Early Identification of School Refusal from Parents’ Perspectives

Abstract

School refusal is a growing concern having potentially long-term consequences for affected children and adolescents. Early intervention is widely recommended, yet there is limited knowledge about early warning signs that could support timely identification. This study explores whether parents’ experiences can contribute to a better understanding of such early indicators. A two-phase survey was conducted among parents of children with school refusal. In the first phase, 196 parents responded to an open-ended question about early signs, which were categorized into 30 types. These categories formed the basis for a structured questionnaire used in the second phase, completed by 509 parents. The analysis revealed significant variation in the occurrence of signs and signals. A factor analysis identified seven distinct types of early indicators, including social difficulties, academic challenges, psychosomatic complaints, and emotional expressions. The findings suggest that parents’ observations can provide valuable insights for early identification. However, many of the signs overlap with indicators of other challenges, such as anxiety or neglect, and must therefore be interpreted within a broader social context. This study highlights the importance of incorporating parental perspectives into early intervention strategies and the need for nuanced interpretation of early signs of school refusal.

1. Introduction

Children and young people who refuse to go to school make up a substantially small proportion of school pupils, but the consequences for those concerned can be large and long-lasting (Kearney, 2008). When school refusal has manifested itself in prolonged absence, it is difficult to return to school. Researchers have therefore advocated that the best measure is to intervene early in the development of school refusal in children and young people (Okuyama et al., 1999; Kearney & Graczyk, 2014; Ingul et al., 2019).

In the literature dealing with early intervention in general, the primary question is whether there are identifiable early signs or signals that can predict future health or behavioral problems in children and adolescents (Meisels & Atkins-Burnett, 2000; Norwegian Directorate of Health, 2018). For this purpose, a number of different mapping tools or lists of warning signals have been developed (Peters & Barlow, 2003; Van der Put et al., 2017; Ozturk et al., 2019). This also includes tools for mapping school refusal. However, a key point in this article is the need for more knowledge about signs and signals that can help detect school refusal at an earlier stage. In this context, the main question in this survey is whether parents’ or guardians’ experiences can contribute to strengthening this knowledge base (in the following, only the term parents will be used).

To address this question, an empirical study was conducted targeting the parents of children with school refusal. The study aimed to identify early signs and signals through a questionnaire distributed to the parents. The survey was carried out in two phases. In the first phase, 196 parents answered an open-ended question about the early signs they had observed. The responses were coded into 30 categories of signs or signals. These categories were subsequently transformed into standardized questions, which were included in the second phase of the survey. In the second phase, 509 parents answered standardized questions about signs and signals in a questionnaire. Based on the collected data, we conducted a descriptive analysis of the reported signs and signals, as well as a factor analysis to uncover latent dimensions within the material. The expected contribution is twofold: first, to expand the empirical knowledge base on early indicators of school refusal; and second, to provide insights that may inform early intervention efforts, foster collaboration between families and schools, and support the development of context-sensitive practices.

2. Previous Research

Categorizing different forms of school absence has been a central theme in the literature (Heyne, 2019; Kearney et al., 2019). Although in practice there may be smooth transitions between different types of absence, in this study we use a definition of school refusal that is in line with what Heyne et al. (2019) call school refusal. This differs from truancy, school withdrawal, and expulsion. Kearney et al. (2019) also suggest that we can distinguish between categorizations and dimensions. This means that within the category of school refusal, one can imagine varying degrees of refusal, ranging from subtle and unclear signals to chronic and long-term absence. This perspective is particularly useful in this study as we aim to identify early signs and signals of school refusal.

This is also in line with Ingul et al. (2019) who draw a distinction between emergent and manifested school refusal. A situation with emerging school refusal will be the phase before the school refusal manifests itself in a long-term absence. Manifested school refusal, on the other hand, means that the school refusal is expressed in a strong reluctance to go to school or in the form of chronic absence. Early intervention is therefore about the opportunities to intervene in the phase of emerging school refusal, and the question is whether there are specific signs that characterize emerging school refusal.

Few studies have carried out systematic research on early signs and signals of school refusal. The topic is included in studies of school refusal in various ways, often related to descriptions of comprehensive early intervention systems. In this literature, either examples of various signs and signals are highlighted, or it is emphasized that the most important sign of school refusal is emerging school absenteeism. In some studies, it is also somewhat unclear whether the topic concerns signs and signals in a phase of emerging school refusal or a phase of manifested school refusal (Gallé-Tessonneau & Heyne, 2020).

An example can be found in a study by Kearney and Graczyk (2014), who propose a comprehensive preventive and intervention system consisting of three levels. The first level consists of general preventive school measures (school climate) and a system for identification. This includes training school personnel in early signs of emerging school refusal. According to the authors, examples of such signals are frequent absences from lessons, pupils leaving school, late arrivals, and frequent visits to the health nurse. In addition, the authors point to risk factors such as transitions between school levels and decline in academic performance. In a recent article, Kearney et al. (2023) argue more specifically that changes in school attendance are an important early warning signal.

Another example can be found in an article by Chu et al. (2019), which describes a system for identification based on analyses of registered absences. In this system, pupils who have a certain level of absence within a given time period (25% absence during a 2-week period or absence for 10 days in a period of 15 weeks) are further surveyed through an online form to identify those at risk. The authors also claim that individual teachers’ observations are important for identifying emerging school refusal, but without explicitly stating what should be observed.

In both the examples above, school absence is considered a central warning sign. There is, however, reason to question whether emerging school absenteeism is a good indicator of future manifested school refusal (Gallé-Tessonneau & Heyne, 2020). The problem with using absenteeism as a warning sign is that students may refuse to go to school without it being reflected in particularly high absenteeism, and it is probably precisely at this stage that early intervention is beneficial. When the child stops going to school, the process has likely progressed to the point where we can rather talk about manifested school refusal (Ingul & Havik, 2021).

Different kinds of mapping tools have also been used to uncover school refusal. Two such tools are SRAS (Kearney, 2007) and SNACK (Heyne et al., 2019). Common to these tools is that they aim to map several types of problematic school absenteeism, not just school refusal. A key question regarding these tools has been whether they can effectively differentiate between various types of problematic absences (primarily truancy and school refusal). According to Gonzálvez et al. (2021), these tools have evolved from being used in clinical settings to being applied in natural settings across the entire population. This shift indicates a growing interest in using these tools to detect school refusal at an earlier stage rather than merely confirming it or determining the appropriate follow-up regime. Consequently, it can be somewhat unclear whether these tools are primarily designed to identify early warning signs or serve as diagnostic instruments to guide necessary follow-up actions (Gallé-Tessonneau & Heyne, 2020; Gonzálvez et al., 2021). In other words, it is unclear whether the tools should provide a basis for answering whether there is reason to act, or whether they should provide an answer to what kind of action is appropriate.

Based on a literature review, Ingul et al. (2019) identified a number of both signs and risk factors for school refusal. These are absenteeism, anxiety and depression, somatic complaints, problematic emotion regulation, and low self-management. In addition, the authors point out that there is an increased risk during the transitions to school and between school levels. In this study early warning signals and risk factors are referred to interchangeably. However, it may be appropriate to draw a distinction between the two terms.

Risk factors are individual characteristics or characteristics of the environment that increase the chances of school refusal. An individual characteristic can be, e.g., autism or ADHD. The rationale is that it is known from various surveys that among school refusers there is a clearly larger proportion with these diagnoses. Similarly, being exposed to bullying will represent a contextual risk factor. There does not have to be a direct causal connection between the risk factor and a negative outcome, but as a minimum, there must be a statistical correlation. Despite many known risk factors, however, it is not necessarily the case that a child shows signs or signals of emerging school refusal. In addition, a child does not need to be in a specific risk group but can still show signs of emerging school refusal.

Early warning signs can be defined as more or less clear expressions, behaviors, or symptoms in the individual child/adolescent, indicating actual emerging school refusal. On the one hand, such signs can be clear, for example, if the child themselves expresses fear of going to school. On the other hand, the signs can be non-verbal and unclear, such as withdrawal and internalized behavior. Between these extremes, we find indirect expressions, such as various somatic complaints. Unlike risk factors, which often (but not necessarily) point to some causal mechanisms underlying the emergence of school refusal, ideally early warning signs are direct indications of emerging school refusal.

The distinction between risk factors and warning signs can have practical significance. Being exposed to a risk factor in itself does not provide a basis for action. However, it may be a reason to be observant, especially if several risk factors are present. On the other hand, if a child/adolescent shows signs and signals, there is reason to conduct further investigations to determine if the warning signs are real and what they might indicate.

An underlying premise of identifying signs and signals of school refusal is that it creates a knowledge base for early interventions. This instrumental understanding implies that action follows naturally from the information that is collected. In practice, however, it turns out that school refusal cases are largely characterized by conflict and disagreements between school and parents (Malcolm et al., 2003; Havik et al., 2015; Amundsen & Møller, 2020). Torrens Armstrong et al. (2011) also show that school personnel have challenges distinguishing between different types of school absenteeism, that they construct various categories, and that these categories influence the reactions of the school personnel. According to the authors, such stereotypical categorizations can contribute to exacerbating the problem, especially in cases where the school personnel have specific perceptions of the individual situation. This means that acquiring knowledge to create a common understanding among the involved parties—such as parents, school personnel, and other support agencies—is just as important as acquiring knowledge to carry out specific actions or measures.

In this study, we have based our approach on parents’ own experiences with early signs and signals. This is based on the assumption that parents are well-positioned to detect early signs of school refusal. The main objective is to investigate whether a systematization of parents’ experiences can contribute to creating a better knowledge base, both to establish a common framework of understanding and to initiate joint measures early.

3. Materials and Methods

This analysis is based on a survey that was carried out among members of a Facebook group for parents whose children struggle with school refusal. The survey was carried out in two stages. In the first round, the respondents were asked the following open-ended question about the children’s early signs of school refusal: “Were there any signals in your child/youth that worried you before school refusal was a fact? If so, what signals were these?” The respondents were thus encouraged to write relevant signals in free text. The open-ended responses were categorized, and the categories were used to design a standardized battery of questions. In the second round, the battery of questions was used in a more extensive survey of the members of the Facebook group.

3.1. First Round of the Survey

In the first round, a link to the survey was posted on the Facebook page in question. A total of 250 respondents participated, with 196 of them answering the open-ended question about early signs and signals. The responses were generally comprehensive, but the number of signs and signals provided by each respondent varied.

The coding of the answers was carried out in three phases. In the first phase, the responses were read multiple times to gain an overview. The purpose was to determine the level of detail in the answers, assess their comprehensiveness, and identify some recurring main categories. The first phase thus laid a foundation for a more systematic coding in the next phase.

In the second phase, each response was reviewed and coded into specific categories. For each response, it was evaluated whether existing codes could be applied or whether new categories had to be created. Each response could be assigned multiple codes depending on its content. For example, a response consisting of multiple signs and signals could be coded with both an existing category and a new category.

The second phase resulted in identifying 32 different signs and signals. On average, each respondent was assigned 2.3 codes, with the number of codes per respondent ranging from 0 to 6. Three respondents were assigned 6 codes. Twelve respondents were not assigned any codes, primarily because they either reported not observing any signs or mentioned environmental factors (e.g., noise in the classroom, a death in the family) rather than characteristics of the child. Consequently, the sample of respondents was reduced to 184 respondents.

In the third phase, the categories were reviewed to consider restructuring. This assessment was based on the frequency of responses in each category, the homogeneity or heterogeneity of the responses within each category, overlap between categories, and whether the categories accurately reflected the responses. As a result, some categories having few responses were either removed or merged with similar categories. For example, categories such as “creepy fantasies”, “preoccupied with other people’s opinions”, and “shyness” were eliminated, and the responses were reclassified into more general categories.

After the final restructuring, we were left with 30 categories of signs and signals. Based on these 30 categories, three sets of question batteries were designed, each with the main question: “Were there any signals in your child that worried you before school refusal became a fact?” For each category, respondents were given three answer options: “no”, “yes, to some extent”, and “yes, to a large extent”.

3.2. Second Round of the Survey

In the second round of the survey, the question batteries about signs and signals were included in a new survey. This was carried out in the same way as the first, i.e., via a link on the Facebook group’s page. The survey was left open for a longer period this time (2 weeks) to allow more people the opportunity to respond. There were 850 who responded to the survey, but only 658 fully answered all the questions in the survey. This means that everyone with incomplete answers was removed from the analysis. The survey responses were also reviewed to identify implausible or unserious responses. Specifically, this involved respondents who had marked an improbable number of signs and signals. A total of 50 respondents were removed because they reported more than 20 signs and signals. This adjustment reduced the sample size to 608 respondents.

In reviewing the data, we found that the number of signs and signals the respondents reported varied somewhat depending on the duration of the school refusal. The longer the time since school refusal began, the more categories they answered (r = 0.12). A possible explanation may be that respondents having longer experiences of school refusal remembered the initial signs less clearly, at the same time as their answers were more influenced by experiences with manifested school refusal. Therefore, respondents who indicated that school refusal started more than 5 years ago (at the time of the survey) were removed. This adjustment resulted in the removal of 99 respondents, leaving us with a final net sample of 509 respondents.

Despite removing respondents having the longest duration since school refusal began, there was still a relationship between the duration of school refusal and the number of signs and signals reported. This likely indicates that the occurrence of some categories was influenced by the duration of school refusal. In other words, it suggests that the registration of these codes may be affected by the characteristics of manifested school refusal. The categories that had a significant correlation were the following:

-

Went home from school/ran away from school (during school hours).

-

Difficulties in handling the transition between kindergarten, primary school and/or secondary school.

-

Protested going to school.

-

Reluctance to/dread of going to school.

-

The child was afraid of the teacher.

3.3. Data Analysis

The data analysis was conducted in two main stages. First, a descriptive analysis was performed to examine the frequency and distribution of the 30 identified signs and signals across the sample. This included comparisons across subgroups such as gender and age at the onset of school refusal. Second, an exploratory factor analysis was carried out to identify potential underlying dimensions among the reported signs and signals. Principal axis factoring with oblique rotation was used, based on the assumption that different types of early indicators may co-occur. The final factor structure was used to interpret the underlying dimensions of early school refusal and was further analyzed using regression to examine associations with gender and age of onset. Further details regarding the factor analysis procedure and results are presented in Section 5.

In addition to the exploratory factor analysis, a confirmatory factor analysis (CFA) was conducted using the same dataset to assess the fit of the proposed factor structure. Given the inductive nature of the study and the lack of an independent sample, the CFA results are reported briefly in the Appendix B.

4. Descriptive Analysis of Signs and Signals

4.1. Sample Caracteristics

The respondents in the sample were parents answering questions about their children. Consequently, the survey primarily focused on the characteristics of the children.

Table 1. Characteristics of the children included in the sample.

	Mean	Std. Dev.	Min	Max
Gender (boy)	0.57	-	0	1
Age at the time of the survey	12.8	2.58	6.5	19
Age when school refusal started	10.1	2.42	0	16
Number of years since school refusal started	2.4	1.29	<1	5

presents some key characteristics of the children in the net sample.

After reducing the sample to 509 respondents, the sample consisted of 57 percent parents of a boy and 43 percent parents of a girl. This was roughly the same as in the gross sample. The average age at which the children began refusing school was 10.1 years, while their average age at the time of the survey was 12.8 years. At the time when the parents completed the survey, they had an average of 2.5 years’ experience with school refusal. However, this experience ranged from less than a year to up to 5 years. As mentioned above, respondents having more than 5 years of experience were excluded from the analysis.

In the survey, we asked parents about the age of their child when school refusal began, without further defining the term “began”. Consequently, the question can be interpreted in at least two ways. One interpretation is the point in time when the child stopped attending school. Although this is not always clear-cut, it is at least easily observable for the parents who responded to the survey. Another interpretation could be the time when the child started resisting going to school, even if they were still physically attending. Several parents noted in the survey that their child continued to attend school despite showing a strong reluctance to attend. Examples of such comments include the following:

-

“The child has always gone to school, but reluctantly.”

-

“[The child] did not want to enter the schoolyard in the morning, and we almost had to force her in. She often came last into the classroom during the last two years of elementary school.”

-

“I answered [one of the survey questions] that the school refusal lasted 2–3 years, but I included the period of reluctance to go to school, even though she was still attending. Actual school refusal was only about 1 year.”

It is uncertain how the parents who responded to the survey interpreted the question about when school refusal began. Since the survey is based on parents’ observations, and it is easier to observe absence than reluctance or resistance to going to school, it is likely that some interpreted the question as referring to the point when the absence started. However, we cannot rule out that some have based their responses on the other interpretation.

Figure 1 shows that there was considerable variation in the timing of when the school refusal began. It was common for school refusal to start between the ages of 9 and 13, with the most frequent age being 10 years. The incidence gradually decreases after age 10, except for a relative increase at age 13. Additionally, a notable proportion of the children began refusing school as early as first grade (6 years old). This early onset at age 6 and the later peak at age 13 are likely related to transitional periods, such as starting school and moving from elementary to middle school. Thus, we observe a normal distribution with the highest frequency at age 10, alongside a transitional effect that increases incidence particularly at ages 6 and 12–13.

In the survey, we also asked the parents about their educational background to identify any potential biases in the sample of the Facebook group members. When compared to the national education level of the population aged 25–40, the results reveal a clear overrepresentation of highly educated individuals in the sample. While 51 percent of the national population has higher education, the corresponding proportion in the sample was 69 percent. On the other hand, 5 percent of the sample had only primary education, compared to 19 percent in the national population. The proportion having only secondary education was also somewhat lower in the sample (26%) than in the population as a whole (31%).

According to a literature review by Leduc et al. (2024), previous research presents contradictory findings regarding parents’ education. Some studies indicate that parents of children with school refusal have higher education levels than parents in general, while others find no significant differences. Therefore, there are no definitive research results with which to compare the findings of this survey. Nevertheless, it is unlikely that the incidence of school refusal is as unevenly distributed by parents’ education as our survey results suggest. It is more plausible that the recruitment process for this survey resulted in a biased sample based on the parents’ educational background.

The question, however, is whether this bias affected how the parents responded to the survey. A test shows that those having education up to high school level reported more signs and signals than parents having higher education. However, the group having only primary education did not differ from those having high school education, and those having short higher education did not differ from those having longer higher education. In other words, the least representative groups (those having the lowest and highest education) did not respond significantly differently from the more representative groups (high school and short higher education). Based on these results and given that we do not know the actual educational level of the population, we have chosen not to weigh the sample. This suggests that we should still consider the possibility that the bias in the sample may have affected the results.

4.2. Occurrence of Signs and Signals

Table 2. The occurrence of signs and signals of school refusal.

		No	Yes, to Some Extent	Yes, to a Large Extent	Score
1	Aversion to/dread of going to school	9%	29%	61%	76%
2	Protested against going to school	11%	34%	55%	72%
3	Reluctance to do homework	16%	26%	58%	71%
4	Had little sense of mastery	17%	35%	44%	66%
5	Sleep difficulties	17%	36%	47%	65%
6	Stomach pain, nausea, vomiting	17%	38%	46%	65%
7	Was sad and bored at home	19%	40%	41%	63%
8	Expressed general dissatisfaction with the school	19%	40%	41%	61%
9	Difficult to get up in the morning	24%	33%	43%	60%
10	Concentration difficulties	22%	38%	40%	59%
11	Refused to participate in special subjects (Norwegian, mathematics, physical education)	25%	33%	42%	59%
12	Expressed that it was unsafe at school	28%	37%	35%	54%
13	The child was sad when he/she came home from school	25%	44%	30%	52%
14	Headache	32%	33%	35%	52%
15	Talked negatively about the school/teachers	30%	36%	34%	52%
16	Got performance anxiety	28%	41%	31%	51%
17	Was angry and acting out at home (toward parents/siblings)	32%	36%	32%	50%
18	Difficulty coping with the transition between kindergarten, primary school and/or secondary school	37%	33%	30%	46%
19	Social withdrawal from other children	32%	44%	25%	46%
20	Got increased requirements for own school performance	39%	38%	23%	42%
21	Difficulty keeping up academically	42%	35%	24%	41%
22	Lost friends	42%	40%	19%	38%
23	Became less concerned with school performance	48%	31%	21%	36%
24	Afraid of other students	47%	37%	16%	34%
25	Lots of inexplicable crying	52%	32%	16%	32%
26	The child expressed being bullied/frozen out	53%	32%	15%	31%
27	The child was afraid of the teacher	58%	29%	13%	28%
28	Went home from school/ran away from school (during school hours)	66%	27%	7%	21%
29	The child argued with the teacher	74%	18%	9%	17%
30	Skipped school (without your parents knowing)	91%	8%	1%	5%

shows the distribution of responses to the 30 signs and signals included in the second round of the survey. In the table, they are ranked according to how frequently they occur. The first three columns show the percentage distribution, while the fourth column provides an overall percentage measure based on the first three columns. The overall indicator in the fourth column is calculated by summing the percentage of those who answered “yes, to a large extent” with half of the percentage of those who answered “yes, to some extent.” This provides an overall measure ranging from 0 to 100 percent. The overall indicator is used in the subsequent tables to compare the occurrence of signs and signals by gender and the timing of the onset of school refusal.

The table shows that some signs and signals occurred relatively frequently. The most common was the child’s expression of anxiety or reluctance to attend school. Other frequently occurring signs included protests against doing homework, sleep difficulties, and psychosomatic symptoms such as stomach pain or headaches. A common characteristic of the most frequently reported signs is that they are directly observable by parents.

Some of these signs may seem obvious, such as protesting against going to school. However, there are at least two reasons for including these and other seemingly obvious signs and signals. First, the signs and signals listed in the table originate from responses to open-ended questions posed to the parents in the first round of the survey. Second, several comments in the survey indicate that the children did not always express their reluctance or fear of school directly. Instead, they often tried to conceal their fear from others. In other words, the signs and signals that seem obvious are not necessarily obvious in all cases.

At the bottom of the table, we find that parents reported to a lesser extent truancy, running away from school, arguing with the teacher, or fear of the teacher. These signs often require parents to be informed by the child or others. For instance, parents would not know if their child argued with a teacher unless the child or school staff informed them. Therefore, the low occurrence of some signs and signals may be due to the fact that they were not directly observable by parents.

The table also includes signs indicating that the child had difficulty coping with the transition from kindergarten to elementary school or between elementary school and middle school. However, transitions themselves are not signs or signals of school refusal. Rather, transitions should be understood as contextual risk factors that increase the likelihood of school refusal, with the potential development of school refusal expressed through various types of signs and signals. Since the list of signs and signals is based on parents’ responses to the open-ended questions in the first round of the survey, we have nevertheless chosen to include this category in the table.

4.3. Gender Differences in Signs and Signals

Table 3. Signs and signals where boys and girls are respectively overrepresented (calculated score).

		Girls	Boys
Signs and signals where boys are overrepresented	(3) Reluctance to do homework	64%	76%
	(28) Went home from school/ran away from school (during school hours)	16%	24%
	(29) The child argued with the teacher	13%	21%
	(2) Protested against going to school	67%	74%
Signs and signals where girls are overrepresented	(25) Lots of inexplicable crying	40%	27%
	(19) Social withdrawal from other children	52%	42%
	(26) The child expressed feelings of being bullied/frozen out	37%	26%
	(14) Headache	58%	48%
	(16) Got performance anxiety	57%	47%
	(22) Lost friends	44%	34%
	(24) Afraid of other students	39%	31%
	(6) Stomachache, nausea, vomiting	69%	61%
	(5) Sleep difficulties	69%	62%
	(9) Difficult to get up in the morning	64%	57%
	(7) Was sad and bored at home	66%	61%

uses the overall indicator to highlight the signs and signals having the greatest differences between boys and girls. The results show that boys are typically overrepresented in signs and signals related to externalizing behavior. In contrast, girls are more often overrepresented in signs and signals related to somatic complaints, sleep difficulties, performance anxiety, crying, and social relationships, which generally point towards internalizing behavior. However, it is important to note that despite these trends, there was still a relatively high incidence of internalizing behavior among boys and externalizing behavior among girls.

4.4. Variation in Signs and Signals by Age of Onset

Table 4. Signs and signals that occur differently depending on the time when school refusal started.

		6 Years	7–9 Years	10–12 Years	>12 Years
Signs and signals having decreasing occurrence	(2) Protested against going to school	88%	78%	70%	55%
	(1) Reluctance to/dread of going to school	85%	79%	76%	65%
	(7) Expressed general dissatisfaction with the school	70%	65%	60%	51%
	(11) Expressed that it was unsafe at school	65%	60%	52%	37%
	(15) Was angry and acting out at home (toward parents/siblings)	60%	64%	48%	38%
	(12) The child was sad when he/she came home from school	59%	56%	53%	41%
	(24) Lots of inexplicable crying	50%	41%	30%	26%
	(23) Afraid of other students	41%	34%	36%	25%
	(26) The child was afraid of the teacher	30%	29%	26%	17%
Signs and signals having increasing occurrence	(8) Difficult to get up in the morning	51%	58%	57%	72%
	(16) Got performance anxiety	49%	42%	49%	62%
	(20) Got increased requirements for own school performance	36%	37%	40%	55%
	(19) Difficulty keeping up academically	34%	42%	40%	47%
Curve-shaped occurrence	(3) Reluctance to do homework	66%	80%	71%	63%
	(22) Became less concerned with school performance	23%	34%	41%	33%
	(27) Went home from school/ran away from school (during school hours)	15%	25%	23%	15%

illustrates that the timing of school refusal onset was associated with different patterns of signs and signals. We can identify three developmental trends: signs and signals having increasing occurrence, those having decreasing occurrence, and those having a curve-shaped occurrence. The latter refers to an occurrence that first increased when school refusal started after the age of 6, then decreased again among those who first refused when they were 13 years or older (forming an A-shaped curve).

Signs and signals having decreasing occurrence are characterized by expressions of reluctance to go to school, fear of school, or negative emotions related to school. Additionally, this category includes emotional expressions such as unexplained crying and anger. Although these signs and signals were most prevalent when school refusal began at an early age, their occurrence remains high even among the oldest children.

The signs and signals that increase in occurrence with age are more closely related to academic challenges. In other words, signs of emerging school refusal among older children tended to manifest as academic struggles. These included a low sense of achievement, increased pressure regarding their own school performance, and difficulties keeping up with schoolwork. Other signs that also increased in occurrence with age are difficulties getting up in the morning and loss of friends.

Finally, we identified three signs/signals that showed a more curvilinear pattern. This means that the occurrence was relatively high among the middle age group and lower when school refusal began either early or late. Two of the three signs included in the table above are related to school matters, namely reluctance to do homework and reduced concern about school performance. Additionally, leaving school or running away during school hours is also included, although these latter two signs were less common.

In general, we can say that various signs and signals were expressed to different degrees depending on when the school refusal started. However, all the signs and signals appeared across all age levels. In other words, it seems that all the signs and signals occurred regardless of the timing of school refusal onset, but their prevalence varied depending on when it began.

5. Factor Analysis of Signs and Signals

A factor analysis was conducted to investigate the presence of latent underlying dimensions in the data. This method is particularly useful for reducing a large number of variables to a smaller set of underlying dimensions (factors). Conducting a factor analysis involves three key considerations. First, the choice of method. Second, determining which variables are appropriate to include in the analysis. Third, deciding how many factors should be extracted. There are no objective criteria for the selection of variables or factors, and this must be assessed based on some subjective judgments.

In the analysis, principal axis factoring was chosen. According to Fabrigar et al. (1999), this method is suitable when variables are not normally distributed and when dealing with categorical variables. In this case, the variables have only three values. Additionally, direct oblique rotation was applied, allowing the underlying factors to correlate. This approach is based on the reasonable assumption that emerging school refusal can be expressed in different ways within the same child.

In the factor analysis, the 30 variables described above were used as the starting point. Since factor analysis is based on the correlation between variables, those having low correlations were excluded from the analysis. According to Eaton et al. (2019), variables included in the analysis should have a communality value between 0.25 and 0.4. For this analysis, a cut-off threshold of 0.3 was chosen, meaning that variables explaining less than 30 percent of the variance were excluded. Additionally, only variables with a factor loading of at least 0.3 on one of the factors were included.

Due to low communality and low factor loadings, a total of six variables were removed from the initial factor analysis. The signs and signals that were excluded were the child going home from school (28), truancy (30), difficulties with transitions (18), difficulty getting up in the morning (9), sleep difficulties (5), and the child becoming less concerned with school performance (23). Although these signs and signals do not fit as well into the factor structure, it is important to emphasize that they remain relevant as a basis for assessing emerging school refusal. The KMO (Kaiser–Meyer–Olkin) test for the 24 remaining variables is 0.86, indicating that the factor analysis is well suited to elucidate patterns in the data.

There are three methods for determining the number of factors to include in a factor analysis. The first is to select factors having an eigenvalue greater than 1. The second is to use visual inspection of scree plots. The third is to conduct a parallel analysis. Both the eigenvalue criterion and visual inspection of the scree plot indicated a seven-factor solution.

Table 5. Pattern matrix (principal axis factoring, oblimin with Kaiser normalization, rotation converged in 15 iterations).

		1	2	3	4	5	6	7
22	Lost friends	0.73
26	The child expressed being bullied/frozen out	0.65
24	Afraid of other students	0.57
19	Social withdrawal from other children	0.55
13	The child was sad when he/she came home from school	0.30
21	Difficulty keeping up academically		0.73
10	Concentration difficulties		0.62
3	Reluctance to do homework		0.54
4	Had little sense of mastery		0.43		0.35
11	Refused to participate in special subjects		0.33
27	The child was afraid of the teacher			−0.71
29	The child argued with the teacher			−0.56
16	Got performance anxiety				0.86
20	Got increased requirements for own school performance				0.61
14	Headache					0.88
6	Stomach pain, nausea, vomiting					0.74
1	Reluctance to/dread of going to school						−0.77
8	Expressed general dissatisfaction with the school						−0.61
2	Protested against going to school						−0.58
15	Talked negatively about the school/teachers			−0.34			−0.50
12	Expressed that it was unsafe at school	0.32					−0.47
7	Was sad and bored at home							0.77
25	Lots of unexplained crying							0.55
17	Was angry and acting out at home							0.51

presents the seven factors extracted from the analysis, including factor loadings of 0.3 or higher. The factor analysis can be evaluated based on two criteria. The first criterion is that the variables in the analysis should not load significantly on multiple factors. The table shows three variables having factor scores above 0.3 on two factors. These variables are retained to satisfy the second criterion, which states that a factor should consist of at least three variables. Although factor 5 does not meet this requirement, it has been retained because it pertains to psychosomatic complaints, which are frequently emphasized in the literature (Ingul et al., 2019).

The seven factors can be labeled as follows:

• Expressed or observed social difficulties;
• Academic difficulties;
• Aversion to adults at school (teachers);
• Performance anxiety;
• Psychosomatic complaints;
• Reluctance towards school or the school environment;
• Emotional expressions in the form of crying, anger, sadness, etc.

A detailed examination using regression analysis was conducted to determine whether the factors varied by gender and by the age at which school refusal began. In the descriptions of the factors below, the results from the regression analysis are also presented. The table containing the regression results can be found in the Appendix C.

(1)

Expressed or observed social difficulties: This factor refers to the child showing or articulating social difficulties with peers. Variables having high loadings on this factor include “the child losing friends” (22), “expressing being bullied or frozen out” (26), “being afraid of other children”, and “withdrawing from other children” (19). Additionally, the variables “sad and upset after school” (13) and “feeling unsafe at school” (12) are included, although with lower factor scores. The variable “feeling unsafe at school” also scores highly on the fifth factor (reluctance towards school or the school environment). Signs and signals of social difficulties were more commonly observed among girls than boys.

(2)

Academic difficulties: The second dimension is labeled academic difficulties. These signs and signals indicate that students have “difficulty keeping up academically” (21), experience “concentration difficulties” (10), or show “reluctance to do homework” (3). Additionally, “low sense of mastery” (4) and “refusal to participate in special subjects” (11) are included, although these two have relatively low factor loadings. Moreover, “low sense of mastery” also correlates with the factor performance anxiety. Boys scored significantly higher on this factor compared to girls.

(3)

Aversion to adults at school (teachers): The third factor is aversion to adults at school. Two variables have relatively high loadings on this dimension: “the child was afraid of the teacher” (27) and “the child argued with the teacher” (29). In addition, the variable “talked negatively about the teacher/school” (15) has a factor loading of 0.34. This variable also scores highly on the sixth factor (reluctance towards school or the school environment). The occurrence of this factor decreased significantly with the age at which school refusal began, indicating that aversion to adults at school was more common among the children who experienced school refusal at an early age.

(4)

Performance anxiety: The fourth factor is named performance anxiety. The two variables having the highest factor loading on this dimension are that the children “experience performance anxiety” (16) and that the child has “increased demands on their own school performance” (20). Additionally, the variable “had little sense of mastery” (4) has a lower factor score (0.34) on this factor. The same variable also loads on the factor “academic difficulties”, suggesting a degree of conceptual overlap. Theoretically, the two factors represent distinct types of early signs. Academic difficulties refer to observable challenges in mastering school subjects—such as concentration problems, poor academic performance, and reluctance to engage in schoolwork. In studies that link school refusal to academic challenges, researchers usually refer to these challenges as academic difficulties (Ingul et al., 2019; Ulaş & Seçer, 2024). Performance anxiety, by contrast, refers to internal emotional responses related to expectations of achievement and fear of failure. This includes worry, nervousness, and self-imposed pressure, often associated with test anxiety and perfectionism (Guerra & Bright, 2021). The results show that performance anxiety is more prevalent in girls than in boys. Additionally, it is more common for performance anxiety to occur when school refusal begins at a later age.

(5)

Psychosomatic complaints: Psychosomatic complaints are often highlighted in the literature as a sign of emerging school refusal. Therefore, it is not surprising that one of the factors consists of psychosomatic complaints. In the table, two variables have particularly high loading in this dimension: headaches (5) and stomach pain, nausea, and vomiting (6). These symptoms are most common in girls and are more prevalent when school refusal begins at an early age.

(6)

Reluctance towards school and the school environment: The sixth dimension is called reluctance towards school and the school environment. The analysis shows that there are five variables having high factor scores on this dimension: “reluctance to/dread of going to school” (1), “general dissatisfaction with the school” (8), “protesting against going to school” (2), “talking negatively about school/teachers” (15), and the child “expressing that it is unsafe at school” (12). Common to all these variables is that the child, in various ways, expresses reluctance towards school or the school environment. This reluctance was most prevalent among the youngest children and decreased with later onset of school refusal. However, it was equally common among both boys and girls.

(7)

Emotional Expressions: The seventh factor is called emotional expressions. This refers to expressions such as anger, crying, and sadness. The variables that score highly on this factor include the child “expressing sadness and bored at home” (6), “lots of unexplained crying” (25), and “angry and acting out at home” (17). These are signs and signals that primarily manifest at home. These emotional expressions are more common in girls than in boys. Additionally, they tend to decrease with later onset of school refusal.

In the factor analysis, a rotation method was used that allows the factors to correlate with each other. This was chosen based on the assumption that children and adolescents can exhibit multiple types of signs and signals simultaneously, not just one type. A correlation analysis also shows that there is a relatively strong positive relationship between some of the factors. In addition to identifying some underlying types of signs and signals, this means that several of these can be expressed by the same child. There is a particularly strong correlation between the factors “social difficulties” and “emotional expressions” (Pearson’s r = 0.45); between the factors “aversion to adults at school” and “reluctance towards school and the school environment” (Pearson’s r = 0.40); and between the factors “psychosomatic complaints” and “emotional expressions” (Pearson’s r = 0.39). In other words, these types of signs and signals often occur together. There are also factors that correlate negatively with each other, meaning it is less common for them to occur simultaneously. This is particularly true between the factor “reluctance towards school and the school environment” on one side, and the three factors “social difficulties” (Pearson’s r = −0.44), “academic difficulties” (Pearson’s r = −0.38), and “emotional expressions” (Pearson’s r = −0.45) on the other. While the results of the factor analysis show that emerging school refusal can appear in different forms (seven factors), the correlation analysis indicates that these factors can occur in varying combinations.

6. Discussion

This article builds on the widely acknowledged consensus in the literature that early intervention is essential to prevent school refusal from becoming manifest and more difficult to address (Okuyama et al., 1999; Kearney & Graczyk, 2014; Ingul et al., 2019). Once school refusal has progressed to prolonged absence, it often becomes significantly more difficult to resolve. Another central premise is that such cases frequently evolve into complex and prolonged situations, characterized by disagreement and conflict between parents and schools (Malcolm et al., 2003; Havik et al., 2015; Amundsen & Møller, 2020). These tensions can delay or even obstruct the provision of necessary support. Establishing a shared understanding between home and school, particularly regarding the early signs and signals expressed by the child, is therefore a prerequisite for effective early intervention. Achieving a common understanding early in the process can help reduce conflict, foster constructive collaboration, and lay the foundation for more targeted and effective support for the child.

6.1. Emergent and Manifested School Refusal

To enable effective early intervention, it is essential to identify school refusal in its early stages—before it becomes entrenched. In this context, it is useful to distinguish between emerging and manifested school refusal. Emerging school refusal refers to a phase in which the child is still attending school but shows reluctance or resistance, often through subtle or ambiguous signs and signals. These may include emotional expressions, somatic complaints, or behavioral changes that are not yet reflected in actual absence. Manifested school refusal, by contrast, involves partial or complete absence from school, typically driven by anxiety, fear, or other emotional distress. While these two phases represent different points along a continuum, the transition between them is often gradual. Recognizing and responding to signs of emerging school refusal is therefore critical—not only to prevent escalation but also to establish a shared understanding between home and school at a point when early intervention is still feasible and potentially more effective.

6.2. Knowledge Gaps in Early Identification

Building on the distinction between emerging and manifested school refusal, a central question addressed in this article is whether we currently possess sufficient knowledge to enable early intervention during the emerging phase. What kinds of signs and signals should we be looking for to act before the situation escalates into prolonged absence? Even though early intervention has long been considered essential in addressing school refusal, recent studies continue to recommend placing more emphasis on early intervention (Boaler & Bond, 2023; Kearney et al., 2023).

While early intervention is widely recommended, there has been limited systematic research into the specific characteristics of emerging school refusal. Previous studies often relied on increasing absenteeism as a key indicator. Recent studies have also suggested using initial absence as the single most important warning signal (Kearney et al., 2023; Sälzer et al., 2024). There are at least two reasons why this is an inadequate indicator. First, according to Gallé-Tessonneau and Heyne (2020) absence is not unique to school refusal. Second, children may exhibit reluctance to attend school long before this is reflected in actual absence (Benoit et al., 2024). By the time absenteeism becomes apparent, the situation may already have progressed into a more entrenched, manifested form.

Although several screening instruments have been developed to assess school refusal, such as the SRAS and SCREEN, it seems that these tools are primarily designed to identify already manifested cases or to evaluate the effectiveness of interventions (Gallé-Tessonneau & Gana, 2019; Heyne et al., 2020). They tend to focus on general symptoms and behavioral patterns associated with manifest school refusal, rather than distinguishing early signs that may emerge before manifested absence occurs. This limits their utility for early identification and preventive efforts.

However, recent research has increasingly sought to develop more comprehensive frameworks for early identification. An example is Ingul et al. (2019), who compiled a literature-based overview of signs and risk factors, including absenteeism, anxiety and depression, somatic complaints, emotional dysregulation, and low self-efficacy. Another example is the study by Boaler et al. (2024), which describes the collaborative development of a locally adapted early identification tool for emotionally based school non-attendance (EBSNA). This tool facilitates structured conversations between schools and families, mapping a wide range of signals that may appear before absenteeism becomes manifest. These studies contribute to a more nuanced and experience-based understanding of early indicators of school refusal. In line with this, the present study adopts a context-sensitive approach and highlights the importance of parental experiences as a key source of insight. By systematically analyzing parents’ own observations, this study demonstrates how subtle behavioral and emotional changes often precede absenteeism. Parents are often the first to notice subtle changes in their child’s behavior, and their insights can enrich the existing knowledge base. In this way, the study provides practically relevant knowledge that can enhance early identification of school refusal.

6.3. Parental Insights as Early Signs and Signals

Building on the conclusion that parents’ experiences offer valuable insights, the next question is what kind of knowledge these contributions entail. Based on responses to an open-ended question, 30 distinct categories of early signs and signals were identified. These ranged from common and overt expressions—such as aversion to or protests against going to school, reluctance to do homework, low sense of mastery, sleep difficulties, and psychosomatic complaints—to less frequently reported signs like skipping school, leaving during school hours, or expressing fear of teachers.

To uncover patterns within this variation, a factor analysis was conducted, revealing seven underlying dimensions: social difficulties, academic difficulties, aversion to adults at school, performance anxiety, psychosomatic complaints, reluctance towards school and the school environment, and emotional expressions. These dimensions underscore the complexity and heterogeneity of early school refusal, illustrating that children frequently display multiple, overlapping signs. As Gallé-Tessonneau and Heyne (2020) point out, school refusal is a heterogeneous phenomenon that cannot be reduced to a single set of symptoms or risk factors.

Importantly, the study also reveals how these signs vary across gender and age. Boys are more likely to show academic difficulties, while girls more often display social and emotional challenges, including psychosomatic symptoms. Furthermore, the timing of school refusal onset influences the expression of signs: early onset is more often associated with emotional and relational difficulties, whereas later onset tends to involve performance-related anxiety. The findings also highlight that school transitions—such as starting school or moving between levels—represent critical risk periods.

Taken together, these results provide a more nuanced and differentiated understanding of emergent school refusal. They show that parents’ observations not only complement existing research but also deepen our knowledge of how early signs manifest across different contexts and developmental stages.

6.4. Limitations of Early Signs and Signals?

While this study identifies a broad range of signs and signals associated with emerging school refusal, an important question remains: are these indicators sufficient to detect school refusal at an early stage—without risking misidentification? The answer is complex. On the one hand, many of the signs reported by parents—such as somatic complaints, sadness, or reluctance to attend school—are not unique to school refusal. These expressions may also be linked to other challenges, including general anxiety, temporary stress, or developmental transitions (Karkhanis & Winsler, 2016). As such, the presence of these signs does not necessarily indicate an emerging school refusal.

On the other hand, several of the signs identified in this study overlap with indicators commonly used to assess neglect or inadequate caregiving (Norwegian Directorate of Health, 2018). This overlap increases the risk of misinterpretation, particularly if school staff or other professionals interpret the signs without considering the broader context. Previous research has shown that a significant number of school refusal cases are reported to child welfare services (Amundsen & Grøgaard, 2023), suggesting that school refusal is sometimes mistaken for neglect. This underscores the necessity of interpreting signs and signals within the broader context of the child, especially school-related factors such as bullying, academic pressure, or poor teacher–student relationships.

It is important to emphasize that individual signs should not be assessed in isolation. A more reliable interpretation requires that signs and signals be considered together, over time, and in light of known risk factors. Some signs, such as psychosomatic complaints or emotional withdrawal, may appear in both school refusal and neglect. However, each condition also has more distinct indicators: academic difficulties may point more clearly to school refusal, while physical injuries may indicate neglect. A comprehensive assessment should therefore consider both shared and distinct signs. This contextual approach is essential to avoid misjudgments and ensure that children receive appropriate support.

The findings of this study underscore the need for caution and contextual sensitivity when interpreting early signs of school refusal. These signs should not be viewed in isolation, but rather as potential indicators that require further exploration—through dialogue with the child, collaboration with parents, and careful assessment of the school environment. By applying such a contextual approach, the knowledge generated by this study may help prevent misdiagnosis, reduce unnecessary conflict between home and school, and support more accurate and constructive responses. As Boaler and Bond (2023) argue, it is important to collaborate with a range of stakeholders (families, professionals, and schools) to identify students’ needs as early as possible.

6.5. Empowering Parents, Informing Schools, and Fostering Collaboration

The signs and signals identified in this study are based on parents’ experiences and observations. As primary caregivers, parents are often well-positioned to detect early signs of school refusal. While most tools for early identification are designed for use in schools, the findings from this study suggest that parental knowledge can be equally valuable. Gaining access to insights from other parents may enhance their capacity to identify emerging school refusal and intervene at an earlier stage. Equipping parents with knowledge about early signs can empower them to act sooner and seek support before the situation escalates.

The results can also serve as a valuable knowledge base for teachers, school counsellors, and other professionals involved in supporting children at risk. By understanding the nuanced and varied early signs identified by parents—ranging from psychosomatic complaints to emotional expressions and academic difficulties—educators and support staff may be better equipped to recognize emerging patterns and respond proactively.

Moreover, the findings underscore the importance of fostering collaborative practices between families and schools. Previous research has shown that school refusal cases are frequently marked by conflict and tension between home and school, particularly regarding the understanding of the situation, responsibility, and appropriate action. Such disagreements may hinder effective intervention. A shared knowledge base, grounded in both parental and professional perspectives, may help foster greater consensus and reduce the extent to which conflict obstructs early support for the child. In this way, the study may contribute not only to individual awareness but also to the development of a common framework for early support (Boaler & Bond, 2023; Boaler et al., 2024).

6.6. Future Research

While this study contributes to a more nuanced understanding of early signs of school refusal, further research is needed to deepen and broaden this knowledge base. Future studies should continue to explore how early signs and signals emerge and take shape during the initial phases of school refusal. A particularly important direction for future research is the investigation of early signs and signals among specific subgroups, especially children and adolescents having neurodevelopmental conditions such as ADHD and autism. These groups are overrepresented among those who experience school refusal, yet there is limited knowledge about how early indicators may differ in form, intensity, or visibility compared to the general population. For instance, emotional expressions or social withdrawal may be interpreted differently in children having autism, while academic stressors may trigger distinct responses in those having ADHD. Understanding these variations is essential for developing more tailored and effective identification tools and intervention strategies.

While parents are often well-positioned to observe early signs of school refusal, their perspectives may differ from those of teachers or the children themselves. Future research should aim to include multiple perspectives to validate and enrich the findings. Combining parental insights with teacher observations and self-reports from children could provide a more holistic and nuanced understanding of the early development of school refusal.

7. Limitations

This study has several limitations that should be taken into account when interpreting the findings. First, the data were collected through a survey distributed to members of a Facebook group for parents of children with school refusal. This recruitment method may have introduced selection bias, particularly with regard to parental education levels. While the analysis did not reveal significant differences in the reporting of signs and signals between educational subgroups, it is still possible that parents having higher education may be more attuned to subtle behavioral changes, more articulate in describing their observations, or more proactive in seeking explanation for their children’s difficulties. These factors could influence both the type and frequency of signs reported. This potential bias has implications for the generalizability of our findings. The patterns and dimensions of the early indicators identified in this study may not fully capture the experiences of families from different educational or socioeconomic backgrounds.

Second, the study relies on retrospective parental reporting, which introduces the risk of recall bias. Parents may unintentionally conflate early signs with characteristics that emerged later, during the period of manifested school refusal. To mitigate this, responses from parents whose children had experienced school refusal for more than 5 years were excluded from the analysis.

A third limitation is that the study is based solely on parental perspectives, without triangulation from other sources such as teachers, clinicians, or the children themselves. This may limit the generalizability of the findings, particularly in cases where parents and schools have divergent views. Despite these limitations, the study provides valuable insights into how parents perceive and interpret early signs of school refusal, and it highlights the importance of including parental perspectives in early identification efforts.

8. Conclusions

School refusal is a complex phenomenon having potentially long-term consequences for children and adolescents. While early intervention is widely emphasized in the literature, there is still a need to expand and refine our understanding of the early signs that may precede more entrenched school refusal. This study contributes to the existing body of research by offering a parent-informed perspective that complements previous findings and highlights the value of systematically capturing parental observations.

The results show that early signs of school refusal are often expressed through a variety of emotional, behavioral, and psychosomatic symptoms. Common examples include reluctance to attend school, sleep difficulties, somatic complaints, and emotional withdrawal. Through factor analysis, seven distinct dimensions of early indicators were identified, including social difficulties, academic challenges, performance anxiety, and aversion to the school environment. These findings highlight the varied ways in which school refusal may initially present itself, before progressing into a manifest phase.

Given the overlap between signs of school refusal and other conditions—such as anxiety, stress, or neglect—it is essential to interpret these indicators through a comprehensive and context-sensitive lens. Individual signs should not be assessed in isolation but rather considered in relation to other risk factors, the child’s developmental stage, and the broader school and home environment. This approach can help reduce the risk of misinterpretation and ensure that children receive appropriate and timely support.

The study also underscores the importance of fostering collaboration between parents and schools. Many cases of school refusal are characterized by misunderstanding and conflict, which can delay intervention. By incorporating parental perspectives into early identification efforts, schools may be better equipped to recognize emerging patterns and respond constructively.