here today, gone tomorrow? revisiting the stability of teachers' achievement goals

Here today gone tomorrow Revisiting the stability of teachersrsquoachievement goalsAnna-Katharina Praetorius a Sebastian Nitsche b Stefan Janke b Oliver Dickhaumluser bKatharina Drexler a Michaela Fasching a Markus Dresel a

a Department of Psychology University of Augsburg Augsburg Germanyb Department of Educational Psychology University of Mannheim Mannheim Germany

A R T I C L E I N F O

Article historyAvailable online 16 October 2014

KeywordsAchievement goalsTeacherStabilityGeneralizability theory

A B S T R A C T

Teachersrsquo achievement goals are typically considered to be stable characteristics although there are ar-guments for both stability and instability Empirical investigations regarding the stability of teachersrsquoachievement goals are rare In this study we investigated the stability of teachersrsquo achievement goals(ie learning performance approach performance avoidance and work avoidance goals) usinggeneralizability theory The sample comprised 166 German mathematics teachers in academic-track sec-ondary schools who completed self-report questionnaires three times over the course of one school yearThe ratio of stable to unstable aspects of teachersrsquo achievement goals varied between 21 and 41 Thenumber of measurement points needed for a reliable measure of the trait aspects of achievement goalsvaried between one and three The results underline the importance of advancing research on teachersrsquoachievement goals both theoretically and methodologically

copy 2014 Elsevier Inc All rights reserved

1 Introduction

Motivational characteristics are a powerful explanation for howand why people think and act as they do Many researchers con-ceptualize motivational characteristics as dispositional characteristicsof persons One famous example is research on motive disposi-tions ie habitual preferences for dealing with certain kinds ofincentives (Heckhausen amp Heckhausen 2010) Several theories andmodels however also point to the influence of occasion-specificcharacteristics on actual motivation (eg model of adaptable learn-ing Boekaerts amp Niemimirta 2000 cognitive-motivational processmodel Vollmeyer amp Rheinberg 1998 for a general overview of therelationship between traits and states see latent-state-trait theorySteyer Schmitt amp Eid 1999)

In research on teacher motivation achievement goals are seenas an important explanation for teachersrsquo perceptions of the envi-ronment and for their actions (Butler 2012 Nitsche DickhaumluserFasching amp Dresel 2011 Retelsdorf amp Guumlnther 2011) It is usuallyassumed (eg Butler 2007 Dresel Fasching Steuer Nitsche ampDickhaumluser 2013 Malmberg 2008 Retelsdorf Butler Streblow ampSchiefele 2010) that these achievement goals can be seen as traits(as ldquogoal orientationsrdquo) and therefore are only influenced byoccasion-specific characteristics to a limited degree Teachers teach-

ing the same class in the same school should thus set themselvessimilar goals across several occasions largely independent of situ-ational circumstances However few investigations have tested thisassumption empirically Taking a contrary position regarding thestability of achievement goals Elliot (2005) stated that the main dif-ference between the achievement goal approach and the classicalachievement motive is that the former has a more specific and con-textual focus Increasing our knowledge about the actual stabilityof teachersrsquo achievement goals will facilitate the development of bothan appropriate theoretical understanding of teachersrsquo achieve-ment goals and subsequently an adequate model of the constructMore concretely knowledge about the stability of teachersrsquo achieve-ment goals is important for the following reasons (a) It helps insurethat investigations will capture the characteristics of interest in anappropriate way (eg choosing a cross-sectional versus a longitu-dinal design) (b) It points out how to construct adequatemeasurement instruments (eg general versus situation-based mea-surements) (c) It helps in the selection of appropriate researchquestions regarding the level of operationalization of the indepen-dent and dependent variables (see Ajzen amp Fishbein 1977) Ifachievement goals are only stable to a small degree effects on ratherstable characteristics (eg teachersrsquo content knowledge) are un-likely However if teachersrsquo achievement goals are stable to a largedegree investigating the effect on variable characteristics (eg in-structional behavior in specific situations) does not seem to bestraightforward (d) Additionally knowledge about the stability ofteachersrsquo achievement goals is useful in deriving appropriate im-plications based on the results of investigations (see also Murphy

Corresponding author Department of Psychology University of AugsburgUniversitaetsstraszlige 10 86159 Augsburg Germany Fax +49 821 598 5289

E-mail address annapraetoriusphiluni-augsburgde (A-K Praetorius)

httpdxdoiorg101016jcedpsych2014100020361-476Xcopy 2014 Elsevier Inc All rights reserved

Contemporary Educational Psychology 39 (2014) 379ndash387

Contents lists available at ScienceDirect

Contemporary Educational Psychology

journal homepage wwwelseviercom locate cedpsych

amp Alexander 2000 Pintrich 2000) The purpose of the study at handis to shed light on this topic by investigating the stability ofteachersrsquo achievement goals as well as the number of measure-ment points necessary to reliably measure these goals acrossoccasions

11 Teachersrsquo achievement goals definition and relevance

Achievement goals explain how and why people behave the waythey do in achievement settings (Dweck amp Leggett 1988 MurayamaElliot amp Friedman 2012) Achievement goal theory differentiatesbetween various goals The achievement goals that are commonlydistinguished when describing and explaining characteristics ofteacher motivation are (a) learning goals (the teacher aims to in-crease his or her own competencies) (b) performance approach goals(the teacher aims to demonstrate high competencies) (c) perfor-mance avoidance goals (the teacher aims to avoid the impressionof low competencies) and (d) work avoidance goals (the teacheraims to reduce his or her workload) Several studies have pointedout the relevance of teachersrsquo achievement goals for the teachingprofession as they have revealed relationships between teachersrsquoachievement goals and various teacher and teaching characteris-tics Associations have been found for example betweenachievement goals and occupational burden or burn-out (NitscheDickhaumluser Fasching amp Dresel 2013 Retelsdorf et al 2010 Toumlnjesamp Dickhaumluser 2009) the perception of help-seeking as beneficialor threatening (Butler 2007 Nitsche et al 2011) participation invocational training programs (Nitsche et al 2013) aspects of in-structional quality (Butler 2012 Butler amp Shibaz 2008 Retelsdorfamp Guumlnther 2011) and the goal structures teachers realize in theirclassrooms ie the extent to which pursuing learning vs perfor-mance goals for students is reinforced by the classroom environment(Butler 2012 Dresel et al 2013 Retelsdorf et al 2010) In all ofthese studies learning goals were positively correlated with vari-ables that are considered to be beneficial (eg attending vocationaltraining programs) and negatively correlated with variables that areregarded as adverse (eg burn-out) For performance avoidance goalsas well as work avoidance goals the relationship pattern was in most

cases reversed The results regarding the effects of performanceapproach goals were mixed (eg positive effects on teacher self-efficacy in a study by Nitsche et al 2011 positive effects on thesocial reference norm in a study by Retelsdorf amp Guumlnther 2011)Most of the relationships identified between teachersrsquo achieve-ment goals and other variables were small some were moderate

12 Teachersrsquo achievement goals stable characteristics

Regarding the conceptualization of achievement goals large differ-ences can be found (for an overview see Pintrich 2000 and Maehr ampZusho 2009 see also Button Mathieu amp Zajac 1996 DeShon amp Gillespie2005) According to some conceptions achievement goals are assumedto be rather stable (eg Silva amp Nicholls 1993) whereas for other con-ceptions they are assumed to be rather unstable (eg Elliott amp Dweck1988) The differences regarding the assumed stability of achievementgoals are important The theoretical conception of achievement goalsinfluences (a) how investigations concerning these goals are con-ducted (eg how many measurement points are used) (b) how theyare measured (eg whether achievement goals are assessed with respectto specific situations) (c) what research questions are appropriate (egwhether it makes sense to investigate effects of achievement goals onstable outcomes) and (d) what implications are derived from the resultsof the investigations (Murphy amp Alexander 2000 Pintrich 2000 for asimilar argumentation regarding intrinsic and extrinsic motivation seeHarter amp Jackson 1992)

Based on the considerations of Fryer and Elliot (2007) and Pintrich(2000) we developed a conceptual framework to explain why achieve-ment goals on the one hand can be assumed to be stable but on theother hand are also assumed to be unstable (see Fig 1) Fryer and Elliot(2007) and Pintrich (2000) do not differentiate between differentachievement goals in their argumentation This implies that differ-ences in the stability of the goals are not expected Additionally noinformation is given regarding the expected magnitude of the stableand the unstable components of achievement goals

Empirical investigations regarding the stability of achievementgoals exist first and foremost for studentsrsquo achievement goals (foran overview see Fryer amp Elliot 2007 Pintrich 2000) According to

Fig 1 Conceptual framework for the stability of achievement goals

380 A-K PraetoriusContemporary Educational Psychology 39 (2014) 379ndash387

Senko Hulleman and Harackiewicz (2011) retest correlations rangebetween r = 40 and r = 70 for both learning and performance goalsamong students

Regarding the stability of teachersrsquo achievement goals very fewconsiderations and empirical investigations exist The assump-tions and empirical results regarding studentsrsquo achievement goalscannot however simply be transferred to teachers as the achieve-ment settings of students and teachers in the school context arecompletely different

Teachersrsquo achievement goals have been conceptualized to dateas rather stable tendencies to adopt certain goals This is especial-ly obvious in publications which use the term ldquogoal orientationrdquo (egButler 2007 Dresel et al 2013 Fasching Dresel Dickhaumluser ampNitsche 2010 Malmberg 2008 Nitsche et al 2011 Retelsdorf et al2010) meaning a general orientation towards goals that includesbeliefs about purposes competence success ability effort errorsand standards (Pintrich 2000) As a consequence teachersrsquo achieve-ment goals are usually assessed using one measurement point withself-report measures that survey general achievement goals withoutconsidering the specific situations teachers are in when reportingthese goals Interpretations of the results of the investigations asa rule deem that teachers have a certain stable and dominantachievement goal

Empirical investigations regarding the degree to which occa-sions influence measures of teachersrsquo achievement goals are rareMost of them focus on student teachers (ie students carrying outtheir university studies) or teacher trainees (ie students in apractice-orientated training phase following university gradua-tion) rather than on teachers The rank-order stability findings ofthese studies are summarized in Table 1 The large variations re-ported for retest correlations (26 le r le 71) indicate that the stabilityof student teachersrsquo teacher traineesrsquo and teachersrsquo achievementgoals is not clear Furthermore the varying correlations point to thefact that different kinds of achievement goals might be differentlystable for teachers One observed tendency is that learning goalsare less stable than performance goals Possible reasons for thesedifferences are not discussed in the studies considered

In the study conducted by Fasching et al (2010) mean-level andintra-individual stabilities of teacher trainees were investigated inaddition to rank-order stabilities Variance analyses revealed thatthe mean levels of learning goals performance approach and per-formance avoidance goals decreased over the course of the twoinvestigated years whereas there was no significant variability withregard to work avoidance goals ICCs were calculated as an esti-mate of intra-individual variability Learning goals showed aconsiderably higher intra-individual variation (1 minus ICC = 72) com-pared to performance approach goals (1 minus ICC = 53) performanceavoidance (1 minus ICC = 58) and work avoidance (1 minus ICC = 52) As thesample consisted solely of teacher trainees it remains unclear as towhat extent these results can be expected with regard to teachersThe contexts of the two groups differ considerably Teacher train-ees are for example assessed several times within the training period

whereas teachers are assessed only very rarely in their daily schoolexperiences

13 Determining the stability of achievement goals usinggeneralizability theory

In most investigations the stability of achievement goals has beencalculated using retest correlations This method has however beencriticized because rank-order stabilities cannot be interpreted un-ambiguously (see Fryer amp Elliot 2007) One disadvantage of analyzingretest correlations is that they confound true state variance (ie truescore variance that is specific for a certain measurement point) andmeasurement error Retest correlations are therefore of restrictedutility when investigating the stability of measures as it is notpossible to obtain information on the magnitude of occasion-specific goals

A solution for this disadvantage is generalizability theory (Gtheory) a statistical framework in which different aspects of sta-bility (eg mean level changes variation in the values of achievementgoals across occasions) can be estimated simultaneously With Gtheory a comparison between the magnitudes of stable and un-stable components of the measurement is also possible This in turngives an impression of how significant the instability of teachersrsquoachievement goals is for the measurement Another advantage ofG theory is that one can estimate the number of measurement pointsnecessary to assess teachersrsquo achievement goals reliably across oc-casions (see Shavelson amp Webb 1991) To date teachersrsquo achievementgoals are assessed using one measurement point without knowingwhether this is sufficient to capture them reliably

Beyond G theory there are several other alternatives one can useto investigate the occasion specificity of measures for example latentstate-trait theory (see Steyer et al 1999) We chose G theory forthis investigation as it not only allows for the separation of trait andstate variance but also helps determine the number of measure-ment points necessary for a reliable measure

14 Research questions and hypotheses

Up to now we have been able to reveal very little concerningthe variability of teachersrsquo achievement goals The few existingstudies have mainly investigated student teachers and teachertrainees

As the existing studies on the stability of student teachersrsquo teachertraineesrsquo and teachersrsquo achievement goals have shown a large vari-ability in the stability of achievement goals we did not derive ahypothesis but pursued the following research question instead Towhat degree do teachersrsquo achievement goals vary across measure-ment points over the course of one school year (research question1) To answer this research question we used G theory to take intoconsideration different aspects of stability

Based on the assumption that teachersrsquo achievement goals arestable characteristics these goals as a rule were collected at one

Table 1Overview of studies analyzing the stability of teachersrsquo achievement goals using retest-correlations

Literature Sample Number oftime points

Retestinterval

Achievement goal

Learning Performanceapproach

Performanceavoidance

Workavoidance

Fasching et al (2010) German teacher trainees 5 6 months 26ndash42 41ndash66 46ndash51 48ndash57Toumlnjes and Dickhaumluser (2009) German teachers 2 3 months 55 61 63 ndashToumlnjes and Dickhaumluser (2009) German teacher trainees 2 12 months 58 71 63 ndashMalmberg (2008) Finnish teacher trainees 5 12 months 37a 41a 35a ndash

Note ldquondashrdquo = This goal was not investigated in the studya These correlations are the mean correlations across the four measurement points None of the reported studies investigated the stability of relational goals

381A-K PraetoriusContemporary Educational Psychology 39 (2014) 379ndash387

measurement point in prior research and were then related to othervariables (eg instructional variables) However no previous em-pirical studies have tested whether one measurement point issufficient to reliably measure teachersrsquo achievement goals We there-fore added the research question How many measurement pointsare needed for a reliable measure of teachersrsquo achievement goals(research question 2)

2 Method

21 Sample and procedure

We invited 288 academic-track secondary schools (ldquoGymnasienrdquo)in the German federal state of Baden-Wuumlrttemberg to participatein the study in the school year 201112 via postal letters 57 of theseschools (46 public schools 11 private schools) decided to take partin the study Of the participating schools 13 were located in urbanand 44 in rural areas The participation rate per school variedbetween one and five teachers as only mathematics teachers teach-ing in 5th grade classrooms were to participate We restricted thestudy to 5th grade classrooms to ensure comparability For the anal-yses we used data from all teachers participating in the study intotal 166 German mathematics teachers (55 female) were as-sessed The mean age of the teachers was 41 years (SD = 13) at thefirst measurement point The teaching experience of the teachersranged from 0 to 40 years (M = 13 SD = 12)

The teachers completed a questionnaire at three points over thecourse of one school year (directly after the summer break in Sep-tember 2011 December 2011 and March 2012)

22 Instruments

In order to measure the achievement goals of teachers a self-report questionnaire developed by Nitsche et al (2011) was usedThis measure was developed based on existing measures (eg theAchievement Goal Questionnaire by Elliot amp Murayama 2008 GoalOrientations for Teaching by Butler 2007) The instrument formu-lated by Nitsche et al (2011) is specifically tailored to the populationof teachers and explicitly considers the different aspects of teach-ersrsquo professional knowledge (eg pedagogic content knowledge) towhich teachersrsquo learning goals can be directed and the different rel-evant others (eg school principal) to which teachers can addresstheir performance goals The instrument is reliable and has beenproven to possess factorial convergentdivergent and predictive va-lidity (Nitsche et al 2011) It has since been used in a number ofstudies (eg Dresel et al 2013 Fasching et al 2010 Nitsche et al2013)

The instrument consists of four achievement goal scales (a) onescale assessing learning goals with three subscales focusing on dif-ferent domains of teacher knowledge (pedagogic knowledge contentknowledge pedagogic content knowledge) each comprising threeitems (eg ldquoIn my vocation I aspire to improve my pedagogic knowl-edge and competencerdquo) (b) a scale for performance approach and(c) a scale for performance avoidance goals each with four subscalesfocusing on different significant addressees (colleagues principalstudents self) and each with three items (eg ldquoIn my vocation Iaspire for my students to realize that I teach better than other teach-ersrdquo [approach] ldquoIn my vocation I aspire to not show my studentswhen I have more trouble meeting the job demands than otherteachersrdquo [avoidance]) and (d) a scale for work avoidance goals withsix items (eg ldquoIn my vocation I aspire to get through the day withlittle effortrdquo) For the work avoidance scale all items conceived byNitsche et al (2011) in their first version of the instrument wereused All items were presented alongside 5-point Likert-type scalesranging from 1 (strongly disagree) to 5 (strongly agree)

23 Analyses

To answer research question 1 regarding the degree of instabil-ity in teachersrsquo achievement goal measures G theory (Shavelson ampWebb 1991) was used This method allows for a separation betweentrait and state true score variance via variance component analy-sis (Brennan 2001 Shavelson amp Webb 1991) The objects ofmeasurement in the G analyses were the persons and their achieve-ment goals Occasions and item parcels were added as facets (iesources of error) As each person was assessed using all item parcelson all occasions the design was fully crossed Persons item parcelsand occasions were treated as random in all of the analyses Thusa two-facet fully crossed random effects design (p times i times o design) wasapplied persons (p) crossed with item parcels (i) crossed withoccasions (o)

We used item parcels for the following reason It cannot beassumed that the subscales of the instruments (eg knowledge facetsaddressed with learning goals) are interchangeable as they focuson different aspects of achievement goals To be able to use randomeffects we used systematically combined item parcels instead ofsubscales (Kishton amp Widaman 1994 Little Cunningham Shaharamp Widaman 2002) One item of each subscale of the respectiveachievement goal was integrated into a parcel Therefore the numberof items per parcel was two for the work avoidance goals three forthe performance approach and performance avoidance goals andfour for the learning goals

G theory designs can be illustrated using Venn diagrams As canbe seen in Fig 2 the design allowed us to separate seven sources ofvariance in the data variance resulting from (a) the persons (σ2

p)(b) the item parcels (σ2

i) (c) the occasions (σ2o) (d) the interaction

between persons and item parcels (σ2pi) (e) the interaction between

persons and occasions (σ2po) (f) the interaction between item parcels

and occasions (σ2io) and (g) the interaction between persons item

parcels and occasions (σ2pioe) The interaction mentioned last is con-

founded with an unspecific error component as it is the highest orderinteraction The variance components can be grouped into (a) stableinter-individual differences regarding achievement goals (= trait vari-ance σ2

p σ2pi) (b) intra-individual differences regarding achievement

goals across measurement points due to specific occasions (= statevariance σ2

o σ2po σ2

io) and (c) residual variance (σ2pioe) as well as

further variance components which are not relevant for the topicin question (σ2

i) In estimating the dependability (ie reliability) ofthe measurement in question two G coefficients are available in Gtheory Both coefficients can be interpreted as being analogous toclassic reliability coefficients Following the rule of thumb in classictest theory 70 is set as a minimum value for a reliable measurein the present investigation The absolute G coefficient (φ) is

po

pi oi

poie

Fig 2 Venn diagram for the persons times item parcels times occasions design


adequate if one is interested in absolute outcomes (eg compe-tencies in large-scale assessment studies) The relative G coefficient(ρ2) is adequate if one focuses on the relative position of persons orvariables and not on the absolute values As this is the case in thepresent investigation the relative G coefficient is reported in thefollowing

The G analyses were conducted with the urGENOVA program(Brennan 2001) version 21 In urGENOVA the implemented esti-mator is the analogous ANOVA procedure A large advantage of thisestimator is that normality assumptions are not required (Brennan2001 see also the simulation study of Shumate Surles Johnsonamp Penny 2007) Missing data are handled in urGENOVA by addingan additional facet to the data (see Brennan 2001) and thus are ex-plicitly considered

To determine the number of measurement points necessary fora reliable measure of teachersrsquo achievement goals (research ques-tion 2) D analyses were conducted These analyses enable researchersto estimate reliability under multiple measurement conditions (egdiffering numbers of measurement points) and thus provide evi-dence regarding how many observations of a certain variable arenecessary to obtain a sufficient reliability (Brennan 2001) Theseanalyses are based on the estimated variance components of theG analyses thus the information from the G analyses is used toestimate the number of necessary measurement points The esti-mation provided by the D analyses works analogously to the

SpearmanndashBrown formula in classical test theory (Webb Shavelsonamp Haertel 2006)

The D analyses were conducted using the GENOVA software (Crickamp Brennan 1983) The implemented estimator was the ANOVAprocedure

3 Results

31 Descriptive analyses

Table 2 separately presents the mean scores standard devia-tions and internal consistencies for the three investigatedmeasurement points The internal consistencies were admissibleThe retest correlations for the five achievement goals can also befound in Table 2 The correlations were on a descriptive level lowestfor learning goals (55 le r le 67) and highest for performance avoid-ance goals (71 le r le 81)

32 Stability of teachersrsquo achievement goals over time

The estimated variance components (see Table 3) can be groupedinto components measuring achievement goals that are occasionunspecific and thus stable and into components that are occa-sion specific and thus unstable The stable components (ie stabledifferences between persons (σ2

p) as well as interactions betweenpersons and item parcels (σ2

pi)) captured between 47 and 64 ofthe total variance The unstable components (ie differences betweenthe occasions (σ2

o) interactions between persons and occasions (σ2po)

and interactions between item parcels and occasions (σ2io)) cap-

tured between 17 and 27 of the variance The proportion of stableand unstable components was then compared for every achieve-ment goal (see the ratios in Table 3) The stable component of themeasure was for all achievement goals larger than the unstable com-ponent however the amount of unstable variance varied to a largedegree for the achievement goals investigated

To understand the reasons for the rather high proportion ofoccasion-specific variance of learning goals it is useful to take a closerlook at the relative variance components in Table 3 This reveals thatthe occasion specificity of this goal type was neither due to changesin the whole group of teachers across occasions (σ2

o) nor due to dif-ferences in the item difficulties across occasions (σ2

oi) In fact nearlyall occasion-specific variance was due to an interaction betweenpersons and occasions (σ2

po) This interaction means that personshad different values for their achievement goals on different occa-sions leading to a changing sequence of the persons between

Table 2Descriptive statistics of teachersrsquo achievement goals

Achievement goals M SD α r12 r13 r23

Learning 55 67 67Measurement point 1 414 044 082Measurement point 2 416 051 088Measurement point 3 415 048 087

Performance approach 74 70 77Measurement point 1 196 074 093Measurement point 2 209 080 095Measurement point 3 218 080 095

Performance avoidance 75 71 81Measurement point 1 249 081 091Measurement point 2 247 084 093Measurement point 3 258 085 094

Work avoidance 61 64 78Measurement point 1 209 071 082Measurement point 2 189 073 087Measurement point 3 182 070 088

Note Retest intervals were 3 months respectively N = 150ndash163 teachers

Table 3G analyses

Learning Performance approach Performance avoidance Work avoidance

Variance component (σ2) VC VC VC VC VC VC VC VCStable components 016 47 051 64 060 64 043 54

Person p 014 41 050 63 058 62 037 47Person times item parcel pi 002 6 001 1 002 2 006 8

Occasion-specific components 009 27 017 21 016 17 016 19Occasion o 0a 0 001 1 0 0 003 3Person times occasion po 009 27 016 20 016 17 013 16Occasion times item parcel oi 0 0 0 0 0a 0 0 0

Item parcel-specific componentItem parcel i 0 0 0 0 004 4 0 0

Error componentPerson times occasion times item parcel poie 008 26 012 16 014 15 021 26

Total variance 033 079 093 080G coefficient Eρ2 075 088 088 081Ratio of stable to occasion-specific components 21 31 41 31

Note VC = absolute variance component VC = variance component relative to the total variance N = 166 teachersa A small negative variance component was estimated due to sampling errors Following the suggestion of Brennan (2001) this negative variance was used for the cal-

culation of the remaining variance components and set to zero afterwards


occasions andor to a changing relation of the personsrsquo valuesbetween occasions

The analysis of the variance components also showed strikingdifferences regarding the residual variances between differentachievement goals (see σ2

poie in Table 3) The amount of residualvariance (ie the amount of variance that could not be explainedwith the variables included) ranged from 15 to 26 indicating thatdifferent teacher goals could be assessed with different degrees ofreliability

33 Number of necessary measurement points

In order to determine how many measurement points are re-quired to assess teachersrsquo achievement goals with sufficient reliability(research question 2) we conducted D analyses with one to ten mea-surement points We fixed the number of teachers and item parcelsto the actual number in the study at hand Figure 3 illustrates theresults for the D analyses for all achievement goals The figure showshow reliably the different achievement goals could be measured witha given number of measurement points For example in order toobtain a G coefficient (ie reliability) larger than 70 one measure-ment point was sufficient for performance approach and avoidancegoals Two measurement points were needed to assess teachersrsquo workavoidance goals with sufficient precision Finally three measure-ment points were required for learning goals to exceed a reliabilityof ρ2 = 70

4 Discussion

According to Pintrich (2000) ldquogoals are not traits in the classicpersonality sense They are cognitive representations and may showboth intraindividual stability as well as contextual sensitivityrdquo (p103) In this study this assumption was tested with regard to teach-ersrsquo achievement goals

41 Stability of teachersrsquo achievement goals and its explanation

The G analyses in the present investigation showed that the ratioof teachersrsquo stable to unstable achievement goals ranged between21 and 51 A closer look at the results revealed that the unstableproportion of variance could be traced back to the main effect ofthe occasions as well as the interaction between persons andoccasions

The existing main effect of occasion may be due to different jobrequirements during the school year At the beginning of the schoolyear (measurement point 1) teachers have to work through a largenumber of different tasks (eg developing an instruction plan at-tending school-year beginning conferences getting to know newstudents) and thus aim to avoid additional work Over the schoolyear the work situation eases enabling teachers to focus on othergoals However if we take a look at the size of the main effects ofoccasions in the G analyses these main effects are very small in com-parison to other existing situation-specific variance components inthe data Mean level differences thus do not seem to be a very im-portant source of variability in teachersrsquo achievement goals

The interaction between persons and occasions was far more im-portant accounting for 16 to 27 of the entire variance in the dataTeachers thus had different values for their achievement goals acrossoccasions This result indicates that measures of teachersrsquo achieve-ment goals contain a large amount of occasion-specific variance ndasheven when these goals are assessed in an occasion-unspecific wayQualitative investigations would be useful to understand what exactlyteachers have in mind when answering achievement goal ques-tionnaires at different measurement points

How can we know based on the results of the present studywhether the degree of occasion specificity regarding teachersrsquoachievement goals is large or small To facilitate the interpreta-tion one can take a look at the number of measurement pointsneeded to measure teachersrsquo achievement goals We argue that mea-surements that require only one measurement point contain a stableproportion that is sufficient enough to characterize the measure-ment as stable If however more than one measurement point isnecessary to capture the characteristic of interest the characteris-tic should not be seen as stable The results of the present studyindicate that different kinds of achievement goals differ in thenumber of measurement points required Based on our results onewould characterize performance approach and performance avoid-ance goals as truly stable goals For work avoidance and learninggoals the trait portions in teachersrsquo achievement goals are not largeenough to permit one measurement point to capture them reli-ably Although operationalized as dispositions measures of someachievement goals of teachers thus also measure ndash at least in theiractual operationalization ndash a considerable part of the specific oc-casions teachers are in when completing surveys this is especiallytrue for learning goals (for a similar result for undergraduate stu-dents see Muis amp Edwards 2009) The term ldquoorientationrdquo thus seemsonly to be appropriate for performance approach and perfor-mance avoidance goals but not for learning and work avoidancegoals

The rather low stability of learning goals is particularly strik-ing These results are in line with study results found for studentsrsquoachievement goals (eg Schoumlne 2008) Several authors have sug-gested that goals change when the environment changesconsiderably (eg Fryer amp Elliot 2007 Nicholls 1984) In the presentstudy teachers were investigated over the course of one school yearteaching the same classes in the same school Considerable changesin the environment thus cannot be assumed So why are teachersrsquolearning goals influenced by situational characteristics over shortperiods of time One possible explanation is based on methodol-ogy As the learning goal scale showed a high mean at the firstmeasurement point the instability could be due to central tenden-cies (Lord 1963) However as there were rarely any mean differencesbetween the measurement points (see σ2

o in the results section)this explanation can be ruled out

Another possible explanation for the instability of learning goalsis related to their definition Learning goals focus on increasingonersquos own competencies The concept of competencies howevercomprises many different aspects It thus can be assumed thatlearning goals vary as a function of the specific aspects under

05

06

07

08

09

1

1 2 3 4 5 6 7 8 9 10

Rel

ativ

e G

co

effi

cien

t

Number of measuring points

Learning goals

Performance approach goals

Performance avoidance goals

Work avoidance goals

01

Fig 3 Decision studies for teachersrsquo achievement goals with 1 to 10 measure-ment points


consideration To operationalize learning goals as validly as possi-ble (see also Groves et al 2009) it seems useful to distinguish thespecific competencies one aims to increase (see Nitsche et al 2011)However the more concretely learning goals are assessed the higherthe probability that their stability will recede as teachers may focustheir professional learning on different things at different times(see the aspects ldquochanging characteristics of the individualrdquo andldquochanging situationalcontextual featuresrdquo in the conceptual frame-work) For example a teacher would probably answer the itemson the learning goal subscale used in the present study focusingon pedagogic content knowledge differently directly after partici-pating in a vocational training program on didactic aspects thanhe or she would a few months later when the training content isno longer fresh in his or her mind Teachers may also adapt theirspecific learning goals over the course of a school year to the classesthey teach After realizing the specific issues present in the classeshe or she is teaching during a specific school year a teacher mayfocus on developing certain aspects of his or her competence (eghisher pedagogic knowledge if heshe has several classes with dis-cipline problems)

Referring to the conceptual framework proposed in the presentstudy there are many more characteristics that could potentiallylead to instability in learning goals Whether the instability is mainlydue to changing personal characteristics to self-regulatory activi-ties or rather to changing situationalcontextual factors is an openquestion which could be addressed using experimental andor in-terview studies in the future

However not all achievement goals showed as much variabili-ty as learning goals According to Fryer and Elliot (2007) differencesin the stability between different achievement goals are not likelyWhy were performance goals so highly stable in the present inves-tigation One possible explanation is that learning goals refer tospecific competencies that vary between occasions (ie an inter-action between changing characteristics of the individual andchanging situationalcontextual features in the conceptual frame-work) Performance goals in contrast refer primarily to theindividuals or groups the demonstration or avoidance is directedtowards (eg students or colleagues see Ziegler Dresel amp Stoeger2008) These significant others rarely vary over the course ofone school year for a teacher teaching the same class(es) in thesame school (ie stable situationalcontextual features in theconceptual framework) If we compare the definition andthe operationalization of performance goals there might be a secondexplanation Performance approach and avoidance goals focus ondemonstrating high competencies or avoiding the impression oflow competencies Thus there are potentially two issues to be so-lidified (a) the significant others to which the demonstration oravoidance is directed (a stable situationalcontextual feature) and(b) the specific competencies (analogous to learning goals a stableor changing characteristic of the person) In the instrument devel-oped by Nitsche et al (2011) which was used in the presentinvestigation the significant others are systematically tapped inthe item formulations whereas the competencies are not (eg ldquoInmy vocation I aspire to demonstrate to my colleagues that I knowmore than other teachersrdquo) The reason for not addressing the spe-cific competencies is that investigations with students have shownthat learning goals vary between different domains whereas per-formance goals rarely vary between domains (Nitsche 2013)However as this reason is related to the achievement goals of stu-dents it seems to be useful to investigate empirically whether theachievement goals of teachers vary with regard to the specific com-petencies under consideration As far as we know no instrumentintending to measure teachersrsquo achievement goals systematicallyvaries significant others and competencies when measuring per-formance goals (see eg Butler 2007 Toumlnjes amp Dickhaumluser 2009Nitsche et al 2011)

42 Implications of the (in-)stability of teachersrsquo achievement goals

The G analyses indicate that some achievement goals of teach-ers seem to be influenced by occasional characteristics to a largerdegree than commonly assumed Therefore researchers should con-sider the theoretical methodological empirical and practicalimplications of this instability

One theoretical implication of the results at hand is the need fora deepened understanding of teachersrsquo achievement goals how theyare generated and what sorts of occasional characteristics are ableto influence them (see also Maehr amp Zusho 2009) ldquoA motiva-tional theory such as goal orientation theory should be able toexplicate the core processes that result in the situational construc-tion of a goal orientation and the role of dispositions in thisconstructionrdquo (Kaplan amp Maehr 2007 p 174) A starting point forinvestigating these aspects could be the conceptual framework pro-posed in the present study To acquire a deepened understandingof teachersrsquo achievement goals it is necessary to reconsider theirconceptualization as primarily dispositional as implied by the useof the term ldquogoal orientationrdquo (see also the discussion regarding stu-dentsrsquo achievement goals for an overview see Maehr amp Zusho 2009)This question has not yet been discussed with regard to teachersrsquoachievement goals The present article is a first step in pursuing thisquestion The results of our study indicate that it might not be ap-propriate to define and investigate teachersrsquo achievement goalsexclusively as traits Indeed our results have revealed that teach-ersrsquo achievement goals ndash even though operationalized as dispositionsndash are influenced considerably by characteristics of the occasions inwhich they are assessed According to Elliot (2005) the partiallyrather low stability of teachersrsquo achievement goals is neither sur-prising nor undesirable the original aim of the achievement goalapproach was to overcome the disadvantages of mere disposi-tional constructs such as the achievement motive by introducingmore contextual information Elliot claimed that the dispositionalfocus of researchers regarding achievement goals is thus rather sur-prising and should be reconsidered

One important methodological implication is that researchersshould be clear about the aspects they want to generalize with theirmeasures When measuring a construct we are often not inter-ested in single performances regarding this construct but in thegeneral construct-related values of persons (Cronbach Gleser Nandaamp Rajaratnam 1972 Lakes amp Hoyt 2008 Shavelson amp Webb 1991)If a measure is largely influenced by occasion characteristics onlystatements regarding this specific measurement point can be madeThis is unproblematic if researchers are solely interested in char-acteristics regarding the specific time points at which they areassessed However if researchers want to draw conclusions beyondthese time points generalizability across time points has to be takeninto account Therefore future studies should empirically survey thenumber of measurement points necessary for a reliable estimateof the construct in question

An empirical implication of the results is that the correlations iden-tified between teachersrsquo goals (measured at one point in time) andother stable variables are influenced by the instability of teach-ersrsquo achievement goals and are in all likelihood underestimationsof the true correlations This is especially true for learning goalsAdditionally teachersrsquo achievement goals are differently influ-enced by occasion characteristics These differences lead to variationregarding the generalizability of the goals Comparisons betweenachievement goals regarding their impact on other variables (eginstructional characteristics) are thus not admissible when achieve-ment goals are only measured once

A practical implication of the partially rather low stability ofachievement goals is that teachersrsquo achievement goals are in prin-ciple modifiable This is good news as it implies that trainingprograms focusing on achievement goals can be successful (Salas


amp Cannon-Bowers 2001) This is again especially true for learninggoals

43 Limitations and further directions

The generalizability of the results across the investigated sampleof teachers and the instruments used are crucial for the implica-tions of these results for research on teachersrsquo achievement goals(see also Kaplan amp Maehr 2007) As is the case with many studieson teachers the present sample was not representative Howeverthe descriptive results (internal consistencies means and stan-dard deviations) were quite similar to those reported by Nitsche et al(2011) It nevertheless remains unclear as to whether the result pat-terns of the study at hand apply for the entire teacher populationbeyond German teachers Regarding the generalizability across in-struments it has already been mentioned that instruments aimingto assess achievement goals differ considerably The (in-)stabilityfound in the study at hand therefore might to a certain degree beinstrument dependent However if we take a look at the retest cor-relations of previous studies learning goals were on a descriptivelevel less stable when compared to performance goals as well

Another critical point regarding the results is that it remainsunclear as to whether the results can be generalized beyond thethree-months retest interval chosen for the present investigationOne can hypothesize that longer retest intervals could lead to lowerstability It can thus be assumed that the time variability found inthe present study is a minimum estimation of change occurring overseveral school years andor more strongly differing contexts An in-dication that this may be true can be derived from the retestcorrelations of former studies investigating teachersrsquo and teachertraineesrsquo achievement goals In most of these studies the retest in-tervals were larger than the one in the present investigation Theretest correlations in these studies were on a descriptive levelsmaller than those found here (see Fasching et al 2010 Malmberg2008 Toumlnjes amp Dickhaumluser 2009) However further studies direct-ly investigating the effects of retest intervals in the context of teacherachievement goals are necessary All in all it seems important toconduct further studies on the stability of teachersrsquo achievementgoals to determine whether the results of the present study can bereplicated using another sample as well as other instruments

A third point that can be criticized is the method used for ana-lyzing the data G theory enables a separation between trait and statevariance as well as a determination of the number of measure-ment points necessary for a reliable assessment of teachersrsquoachievement goals Applying G theory is therefore very useful foranalyzing stability questions Nevertheless the variance compo-nents provided are subject to sampling variability (Brennan 2001)To verify the credibility of the results standard errors andor con-fidence intervals of the variance components would be a usefulmeasure (see Hoyt amp Melby 1999) Computing them is howevernot straightforward as this would require distributional assump-tions (eg normality assumptions) which cannot be assumed indesigns such as the one used A replication of the results at handis thus also important for this reason Here this is even more im-portant as teachers are nested within schools in the data Thisdependency could however not be accounted for in the G analy-ses as the design would have been too complex and the variancecomponent estimates less trustworthy Not taking into account thenested structure can have an additional impact on the standarderrors However as previous research has shown that school effectsare rarely relevant for teachersrsquo achievement goals (see egDickhaumluser Nitsche Fasching amp Dresel 2012) this impact can beassumed to be rather small With regard to the interpretation of theanalyses conducted two additional points can be criticized Firstwith the chosen method it is not possible to separate variability anddevelopmental change Second time stability and trans-situational

stability may be confounded as situations were not systematicallyvaried or held constant

In addition to questions regarding the generalizability of theresults and the method used for data analysis further questionsremain for future research One of them is related to factors un-derlying potential differences between teachers in their stability ofgoals One could for example hypothesize that more experiencedteachers have more stable goals as they may have identified long-term learning goals which do not change within short time periods

5 Conclusions

Teachersrsquo achievement goals are often assumed to be stable char-acteristics The study at hand provides an indication that this is notnecessarily the case Learning goals in particular seem to be con-siderably influenced by time-variable characteristics The results showthat stability questions are largely relevant for the interpretation ofteachersrsquo achievement goals and that theoretical assumptions re-garding stability do not suffice In the short run it is necessary toinclude a sufficient number of measurement points in a study inorder to capture the constructs in question In the long run the resultsunderline the importance of advancing research on teachersrsquo achieve-ment goals both theoretically and methodologically

Acknowledgments

This research was funded by a grant from the German FederalMinistry of Education and Research to Oliver Dickhaumluser (01 HJ 0901)and Markus Dresel (01 HJ 0902)

References

Ajzen I amp Fishbein M (1977) Attitude-behavior relations A theoretical analysisand review of empirical research Psychological Bulletin 84 888ndash918

Boekaerts M amp Niemimirta M (2000) Self-regulation in learning Finding a balancebetween learning- and ego-protective goals In M Boekaerts P R Pintrich amp MZeidner (Eds) Handbook of self-regulation (pp 417ndash450) San Diego CA AcademicPress

Brennan R L (2001) Manual for urGENOVA Iowa City IA Iowa Testing ProgramsUniversity of Iowa

Butler R (2007) Teachersrsquo achievement goals and associations with teachersrsquo helpseeking Examination of a novel approach to teacher motivation Journal ofEducational Psychology 99 241ndash252

Butler R (2012) Striving to connect Extending an achievement goal approach toteacher motivation to include relational goals for teaching Journal of EducationalPsychology 104 726ndash742 doi101037a0028613

Butler R amp Shibaz L (2008) Achievement goals for teaching as predictors of studentsrsquoperceptions of instructional practices and studentsrsquo help seeking and cheatingLearning and Instruction 18 453ndash467 doi101016jlearninstruc200806004

Button S Mathieu J amp Zajac D (1996) Goal orientation in organizational researchA conceptual and empirical foundation Organizational Behavior and HumanDecision Processes 67 26ndash48

Crick J E amp Brennan R L (1983) Manual for GENOVA A generalized analyses ofvariance system [Computer software and manual] University of Iowa

Cronbach L J Gleser G C Nanda H amp Rajaratnam N (1972) The dependabilityof behavioral measurements New York John Wiley

DeShon R P amp Gillespie J Z (2005) A motivated action theory account of goalorientation Journal of Applied Psychology 90 1096ndash1127 doi1010370021-90109061096

Dickhaumluser O Nitsche S Fasching M S amp Dresel M (2012) Kommt es auf dieSchule an Zu Bedeutung von wahrgenommenen Kontextmerkmalen fuumlr dieberuflichen Zielorientierungen von Lehrkraumlften [Does the school matter Therelevance of perceived context characteristics for teachersrsquo goal orientations]In A Philipp amp M Kunter (Chairs) Macht die Schule den Unterschied Zur Rolledes Schulkontexts fuumlr das Verhalten und Erleben von Lehrkraumlften Symposiumconducted at the meeting of the German Psychology Society Bielefeld Germany

Dresel M Fasching M S Steuer G Nitsche S amp Dickhaumluser O (2013) Relationsbetween teachersrsquo goal orientations their instructional practices and studentmotivation Psychology (Savannah Ga) 7 572ndash584 doi104236psych201347083

Dweck C S amp Leggett E L (1988) A social-cognitive approach to motivation andpersonality Psychological Review 95 256ndash273

Elliot A J (2005) A conceptual history of the achievement goal construct In A Elliotamp C Dweck (Eds) Handbook of competence and motivation (pp 52ndash72) New YorkGuilford Press


Elliott E S amp Dweck C S (1988) Goals An approach to motivation and achievementJournal of Personality and Social Psychology 54 5ndash12

Elliot A J amp Murayama K (2008) On the measurement of achievement Goalscritique Illustration and application Journal of Educational Psychology 100613ndash628

Fasching M S Dresel M Dickhaumluser O amp Nitsche S (2010) Achievement goalsof teacher trainees Longitudinal analysis of magnitude change and relevanceJournal of Educational Research Online 2 9ndash33

Fryer J W amp Elliot A J (2007) Stability and change in achievement goals Journalof Educational Psychology 99 700ndash714

Groves R M Fowler F J Jr Couper M P Lepkowski J M Singer E amp TourangeauR (2009) Survey methodoloy (2nd ed) Hoboken NJ John Wiley amp Sons

Harter S amp Jackson B K (1992) Trait vs nontrait conceptualizations of intrinsicextrinsic motivational orientation Motivation and Emotion 16 209ndash230

Heckhausen J amp Heckhausen H (2010) Motivation and action introduction andoverview In J Heckhausen amp H Heckhausen (Eds) Motivation and action (2ndedition pp 1ndash9) Cambridge Cambridge University Press

Hoyt W T amp Melby J N (1999) Dependability of measurement in counseling Anintroduction to generalizability theory The Counseling Psychologist 27 325ndash352

Kaplan A amp Maehr M L (2007) The contributions and prospects of goal orientationtheory Educational Psychology Review 19 141ndash184 doi101007s10648-006-9012-5

Kishton J M amp Widaman K F (1994) Unidimensional versus domain representativeparceling of questionnaire items An empirical example Educational andPsychological Measurement 54 757ndash765

Lakes K D amp Hoyt W T (2008) What sources contribute to variance in observerratings Using generalizability theory to assess construct validity of psychologicalmeasures Infant and Child Development 17 269ndash284

Little T D Cunningham W A Shahar G amp Widaman K F (2002) To parcel ornot to parcel Exploring the question weighing the merits Structural EquationModeling 9 151ndash173

Lord F M (1963) Elementary models for measuring change In C W Harris (Ed)Problems in measuring change Madison University of Wisconsin Press

Maehr M L amp Zusho A (2009) Achievement goal theory The past present andfuture In K R Wentzel amp A Wigfield (Eds) Handbook of motivation at school(pp 77ndash104) New York Routledge

Malmberg L-E (2008) Student teachersrsquo achievement goals during teacher studiesAntecedents correlates and outcomes Learning and Instruction 18 438ndash452

Muis K R amp Edwards O (2009) Examining the stability of achievement goalorientation Contemporary Educational Psychology 34 265ndash277 doi101016jcedpsych200906003

Murayama K Elliot A J amp Friedman R (2012) Achievement goals and approach-avoidance motivation In R M Ryan (Ed) The Oxford handbook of humanmotivation (pp 191ndash207) Oxford Oxford University Press

Murphy P K amp Alexander P A (2000) A motivated exploration of motivationterminology Contemporary Educational Psychology 25 3ndash53 doi101006ceps19991019

Nicholls J G (1984) Achievement motivation Conceptions of ability subjectiveexperience task choice and performance Psychological Review 91 328ndash346

Nitsche S (2013) Zielorientierungen von Lehrkraumlften und ihre Bedeutung fuumlr dasberufliche Lern- und Fortbildungsverhalten [Goal orientations among teachers and

their significance for occupational training choices] (Unpublished doctoraldissertation) University of Mannheim Germany

Nitsche S Dickhaumluser O Fasching M S amp Dresel M (2011) Rethinking teachersrsquogoal orientations Conceptual and methodological enhancements Learning andInstruction 21 574ndash586

Nitsche S Dickhaumluser O Fasching M S amp Dresel M (2013) Teachersrsquo professionalgoal orientations Importance for further training and sick leave Learning andIndividual Differences 23 272ndash278 doi101016jlindif201207017

Pintrich P R (2000) Multiple goals multiple pathways The role of goal orientationin learning and achievement Journal of Educational Psychology 92 544ndash555

Retelsdorf J Butler R Streblow L amp Schiefele U (2010) Teachersrsquo goal orienta-tions for teaching Associations with instructional practices interest inteaching and burnout Learning and Instruction 20 30ndash46 doi101016jlearninstruc200901001

Retelsdorf J amp Guumlnther C (2011) Achievement goals for teaching and teachersrsquoreference norms Relations with instructional practices Teaching and TeacherEducation 27 1111ndash1119 doi101016jtate201105007

Salas E amp Cannon-Bowers J A (2001) The science of training A decade of progressAnnual Review 52 471ndash499 doi0066-4308010201-0471$1400

Schoumlne C (2008) Zielorientierung und Bezugsnormpraumlferenzen in Lern- undLeistungssituationen [Goal orientation and reference norm preferences in learningand achievement situations] (Unpublished dissertation) University of GieszligenGermany

Senko C Hulleman C S amp Harackiewicz J M (2011) Achievement goal theory atthe crossroads Old controversies current challenges and new directionsEducational Psychologist 46 26ndash47

Shavelson R J amp Webb N M (1991) Generalizability theory A primer Newbury ParkCA Sage

Shumate S R Surles J Johnson R L amp Penny J (2007) The effects of the numberof scale points and non-normality on the generalizability coefficient A MonteCarlo study Applied Measurement in Education 20 357ndash376

Silva T amp Nicholls J (1993) College students as writing theorists Goals and beliefsabout the causes of success Contemporary Educational Psychology 18 281ndash293

Steyer R Schmitt M amp Eid M (1999) Latent state-trait theory and research inpersonality and individual differences European Journal of Personality 13 389ndash408

Toumlnjes B amp Dickhaumluser O (2009)Laumlngsschnittliche Effekte von Zielorientierungenauf Dimensionen des beruflichen Belastungserlebens im Lehrerberuf[Longitudinal effects of achievement goals on factors of occupational burden inthe teaching profession] Zeitschrift fuumlr Paumldagogische Psychologie 41 79ndash86

Vollmeyer R amp Rheinberg F (1998) Motivationale Einfluumlsse auf Erwerb undAnwendung von Wissen in einem computersimulierten System [Motivationalinfluences on the acquisition and application of knowledge in a computer-simulated system] Zeitschrift fuumlr Paumldagogische Psychologie 12 11ndash24

Webb N M Shavelson R J amp Haertel E H (2006) Reliability coefficients andgeneralizability theory In C Rao amp S Sinharay (Eds) Handbook of statistics (Vol26 pp 81ndash124) Amsterdam The Netherlands Elsevier

Ziegler A Dresel M amp Stoeger H (2008) Addressees of performance goals Journalof Educational Psychology 100 643ndash654 doi1010370022-06631003643


Here today gone tomorrow Revisiting the stability of teachers achievement goals

Introduction

Teachers achievement goals definition and relevance

Teachers achievement goals stable characteristics

Determining the stability of achievement goals using generalizability theory

Research questions and hypotheses

Method

Sample and procedure

Instruments

Analyses

Results

Descriptive analyses

Stability of teachers achievement goals over time

Number of necessary measurement points

Discussion

Stability of teachers achievement goals and its explanation

Implications of the (in-)stability of teachers achievement goals

Limitations and further directions

Conclusions

Acknowledgments

References

amp Alexander 2000 Pintrich 2000) The purpose of the study at handis to shed light on this topic by investigating the stability ofteachersrsquo achievement goals as well as the number of measure-ment points necessary to reliably measure these goals acrossoccasions

11 Teachersrsquo achievement goals definition and relevance

Achievement goals explain how and why people behave the waythey do in achievement settings (Dweck amp Leggett 1988 MurayamaElliot amp Friedman 2012) Achievement goal theory differentiatesbetween various goals The achievement goals that are commonlydistinguished when describing and explaining characteristics ofteacher motivation are (a) learning goals (the teacher aims to in-crease his or her own competencies) (b) performance approach goals(the teacher aims to demonstrate high competencies) (c) perfor-mance avoidance goals (the teacher aims to avoid the impressionof low competencies) and (d) work avoidance goals (the teacheraims to reduce his or her workload) Several studies have pointedout the relevance of teachersrsquo achievement goals for the teachingprofession as they have revealed relationships between teachersrsquoachievement goals and various teacher and teaching characteris-tics Associations have been found for example betweenachievement goals and occupational burden or burn-out (NitscheDickhaumluser Fasching amp Dresel 2013 Retelsdorf et al 2010 Toumlnjesamp Dickhaumluser 2009) the perception of help-seeking as beneficialor threatening (Butler 2007 Nitsche et al 2011) participation invocational training programs (Nitsche et al 2013) aspects of in-structional quality (Butler 2012 Butler amp Shibaz 2008 Retelsdorfamp Guumlnther 2011) and the goal structures teachers realize in theirclassrooms ie the extent to which pursuing learning vs perfor-mance goals for students is reinforced by the classroom environment(Butler 2012 Dresel et al 2013 Retelsdorf et al 2010) In all ofthese studies learning goals were positively correlated with vari-ables that are considered to be beneficial (eg attending vocationaltraining programs) and negatively correlated with variables that areregarded as adverse (eg burn-out) For performance avoidance goalsas well as work avoidance goals the relationship pattern was in most

cases reversed The results regarding the effects of performanceapproach goals were mixed (eg positive effects on teacher self-efficacy in a study by Nitsche et al 2011 positive effects on thesocial reference norm in a study by Retelsdorf amp Guumlnther 2011)Most of the relationships identified between teachersrsquo achieve-ment goals and other variables were small some were moderate

12 Teachersrsquo achievement goals stable characteristics

Regarding the conceptualization of achievement goals large differ-ences can be found (for an overview see Pintrich 2000 and Maehr ampZusho 2009 see also Button Mathieu amp Zajac 1996 DeShon amp Gillespie2005) According to some conceptions achievement goals are assumedto be rather stable (eg Silva amp Nicholls 1993) whereas for other con-ceptions they are assumed to be rather unstable (eg Elliott amp Dweck1988) The differences regarding the assumed stability of achievementgoals are important The theoretical conception of achievement goalsinfluences (a) how investigations concerning these goals are con-ducted (eg how many measurement points are used) (b) how theyare measured (eg whether achievement goals are assessed with respectto specific situations) (c) what research questions are appropriate (egwhether it makes sense to investigate effects of achievement goals onstable outcomes) and (d) what implications are derived from the resultsof the investigations (Murphy amp Alexander 2000 Pintrich 2000 for asimilar argumentation regarding intrinsic and extrinsic motivation seeHarter amp Jackson 1992)

Based on the considerations of Fryer and Elliot (2007) and Pintrich(2000) we developed a conceptual framework to explain why achieve-ment goals on the one hand can be assumed to be stable but on theother hand are also assumed to be unstable (see Fig 1) Fryer and Elliot(2007) and Pintrich (2000) do not differentiate between differentachievement goals in their argumentation This implies that differ-ences in the stability of the goals are not expected Additionally noinformation is given regarding the expected magnitude of the stableand the unstable components of achievement goals

Empirical investigations regarding the stability of achievementgoals exist first and foremost for studentsrsquo achievement goals (foran overview see Fryer amp Elliot 2007 Pintrich 2000) According to

Fig 1 Conceptual framework for the stability of achievement goals


















Retestinterval

Achievement goal



Workavoidance





2 Method




22 Instruments



23 Analyses













o σ2po σ2




po

pi oi

poie








3 Results





















Table 3G analyses
















4 Discussion










05

06

07

08

09

1

1 2 3 4 5 6 7 8 9 10

Rel

ativ

e G

co

effi

cien

t


Learning goals




01




















5 Conclusions


Acknowledgments


References























































Introduction





Method


Instruments

Analyses

Results




Discussion




Conclusions

Acknowledgments

References

















Retestinterval

Achievement goal



Workavoidance





2 Method




22 Instruments



23 Analyses













o σ2po σ2




po

pi oi

poie








3 Results





















Table 3G analyses
















4 Discussion










05

06

07

08

09

1

1 2 3 4 5 6 7 8 9 10

Rel

ativ

e G

co

effi

cien

t


Learning goals




01




















5 Conclusions


Acknowledgments


References























































Introduction





Method


Instruments

Analyses

Results




Discussion




Conclusions

Acknowledgments

References


2 Method




22 Instruments



23 Analyses













o σ2po σ2




po

pi oi

poie








3 Results





















Table 3G analyses
















4 Discussion










05

06

07

08

09

1

1 2 3 4 5 6 7 8 9 10

Rel

ativ

e G

co

effi

cien

t


Learning goals




01




















5 Conclusions


Acknowledgments


References























































Introduction





Method


Instruments

Analyses

Results




Discussion




Conclusions

Acknowledgments

References






3 Results





















Table 3G analyses
















4 Discussion










05

06

07

08

09

1

1 2 3 4 5 6 7 8 9 10

Rel

ativ

e G

co

effi

cien

t


Learning goals




01




















5 Conclusions


Acknowledgments


References























































Introduction





Method


Instruments

Analyses

Results




Discussion




Conclusions

Acknowledgments

References






4 Discussion










05

06

07

08

09

1

1 2 3 4 5 6 7 8 9 10

Rel

ativ

e G

co

effi

cien

t


Learning goals




01




















5 Conclusions


Acknowledgments


References























































Introduction





Method


Instruments

Analyses

Results




Discussion




Conclusions

Acknowledgments

References


















5 Conclusions


Acknowledgments


References























































Introduction





Method


Instruments

Analyses

Results




Discussion




Conclusions

Acknowledgments

References








































Introduction





Method


Instruments

Analyses

Results




Discussion




Conclusions

Acknowledgments

References

here today, gone tomorrow? revisiting the stability of teachers' achievement goals

Documents