4_2016_Schiefele & Schaffner - New Instrument Reading Motivation Questionnaire

Factorial and Construct Validity of a New Instrument for the Assessment of Reading Motivation Ulrich Schiefel S chiefele e

ABSTRACT

Ellen Schaffner

Reading motivation motivation has been defined consistently as a multidi multidimensional mensional construct. However, there is some disagreement regarding the number and nature of the dimensions of reading motivation. In particular, there is a lack of studies investigating investigating the dimensional structure and measurement invariance (e.g. (e. g.,, across gender) of reading motivation questionnaires. Bas ed on earlier instruments, qualitative findings referring to students’ reasons for read ing, and theoretical considerations, we developed the Reading Motivation Questionnaire (RMQ). A sample of 883 sixth-grade sixth- grade students was presented with 34 reading motivation items pertaining to seven dimensions. Five of these dimensions (i.e., curiosity, involvement, grades, competition, social recognition) referred to Wigfield and Guthrie’s Motivations for Reading Questionnaire, whereas two dimensions (i.e., emotional regulation, relief from boredom) were based on recent qualitative findings. The results from confirmatory factor analyses supported the hypothesized factor structure. In addition, three higher order factors were identified: intrinsic, extrinsic, and regulatory reading motivation. Moreover, strict measurement invari ance across female and male students and across groups with low versus high reading reading competence was established. Construct validity of the R MQ was supported by the contributions of the RMQ factors to reading amount, fluency, and comprehension and by the predicted gender differences in the dimensions of reading motivation.

University of Potsdam, Germany

T

Reading Research Quarterly , 0(0)

pp. 1–17 | doi:10.1002/rrq.134 © 2016 International Literacy Association.

he ability to comprehend texts represents an indispensable prerequisite of academic success (e.g., Chapman, Tunmer, & Prochnow, 2000). In particular, learning at school relies to a large extent on written materials. For that reason, it becomes important to study those factors that facilitate the development of reading competence. In addition to cognitive factors, such as working memory capacity, reasoning ability, or prior knowledge (e.g., Alloway & Gregory, 2013; Kendeou & van den Broek, 2007; Kintsch, 1998; Tighe & Schatschneider, 2014), aspects of reading motivation have been shown to be significantly associated with various indicators of reading comprehension comprehension (e.g., Guthrie & Wigfield, 1999; Park, 201 2011; 1; Unrau & Schlackman, 2006). Despite this positive evidence, previous assessments of reading motivation suffer from several shortcomings. First, there is only partial agreement on the number and nature of the primary factors of reading motivation. For example, the influential Motivations for Reading Questionnaire (MRQ; Wigfield & Guthrie, 1997) consists of 11 dimensions, whereas other approaches involve fewer dimensions of reading motivation (e.g., Greaney & Neuman, 1990; Sainsbury & Schagen, 2004; Schutte & Malouff, 2007;

1

Watkins & Coffey, 2004) or proposed one-dimensional measures (e.g., McKenna, Kear, & Ellsworth, 1995). Second, researchers using multidimensional questionnaires have created varying composite scores to capture intrinsic and extrinsic reading motivation without pro viding empirical evidence for secondary factors (e.g., Andreassen & Bråten, 2010; Guthrie, Wigfield, Metsala, & Cox, 1999; Wigfield & Guthrie, 1997). Third, although measures of reading motivation have been widely applied to female and male students, tests of measurement invariance across gender have not been conducted. This also applies to student groups with varying ages or developmental statuses, different levels of reading competence, and different socioeconomic and ethnic affiliations. Fourth, qualitative studies on reading motivation (e.g., Guthrie, Van Meter, McCann, & Wigfield, 1996; Nolen, 2007; Schiefele & Schaffner, 2013) have suggested dimensions of reading motivation that have not been included in previous questionnaires. Among those dimensions, the motivations to use reading as a means of coping with negative emotions and overcoming boredom appear to be of particular importance. For the purpose of overcoming the deficits of previous research, we intended to develop a new multidimensional questionnaire of reading motivation in a sample of sixth-grade students, analyze its structure of primary and secondary factors, provide tests of measurement invariance across gender and groups with low versus high reading competence, and examine its relations with various validation variables (e.g., reading fluency).

Dimensions of Reading Motivation Motivation to read can be conceptualized at the level of current or habitual reading motivation (cf. Pekrun, 1993; Schiefele, Schaffner, Möller, & Wigfield, 2012). Current motivation to read refers to the strength of a person’s intention to read a specific text in a given situation. For example, someone very eager to read a particular book at home shows strong current reading motivation. In contrast, an individual who is repeatedly motivated to read can be ascribed a certain amount of habitual reading motivation. Thus, habitual reading motivation denotes the relatively stable readiness of a person to initiate reading activities (Schiefele et al., 2012). Reading motivation inventories, such as Wig field and Guthrie’s (1997) MRQ and the present Reading Motivation Questionnaire (RMQ), usually assess habitual forms of motivation. Quantitative and qualitative studies suggest multiple dimensions of reading motivation (e.g., Guthrie

2 | Reading Research Quarterly , 0(0)

et al., 1996; Nolen, 2007; Watkins & Coffey, 2004; Wigfield & Guthrie, 1997). These dimensions correspond to different incentives of reading that may be subsumed under two higher order categories: intrinsic and extrinsic reading motivation (Schiefele et al., 2012; Wigfield & Guthrie, 1997). Intrinsic reading motivation refers to the willingness to read because reading is satisfying or rewarding in its own right. According to a distinction suggested by Schiefele (1999, 2009), there are two forms of intrinsic motivation to read. In the case of object-oriented intrinsic motivation (labeled “curiosity” in the MRQ and RMQ), reading is motivated by thematic interests. In the case of experience-oriented intrinsic motivation (labeled “involvement” in the MRQ and RMQ), reading is motivated by positive experiences, such as becoming absorbed by a story. In contrast, extrinsic reading motivation refers to reasons that are external to the activity of reading and the text content. The extrinsically motivated reader strives to attain particular outcomes of reading, such as improving one’s performance in school or being praised by one’s parents (Wigfield & Guthrie, 1997). A further distinction refers to academic (or schoolrelated) and recreational reading motivation (De Naeghel & Van Keer, 2013; De Naeghel, Van Keer, Vansteenkiste, & Rosseel, 2012). Academic reading is defined as reading at school and for homework, whereas recreational reading involves reading in one’s leisure time. The RMQ is directed at recreational reading motivation for two reasons: First, it has been repeatedly found that out-of-school reading amount contributes more strongly to the development of reading competence than school-related reading amount (cf. Schiefele et al., 2012). Thus, the motivation to read in one’s free time appears to be more important for reading comprehension than academic reading motivation. This assumption was confirmed by De Naeghel et al. (2012), who reported significant effects on reading comprehension only for recreational reading motivation, not academic. A second reason for focusing on recreational reading motivation refers to the overlap between motivation to learn and motivation to read at school and for homework. School-related reading may often coincide with school-related learning, and thus, measures of academic reading motivation probably reflect the more general motivations of students to learn or to achieve in school (e.g., achievement goal orientations; e.g., Elliot, 2005). Consequently, academic reading motivation might be confounded with more general motivational orientations pertaining to students’ school-related learning. A wide variety of dimensions of reading motivation have been suggested (cf. Schiefele et al., 2012). The most comprehensive approach is represented by Wigfield and Guthrie’s (1997) MRQ with 11 dimensions: curiosity,

involvement, grades, competition, recognition, compliance, challenge, importance, work avoidance, social reasons, and efficacy. In our view, not all of these dimensions denote forms of reading motivation in a more narrow sense (see also Schiefele et al., 2012). Instead, some constructs indicate antecedents and/or consequences of reading motivation (e.g., reading efficacy, importance of reading, preference for challenging reading materials). Moreover, Watkins and Coffey (2004) could not identify the MRQ scales of importance and challenge as separate factors by means of confirmatory factor analyses (CFAs). According to the expectancy–value approach (e.g., Heckhausen, 1991; Weiner, 1989; Wigfield & Eccles, 2000), a given motivation (e.g., striving to read a particular book) depends on expectancy beliefs (e.g., reading efficacy: “I will be able to understand the book”) and value beliefs (e.g., importance of reading: “To be a good reader is important to me”). In the present case, the motivation to read a particular book would be strong if the person believes that he or she is able to comprehend the book and that reading (or being a good reader) is personally important. Also in accordance with expectancy–value theory, the preference for challenging reading materials may be regarded as an example of choice behavior. There is a long tradition in research that views choice behavior (e.g., choice of difficulty levels of a given task) as an important outcome of motivation (Atkinson, 1964; Eccles & Wigfield, 2002; Zimmerman, 2000). For example, students high in achievement motivation have been found to choose challenging tasks (Atkinson, 1964). In contrast to the expectancy–value approach, it seems plausible to conceptualize reading efficacy and importance of reading not only as antecedents but also as consequences of reading motivation (pertaining to the relation between self-efficacy and motivation, see Schunk, 1991; Zimmerman, 2000). For example, intrinsic reading motivation likely enhances the amount of reading (e.g., Becker, McElvany, & Kortenbruck, 2010), which in turn facilitates students’ feelings of efficacy and importance of reading through enhanced comprehension ability (Schunk, 1991). Moreover, the preference for challenging reading tasks may be regarded not only as an outcome but also as an antecedent of reading motivation. Specifically, the choice of challenging reading tasks is likely to contribute to comprehension ability and feelings of efficacy, which in turn increase the motivation to read (Becker et al., 2010). Thus, taken together, it seems appropriate to assume reciprocal relations among reading motivation, reading efficacy, reading importance, and preference for challenging reading tasks. The status of work avoidance, social reasons, and compliance as dimensions of (recreational) reading

motivation seems also debatable. Inspection of the corresponding MRQ items shows that work avoidance refers to disliking of complex and difficult stories. As such, strong work avoidance may be regarded as a consequence of lacking motivation to read and/or of low levels of reading efficacy beliefs. In accordance with our view, Guthrie and Klauda (2016) conceptualized the avoidance of reading as a negative aspect of engagement that depends on motivation. For example, empirical evidence suggests that students’ intrinsic reading motivation is linked positively to reading engagement (e.g., amount of reading; Becker et al., 2010) and negatively to reading avoidance (Guthrie, Klauda, & Ho, 2013). The item contents of the social reasons for reading scale of the MRQ express both preference and frequency pertaining to literary practices within the family and the peer group (e.g., visiting a library, talking about books). Four out of seven items state particular activities that refer to reading in a social context (e.g., “I often read to my brother or my sister”). Because the reasons for these activities are not addressed by the items, it seems diff icult to infer social motivations (e.g., Wentzel, 2005). For example, a child might read to his or her younger siblings because it is an enjoyable experience or because the parents demand it. Three items of the social reasons scale, however, refer to preferences (e.g., “My friends and I like to trade things to read”) and indicate at least implicitly social reasons for reading. Because of the unclear nature of both the scale and the construct of social reasons for reading, we decided not to include this dimension in our questionnaire. Finally, compliance was defined by Wigfield and Guthrie (1997) as the motivation to read because it is required by school or a teacher. For example, students high in compliance strive to finish every reading assignment or to do their reading work exactly as the teacher wants it. Obviously, this dimension of reading motivation addresses only reading for school but not reading as part of leisure time activities. In that regard, it is important to note that the dimensions of grades and competition differ from compliance because they ask for students’ motivation to read in their free time to get better in school (or reading) and outperform their classmates. Thus, based on the preceding considerations, we propose the following dimensions as facets of reading motivation in a more narrow sense: curiosity (to learn more about topics of one’s interest), involvement (to get lost in a story or experience imaginative actions), grades (to improve one’s grades, particularly in reading), competition (to reach higher levels of school achievement, particularly in reading, than other students), and recognition (to get praise for good reading performance).

Factorial and Construct Validity of a New Instrument for the Assessment of Reading Motivation | 3

The Current Study The main goal of the present study was to develop and evaluate a multidimensional reading motivation questionnaire that should be applicable to a wide range of students starting at the upper elementary level (grade 4). In our view, most students at that level should have developed basic reading and comprehension skills (Foorman & Connor, 2011) and the cognitive ability to answer specific questions regarding their own motivation (Kuhn & Franklin, 2006). In the present study, we chose a sample of sixth-grade students to examine the questionnaire. The selection of seven dimensions of the RMQ was based on theoretical considerations and previous qualitative and quantitative studies (see the preceding section). The dimensions of curiosity, involvement, grades, competition, and social recognition correspond conceptually to scales of the MRQ (Wigfield & Guthrie, 1997; see also Schaffner & Schiefele, 2007; Schaffner, Schiefele, & Ulferts, 2013). As sources of items for these dimensions, we referred not only to the MRQ but also to other instruments and results from qualitative studies (Becker et al., 2010; Greaney & Neuman, 1990; Guthrie et al., 1996; Möller & Bonerad, 2007; Nolen, 2007; Schiefele & Schaffner, 2013; Wigfield & Guthrie, 1997). In addition to curiosity, involvement, grades, competition, and recognition, we included two new dimensions that have not been considered before as scales of reading motivation instruments and were derived from qualitative studies (Greaney & Neuman, 1990; Guthrie et al., 1996; Nolen, 2007; Schiefele & Schaffner, 2013). The findings of these studies suggest that emotional regulation (reading to cope with negative emotions) and relief from boredom (reading to overcome boredom and to fill in time because more preferred activities are not available) represent possibly relevant dimensions of reading motivation. In our own research (Schiefele & Schaffner, 2013), these dimensions were rather frequently mentioned by students when asking them for their motivation to read in their free time. The analysis of the RMQ focused on three issues: the structure of primary and secondary factors, measurement invariance across gender and competence groups, and construct validity. In particular, the first two issues have been neglected by previous research. Specifically, researchers applying the MRQ have used varying composites for intrinsic and extrinsic reading motivation (for an overview, see Schiefele et al., 2012) without providing evidence for second-order factors that would justify the combination of particular dimensions. Moreover, in light of t he repeatedly found gender differences in reading motivation (Chiu & McBrideChang, 2006; Logan & Johnston, 2009, 2010; Mullis, Martin, Kennedy, & Foy, 2007), it seems important to


establish measurement invariance across gender to be able to compare mean values or correlative associations between female and male students. In addition, there is also a need to demonstrate measurement invariance across groups with varying reading competence because poor and good readers differ in their reading motivations and the strengths of their relations among reading motivation, reading behavior, and reading comprehension (Lau & Chan, 2003; Logan, Medford, & Hughes, 2011). For the purpose of providing empirical support for the validity of the RMQ scales, we included measures of reading amount, reading fluency, and reading comprehension. Positive relations with the validation variables were expected for the intrinsic dimensions (curiosity and involvement) but not for the extrinsic dimensions (grades, competition, and recognition). In line with previous findings (e.g., Schaffner et al., 2013), we hypothesized nonsignificant or negative relations between the extrinsic dimensions and the validation variables. According to Schaffner et al., reading is largely a leisure time activity and as such is more strongly controlled by intrinsic incentives. Extrinsically motivated readers may tend to read only when they have to (e.g., to better achieve in school, to please their parents), and thus, the amount of leisure time reading and the development of reading skills will not be enhanced or even reduced (Becker et al., 2010). Moreover, it is likely that extrinsically motivated readers are more strongly concerned with future outcomes of their reading activities than with understanding a given text (Brophy, 2005; Hulleman, Durik, Schweigert, & Harackiewicz, 2008). Thus, extrinsic reading motivation may interfere with the processes necessary for indepth text comprehension, such as inference making and identifying main ideas (Pintrich & Schrauben, 1992; Wang & Guthrie, 2004). The status and role of emotional regulation and relief from boredom were more difficult to determine. Because of their ambiguous nature, we did not categorize these dimensions as either intrinsic or extrinsic. Although they possess an instrumental aspect (e.g., reading as a means to overcome negative emotions), they seem also closely associated with the positive experience involved in reading. Specifically, both dimensions presuppose that reading is intrinsically rewarding and, thus, facilitates positive emotional states or averts boredom. Because of their assumed close relation with intrinsic reading motivation, we anticipated positive contributions of emotional regulation and relief from boredom to the validation variables. As additional support for the construct validity of the RMQ, we analyzed gender differences in reading motivation. Based on previous research, it was expected that girls tend to exhibit higher intrinsic

reading motivation than do boys (Logan & Johnston, 2009, 2010; Wigfield & Guthrie, 1997). Whereas prior findings do not suggest gender effects on most of the dimensions of extrinsic reading motivation (McGeown, Goodwin, Henderson, & Wright, 2012; Schaffner et al., 2013; Wigfield & Guthrie, 1997), there is some evidence for boys being higher in competitionoriented reading motivation (Stutz, Schaffner, & Schiefele, 2016; Unrau & Schlackman, 2006; Wigfield & Guthrie, 1997; however, see Schaffner et al., 2013; Wang & Guthrie, 2004). Finally, qualitative findings (Schiefele & Schaffner, 2013) suggest that reading for the purpose of emotional regulation is more pronounced in girls than in boys.

Method Sample and Procedure The present sample comprised 883 sixth-grade students (442 girls, 441 boys) from 48 classes within 24 elementary schools1 in and around a large city in the northeast of Germany. The schools were selected to represent the population in the area, thereby including rural and urban areas of different socioeconomic backgrounds. The average age of the sample was 11.33 years (standard deviation [SD] = 0.57). Moreover, about 7% of the participating students reported that both of their parents were born in a foreign country. This percentage is below the average of students in Germany with both parents born in a foreign country (about 15%), but it is typical for student populations in the “new” German states t hat formerly belonged to East Germany (cf. Baumert & Schümer, 2001; Tarelli, Schwippert, & Stubbe, 2012). Finally, the average norm-referenced scores (T - values) for reading fluency (mean [ M ] = 46.9, SD = 8.32) and reading comprehension ( M = 48.1, SD = 7.66) in the present sample were only slightly below the population mean for sixth-grade students in Germany across all school tracks ( M = 50.0, SD = 10.0). For purposes of the present study, the total sample was randomly divided into two subsamples. Sample 1 involved 438 students (221 girls, 217 boys; M age = 11.34 years, SD = 0.59), and sample 2 entailed 445 students (221 girls, 224 boys; M age = 11.32 years, SD = 0.56). The two samples did not differ significantly with respect to the included study variables. The participants were tested during regular class time. They first answered the questionnaires on reading motivation and reading amount and then were presented with the tests on reading fluency and comprehension. The whole session took about 45 minutes. During the session, a teacher was present. The study was explained to the students as being concerned with their attitudes toward reading. They were assured that

the study was anonymous and that their results would not be conveyed to their teachers or parents.

Reading Motivation The RMQ comprised the following dimensions (see Table 1): • Curiosity: To learn more about topics of one’s interest • Involvement: To experience positive states of feeling, such as getting lost in a story or experiencing imaginative actions • Grades: To improve one’s grades or achievement in school • Competition: To outperform one’s classmates in school • Social recognition: To get praise for frequent reading • Emotional regulation: To cope with negative emotions, such as anger or sadness • Relief from boredom: To overcome boredom and to fill in time because other, more preferred activities are not available All items had to be answered on 4-point rating scales ranging from 1 (not at all true) to 4 (very true). In each case, a higher score indicates a higher level of motivation. In a short introduction, the participants were asked to rate possible reasons they have for reading in their leisure time: What are usually the reasons for you to read in your free time when you are not at school and not doing schoolwork? Reading activities in your free ti me may include reading of books, journals or magazines, and texts in the Internet. The following statements describe a number of possible reasons for reading in your free time. Please, indicate to what extent these reasons are true for you.

Validation Measures Reading Amount The scale to assess reading amount was based on previous instruments (e.g., Becker et al., 2010; Schaffner et al., 2013). In accordance with Becker et al., the reading amount scale addressed both the frequency and the length of reading. To assess reading frequency, the participants were asked how many books they read in their spare time during the last 12 months (1 = 0 books; 2 = 1–5 books; 3 = 6–10 books; 4 = 11–20 books; 5 = more than 20 books), how often they read in their spare time (1 = about once a month; 2 = about once a week; 3 = several times a week; 4 = daily; 5 = several times a day), and how often they read during school vacations (1 = never; 2 = rarely; 3 = sometimes;


TABLE 1 Reading Motivation Questionnaire: Items and Factor Loadings Resulting From Confirmatory Factor Analyses Factor loadings Sample 1 Itema (“I read…”)

Version 1

Sample 2

Version 2

Version 2

Version 3

Curiosity (factor 1) 1. because I like to think about particular topics.

0.63

0.62

0.65

0.67

2. because texts or books on particular topics are exciting.

0.80

0.80

0.81

—

3. because I can deal with personally important topics.

0.75

0.75

0.74

0.76

4. because I can learn more about things of interest to me.

0.79

0.79

0.77

0.78

5. because I like to be involved with particular topics

0.81

0.81

0.80

0.81

6. because sometimes I can forget everything around me.

0.82

—

—

—

7. because I like to identify with the main character of a good story.

0.84

0.86

0.85

0.85

8. because some stories stimulate my imagination.

0.77

0.78

0.79

0.80

9. because I can experience real adventures in my mind.

0.85

0.87

0.87

0.84

0.86

0.88

0.88

0.89

11. in order to get better grades in school.

0.79

0.79

0.82

0.83

12. because it helps me perform well in school.

0.82

0.82

0.86

0.86

13. because it is important for my achievement in some subjects.

0.82

0.81

0.84

0.84

14. because it helps me get better in school.

0.79

0.79

0.77

0.78

15. because it helps me perform better in school than my classmates.

0.77

0.78

0.75

0.75

16. because it is important to me to understand texts better than other students.

0.85

0.85

0.79

0.79

17. because it is important to me to be among the best students.

0.76

0.76

0.77

0.77

18. because it is important to me to know more than other students.

0.83

0.83

0.83

0.83

19. because other people say it is important.

0.67

0.70

0.72

0.72

20. because I know that my friends also read a lot.

0.64

0.65

0.63

0.63

21. because one gets praise for frequent reading.

0.78

0.80

0.71

0.71

22. because I like it when other people think that I am a diligent reader.

0.77

—

—

—

23. because my parents think that it is important that I read a lot.

0.56

0.59

0.69

0.69

24. because I want my parents to be proud of me.

0.69

0.71

0.78

0.78

25. because it helps me when I am sad.

0.87

0.87

0.88

0.88

26. because it helps me when I am angry.

0.87

0.87

0.88

0.88

27. in order to cheer me up when I am in a bad mood.

0.89

0.89

0.90

0.89

28. in order to distract myself after a quarrel.

0.85

0.84

0.84

0.84

29. when I am furious and need to calm down.

0.88

0.88

0.83

0.83

Involvement (factor 2)

10. because it allows me to imagine everything so well. Grades (factor 3)

Competition (factor 4)

Social recognition (factor 5)

Emotional regulation (factor 6)

(continued)


TABLE 1 Reading Motivation Questionnaire: Items and Factor Loadings Resulting From Confirmatory Factor Analyses (continued) Factor loadings Sample 1 Itema (“I read…”)

Version 1

Sample 2

Version 2

Version 2

Version 3

Relief from boredom (factor 7) 30. in order to avoid being bored.

0.72

0.69

0.66

0.66

31. if there is nothing better to do.

0.55

—

—

—

32. if there is nothing interesting on television.

0.74

—

—

—

33. in order to have something to do.

0.78

0.80

0.77

0.77

34. because it helps me pass the time.

0.84

0.86

0.84

0.84

Note. Version 1 = initial version, 34 items. Version 2 = adaptation of version 1, 30 items. Version 3 = adaptation of Version 2, 29 items. a These are translations of the German-language items. They aim to convey the content rather than serve as items for direct use in English.

4 = often; 5 = very often). Reading length was captured by asking the respondents how long they usually read every day (1 = not at all; 2 = less than 30 minutes; 3 = 30–60 minutes; 4 = between 1 and 2 hours; 5 = more than 2 hours) and how long they usually read a book without taking a break (1 = 5 minutes; 2 = 15 minutes; 3 = 30 minutes; 4 = 60 minutes; 5 = more than 60 minutes). The reading amount scale was highly consistent (Cronbach’s α = .86). In contrast to the Reading Activities Inventory (Guthrie, McGough, & Wigfield, 1994), our scale involved not only the frequency but also the length of reading. However, it did not differentiate between different themes. In addition, the present scale referred only to leisure time reading (reading for enjoyment) because this aspect of reading amount has been shown to be more strongly associated with reading motivation and reading comprehension than school-related reading (Cox & Guthrie, 2001; Wang & Guthrie, 2004). Significant associations with aspects of the home literacy environment (e.g., number of books at home) and intrinsic (but not extrinsic) reading motivation (Schaffner et al., 2013) support the validity of the present reading amount scale.

Reading Fluency and Comprehension

alternatives. After four minutes, the participants mark the point they were able to reach in the text. Reading fluency is captured by the number of words read, whereas reading comprehension is indicated by the number of correctly selected words. The scoring of the word selection task is a s follows: 2 points for each correctly selected word, 0 points for not choosing a word, and −1 point for selecting the wrong word (as a means of correction for guessing). Schneider et al. (2007) reported high test–retest reliabilities for both tests (reading fluency: .84; reading comprehension: .87). It should be noted that the two indicators of Schneider et al.’s (2007) test may be alternatively interpreted as different aspects of decoding skills (or reading fluency). Accordingly, the number of words read refers to the speed of decoding, whereas the number of correctly selected words indicates the accuracy of decoding (cf. Rasinski, Reutzel, Chard, & Linan-Thompson, 2011; Roberts, Christo, & Shefelbine, 2011). Individual scores for reading fluency and reading comprehension were determined by transforming the obtained raw scores (number of words read and number of correctly selected words) into standardized values (i.e., T - values; M = 50, SD = 10). In addition, individual T - values were linearly transformed by dividing the scores by 10 to reduce their absolute size. Large differences between the sizes of values of different variables are likely to cause computational problems when applying the statistical software M plus that we used to conduct the present analyses.

These variables were assessed by means of a test developed by Schneider, Schlagmüller, and Ennemoser (2007). This test is based on the maze technique that is commonly used to measure reading fluency (Tichá, Espin, & Wayman, 2009). Respondents are asked to read a lengthy and highly coherent narrative text (1,727 words) as quickly as possible within four minutes. In Missing Values addition, the text contains 23 blanks that have to be The initial sample comprised 892 students. Nine stufilled in by selecting the correct word out of three dents had large amounts of missing data (>30%) and


were thus excluded from the analyses. Apart from this small group of students, we had only few missing data for all variables (<1.0%). Moreover, only 17.4% of the participants showed missing values. In these cases, the percentage of missing values did not exceed 10%. For the purpose of descriptive analyses, missing values were replaced by expectation–maximization estimates (generated by NORM 2.03; Schafer, 1997, 1999). When conducting structural equation analyses with M plus, missing data were accounted for by maximum likelihood estimation (Asparouhov & Muthén, 2010; Graham, 2009).

Data Analysis To examine the new version of the RMQ, we conducted CFAs. Confirmatory instead of exploratory factor analysis was chosen because the assumed factor structure of the RMQ was theoretically and empirically well grounded (see also Baker & Wigfield, 1999). In the first step, CFA was applied to examine the R MQ in sample 1. The results of these analyses were then validated in sample 2. All factor models were estimated by means of M plus 7.3 (Muthén & Muthén, 1998–2014). In addition, individual items were defined as ordered categorical variables instead of continuous variables (cf. Carifio & Perla, 2007; Lubke & Muthén, 2004). Consequently, a weighted least squares estimator (WLSMV) was used for these analyses (default option in M plus). Moreover, we accounted for the multilevel structure of our data (the students in our sample were clustered in 48 classrooms) by using the TYPE = COMPLEX option in M plus.

Results Factor Structure of the RMQ CFA in Sample 1

TLI = 0.973, RMSEA = 0.038. All factor loadings were above 0.50 (see Table 1, Sample 1/Version 2).

CFA in Sample 2 The good fit of the CFA model of the RMQ in sample 1 was confirmed in sample 2, χ 2 = 619.31, df = 384, p < .001, CFI = 0.973, TLI = 0.970, RMSEA = 0.037. Inspection of modification indexes (>20) revealed one problematic item from the curiosity factor (see Table 1, item number 2). This item showed a substantial loading on the dimension of involvement and, therefore, was excluded. The fit indexes of the modified RMQ scale with 29 items were as follows: χ 2 = 552.67, df = 356, p < .001, CFI = 0.977, TLI = 0.973, RMSEA = 0.035. Again, all factor loadings were above 0.50 (see Table 1, Sample 2/Version 3).

Descriptive Statistics and Intercorrelations Table 2 represents mean values, standard deviations, reliabilities, and latent correlations for the final version of the RMQ based on the total sample. For all subscales, acceptable or high reliabilities (Cronbach’s α) were obtained. In accordance with previous research, substantial positive correlations among the dimensions of reading motivation were found (Schaffner et al., 2013; Wang & Guthrie, 2004; Wigfield & Guthrie, 1997). Also in line with previous research, gender was significantly related to reading amount. Accordingly, the girls in our sample tended to read more than boys (cf. Baker & Wigfield, 1999; Logan & Johnston, 2009, 2010). In contrast, the relations between gender and reading competence were either nonsignificant (reading fluency) or small (reading comprehension). These findings are not unusual. Although higher scores for girls in reading performance have often been found, these differences tend to be rather small (for an overview, see Logan & Johnston, 2009). In the case of the standardized test used in our study, results based on large samples at different grade levels confirm negligible or small gender differences (<1% explained variance) in reading fluency and comprehension (see Schneider et al., 2007). This is in line with findings pertaining to other German standardized tests of reading comprehension (cf. Lenhard, 2013).

The hypothesized seven-factor model of the RMQ in volving 34 items showed a good fit with the data, χ 2 = 879.97, df = 506, p < .001, comparative fit index [CFI] = 0.964, Tucker–Lewis Index [TLI] = 0.960, root mean square error of approximation [RMSEA] = 0.041. To detect inappropriate items with substantial double or multiple loadings, we consulted modification indexes. For that purpose, only modification indexes > 20 Measurement Invariance were considered to be critical because of the good fit of Because differences in reading motivation due to genthe overall model and to reduce the risk of sample- der and reading competence have been consistently dependent decisions. According to this criterion, four documented (e.g., Lau & Chan, 2003; Logan & Johnston, items (cf. Table 1) were removed from the model in a 2010; Logan et al., 2011; Mullis et al., 2007), it seemed stepwise procedure starting with the item that pro- necessary to provide evidence that our measure of readduced the highest modification index. As expected, the ing motivation applies equally to female and male as adapted CFA model with 30 items exhibited a high level well as low- and high-competence students. To deterof fit, χ 2 = 622.15, df = 384, p < .001, CFI = 0.976, mine groups of students with low and high levels of


TABLE 2 Latent Correlations and Descriptive Statistics for the Dimensions of Reading Motivation (Reading Motivation Questionnaire, Final Version) and the Validation Variables Variable 1. Gendera

1

2

3

4

5

6

7

8

9

10

11

—

Reading motivation 2. Curiosity

.16**

—

3. Involvement

.27***

.76***

—

4. Grades

−.02

.38***

.17***

—

5. Competition

−.14**

.37***

.21***

.76***

—

−.05

.37***

.24***

.79***

.80***

—

6. Social recognition 7. Emotional regulation

.28***

.59***

.63***

.18***

.30***

.32***

—

8. Relief from boredom

.05

.50***

.54***

.19***

.32***

.40***

.65***

—

.25***

.60***

.71***

.03

.13***

.10*

.55***

.50***

—

−.01

.20***

.26***

−.01

.06

−.06

.18***

.13***

.39***

—

.08*

.18***

.23***

−.08*

−.04

−.09*

.17***

.10*

.33***

.44***

Validation variables 9. Reading amount 10. Reading fluency 11. Reading comprehension

—

Mean

—

2.88

3.03

2.69

2.07

2.19

2.34

2.57

3.06

4.69

4.81

Standard deviation

—

0.76

0.89

0.83

0.80

0.74

0.98

0.82

0.93

0.83

0.77

Cronbach’s α

—

.79

.87

.85

.83

.77

.91

.76

.86

—

—

Note. All variables are latent except gender, reading fluency, and reading comprehension. a Scoring of gender: 1 = male; 2 = female. * p < .05, two-tailed. ** p < .01, two-tailed. *** p < .001, two-tailed.

reading competence, we combined their scores in reading fluency and reading comprehension and then created two groups by means of a median split (low competence: n = 417; high competence: n = 466). The confirmation of measurement invariance across groups represents an important prerequisite for conducting meaningful cross-group comparisons (e.g., between mean values and structural path coefficients; Vandenberg & Lance, 2000). Measurement invariance implies that the probability of attaining a specific observed score given a particular level of the underlying disposition is independent of group membership (Yoon & Millsap, 2007). The assessment of measurement invariance typically involves four hierarchical steps (e.g., Millsap, 2011; Vandenberg & Lance, 2000; Wu, Li, & Zumbo, 2007). First, configural invariance represents the lowest level of invariance and refers to whether the general factor structure (assignment of items to latent factors) is the same across groups. Second, metric in variance involves all factor loadings being equal across groups. As a consequence of metric invariance, the unit of measurement of the latent factor is the same for all groups. Third, scalar invariance is tested by imposing equality constraints on item intercepts or thresholds, in

addition to the previous restrictions. If scalar invariance is established, the same observed score is transformed into the same factor score in different groups. This allows, for example, comparisons between latent factor means across different groups. Finally, strict in variance involves cross-group equality constraints on the item residuals (error variances) that are unexplained by the latent factor. If strict measurement invariance can be confirmed, the items of a factor model are equally reliable across different groups of respondents. Strict measurement invariance is only necessary when manifest means and correlations are to be compared across groups. Differences between nested models are usually determined by means of changes in χ 2 (Δχ 2). However, Δχ 2 is directly affected by sample size, and thus, for larger samples, even trivial differences between models might become significant. This problem applies in particular to the comparison among models representing different levels of measurement invariance (Chen, 2007). Therefore, it has been suggested that one should utilize ΔCFI and ΔRMSEA as the main criteria for evaluating differences between levels of measurement invariance (cf. Chen, 2007; Cheung & Rensvold, 2002; Laverdière,


Morin, & St-Hilaire, 2013). Accordingly, a higher level of measurement invariance can be regarded as confirmed when the model specifying a higher level of measurement invariance does not substantially differ from the model specifying a lower level of measurement invariance with respect to values of CFI and RMSEA. Specifically, the decrease in fit from the lower to the higher level of invariance should be less than 0.01 for ΔCFI and less than 0.015 for ΔRMSEA (cf. Chen, 2007). As a precondition for testing measurement invariance of the RMQ across different groups, it is necessary to demonstrate adequate model fit separately for these groups (Wu et al., 2007). In the present case, the results showed a good level of fit for the seven-factor CFA model for girls, χ 2 = 578.56, df = 356, p < .001, CFI = 0.976, TLI = 0.972, RMSEA = 0.038; for boys, χ 2 = 543.20, df = 356, p < .001, CFI = 0.981, TLI = 0.979, RMSEA = 0.035; for students low in reading competence, χ 2 = 534.15, df = 356, p < .001, CFI = 0.979, TLI = 0.976, RMSEA = 0.035; and for students high in reading competence, χ 2 = 584.24, df = 356, p < .001, CFI = 0.980, TLI = 0.977, RMSEA = 0.037. Consequently, we proceeded to evaluate the four levels of measurement invariance of the RMQ across gender and competence groups. The results (see Table 3) revealed that all models specifying a particular level of measurement invariance showed good fit indexes. In addition, the results confirmed the highest level of measurement invariance (i.e., strict invariance) for the seven-factor model of the RMQ across gender and across competence groups. Interestingly, the comparison across

models in Table 3 shows that the increasing amount of restrictions did not lead to a reduction of model fit as indicated by ΔCFI and ΔRMSEA. Instead, the highest level of fit was observed for the strict invariance models. This suggests that the latent factors of the RMQ are well specified and that their means and interrelations can be directly compared between girls and boys and between low- and high-competence students.

Second-Order Factors The pattern of correlations among RMQ dimensions suggests the existence of second-order factors. More specifically, two alternative models of the RMQ with second-order factors seem possible. The first model includes two second-order factors: One factor comprises intrinsic reading motivation (curiosity and in volvement) and regulatory reading motivation (emotional regulation and relief from boredom), and the other factor represents extrinsic reading motivation (grades, competition, and social recognition). This hypothetical model corresponds closely to the obtained correlation pattern (see Table 2). However, from a theoretical point of view, intrinsic reading motivation should be distinguishable from motivation to read to regulate one’s feelings. Therefore, we also examined an alternative model with three secondorder factors: intrinsic (curiosity and involvement), extrinsic (grades, competition, and social recognition), and regulatory reading motivation (emotional regulation and relief from boredom).

TABLE 3 Measurement Invariance of the Reading Motivation Questionnaire Across Gender and Competence Groups Measurement invariance model

Model fit χ2

df

CFI

TLI

RMSEA

Comparison of models

Δ χ2 (df )a

ΔCFI b

ΔRM SEA c

Measurement invariance across gender Configural

1,116.12***

712

0.979

0.976

0.036

Metric

1,134.43***

734

0.979

0.977

0.035

Metric vs. configural

23.13 (22)

Scalar

1,192.53***

785

0.978

0.978

0.034

Scalar vs. metric

91.37 (51)***

Strict

1,195.58***

814

0.980

0.980

0.033

Strict vs. scalar

0.000

−0.001

−0.001

−0.001

38.65 (29)

0.002

−0.001

Measurement invariance across students low and high in reading competence Configural

1,111.61***

712

0.980

0.977

0.036

Metric

1,127.64***

734

0.980

0.978

0.035

Metric vs. configural

19.36 (22)

0.000

−0.001

Scalar

1,173.06***

785

0.980

0.979

0.033

Scalar vs. metric

65.27 (51)

0.000

−0.002

Strict

1,187.12***

814

0.981

0.981

0.032

Strict vs. scalar

48.21 (29)*

0.001

−0.001

Note. CFI = comparative fit index. RMSEA = root mean square error of approximation. TLI = Tucker – Lewis Index. a For ordered categorical data, the χ2 values of different models cannot be directly compared, and therefore, Δ χ2 was estimated by using the DIFFTEST option in M plus. bRejection of the higher level of measurement invariance if ΔCFI ≤ −0.01. cRejection of the higher level of measurement invariance if ΔRMSEA ≥ 0.015. * p < .05. *** p < .001.


We compared the two alternative models with each other and with a baseline model that assumed that all dimensions of the RMQ are loading on the same higher order factor. For these comparisons, we relied mainly on the χ 2-difference test because there are no general guidelines for the use of alternative cutoff criteria referring to ΔCFI or ΔRMSEA when comparing nested single-group models (in contrast to comparing nested multiple-group models involved in testing measurement invariance). The results in Table 4 reveal that only the models with two or three second-order factors attained acceptable fit indexes. Accordingly, the results indicate that the model with two second-order factors performed significantly better than the baseline model. However, the model with three second-order factors showed better fit than the model with two. Thus, it seems justified to distinguish between intrinsic and regulatory reading motivation, although these factors were highly associated (r latent = .82). The latent correlations between extrinsic reading motivation and both intrinsic (r latent = .37) and regulatory reading motivation (r latent = .38) were less strong. As explained earlier, the regulation of emotions and the avoidance of boredom by means of reading should be mainly successful if reading is intrinsically motivated. Accordingly, the obtained overlap of intrinsic and regulatory reading moti vation appears to be theoretically meaningful. It should be noted, however, that despite a significant χ 2 difference, the models with two and three second-order factors showed almost identical CFI values. Thus, both models performed almost equally well, and more evidence is needed to confirm their distinction. We also examined whether the obtained structure of the RMQ with second-order factors applied to both female and male students and to both low- and highcompetence groups. For that purpose, the comparisons among models with one, two, or three second-order

factors were conducted separately for each group. This procedure was chosen because it is not possible to include second-order factors in tests of measurement in variance. The results showed for all four groups that the model with two second-order factors performed better than the model with one second-order factor, for girls: Δχ 2(2) = 221.92, p < .001; for boys: Δχ 2(2) = 157.96, p < .001, for low-competence students: Δχ 2(2) = 238.96, p < .001; and for high-competence students: Δχ 2(2) = 204.07, p < 0.001, and the model with three second-order factors performed better than the model with two second-order factors, for girls: Δχ 2(1) = 28.77, p < .001; for boys, Δχ 2(1) = 21.57, p < .001; for lowcompetence students: Δχ 2(2) = 17.65, p < .001; and for high-competence students: Δχ 2(2) = 28.40, p < .001. Thus, the factor structure with three second-order factors applies equally to girls and boys and to low- and high-competence students.

Construct Validity To examine the construct validity of the RMQ, we first analyzed the latent correlations between the dimensions of the RMQ and reading amount, fluency, and comprehension. Significant positive correlations with the validation variables were expected for curiosity, in volvement, emotional regulation, and relief from boredom. In contrast, we hypothesized nonsignificant or significant negative correlations between the validation variables and grades, competition, and social recognition. The findings (see Table 2) were mostly in line with the hypotheses. However, in contrast to expectations, small but significant positive correlations between reading amount and both competition and social recognition were observed. These positive associations might be due to the relatively high correlations of competition and social recognition with the intrinsic dimensions of curiosity and involvement (see Table 2).

TABLE 4 Models of the Reading Motivation Questionnaire With Second-Order Factors

df

Comparative fit index

Tucker– Lewis Index

Root mean square error of approximation

3,033.52***

370

0.858

0.844

0.090

Model 2: Two second-order factors

891.68***

368

0.972

0.969

Model 3: Three second-order factors

835.46***

367

0.975

0.972

Model Model 1: One second-order factor

χ2

Comparison

Δ χ2 (df )a

0.040

Model 1 vs. model 2

362.92 (2)***

0.038

Model 2 vs. model 3

39.08 (1)***

Note. Model 2: Second-order factors are intrinsic/regulatory reading motivation (curiosity, involvement, emotional regulation, and relief from boredom) and extrinsic reading motivation (grades, competition, and social recognition). Model 3: Second-order factors are intrinsic (curiosity and involvement), regulatory (emotional regulation and relief from boredom), and extrinsic reading motiv ation (see model 2). a For ordered categorical data, the χ2 values of different models cannot be directly compared, and therefore, Δ χ2 was estimated by using the DIFFTEST option in M plus. *** p < .001.


In the same vein, the hypothesized significant positive correlations between the validation variables and both emotional regulation and relief from boredom are likely to be explained by the rather high associations between the intrinsic and regulatory dimensions of reading motivation. To determine their unique contributions, we examined the RMQ dimensions simultaneously as predictors of the validation variables. In this case, because of the high correlations among the dimensions of reading motivation and to avoid artifacts due to multicollinearity, we used the second-order factors of the RMQ as predictors. More specifically, a model was tested in which the second-order factors served as predictors of reading amount, fluency, and comprehension. The secondorder factors were indicated by their corresponding primary factors, which in turn were represented at the item level. Reading amount was indicated by its corresponding items, whereas reading fluency and comprehension were manifest variables. To control for gender differences, we included direct paths from gender to all variables in the model. According to expectations, the results revealed significant and substantial associations between intrinsic reading motivation and the validation variables (see Table 5), whereas the contributions of extrinsic reading motivation all became significantly negative. The latter finding is in line with previous studies showing that extrinsic reading motivation contributes negatively to reading amount and reading competence when controlling for intrinsic reading motivation (e.g., Schaffner et al., 2013; Unrau & Schlackman, 2006; Wang & Guthrie, 2004). In contrast, the findings did not confirm unique contributions of regulatory reading moti vation. There was, however, a marginal significant association between regulatory reading motivation and TABLE 5 Prediction of the Validation Variables by Second- Order Factors of Reading Motivation (RM) Reading amount

Reading fluency

Gendera

.02

−.08

Intrinsic RM

.71***

.32***

.28***

Extrinsic RM

−.22***

−.13**

−.19***

Regulatory RM

.16†

.01

.04

R2

.62

.09

.09

Predictor

Reading comprehension .07

Note. The parameters (standardized path coefficients) were estimated within a single model, χ2 = 1,119.02, df = 604, p < .001, comparative fit index = 0.972, Tucker– Lewis Index = 0.969, root mean square error of approximation = 0.031. All variables are latent except reading f luency and reading comprehension. a Scoring of gender: 1 = male; 2 = female. † ** p < .01. *** p < .001. p = .052.


reading amount. Also noteworthy, the effects of gender on the validation variables were not significant. This finding is in line with studies showing that gender differences in reading behavior or comprehension disappear when controlling for intrinsic reading motivation (Artelt, Naumann, & Schneider, 2010). Additional evidence for the construct validity of the RMQ was provided by examining assumed gender differences in the dimensions of reading motivation. Specifically, we expected that girls are higher in curiosity, involvement, and emotional regulation and lower in competition than boys. The remaining dimensions of reading motivation should not be affected by gender. The correlational findings in Table 2 were in accordance with our assumptions and, thus, support the validity of the RMQ.

Discussion This study was designed to develop and evaluate the multidimensional RMQ for students at the higher elementary and secondary school levels. Based on Wigfield and Guthrie’s (1997) framework, previous research (e.g., Greaney & Neuman, 1990; Schutte & Malouff, 2007; Watkins & Coffey, 2004), and theoretical considerations, the following dimensions of reading motivation were proposed: curiosity, involvement, grades, competition, social recognition, emotional regulation, and relief from boredom. The factorial structure of the RMQ was examined by applying an item-level CFA. Also, in extending previous research, the present study (a) introduced two new dimensions of reading motivation (reading to regulate negative emotions and to avoid boredom), (b) tested the measurement invariance of the proposed reading motivation instrument across gender and across competence groups, and (c) analyzed the second-order factor structure of the RMQ for the purpose of providing an empirical basis for the use of composite scores (e.g., for intrinsic reading motivation). First of all, the measurement model of the RMQ pertaining to the primary factors showed good fit indexes and thus confirmed the proposed dimensions of reading motivation. In addition, strict measurement in variance for the RMQ across girls and boys and across low- and high-competence students was established. This suggests that the latent factors of the RMQ are in variant across groups referring to their conceptual meaning, the unit of measurement, the correspondence between observed scores and factor scores, and the item residuals. Although strict measurement invariance is not required when latent variables are analyzed, its confirmation underlines the fact that the RMQ is equally applicable to female and male students and to students who are low and high in reading competence.

Another important result of the present study pertains to the structure of higher order or secondary factors. Previous studies have usually created composite scores for intrinsic and extrinsic reading motivation with varying dimensions being involved (e.g., Andreassen & Bråten, 2010; Guthrie et al., 1999; Wigfield & Guthrie, 1997). Such scores are useful to avoid multicollinearity that is likely to occur when highly correlated variables are simultaneously included as predictors in regression models. However, empirical evidence for such composite measures has not been provided (cf. Schiefele et al., 2012). The obtained second-order factors in the present study support the use of composite scores for intrinsic and extrinsic reading motivation. Specifically, the present findings confirm the use of (a) curiosity and involvement as components of a composite score for intrinsic reading motivation (Guthrie et al., 1999; Wang & Guthrie, 2004) and (b) grades, competition, and recognition as components of a composite score for extrinsic reading motivation (Wigfield & Guthrie, 1997). Moreover, we were able to demonstrate that the obtained second-order structure of the RMQ is valid across gender and competence groups. In addition to intrinsic and extrinsic motivation, the present study suggests regulatory reading motivation as a third higher-order factor encompassing emotional regulation and relief from boredom. Although the higher order factors of intrinsic and regulatory reading motivation were highly correlated, they seem to represent distinct factors. However, because of the small difference in the goodness of fit between the models with and without differentiation between intrinsic and regulatory reading motivation, more research is necessary to further substantiate the empirical distinction between these dimensions of reading motivation. Our assumptions regarding the construct validity of the dimensions of reading motivation were largely confirmed. In line with previous studies, the analysis of bivariate correlations revealed that girls’ scores for curiosity, involvement, and emotional regulation were higher than boys’ scores (cf. Logan & Johnston, 2009, 2010; Schiefele & Schaffner, 2013; Wigfield & Guthrie, 1997), whereas boys tended to be more competitively motivated than girls (cf. Stutz et al., 2016; Unrau & Schlackman, 2006; Wigfield & Guthrie, 1997). Moreover, the dimensions of intrinsic reading motivation were positively correlated with the validation variables, thus confirming previous research (Andreassen & Bråten, 2010; Becker et al., 2010; Law, 2009; Unrau & Schlackman, 2006; Wang & Guthrie, 2004). Also in line with our assumptions, the extrinsic dimensions were either nonsignificantly or negatively related to the validation variables (see also Law, 2009; Unrau & Schlackman, 2006; Wang & Guthrie, 2004).

In contrast to expectations, competition and social recognition showed small but significant positive correlations with reading amount. Previous findings, however, are not quite consistent pertaining to these particular relations. Specifically, in line with the present results, Wang and Guthrie (2004) and Schaffner et al. (2013) reported a significant correlation between competition (but not recognition) and reading amount, whereas Wigfield and Guthrie (1997) observed a significant correlation between recognition (but not competition) and reading amount. However, all of these associations were rather low and did not ta ke other predictors into account. The analysis of a structural equation model involving second-order dimensions of reading motivation as predictors of the validation variables was only partly consistent with the correlational findings. Whereas the positive correlations between intrinsic reading motivation and the validation variables were clearly confirmed, the coefficients for the contributions of extrinsic reading motivation all became significantly negative. The latter finding is in accordance with the hypotheses and demonstrates that the negative effects of extrinsic reading motivation are more pronounced when controlling for intrinsic reading motivation. This more pronounced effect is the result of reciprocal suppression between intrinsic and extrinsic reading motivation (Schaffner et al., 2013; Unrau & Schlackman, 2006; Wang & Guthrie, 2004). More specifically, the positive correlation between intrinsic and extrinsic motivation induces positive spurious effects of extrinsic motivation and negative spurious effects of intrinsic motivation on outcome variables such as reading amount and reading competence. As a consequence, when testing the two components of reading motivation as simultaneous predictors, the nonsignificant (or small positive) contributions of extrinsic motivation become significantly negative, whereas the positive contributions of intrinsic motivation become even more positive. In contrast to our assumptions, the obtained positive correlations between regulatory reading motivation and the validation variables did not hold in the structural equation model with multiple predictors. Thus, it can be concluded that the positive bivariate correlations between the regulatory dimensions and the validation variables are mostly due to the strong association between regulatory and intrinsic reading motivation. The effective variable appears to be intrinsic reading moti vation. Only with respect to reading amount, the findings suggested a small but unique contribution of regulatory reading motivation. Despite the limited predictive validity of regulatory reading motivation, we suggest that this construct should be followed up in the future. On t he one hand, regulatory motivation might be interesting in itself.


For example, regulatory use of reading could be studied in greater depth as a mechanism of emotion regulation, thereby focusing on its effectiveness and relation to other forms of emotion regulation (Holodynski, & Friedlmeier, 2006; Kopp, 1989; Mantzicopoulos, 1997). On the other hand, reading for the purpose of regulating one’s feelings may turn out to be an effective means of increasing children’s intrinsic reading motivation.

Limitations and Future Research Although the present study provided evidence for the multidimensional structure of the RMQ, its measurement invariance across gender and across competence groups, and the reliability and validity of the proposed dimensions of reading motivation, a few limitations have to be stated. First, it may be criticized that we decided to exclude those dimensions that are likely to be closely related to reading motivation but do not themselves represent forms of motivation. This refers in particular to work avoidance, challenge, importance, and efficacy (cf. Wigfield & Guthrie, 1997). We would like to point out that we do not regard these dimensions as unimportant. To the contrary, these dimensions represent highly relevant antecedents and consequences of reading motivation. Therefore, it is justified and useful to include these dimensions in questionnaires on reading motivation such as the MRQ. In the case of our questionnaire, however, we preferred to cover only those dimensions that represent reading motivation in a more narrow sense. Second, the present analyses were based on sixthgrade students only. Thus, it remains unclarified whether the confirmed factor structure and measurement invariance also apply to fourth and fifth graders and to students above sixth grade. As a consequence, future research should test the RMQ in other age groups or use longitudinal designs to examine the measurement invariance of the RMQ across time. Third, there are further important goals for future research referring to the dimensions of reading motivation. On the one hand, it seems necessary to reconsider those dimensions that are strongly related to school (grades and competition). Although there is some evidence from qualitative research that students in fact use their leisure time reading as a means to improve their achievement in school and to outperform others, more evidence is needed to show that students in fact refer their ratings of the grades and competition scales to leisure time reading and not to reading for school (e.g., reading of school books). On the other hand, further relevant dimensions that are not represented by the RMQ might exist.


Among those is certainly the dimension of social reading motivation or social reasons for reading. Although we have criticized the MRQ scale of social reasons, it seems quite meaningful, in accordance with research on social goals (e.g., Wentzel, 2000, 2005), to suggest a form of reading motivation that is directed at social exchange or relatedness (e.g., talking about books with others, reading to others). However, in contrast to the MRQ, items measuring social reading motivation should take the form of directly stating social reasons for reading (e.g., “I read because it allows me to talk about it with friends”) instead of merely listing literacy-related activities (e.g., “I visit the library often with my family”; Wigfield & Guthrie, 1997). NOTES

We wish to thank Kathari na Kulisz for her contribution to the collection of data. 1 In most states (Bundesländer) in Germany, elementary schools only include four grades, whereas in a few states, such as Berli n or Brandenburg, elementary schools comprise grades 1–6. REFERENCES

Alloway, T.P., & Gregory, D. (2013). The predictive ability of IQ and working memory scores in literacy in an adult population. International Journal of Educational Research , 57 , 51–56. doi:10.1016/j.ijer.2012.10.004 Andreassen, R., & Bråten, I. (2010). Examining the prediction of reading comprehension on different multiple-choice tests. Journal of Research in Reading , 33(3), 263–283. doi:10.1111/j.1467-9817 .2009.01413.x Artelt, C., Naumann, J., & Schneider, W. (2010). Lesemotivation und Lernstrategien [Reading motivation and reading strategies]. In E. Klieme, C. Artelt, J. Hartig, N. Jude, O. Köller, M. Prenzel, … P. Stanat (Eds.), PISA 2009: Bilanz nach einem Jahrzehnt [PISA 2009: Balance after a decade] (pp. 73–112). Münster, Germany: Waxmann. Asparouhov, T., & Muthén, B. (2010). Weighted least squares estimation with missing data [Technical appendix]. Los Angeles, CA: Muthén & Muthén. Retrieved from www.statmodel.com/techappen.shtml Atkinson, J.W. (1964). An introduction to mot ivation . Princeton, NJ: Van Nostrand. Baker, L., & Wigfield, A. (1999). Dimensions of children’s motivation for reading and their relations to reading activity and reading achievement. Reading Research Quarterly , 34(4), 452–477. doi:10.1598/RRQ.34.4.4 Baumert, J., & Schümer, G. (2001). Schulformen als selektionsbedingte Lernmil ieus [School forms as selection-based learning contexts]. In J. Baumert, E. Klieme, M. Neubrand, M. Prenzel, U. Schiefele, W. Schneider, … M. Weiß (Eds.), PISA 2000: Basiskompetenzen von Schülerinnen und Schülern im internationalen Vergleich [PISA 2000: Basic competencies of students in an international comparison] (pp. 454–467). Opladen, Germany: Leske + Budrich. Becker, M., McElvany, N., & Kortenbruck, M. (2010). Intrinsic and extrinsic reading motivation as predictors of reading literacy: A longitudinal study. Journal of Educationa l Psychology , 102(4), 773–785. doi:10.1037/a0020084 Brophy, J.E. (2005). Goal theorists should move on from performance goals. Educational Psychologist , 40 (3), 167–176. doi:10.1207/s15326985ep4003_3

Carifio, J., & Perla, R.J. (2007). Ten common misunderstandings, misconceptions, persistent myths a nd urban legends about Li kert scales and Li kert response formats and their antidotes. Journal of Social Sciences , 3(3), 106–116. doi:10.3844/jssp.2007.106.116 Chapman, J.W., Tunmer, W.E., & Prochnow, J.E. (2000). Early reading-related skills and performance, reading self-concept, and the development of academic self-concept: A longitudinal study. Journal of Educational Psychology , 92(4), 703–708. doi:10.1037/ 0022-0663.92.4.703 Chen, F.F. (2007). Sensitivity of goodness of fit indexes to lack of measurement invariance. Structural Equation Modeling , 14(3), 464–504. doi:10.1080/10705510701301834 Cheung, G.W., & Rensvold, R.B. (2002). Evaluating goodness-of-fit indexes for testing measurement invariance. Structural Equation Modeling , 9 (2), 233–255. doi:10.1207/S15328007SEM0902_5 Chiu, M.M., & McBride-Chang, C. (2006). Gender, context, and reading: A comparison of students in 43 countries. Scientific Studies of Reading , 10(4), 331–362. doi:10.1207/s1532799xssr1004_1 Cox, K.E., & Guthrie, J.T. (2001). Motivational and cognitive contributions to students’ amount of reading. Contemporary Educational Psychology , 26 (1), 116–131. doi:10.1006/ceps.1999.1044 De Naeghel, J., & Van Keer, H. (2013). The relation of student and class-level characteristics to primary school students’ autonomous reading motivation: A multi-level approach. Journal of Research in Reading , 36 (4), 351–370. doi:10.1111/jrir.12000 De Naeghel, J., Van Keer, H., Vansteenkiste, M., & Rosseel, Y. (2012). The relation between elementary students’ recreational and academic reading motivation, reading frequency, engagement, and comprehension: A self-determination theory perspective. Journal of Educational Psychology , 104(4), 1006–1021. doi:10.1037/a0027800 Eccles, J.S., & Wigf ield, A. (2002). Motivational beliefs, values, and goals. Annual Review of Psychology , 53, 109–132. doi:10.1146/ annurev.psych.53.100901.135153 Elliot, A.J. (2005). A conceptual history of the achievement goal construct. In A.J. Elliot & C.S. Dweck (Eds.), Handbook of competence and motivation (pp. 52–72). New York, NY: Guilford. Foorman, B.R., & Connor, C.M. (2011). Primary grade reading. In M.L. Kamil, P.D. Pearson, E.B. Moje, & P. Afflerbach (Eds.), Handbook of reading research (Vol. 4, pp. 136–156). New York, NY: Routledge. Graham, J.W. (2009). Missing data analysis: Making it work in the real world. Annual Review of Psychology , 60, 549–576. doi:10.1146/ annurev.psych.58.110405.085530 Greaney, V., & Neuman, S.B. (1990). The functions of reading: A cross-cultural perspective. Reading Research Quarterly , 25(3), 172–195. doi:10.2307/748001 Guthrie, J.T., & Klauda, S.L. (2016). Engagement and motivational processes in reading. In P. Afflerbach (Ed.), Handbook of individual differences in reading: Reader, text, and context (pp. 41– 53). New York, NY: Routledge. Guthrie, J.T., Klauda, S.L., & Ho, A.N. (2013). Modeling the relationships among reading instruction, motivation, engagement, and achievement for adolescents. Reading Research Quarterly , 48 (1), 9–26. doi:10.1002/rrq.035 Guthrie, J.T., McGough, K., & Wigfield, A. (1994). Measuring reading activity: An inventory (Instructional Resource No. 4). Athens, GA: National Reading Research Center. Guthrie, J.T., Van Meter, P., McCann, A.D., & Wigfield, A. (1996). Growth of literacy engagement: Changes in motivations and strategies during Concept-Oriented Reading Instruction. Reading Research Quarterly , 31(3), 306–332. doi:10.1598/RRQ.31.3.5 Guthrie, J.T., & Wigfield, A. (Eds.). (1999). How motivation fits into a science of reading [Special issue]. Scientific Studies of Reading , 3(3). Guthrie, J.T., Wigfield, A., Metsala, J.L., & Cox, K.E. (1999). Motivational and cognitive predictors of text comprehension and

reading amount. Scientific Studies of Reading , 3(3), 231–256. doi:10.1207/s1532799xssr0303_3 Heckhausen, H. (1991). Motivation and action . New York, NY: Springer. Holodynski, M., & Friedlmeier, W. (2006). Development of emotions and emotion regulation . New York, NY: Springer. Hulleman, C.S., Durik, A.M., Schweigert, S.A., & Harackiewicz, J.M. (2008). Task values, achievement goals, and interest: An integrative analysis. Journal of Educatio nal Psychology , 100(2), 398– 416. doi:10.1037/0022-0663.100.2.398 Kendeou, P., & van den Broek, P. (2007). The effects of prior knowledge and text structure on comprehension processes during reading of scientific texts. Memory & Cognition , 35(7), 1567–1577. doi:10.3758/BF03193491 Kintsch, W. (1998). Comprehension: A paradigm for cognition . Cambridge, UK: Ca mbridge University Press. Kopp, C. (1989). Regulation of distress and negative emotions: A de velopmental view. Developmental Psychology , 25(3), 343–354. doi:10.1037/0012-1649.25.3.343 Kuhn, D., & Frank lin, S. (2006). The second decade: What develops (and how). In D. Kuhn & R. Siegler (Eds.), Handbook of child psychology: Vol. 2. Cognition, perception, and language (pp. 953– 993). Hoboken, NJ: Wiley. Lau, K., & Chan, D.W. (2003). Reading strategy use and motivation among Chinese good and poor readers in Hong Kong. Journal o f Research in Reading , 26 (2), 177–190. doi:10.1111/1467-9817.00195 Laverdière, O., Morin, A.J., & St-Hilaire, F. (2013). Factor structure and measurement invariance of a short measure of the Big Five personality traits. Personality and Individual Differences , 55(7), 739–743. doi:10.1016/j.paid.2013.06.008 Law, Y. (2009). The role of attribution beliefs, motivation and strategy use in Chinese fifth-graders’ reading comprehension. Educational Research , 51(1), 77–95. doi:10.1080/001318808027 04764 Lenhard, W. (2013). Leseverständnis und Lesekompetenz [Reading comprehension and reading competence]. Stuttgart, Germany: Kohlhammer. Logan, S., & Johnston, R.S. (2009). Gender differences in reading ability and attitudes: Examining where these differences lie. Journal of Research in Reading , 32(2), 199–214. doi:10.1111/ j.1467-9817.2008.01389.x Logan, S., & Johnston, R.S. (2010). Investigating gender differences in reading. Educational Review, 62(2), 175–187. doi:10.1080/ 00131911003637006 Logan, S., Medford, E., & Hughes, N. (2011). The importance of intrinsic motivation for high and low ability readers’ reading comprehension performance. Learning and Individual Differences , 21(1), 124–128. doi:10.1016/j.lindif.2010.09.011 Lubke, G.H., & Muthén, B.O. (2004). Applying multigroup confirmatory factor models for continuous outcomes to Likert scale data complicates meaningfu l group comparisons. Structural Equation Modeling , 11(4), 514–534. doi:10.1207/s15328007 sem1104_2 Mantzicopoulos, P. (1997). How do children cope with school fa ilure? A study of social/emotional factors related to children’s coping strategies Psychology in the Schools , 34(3), 229–237. doi:10.1002/ (SICI)1520-6807(199707)34:3<229::AID-PITS5>3.0.CO;2-J McGeown, S., Goodwin, H., Henderson, N., & Wright, P. (2012). Gender differences in reading motivation: Does sex or gender identity provide a better account? Journal of Re search in Reading , 35(3), 328–336. doi:10.1111/j.1467-9817.2010.01481.x McKenna, M.C., Kear, D.J., & Ellsworth, R.A. (1995). Children’s attitudes toward reading: A national survey. Reading Research Quarterly , 30 (4), 934–956. doi:10.2307/748205 Millsap, R.E. (2011). Statistical approaches to measurement invariance. New York, NY: Taylor & Francis.


Möller, J., & Bonerad, E.-M. (2007). Fragebogen zur habituellen Lesemotivation [Habitual reading motivation questionnaire]. Psychologie in Erziehung und Unterricht , 54(4), 259–267. Mullis, I.V.S., Martin, M.O., Kennedy, A.M., & Foy, P. (2007). PIRLS 2006 international report: IEA’s Progress in International Reading Literacy Study in primary schools in 40 countries . Chestnut Hill, MA: TIMSS & PIRLS International Study Center, Lynch School of Education, Boston College. Muthén, L.K., & Muthén, B.O. (1998–2014). Mplus u ser’s guide . Los Angeles, CA: Muthén & Muthén. Nolen, S.B. (2007). Young children’s motivation to read and write: Development in social contexts. Cognition and Instruction , 25(2/3), 219–270. doi:10.1080/07370000701301174 Park, Y. (2011). How motivational constructs interact to predict elementary students’ reading performance: Examples from attitudes and self-concept in reading. Learning and Individual Differences , 21(4), 347–358. doi:10.1016/j.lindif.2011.02.009 Pekrun, R. (1993). Facets of adolescents’ academic motivation: A longitudinal expectancy–value approach. In M.L. Maehr & P.R. Pintrich (Eds.), Advances in motivation an d achievement (Vol. 8, pp. 139–189). Greenwich, CT: JAI. Pintrich, P.R., & Schrauben, B. (1992). Students’ motivational beliefs and their cognitive engagement in classroom academic tasks. In D.H. Schunk & J.L. Meece (Eds.), Student perceptions in the classroom (pp. 149–183). Hillsdale, NJ: Erlbaum. Rasinski, T.V., Reutzel, D.R., Chard, D., & Linan-Thompson, S. (2011). Reading fluency. In M.L. Kamil, P.D. Pearson, E.B. Moje, & P. Afflerbach (Eds.), Handbook of reading research (Vol. 4, pp. 286–319). New York, NY: Routledge. Roberts, T.A., Christo, C., & Shefelbine, J.A. (2011). Word recognition. In M.L. Kamil, P.D. Pearson, E.B. Moje, & P. Afflerbach (Eds.), Handbook of reading research (Vol. 4, pp. 229–258). New York, NY: Routledge. Sainsbury, S., & Schagen, I. (2004). Attitudes to reading at ages nine and eleven. Journal of Research in Reading , 27 (4), 373–386. doi:10.1111/j.1467-9817.2004.00240.x Schafer, J.L. (1997). Analysis of incomplete multivariate data . New York, NY: Chapman & Hall. Schafer, J.L. (1999). NORM 2.03 for Windows 95/98/NT [Computer software]. University Park: Department of Statistics, Pennsylvania State University. Schaffner, E., & Schiefele, U. (2007). Auswirkungen habitueller Lesemotivation auf die situative Textrepräsentation [Effects of habitual reading motivation on the situational representation of text]. Psychologie in Erziehung und Unterricht , 54 (4), 268–286. Schaffner, E., Schiefele, U., & Ulferts, H. (2013). Reading amount as a mediator of the effects of intrinsic and extrinsic reading moti vation on reading comprehension. Reading Research Quarterly , 48 (4), 369–385. doi:10.1002/rrq.52 Schiefele, U. (1999). Interest and learning from text. Scientific Studies of Reading , 3(3),257–279. doi:10.1207/s1532799xssr0303_ 4 Schiefele, U. (2009). Situational and individual interest. In K.R. Wentzel & A. Wigfield (Eds.), Handbook of motivation at school (pp. 197–222). New York, NY: Routledge. Schiefele, U., & Schaffner, E. (2013). Die Lesemotivation von Grundschülerinnen und -schülern der sechsten Klassenstufe: Ergebnisse einer Interviewstudie [Reading motivation of elementary school students: Results from an interview study]. Psychologie in Erziehung und Unterricht , 60 (3), 214–233. Schiefele, U., Schaffner, E., Möller, J., & Wigfield, A. (2012). Dimensions of reading motivation and their relation to reading behavior and competence. Reading Research Quarterly , 47 (4), 427–463. doi:10.1002/RRQ.030 Schneider, W., Schlagmüller, M., & Ennemoser, M. (2007). LGVT 6–12: Lesegeschwindigkeits- und -verständnistest für die


Klassenstufen 6 –12 [Reading f luency and reading comprehension test for grades 6–12]. Göttingen, Germany: Hogrefe. Schunk, D.H. (1991). Self-efficacy and academic motivation. Educational Psychologist , 26 (3/4), 207–231. Schutte, N.S., & Malouff, J.M. (2007). Dimensions of reading motivation: Development of an adult reading motivation scale. Reading Psychology , 28(5), 469–489. doi:10.1080/02702710701568991 Stutz, F., Schaffner, E., & Schiefele, U. (2016). Relations among reading motivation, reading amount, and reading comprehension in the early elementary grades. Learning and Individual Differences, 45, 101–113. doi:10.1016/j.lindif.2015.11.022 Tarelli, I., Schwippert, K., & Stubbe, T.C. (2012). Mathematische und naturwissenschaft liche Kompetenzen von Schülerinnen und Schülern mit Migrationshintergrund [Competencies of students with an immigration background in mathematics and science]. In W. Bos, H. Wendt, O. Köller, & C. Selter (Eds.), TIMSS 2011: Mathematische und naturwissenschaftliche Kompetenzen von Grundschulkindern in Deutschland im internationalen Vergleich [TIMSS 2011: Mathematical and scientific skills of primary school children in Germany in an international comparison] (pp. 247–267). Münster, Germany: Waxmann. Tichá, R., Espin, C.A., & Wayman, M.M. (2009). Reading progress monitoring for secondary-school students: Reliability, validity, and sensitivity to growth of reading-aloud and maze-selection measures. Learning Disabilities Research & Practice , 24(3), 132– 142. doi:10.1111/j.1540-5826.2009.00287.x Tighe, E.L., & Schatschneider, C. (2014). A dominance analysis approach to determining predictor importance in third, seventh, and tenth grade reading comprehension skills. Reading and Writing , 27 (1), 101–127. doi:10.1007/s11145-013-9435-6 Unrau, N., & Schlackman, J. (2006). Motivation and its relationship with reading achievement in an urban middle school. The Journal of Educational Research , 100(2), 81–101. doi:10.3200/JOER.100.2.81-101 Vandenberg, R.J., & Lance, C.E. (2000). A review and synthesis of t he measurement invariance literature: Suggestions, practices, and recommendations for organizational research. Organizational Research Methods , 3(1), 4–70. doi:10.1177/109442810031002 Wang, J.H., & Guthrie, J.T. (2004). Modeling the effects of intrinsic motivation, extrinsic motivation, amount of reading, and past reading achievement on text comprehension between U.S. and Chinese students. Reading Research Quarterly , 39(2), 162–186. doi:10.1598/RRQ.39.2.2 Watkins, M.W., & Coffey, D.Y. (2004). Reading motivation: Multidimensional and indeterminate. Journal of Educational Psychology , 96 (1), 110–118. doi:10.1037/0022-0663.96.1.110 Weiner, B. (1989). Human motivation . Mahwah, NJ: Erlbaum. Wentzel, K.R. (2000). What is it that I’m trying to achieve? Classroom goals from a content perspective Contemporary Educational Psychology , 25(1), 105–115. doi:10.1006/ceps.1999.1021 Wentzel, K.R. (2005). Peer relationships, motivation, and academic performance at school. In A.J. Elliot & C.S. Dweck (Eds.), Handbook of competence and motivation (pp. 279–296). New York, NY: Guilford. Wigfield, A., & Eccles, J.S. (2000). Expectancy–value theory of achievement motivation. Contemporary Educational Psychology , 25(1), 68–81. doi:10.1006/ceps.1999.1015 Wigfield, A., & Guthrie, J.T. (1997). Relations of children’s motivation for reading to the amount and breadth of their reading. Journal of Educational Psychology , 89(3), 420–432. doi:10.1037/ 0022-0663.89.3.420 Wu, A.D., Li, Z., & Zumbo, B.D. (2007). Decoding the meaning of factorial invariance and updating the practice of multi-group confirmatory factor analysis: A demonstration with TIMSS data. Practical Assessment, Research & Evaluation , 12(3), 1–26. Yoon, M., & Millsap, R.E. (2007). Detecting violations of factorial invariance using data-based specification searches: A Monte

Carlo study. Structural Equation Modeling , 14(3), 435–463. doi:10.1080/10705510701301677 Zimmerman, B.J. (2000). Self-efficacy: An essential motive to learn. Contemporary Educational Psychology , 25(1), 82–91. doi:10.1006/ ceps.1999.1016 Submitted April 8, 2015 Final revision received October 29, 2015 Accepted November 13, 2015

ULRICH SCHIEFELE (corresponding author) is a professor in the Department of Psychology at the University of Potsdam, Germany; e-mail [email protected]. ELLEN SCHAFFNER was a research scientist in the Department of Psychology at the University of Potsdam, Germany, at the time of this study; e- mail ellen.schaffner@ uni-potsdam.de.


4_2016_Schiefele & Schaffner - New Instrument Reading Motivation Questionnaire

Recommend Documents