2007). Autonomousmotivation for teaching: How self-determined teaching may lead to self-determined learning. Journal of Educational Psychology, 99, 761774.

Roth, G., Assor, A., Niemiec, C. P., Ryan, R. M., & Deci, E. L. (2009). Theemotional and academic consequences of parental conditional regard:Comparing conditional positive regard, conditional negative regard, andautonomy support as parenting practices. Developmental Psychology,45, 11191142.

Rozendaal, J. S., Minnaert, A., & Boekaerts, M. (2005). The influence of teacher-perceived administration of self-regulated learning on students motivation andinformation processing. Learning and Instruction, 15, 141160.

Ryan, R. M. (1982). Control and information in the intrapersonal sphere:An extension of cognitive evaluation theory. Journal of Personality andSocial Psychology, 43, 450461.

Ryan, R. M., & Connell, J. P. (1989). Perceived locus of causality andinternalization: Examining reasons for acting in two domains. Journal ofPersonality and Social Psychology, 57, 749761.

Ryan, R. M., Connell, J. P., & Grolnick, W. S. (1992). When achievementis not intrinsically motivated: A theory of internalization and self-regulation in school. In A. K. Boggiano & T. S. Pittman (Eds.), Achieve-ment and motivation (pp. 167188). New York, NY: Cambridge Uni-versity Press.

Schaefer, E. S. (1965). Childrens reports of parental behavior: An inven-tory. Child Development, 36, 413424.

Skinner, E. A., & Belmont, M. J. (1993). Motivation in the classroom:Reciprocal effects of teacher behavior and student engagement acrossthe school year. Journal of Educational Psychology, 85, 571581.

Sobel, M. E. (1982). Asymptotic intervals for indirect effects in structuralequations models. In S. Leinhart (Ed.), Sociological methodology (pp.290312). San Francisco, CA: Jossey-Bass.

Soenens, B., Luyckx, K., Vansteenkiste, M., Duriez, B., & Goossens, L. (2008).Clarifying the link between parental psychological control and adolescentsdepressive symptoms. Merrill-Palmer Quarterly, 54, 411444.

Soenens, B., & Vansteenkiste, M. (2005). Antecedents and outcomes of

self-determination in three life-domains: The role of parents and teach-ers autonomy support. Journal of Youth and Adolescence, 34, 589604.

Soenens, B., & Vansteenkiste, M. (2010). A theoretical upgrade of theconcept of parental psychological control: Proposing new insights on thebasis of self-determination theory. Developmental Review, 30, 7499.

Stoeber, J., & Rennert, D. (2008). Perfectionism in school teachers: Rela-tions with stress appraisals, coping styles, and burnout. Anxiety, Stress,and Coping, 21, 3753.

Taylor, I. M., & Ntoumanis, N. (2007). Teacher motivational strategies andstudent self-determination in physical education. Journal of EducationalPsychology, 99, 747760.

Taylor, I. M., Ntoumanis, N., & Standage, M. (2008). A self-determinationtheory approach to understanding the antecedents of teachers motiva-tional strategies in physical education. Journal of Sport and ExercisePsychology, 30, 7594.

Vallerand, R. J., & Bissonnette, R. (1992). Intrinsic, extrinsic, and amoti-vational styles as predictors of behavior: A prospective study. Journal ofPersonality, 60, 599620.

Vansteenkiste, M., & Lens, W. (2006). Intrinsic versus extrinsic goalcontents in self-determination theory: Another look at the quality ofacademic motivation. Educational Psychologist, 41, 1931.

Vansteenkiste, M., Lens, W., & Deci, E. L. (2006). Intrinsic versus extrin-sic goal contents in self-determination theory: Another look at thequality of academic motivation. Educational Psychologist, 41, 1931.

Vansteenkiste, M., Mouratidis, A., & Lens, W. (2010). Detaching reasonsfrom aims: Fair play and well-being in soccer as a function of pursuingperformance-approach goals for autonomous or controlling reasons.Journal of Sport and Exercise Psychology, 32, 217232.

Vansteenkiste, M., Niemiec, C. P., & Soenens, B. (2010). The five mini-theories of self-determination theory: An historical overview, emergingtrends, and future directions. In T. C. Urdan & S. A. Karabenick (Eds.),Advances in motivation and achievement: Vol. 16A. The decade ahead:Theoretical perspectives on motivation and achievement (1st ed., pp.105166). Bingley, England: Emerald.

Vansteenkiste, M., Sierens, E., Soenens, B., Luyckx, K., & Lens, W.(2009). Motivational profiles from a self-determination perspective: Thequality of motivation matters. Journal of Educational Psychology, 101,671688.

Vansteenkiste, M., Simons, J., Lens, W., Soenens, B., & Matos, L. (2005).Examining the motivational impact of intrinsic versus extrinsic goalframing and autonomy-supportive versus internally controlling commu-nication style on early adolescents academic achievement. Child De-velopment, 76, 483501.

Vansteenkiste, M., Zhou, M. M., Lens, W., & Soenens, B. (2005). Expe-riences of autonomy and control among Chinese learners: Vitalizing orimmobilizing? Journal of Educational Psychology, 97, 468483.

Vrugt, A., & Oort, F. J. (2008). Metacognition, achievement goals, studystrategies and academic achievement: Pathways to achievement. Meta-cognition and Learning, 30, 123146.

Winne, P. H. (1995). Inherent details in self-regulated learning. Educa-tional Psychologist, 30, 173187.

Wolters, C. A., & Pintrich, P. R. (1998). Contextual differences in studentmotivation and self-regulated learning in mathematics, English, andsocial studies classrooms. Instructional Science, 26, 2747.

Zimmerman, B. J. (2008). Investigating self-regulation and motivation:Historical background, methodological developments, and future pros-pects. American Educational Research Journal, 45, 166183.

Received August 18, 2009Revision received July 7, 2011

Accepted July 7, 2011

120 SOENENS ET AL.