What Teachers Think and Know about ADHD: Validation of the ADHD-school-expectation Questionnaire (ASE)

ABSTRACT

Although an average of one to two children per classroom suffer from ADHD, empirically supported classroom interventions are not yet implemented possibly because of teachers’ lack of knowledge or negative attitude towards them. To investigate this science-practitioner gap, we need an instrument assessing knowledge, attitude and the use of ADHD-related interventions. The self-report ADHD Questionnaire by Kos and the ADHD specific knowledge and attitudes scale for teachers by Mulholland et al. are examples of such instruments . Yet, these instruments have primary weaknesses concerning their content validity. Our study validated a newly developed instrument. The ADHD-school-expectation questionnaire (ASE) consists of an ADHD knowledge (24 items), attitude (33 items), and intervention scale (27 items). Attitudes are formed by expectations and related ratings according to Fishbein’s and Ajzen’s rational choice theory and Ajzen’s theory of planned behaviour. Our study results revealed theory-based content validity and good reliability. Therefore, the ASE can be used for future research in the field of ADHD in the classroom.

KEYWORDS:

• ADHD
• attitude
• classroom
• interventions
• knowledge

Introduction

Attention deficit hyperactivity disorder (ADHD) is a neurodevelopmental disorder comprising the three core symptoms inattention, hyperactivity and impulsivity that are required to present cross situationally, i.e. at home and in the kindergarten or school context before the age of six according to the ICD-10 (Remschmidt, Schmidt, & Poustka, 2017) or twelve years according to the DSM-5 (American Psychiatric Association, 2013). ADHD is most often diagnosed when school starts as teachers are faced with those core symptoms respectively the associated behavioural problems in school e.g. children not following teachers’ instructions, fidgeting in their chairs, blurting out in the classroom (Campbell, Halperin, & Sonuga-Barke, 2015).

ADHD affects 3% – 5% of children or adolescents worldwide (Polanczyk, Willcutt, Salum, Kieling, & Rohde, 2014), making them present in almost every classroom. To address the associated behavioural problems, classroom interventions have been developed and validated. Although such interventions reveal large effect sizes (Gaastra, Groen, Tucha, & Tucha, 2016) that are comparable to psychostimulant medication or psychotherapy (Catalá-López et al., 2017), they are not translated into everyday practice of schools (Ruhmland & Christiansen, 2017). What causes this science-practitioner gap?

Possible Reasons for the Science-practitioner Gap

Lack of Knowledge about ADHD

One relevant factor concerning this science-practitioner gap is teachers’ knowledge about ADHD. Knowledge can be defined as the amount of correct information a person possesses (Ajzen, Joyce, Sheikh, & Cote, 2011). So far, only few studies have addressed the relation between knowledge of ADHD and use of classroom interventions. Ohan, Cormier, Hepp, Visser, and Strain (2008) found that teachers with moderate and high knowledge levels of ADHD favoured educational and domestic support; these teachers were also found to perceive more benefit from changes in the classroom than those with low knowledge level. Unfortunately, several investigations have demonstrated that teachers’ knowledge of ADHD is often poor and based on false information, which, in turn, contributes to negative perceptions of pupils with ADHD and counteracts the implementation of effective interventions (Ruhmland & Christiansen, 2017; Soroa, Gorostiaga, & Balluerka, 2016). Some teachers and schools participate in knowledge-based training sessions, but their participation does not seem to have resulted in the greater implementation of effective classroom interventions (Moore, Russell, Arnell, & Ford, 2017). A possible reason for this situation can be the type of knowledge that is imparted. General knowledge does not predict a specific behaviour (Ajzen et al., 2011). Correspondingly, knowledge that is related to the behaviour of interest must be communicated. However, the accurate knowledge of ADHD treatment, which is probably the most cogent knowledge regarding classroom interventions, has been often lacking (Ohan et al., 2008).

At present, the role of knowledge of ADHD in implementing evidence-based classroom interventions is difficult to test using available scales, as the latter offers no information on scale construction, factorial, or content validity. These scales fail to mention the underlying relevance of content items in a school context (Kos, 2004; Mulholland, 2016; Sciutto, Terjesen, & Frank, 2000). However, such relevance to school context is crucial in investigating the relation between knowledge and appropriate use of interventions (Ajzen et al., 2011).

Negative Attitude Towards Children with ADHD

Attitude is another relevant factor when considering the use of interventions (Lübke, Meyer, & Christiansen, 2016). Attitude represents individuals’ expectations and related rating towards an object. The rational choice theory of Fishbein and Ajzen (1975) suggests that attitude is the sum of expectations multiplied by the associated individual ratings. Attitude consists of a cognitive (beliefs that a person associates with a certain object), an affective (affective reactions that a person expects to be elicited by an object), and a behavioural (former or anticipated future behaviours related to an object) component. An example of the components’ interrelationship is: a positive cognitive and affective attitude is often accompanied by a positive behavioural attitude (Haddock & Maio, 2014). A positive behavioural attitude is crucial in demonstrating a specific behaviour, according to Ajzen’s theory of planned behaviour (TPB; 1991).

In TPB, the essential factors to predict behavioural intentions and the occurrence of a specific behaviour are: attitude towards the behaviour (a person’s belief that executing a certain behaviour will lead to a desirable consequence), subjective norm (whether people who are important to a person support the intended behaviour and whether that person appreciates what those people think about him/her), and perceived control (whether a person is able to reveal the intended behaviour and/or handle a related situation). The neglect of factors like attitude could explain why the latest scientific knowledge has not been put into practice. Therefore, the assessment of teachers’ attitudes towards children with ADHD is probably crucial in implementing evidence-based classroom interventions.

Some working groups have assessed teachers’ attitudes towards pupils with ADHD using qualitative open-ended questions (Anderson, Watt, Noble, & Shanley, 2012) or quantitative surveys measuring the extent of (dis-)agreement to items (Kos, 2004; Mulholland, 2016). A problem in these instruments is that the investigator decides whether a cognition is positive or negative when evaluating answers or formulating items. According to Fishbein and Ajzen (1975), each person may rate a given aspect differently, as this is already part of his/her attitude. Researchers have suggested different nuances within positive vs. negative ratings that interviewees choose, ranging from −3 to +3.

Additionally, different aspects in existing scales appear confounded. These scales often contain a mix of items on knowledge and beliefs, assess the extent of agreement with prejudices, and/or include items that measure perceived control or subjective norm (Kos, 2004; Mulholland, 2016). This mixture leads to an unclear content-related structure. To our knowledge, no existing measure sets teachers’ attitudes towards children with ADHD as a distinct variable.

The Present Study

To address the science-practitioner gap and implement evidence-based classroom interventions, knowledge of and attitudes towards pupils with ADHD needs to be investigated. Developing an instrument that assesses these factors accurately is the first step to improving the treatment of pupils with ADHD.

The newly constructed ADHD-school-expectation questionnaire (ASE) should thus contain knowledge, attitude, and intervention scales. The knowledge scale should focus on knowledge of symptoms, aetiology, diagnostics, prevalence, and interventions that enable teachers to identify pupils with ADHD and treat them effectively. The attitude scale should assess expectations and related ratings to operationalise teachers’ attitudes towards pupils with ADHD following Fishbein and Ajzen (1975) rational choice theory. This scale should also include the cognitive and affective components of attitude. The cognitive part should focus on situations in a classroom setting, which the interventions are addressing. Finally, the intervention scale should assess the use of interventions and their effectiveness ratings. The latter represents the expectation of a positive consequence and behavioural component of attitude.

Overall, this study validated the new ASE to determine whether it is psychometrically sound and can be used to assess teachers’ knowledge and attitude towards children with ADHD, as well as teachers’ attitudes towards and their use of evidence-based classroom interventions in this context.

Methods

Study Design and Procedure

This validation study was conducted without blinding or randomisation, via an online survey on https://www.soscisurvey.de. It was administered to pre-service teachers in the first step and in-service ones in the second step.

Existing scales on teachers’ knowledge and attitudes of ADHD and affected children were reviewed. The ASE’s initial version was piloted and then rated by two child and youth psychology experts and three in-service teachers. Then minor changes with respect to wording were made. The ASE’s final version was validated among pre-service teachers, as they were most likely to receive updated information. A cover letter providing a detailed study information and link to the online survey was disseminated among German university email lists and Facebook. After three weeks, a reminder was sent. Data collection lasted for two more weeks. A Nintendo Switch (value about EUR 350) and three vouchers worth EUR 50 were raffled to encourage participation in this study.

The examination of the psychometric properties of the ASE was afterwards replicated with in-service teachers.

Participants

Data from N = 1,086 pre-service teachers, who completed at least 75% of the items per scale, were used in the analysis (N = 1741, drop out: n = 655; 37.62%). Participants’ mean age was 23.22 years (SD = 3.93 years). There were 30.57% male, 68.97% female, and 0.46% other. The distribution of school types for which they were educated is illustrated in Table 1 (supplemental material). All participants provided written and informed consent after study information. Participation was completely anonymous and voluntary. Participants could always opt out without stating the reason.

Replication Study

The study was replicated with data from N = 599 in-service teachers, their mean age was 41.33 years (SD = 10.01 years). There were 17.67% male and 82.30% female participants.

ASE Development

The ASE comprised three scales: 1) knowledge of, 2) attitude towards, and 3) school-based interventions for pupils with ADHD.

Knowledge Scale

We aimed to document the knowledge scale’s content validity. All items were therefore related to teachers’ use of interventions.

Choice of Categories

Options for categorising knowledge were researched. We decided on knowledge of symptoms, aetiology, diagnostics, prevalence, and interventions to be important for teachers to know a) whether a child is affected by ADHD, and if so, b) that there is nobody to blame, c) that a valid diagnosis is not given intuitively or inflationary, and d) that interventions are useful. This knowledge is expected to be the basis for their willingness to use evidence-based classroom interventions for pupils with ADHD. Diagnostics and prevalence were combined as both are related to how and how often ADHD is diagnosed.

Item Construction

After item construction, the number of items per category was set to six, including the most important aspects, according to expert judgements. Items were designed for laypeople and to be easy to answer. Only correct knowledge of ADHD according to evidence, current guidelines, and common ADHD-related myths among teachers mentioned in other studies (Ruhmland & Christiansen, 2017) were integrated. Items stated accurately and falsely were balanced for every category. The order of all items was randomised.

The knowledge scale contained 24 items from the symptoms, aetiology, diagnostic and prevalence, and intervention categories. After conducting a pilot study, two items of the intervention category were rephrased in accordance with the feedback. As a result, one item that was stated falsely and five items that were stated accurately were left in this category.

Answer Format

A visual true-or-false analogue scale (VAS) was used to enhance compliance. This answering format is known to be valid and reliable (Ahearn, 1997). The VAS was subdivided into 12 sections so that the cursor could move smoothly. Only a correct answer within the first sixth of the VAS was granted one knowledge point, as our main focus was to examine what items were certainly correctly answered. Participants could earn 24 knowledge points.

Attitude Scale

This scale should assess teachers’ attitude towards pupils with ADHD based on Fishbein and Ajzen (1975) rational choice theory, including expectations and their related ratings. We included the cognitive and affective aspects of attitude. Assessing the behavioural component to address previous behaviours is pointless when investigating changes in attitude. The anticipated future behaviour was incorporated into our effectiveness ratings to represent the expectation that the use of interventions will trigger a positive consequence. It was integrated into our intervention scale.

Cognitive Component

The cognitive part focused on teachers’ expectations concerning pupils with ADHD, especially during classroom instructions. We assessed their attitudes towards academic studies (22 items: main focus during classroom instruction) and social behaviour (five items) in class. We based the ASE attitude scale on ADHD symptoms and current myths (Ruhmland & Christiansen, 2017), which were balanced as positive and negative statements. Teachers were asked to rate how likely they were to expect pupils with ADHD to exhibit the specified characteristics and how they would rate those characteristics in general.

Affective Component

The affective part assessed the emotions that teachers generally feel in their job (Lee et al., 2016). The items were balanced between positive and negative emotions. Teachers rated how likely for pupils with ADHD to elicit such emotions in them and how they would rate those emotions in general.

Answer Format

The total scale included 33 items (6 affective and 27 cognitive ones, with the latter targeting academic studies [22 items] and social behaviour [five items]). The cognitive items were answered first; they represented the most important influence on attitude. The affective items were answered next. The order of presentation of cognitive and affective items was randomised. The expectation of likelihood (VAS scale from 0 = unlikely to 1 = likely) and positive or negative ratings (VAS scale from −3 = negative to 3 = positive) were required in all items. The VAS ratings were clustered into six sections. To operationalise the variable attitude, every expectation was multiplied by its related rating. All results were added up.

Intervention Scale

The intervention scale assessed teachers’ attitudes towards and the use of school-based interventions for pupils with ADHD. Items were formulated neutrally. The scale contained 27 items addressing intervention strategies; these were divided into 15 effective and 12 ineffective strategies based on previous studies (Dupaul, Eckert, & Vilardo, 2012; Miranda, Presentación, & Soriano, 2002). The order of presentation was randomised. The (estimated) use of the intervention strategy (VAS scale from 0 = never to 1 = very often) and its estimated effectiveness (VAS scale from 0 = not effective at all to 1 = very effective), which represented the attitude towards the behaviour, were required in all items. The VAS ratings were clustered into six sections.

Data Analysis

Data were downloaded from https://www.soscisurvey.de as an IBM SPSS Statistics file. Analyses were conducted using IBM SPSS Statistics 24 (2016) and IBM SPSS AMOS 24 (2016). After excluding the questionnaires with at least 25% incomplete answers per scale, recoding the items, and calculating the knowledge points, attitude values, and intervention use, a descriptive statistical analysis was conducted, followed by a confirmatory factor analysis (CFA). If the CFA did not reveal satisfactory model fits, an exploratory factor analysis (EFA) was performed to establish the scale’s factor structure. No factor analysis was conducted for the intervention scale, as its construction was not theory-based but represented findings from earlier studies. As an indicator for discriminant validity, the correlations between various scales were calculated. Finally, the scales’ psychometric properties were computed.

CFA

The CFA factors were allowed to correlate, which is theoretically plausible. To identify the model, variances of the latent variables were fixed to one. The maximum-likelihood-method was used to estimate the parameters. The χ2/degrees of freedom (df) value was used to assess the model-fit inferential; χ2/df values between 0.000 and 2.000 and between 2.010 and 3.000 represented good and acceptable model fit, respectively. As the χ2/df value can be influenced by a large sample size, the root mean square error of approximation (RMSEA), comparative fit index (CFI), and normed fit index (NFI) as descriptive measures were considered. RMSEA values <.050 represent good and RMSEA values between.051 and .080 acceptable fit. CFI values between 0.970 and 1.000 and NFI values between 0.950 and 1.000 were interpreted as good model fit; CFI values between 0.950 and 0.969 and NFI values between 0.900 and 0.949 indicated acceptable model fit (Moosbrugger & Schermelleh-Engel, 2012).

Knowledge Scale: CFA Models

We tested three different models for the knowledge scale. Model 1 assumed the four factors (i.e., symptoms, aetiology, diagnostic and prevalence, and intervention) as analogous to the chosen categories. Model 2 subdivided each of these factors into one representing information that was stated falsely and another that was stated accurately, and assumed eight factors. Model 3 represented an even more detailed bi-factor model, assuming the eight factors of model 2 and a general ADHD knowledge factor.

Attitude Scale: CFA Models

Two different models were tested for the attitude scale. Model 1 assumed academic behaviour, social behaviour, and elicited emotions factors. Model 2 subdivided each factor into a positive and negative element and assumed six factors.

EFA

We applied principal component extraction with Varimax rotation. The Kaiser-criterion, which suggests extracting factors with an Eigenvalue > 1, was used. As this criterion often overestimates the number of relevant factors, we also considered the scree test (Moosbrugger & Schermelleh-Engel, 2012).

Results

We present results of the main study with pre-service teachers first; those are followed by the replication with in-service teachers.

Descriptive Results

On average, participants earned 7.23 knowledge points (SD = 4.21, maximum = 22.00, minimum = 0.00).

In the 33-item attitude scale, all expectations were multiplied by their related ratings and then added together. The mean attitude score was −11.60 (SD = 13.92, maximum = 44.80, minimum = −72.60). Table 2 (supplemental material) visualises the expectation probability and average ratings of items.

Participants’ mean level use of all interventions was 0.55 (SD = 0.09), of effective interventions was 0.72 (SD = 0.12), and of ineffective interventions was 0.35 (SD = 0.14). The average effectiveness rating for all interventions was 0.52 (SD = 0.09), for effective interventions was 0.73 (SD = 0.13), and for ineffective interventions was 0.26 (SD = 0.012). Table 3 (supplemental material) presents the results of each item.

Knowledge Scale

The CFA for model 1 revealed χ2/df = 4.76, RMSEA = 0.06, CFI = 0.73 and NFI = 0.68. Only the RMSEA value suggested an acceptable model fit.

Model 2 resulted in χ2/df = 2.81, RMSEA = 0.04, CFI = 0.88 and NFI = 0.83. The χ2/df and RMSEA values indicated an acceptable and good model fit, respectively. The CFI and NFI values did not show good fit.

The bi-factor model resulted in an overall good model fit: χ2/df = 1.88, RMSEA = 0.03, CFI = 0.95 and NFI = 0.90. This model covered our ideas of the knowledge scale development. Thus, this result indicated good content and construct validity.

Cronbach’s α of .80 indicated good internal consistency and reliability for the knowledge scale. Each item difficulty and discriminatory power is presented in Table 4 (supplemental material). The values fall within a methodologically acceptable range.

Attitude Scale

The CFA of model 1 revealed an overall unsatisfactory model fit; χ2/df = 14.50, RMSEA = 0.11, CFI = 0.45 and NFI = 0.41.

Model 2 resulted in χ2/df = 3.92, RMSEA = 0.05, CFI = 0.88 and NFI = 0.84. Only the RMSEA value indicated an acceptable model fit. Both CFA results did not support our primary assumptions of the attitude scale development.

A subsequent EFA suggested six factors according to the Kaiser-criterion with Eigenvalues of 6.05, 5.68, 1.80, 1.68, 1.04, and 1.01. The rotated factor matrix is shown in Table 5 (supplemental material). All items tended to load on factors that distinguished between positive and negative aspects: factor 1, positive behaviour; factor 2, negative behaviour; factor 3, negative emotions; factor 4, outstanding negative behaviour; factor 5, outstanding positive behaviour; and factor 6, positive emotions. The scree test suggested mainly two factors. Therefore, we performed a second EFA with two fixed factors (i.e., negative and positive aspects), with eigenvalues of 6.05 and 5.68 (see Table 6 [supplemental material] for details).

Both EFA solutions represented rating nuances that were a core aspect in developing the attitude scale. The more detailed solution also differentiated cognitive and affective factors. The results therefore indicated content and construct validity for the ASE’s attitude scale. The two-factor-solution was less generous and should be preferred.

The attitude scale’s internal consistency revealed a Cronbach’s α of .85, representing good internal consistency and reliability. The negative and positive aspects of the subscale’s internal consistency accounted for a Cronbach’s α of .87. The items’ discriminatory power is shown in Table 2 (supplemental material).

Intervention Scale

We calculated the intervention scale’s psychometric properties separately for the use of and estimated effectiveness of the interventions. Both calculations revealed a Cronbach’s α of .73, representing good internal consistency and reliability. The items’ discriminatory power is presented in Table 3 (supplemental material).

Inter-scale Correlations

The correlation between knowledge and attitude resulted in r = −.09. The use of interventions correlated to r = .12 with knowledge and to r = −.18 with attitude. The rating of interventions correlated with knowledge to r = .09 and with attitude to r = .09. These low inter-correlation results indicated distinct variables. Detailed results for the subscales’ correlations are presented in Table 7 (supplementary material).

Results of the Replication Study with In-service Teachers

Descriptive analysis for the group of in-service teachers resulted in M = 9.25 (SD = 4.29) knowledge points, and an average attitude score of −13.12 (SD = 14.66). Their mean use of all interventions was 0.57 (SD = 0.09), 0.74 (SD = 0.13) of effective, and 0.36 (SD = 0.14) of ineffective interventions. The average effectiveness rating for all interventions was 0.54 (SD = 0.09), 0.76 (SD = 0.14) for effective, and 0.26 (SD = 0.13) for ineffective interventions.

The CFA for the bi-factor model of the knowledge scale resulted in an overall acceptable model fit and good internal consistency: χ2/2 = 1.59, RMSEA = 0.03, CFI = 0.93 and NFI = 0.85, and Cronbach’s α = .78.

For the attitude scale, the scree test suggested two factors that were analogous to the main study, with Eigenvalues of 5.89 and 4.70. Cronbach’s α was .85 for the total attitude scale, .86 for the negative subscale, and .84 for the positive subscale. All values represented good internal consistency and reliability.

Cronbach’s α for the intervention scale was .72 for the use and .75 for the estimated effectiveness of the interventions, indicating good internal consistency and reliability.

In summary, the results of the study with pre-service teachers could be replicated by the sample of in-service teachers. This further supports the ASE’S content and construct validity as well as reliability.

Discussion

The present study demonstrated how the ASE assesses pre- and in-service teachers’ knowledge of, attitudes towards, and use of school-based interventions for pupils with ADHD and that it revealed solid psychometric properties.

Knowledge Scale

The knowledge scale resulted in a bi-factor model, with the symptoms, aetiology, diagnostic and prevalence, and intervention categories subdivided into stated accurately and stated falsely, and a general ADHD knowledge factor. This bi-factor model was complex but covered the central aspects of the scale development process, revealing content and construct validity. The Cronbach’s α value confirmed the knowledge scale’s good reliability.

The descriptive results of this study demonstrated that pre- and in-service teachers seem to have limited knowledge of ADHD. Items might have been too difficult, as none of the participants could answer all questions correctly (on average, not even half of the questions were answered correctly). However, item difficulties were predominantly methodologically acceptable as the values ranged (except for one item) between .05 and .95, with an average value of .30. Moreover, teachers’ poor knowledge of ADHD is consistent with previous findings in this field (Ruhmland & Christiansen, 2017; Soroa et al., 2016). Nevertheless, a study with child and adolescent psychotherapists, who can be expected to possess solid knowledge of ADHD, supported methodologically acceptable item difficulties (Dort et al., in preparation). Some knowledge items also did not reveal good discriminatory power values. However, this can be attributed to associated item difficulties, which were mainly items that were answered incorrectly by nearly all participants, e.g. the item ‘Approximately 3–4 pupils per class* are affected by ADHD’ resulted in a low item difficulty value of .08 with a low discriminatory power of 0.19.

Attitude Scale

The attitude scale was assumed to load on academic behaviour and social behaviour factors (cognitive aspect) and elicited emotions (affective aspect). The study results did not support such a factor structure. Instead, the exploratory analysis revealed that the items mainly load on one factor representing positive aspects and another representing negative aspects. A more differentiated EFA revealed the six factors positive behaviour, negative behaviour, negative emotions, outstanding negative behaviour, outstanding positive behaviour, and positive emotions. Those factors demonstrated different nuances of positive and negative ratings and that the cognitive and affective items load on different factors. This pattern of results is consistent with the developmental process of the scale, as we distinguished between cognitive and affective aspects as well as positive and negative ones. Furthermore, different nuances of positive and negative aspects fit the rational choice approach (Fishbein & Ajzen, 1975) that the scale was based on. Thus, the scale is meeting content validity criteria. Our analysis of internal consistency revealed good reliability for the attitude scale as a whole and subscales on attitude towards positive aspects and attitude towards negative aspects. Considering the attitude scale as a whole, some discriminatory power values were quite low. However, within the corresponding subscale, all discriminatory power values fell within the medium-to-high range, e.g. the item ‘is creative’ of the attitude scale only obtained a discriminatory power value of 0.19, but the discriminatory power of the corresponding attitude scale ‘attitude towards positive aspects’ obtained a value of 0.35.

The ASE improves the critical aspects (e.g., confounding different variables) of existing instruments assessing attitude (Kos, 2004; Mulholland, 2016). The ASE attitude scale has the advantage of differentiating attitude’s aspects. By measuring expectations and their associated ratings separately, as well as cognitive and affective aspects, we can investigate which of these elements are particularly important when trying to change teachers’ attitudes towards pupils with ADHD.

Intervention Scale

We did not investigate the intervention scale’s content validity, as the construction was based not on theoretical assumptions but previous research findings. The scale revealed good reliability and allows for assessing the use of interventions and their estimated effectiveness. The latter represents the attitude towards the use of interventions as it exploits the expectation that this would have a positive consequence.

The descriptive results of this study suggested that pre- and in-service teachers can estimate the effectiveness of interventions correctly. Both groups are assumed to employ ineffective interventions. This finding is congruent with results of previous studies that demonstrated the use of ineffective interventions (e.g. Ruhmland & Christiansen, 2017). However, so far it was unknown that the teachers themselves do not believe in the effectiveness of the identified ineffective interventions.

Implications

The present study introduces the ASE as an instrument to measure knowledge of, attitudes towards pupils with ADHD, and attitudes towards and the use of classroom interventions for pupils with ADHD. Such an instrument builds the basis for further investigation of the science-practitioner gap regarding the implementation of evidence-based classroom interventions for pupils with ADHD. This fact represents the motivation for the development of the ASE.

So far, studies predominantly highlighted the existence of the mentioned science-practitioner gap (e.g. Ruhmland & Christiansen, 2017). Though reasons for such a gap as well as possibilities to close it have rarely been addressed yet (Dort et al., 2020). With the ASE, first examinations could be conducted regarding the role of the variables knowledge and attitude in this context as both are significant factors moderating the use of classroom based interventions (Strelow, Dort, Schwinger, & Christiansen, 2020). Future studies should continue addressing implementation possibilities as well potential barriers. In a recently conducted bibliometric review we could show that this topic is rather neglected in research regarding the use of classroom interventions for pupils with ADHD (Dort et al., 2020). In this regard, the ASE could be used to evaluate intra-individual changes or the effects of implementation programs for example.

Further, as the ASE’s instruction can easily be adapted, it could also be used to assess attitudes towards other pupils. A study comparing teachers’ attitude towards pupils with ADHD, average pupils and ideal pupils to find out difference between the teachers’ current and desired classroom situation is currently in progress (Dort et al., in preparation).

Limitations

As a first limitation, we have to state that the pre-service teachers were heterogeneous in terms of their amount of classroom experience, so that only the estimated use of interventions could be assessed. Our scale’s properties and structure had to be confirmed in a sample of in-service teachers. Although this revealed comparable results, future research should investigate the relation among variables, such as experience, knowledge, attitude, and intervention use. A study that addresses this topic and aims to present a model of the relationships of the variable has just been published (Strelow et al., 2020).

Another limitation is the knowledge scale’s bi-factorial solution. Although this final structure is more reasonable than the basic factors alone, it limits the possibility of creating differentiated subscales. Owing to scoring only a correct answer within the first sixth of the VAS with one knowledge point, the information content is also rather low. Our main focus was to reduce guessing and determine whether teachers know a correct answer for sure. Using the VAS provides the advantage that scoring can be adjusted and more information on answering behaviour can be gained if desired.

We assumed some general knowledge of ADHD to be potentially a relevant basis for the use of evidence-based classroom interventions for pupils with ADHD. Future research must identify whether this is true or only specific knowledge of interventions is enough. So far, a model by Strelow et al. (2020) suggests a significant influence of general knowledge measured by the ASE on the effectiveness rating and intention to use effective interventions.

Additionally, the ASE’s cognitive items mainly asked for expectations related to pupils’ behaviours during lessons, as classroom interventions focus on such situations. Expectations considering breaks or homework situations, among others, could conceivably also affect teachers’ attitudes towards pupils with ADHD. Asking for cognitive expectations first could have had an effect on affective expectations. Queries about emotions first might be taken more personally and influence the willingness to complete the questionnaire.

The correlations of ASE scales’ variables were not compared with those of the same variables in other similar instruments, so that, inter-correlations cannot be interpreted exhaustively. However, the low inter-correlations suggest distinct variables. More information on the relation of knowledge to attitude scale values and other measures or external criteria would be desirable and must be addressed in future research. Nonetheless, the results of this study supported the theoretically underlying assumptions of the scale constructions indicating content validity.

Conclusion

In conclusion, we described the development and validation of a comprehensive instrument to measure pre- and in-service teachers’ knowledge of, attitudes towards pupils with ADHD, and attitudes towards and the use of classroom interventions for pupils with ADHD. Indications of the ASE’s content validity and reliability were successfully confirmed. Therefore, this study provides a theory-based instrument that improves on the limitations of previous research. The ASE demonstrated its usability in another study (Strelow et al. (2020) and can now be used to investigate the relation of the variables included in this questionnaire and the potential to influence them (e.g. with expectation violation) as further steps in closing the science-practitioner gap concerning classroom interventions for pupils with ADHD.

Supplemental Material

Supplemental data for this article can be accessed here.

Disclosure Statement

No potential conflict of interest was reported by the authors.

Additional information

Funding

This work was supported by the German Research Foundation (Deutsche Forschungsgemeinschaft – DFG) project number 290878970-GRK 2271, project 1.