The relationship between I- and D-types of curiosity and school performance By Noor Galesloot (10587063)
Bachelor project UvA
University of Amsterdam Bachelor Project Curiosity Mentor: Brenda Jansen 5132 words
2
Table of Contents
Abstract ... 3
Introduction ... 3
D- and I-type curiosity ... 4
Measuring D-type curiosity ... 5
The current study ... 6
Method ... 7
Participants ... 7
Measures ... 8
Procedure ... 10
Results ... 10
Relationship between the I- and D-scale and school performance ... 10
Psychometric properties of the CTD task ... 13
Discussion ... 15
References ... 17
Appendix 1. Teacher questionnaire ... 19
Appendix 2. Instructions CTD task ... 19
3 Abstract
This study focused on whether D type curiosity is a better predictor of school performance than I-type curiosity. D-type curiosity is motivated by an information gap, whereas I-type is motivated by the pleasure of learning. Also, psychometric properties of a revised version of the CTD task, which measures D-type curiosity, were assessed. The CTD task was
administrated to 191 first and second graders. School performance was measured by means of teacher’s ratings, I- and D-type curiosity by means of a parent questionnaire and an indication of intelligence using the RCPM. In constrast to expectations, the I-scale of the parent
questionnaire was a positive predictor of school performance, whereas the D-scale was not a predictor of school performance. Scores on the CTD taks were not normally distributed. Internal consistency of the CTD task was high. Convergent validity was insufficient, since CTD scores were only related to teacher’s ratings of curiosity and not to the D-scale of the parent questionnaire. The CTD task was not related to age, however was related to gender, since boys obtained higher scores than girls, and intelligence. In conclusion, I-type curiosity is positively associated with school performance, whereas D-type curiosity is not, and
psychometric properties of the CTD task are insufficient.
Introduction
Curiosity, the motive to seek new knowledge (Lauriola, Litman, Mussel, De Santis, Crowson & Hoffman, 2015), is thought to be the force behind learning and academic achievement (Grossnickle, 2014). High-curious children are more inclined to explore their environment and actively seek out new information than low-curious children, which gives them more opportunities for learning (Arnone, Grabowski & Rynd, 1994). Research has shown that curiosity enhances memory for new information, is linked to more question asking and stimulates persistence (Grossnickle). Because of this, curiosity could play an important role in academic achievement (Von Stumm, Hell & Chamorro-Premuzic, 2011).
Knowledge about the relationship between curiosity and learning is important for teachers to be able to adjust education to an individual’s level of curiosity (Grossnickle, 2014). For instance, research has suggested that high-curious children thrive academically when given more freedom and control of their environment (Jirout & Klahr, 2012). However, clarity lacks in several areas, which are outlined below.
First of all, no conclusive evidence has been found regarding the relationship between curiosity and intelligence. Intuitively, it is logical to assume that people who are more
4
information have more opportunities to learn and therefore would be more intelligent. Some researchers have indeed found a correlation between curiosity and intelligence in adults (Von Stumm, Hell & Chamorro-Premuzic, 2011), whereas others have failed to find this
relationship in young children (Jirout & Klahr, 2012, Henderson & Wilson, 1991). Other areas, important for the relationship between learning and curiosity, but lacking, are outlined below.
D- and I-type curiosity
According to Loewenstein (1994), curiosity is caused by a feeling of not knowing something, or an information gap. This feeling of deprivation is unpleasant and thus motivates information seeking to fill the information gap and reduce the negative feelings (Jirout & Klahr, 2012). Apart from this deprivation (D) type of curiosity, another type has been
identified: the interest (I) type. This type of curiosity is defined by seeking information purely for the pleasure of learning (Litman, 2008). The I- and D-types are relatively stable traits (Litman et al., 2010), which can be measured in early childhood (Piotrowski, Litman & Valkenburg, 2014).
To date, the only research that tried to identify the relationship of I- and D-type curiosity with learning was conducted with students. For instance, I-type curiosity is found to be higher for those with a higher education level compared to people with a lower education, but it is not clear whether I-type curiosity motivates one to strive for higher education or whether education stimulates I-type curiosity (Litman et al., 2010). There are however several arguments why D-type curiosity could be a better predictor of academic outcomes than I-type curiosity.
First of all D-type curiosity motivates a higher level of explorative behaviour than I-type curiosity. Research has shown that curiosity caused by a perceived information gap is more intense, but also involves more negative feelings of tension, than curiosity caused by general interest associated with I-type curiosity, resulting in more exploration (Litman, Hutchins & Russon, 2005).
Secondly, the information sought after by D-type curiosity could be better suited for learning in an educational setting. I-type curiosity only needs information to be engaging in order to be rewarding, whereas this information has to be accurate and useful to satisfy D-type curiosity, since it has to fill the specific information gap (Lauriola et al., 2015). Seeking accurate and useful information in order to reduce uncertainty and ignorance might be more beneficial for school performance than seeking engaging information.
5
Third, related to the previous argument, the motivation associated with D-type curiosity could also be more suitable for education. I-type curiosity is linked to intrinsic motivation, which means doing something because it’s interesting or enjoyable. Since D-type curiosity needs accurate and useful information, it is also associated with concerns about performance and is therefore linked to both intrinsic and extrinsic motivation, which means doing something because it leads to a good outcome (Ryan & Deci, 2000). In contrast to what was previously thought, extrinsic motivation can also be valuable to education, since it is related to engagement, high school performance and high quality learning (Ryan & Deci). Educators cannot always rely on intrinsic motivation, since the tasks used for education are not always very interesting or enjoyable for pupils (Ryan & Deci). Also, intrinsic motivation tends to decline with age, whereas extrinsic motivation is more stable (Lepper, Corpus & Iyengar, 2005).
Fourth, D-type curiosity motivates seeking specific knowledge for the experienced information gap (Piotrowski et al., 2014), whereas I-type curiosity motivates one to seek a range of different sources of intellectual stimulation. Therefore, D-type curiosity is more linked to focussing attention on one specific source of intellectual stimulation (Piotrowski et al.). I-type curiosity is not correlated with concentration and may even have a negative impact on concentration, since it could motivate shifting one’s attention to something more
interesting than the task at hand. In student samples, D-type curiosity is also more than I-type curiosity associated with persistence related to academic achievement (Richards, Litman & Roberts, 2013) and higher grades (Elliot & Church, 1997).
It may seem counter intuitive to some that D-type curiosity would be associated with better learning outcomes, since it is caused by feelings of tension and not by the pleasure of learning. However, D-type curiosity is still rewarding, because when satisfied the feelings of tension reduce (Litman, 2010). However, due to a lack of research, it is still unclear which type of curiosity is more important for learning.
Measuring D-type curiosity
Clarity is also lacking in the measurement of children’s D-type curiosity. Piotrowski et al. (2014) developed a parent questionnaire measuring both types, but recently researchers have begun to develop more objective tasks measuring D-type curiosity in children. In these tasks, uncertainty is manipulated since this is thought to elicit an information gap, which causes curiosity (Jirout & Klahr, 2012). One such tasks is the Connect The Dots (CTD) task. Children are shown two connect-the-dots puzzles. Each puzzle consists of a number of dots
6
which have to be connected to form an image. Children are asked to choose the puzzle they would like to solve. Uncertainty is manipulated by giving either concrete hints about the image, abstract hints or no hints at all. Children who are more curious are expected to prefer the most uncertain option (Baesjou, Boonstra, ten Nijenhuis, van Steijn & van Trommel, 2013).
There are however a few problems with this task. First of all, the internal consistency of the CTD was insufficient. Secondly, reading the hints is too difficult for younger children, such as first grade pupils.
The current study
In short, no studies have been done to explore the relationship between the types of curiosity and learning outcomes. It is however conceivable that D-type curiosity would be a better predictor for learning than I-type curiosity. The question of the present study is to test whether I-type curiosity or D-type curiosity has a stronger relation with learning outcomes. We use parent-reports for both types of curiosity but also include a D-type curiosity tasks, to be administered to the children themselves. Revisions of this original task, CTD, were performed. The psychometric properties of a revised version of the CTD task are assessed. Finally, the relationship between intelligence and curiosity is unclear, which makes it necessary to control.
The revised version of the CTD task will be administered to first and second graders to measure D-type curiosity. I- and D-type curiosity will be assessed by a parent questionnaire. School performance will be measured by teacher’s ratings and an indication of intelligence by administering the Raven Coloured Progessive Matrices (Raven, 1958). The following
hypotheses will be tested:
1. D-type curiosity is positively associated with school performance, whereas I-type curiosity is not. It is expected that the D-scale of the parent questionnaire will be a positive predictor of ratings of school performance after controlling for RCPM scores, whereas the I-scale will not predict school performance.
2. It is expected that the pupils who perform poorly, average or above average at school (according to their teachers) , but not in their scores on the I-scale.
3. The psychometric properties of the CTD task are acceptable or good. a. Scores on the CTD task are normally distributed.
7
b. The CTD task is reliable. It is expected that Cronbach’s alpha, an indication of internal consistency, is at least acceptable, meaning .7 or higher.
c. The CTD task is valid, with convergent and divergent validity. It is expected that the correlation between the CTD task and the D-scale of the parental questionnaire is significantly higher than the correlation between the CTD task and the I-scale, and that CTD scores and curiosity ratings by teachers are significantly related. With regards to divergent validity, it is expected that CTD scores are not related to age, sex or RCPM scores.
Method Participants
A total of 191 children, including 76 first graders and 115 second graders, was recruited from seven elementary schools in the Netherlands. Of the 191 children, 102 were girls, 88 were boys and of one participant the sex was unknown. The children’s age ranged from 6.1 to 8.9 years old (M = 7.4, SE = 0.7).
The research project was approved by the Ethics Review Board. Depending on the school’s preference, either a passive or active consent procedure was used. Using passive consent, parents were sent a letter containing information about the study. Unless the parents refuse to let their child participate, consent was assumed. When an active consent procedure was used, children received the letter to give it to their parents. In this case the parents had to give written permission for their child to participate in the study.
Out of the 191 participants, five participants did not complete all fifteen items of the CTD task. Their data were excluded from any analyses concerning the CTD task. Of one child the sex was unknown. The parental questionnaire measuring I-and D-type curiosity was returned for 61 children, which corresponds with a response rate of 32%. Of six participants teacher’s ratings were missing, and of sixty participants the date of birth was missing.
8 Measures
D-type curiosity
For this study the original CTD task (Baesjou et al., 2013), was revised. Children were presented with a booklet, containing 15 pairs of connect-the-dots puzzles, one of which can be seen in Figure 1. The instruction and the booklet are included in the Appendix. Each pair consisted of one puzzle in which it was very clear to predict what image would appear if the puzzle was solved. The other puzzle was an ambiguous cloud of dots, which was designed to create uncertainty. Children were asked to choose which puzzle they most likely would want to solve by ticking the box next to it, without actually being allowed to solve the puzzles. First
the researchers explained the task groupwise by one example item. To make sure the children were motivated, an actual connect-the-dots puzzle was given after completion of the task (Baesjou et al., 2013).
The task was scored by giving each choice for an unambigous and ambiguous puzzle a score of 0 and 1 respectively. The minimum score for the CTD task is 0 and the maximum score is 15, with a higher score meaning higher D-type curiosity.
I- and D-type curiosity
A questionnaire, developed by Piotrowski et al (2014), was administrated to parents either electronically or by paper-and-pencil. The questionnaire consisted of ten statements about curious characteristics or behaviours. The parents were asked to indicate on a scale of 1 (almost never) to 4 (almost always) how often their child showed these behaviours. Half of the items were behaviours associated with I-type curiosity, such as ‘My child shows visible
9
enjoyment when discovering something new’, and the other half characterize D-type
curiosity, such as ‘My child will work for a long time to solve a problem because s/he wants to know the answer’. Research has shown that both scales have acceptable psychometric properties (Piotrowski et al., 2014).
A separate score for I- and D-type curiosity was computed. The minimum score for both scales was 5 and the maximum score was 20, with a higher score meaning higher curiosity.
Intelligence
For an indication of intelligence, the Raven Coloured Progressive Matrices (RCPM, Raven, 1958) was administrated. This task can be used to measure intelligence in children ranging from 4 to 9 years of age. The RCPM is a booklet containing 36 matrices, each with one piece missing. The child is asked to choose the piece that would fill the gap out of six options and to circle this piece (Raven). The standard procedure in administrating the RCPM is to give children unlimited time to finish the task, however in this study a time limit of ten minutes was used. Reliability and validity of the RCPM have been rated as sufficient by the COTAN (Egberink, Janssen & Vermeulen, 1985).
The RCPM is scored by giving each correct answer a score of 1 and each wrong answer a score of 0. The first three items are examples. The minimum score is 0 and the maximum score 33, with a higher score indicative of higher intelligence.
School performance and curiosity
School performance and curiosity were rated electronically by teachers. Teachers were presented with a list of names of their pupils and were asked to rate each child on four
abilities: language, math, creativity and curiosity. Teachers gave pupils either a score of 1 (performing poorly), 2 (average) or 3 (above average) for each ability. In total this
questionnaire lasted for approximately fifteen minutes. The questionnaire is included in the Appendix.
Teacher’s ratings of math, language and creativity were combined into one general measure of school performance. The minimum score was 3 and the maximum score was 9, with a higher score meaning better school performance. For the ratings of curiosity, the minimum score was 1 and the maximum score 3, with a higher score indicative of more curiosity.
10 Procedure
Three researchers visited each school class. All tasks were administrated groupwise, preceded by a general instruction in which the researchers explained that the children would take part in three tasks. The children whose parents did not give permission for them to participate were given colouring pages and connect-the-dots puzzles.
First, the CTD task was administrated. The instruction was given groupwise, by means of an example item. The researchers explained that the children could choose which puzzle they would like to solve by ticking the box next to it. It was announced that the children had to choose without actually solving the puzzles, but that they were allowed to solve the last puzzle.
After the CTD task was completed, the RCPM was explained to the class using three example items. Researchers explained that the children had to fill in the missing gap, by choosing one piece out of the six options and circling this piece. The time limit of this task is 10 minutes, but it was said that it was not a problem if they were not able to finish the task. Most children finished the task in under ten minutes.
For a separate research question, reported elsewhere, it was investigated if curiosity can be stimulated. Children were asked various open questions about one animal or a story was read. Before and after these questions they were asked to indicate their interest in a number of animals. Further details will not be given in this paper.
In total the study lasted for fifty to sixty minutes. After the study was completed, the children were thanked for their participation.
During the administration in the classroom, the teacher was asked to fill in the questionnaire measuring academic performance. If the regular teacher was not present, the questionnaire was sent to him/her by e-mail. Teachers were asked to rate each child as performing poorly, average or above average on language, math, creativity and curiosity.
Parents were sent the questionnaire measuring I- and D-type curiosity by e-mail or received a paper-and-pencil version. They were asked to fill in their child’s name, but they were assured that these names would be transferred to codes before analysing the data to ensure anonymity.
Results
Relationship between I- and D-scale and school performance
To assess the relationship between I- and D-type curiosity as rated by parents and teacher’s ratings of school performance, a stepwise regression analysis was conducted whilst
11
controlling for RCPM scores. In this analysis parent ratings were used instead of the CTD, since the first measured both types of curiosity.
Firstly it was tested whether there was a difference in school performance or RCPM scores between children who’se parents did not return the questionnaire and children who’se parents did return the questionnaire. Levene’s test showed that for ratings of school
performance, variances were equal for both groups, F(1, 183) = 0.63, p = 0.429, but variances were not equal for RCPM scores, F(1, 183) = 6.78, p = 0.010. The groups did not differ on school performance, t(183) = -0.27, p = 0.784, or RCPM scores, t(158.56) = -1.91, p = 0.058.
The It was expected that D-type curiosity would be a better predictor of school performance than I-type curiosity and would therefore explain additional variance to the variance explained by the RCPM scores. Scores on the RCPM were converted into z-scores, separately for first- and second-graders, which makes it possible to control for each pupil’s intelligence score relative to the mean of their grade. First, assumptions for a stepwise regression analysis were tested. A Kolmogorov-Smirnov test was used to assess whether teacher’s ratings of school performance and scores on the RCPM were normally distributed. The teacher’s ratings, D(185) = 0.14, p < 0.001, and RCPM scores, D(186) = 0.08, p = 0.003, deviated significantly from normal. However, since the sample is larger than 30, normality was assumed (Field, 2009). To check for multicollinearity, VIF values were tested. All VIF values were less than 10, so the assumption of multicollinearity was not violated. Using a stepwise analysis, two models were tested. In both models, z-scores for the RCPM scores were introduced first, since this is the control variable. In the first model, I-type curiosity was entered in the model after the RCPM scores, followed by D-type curiosity. Table 1 shows the results of the first stepwise regression analysis. First, RCPM scores were a good predictor of school performance. Second, I-type curiosity explained a significant amount of variance in ratings school performance, since higher I-type curiosity was positively associated with school performance. However, in contrast to expectations, D-type curiosity did not explain additional variance in school performance ratings to the variance explained by I-type curiosity and RCPM scores. In the second model tested, D-type curiosity was added before I-type curiosity. The results of the second stepwise regression analysis can be seen in Table 1. Adding D-type curiosity to the model did not improve the model significantly. In conclusion, I-type curiosity proved to be a predictor of school performance, after controlling for
12
Table 1
Linear model of predictors of CTD scores, with 95% confidence intervals reported in parentheses I-type entered before D-type
Step 1 b SE B β p Constant 6.56 (6.16, 7.00) 0.20 p < 0.001 RCPM scores 0.62 (0.13, 1.12) 0.25 0.32 p = 0.015 Step 2 Constant 2.88 (0.02, 5.74) 1.43 p = 0.049 RCPM scores 0.59 (0.12, 1.06) 0.24 0.30 p = 0.015 I-type 0.21 (0.05, 0.38) 0.08 0.31 p = 0.012 Step 3 Constant 2.87 (-0.02, 5.76) 1.44 p = 0.051 RCPM scores 0.59 (0.11, 1.07) 0.24 0.30 p = 0.017 I-type 0.20 (-0.01, 0.42) 0.11 0.29 p = 0.062 D-type 0.02 (-0.15, 0.19) 0.09 0.03 p = 0.853
D-type entered before I-type Step 2 Constant 4.97 (3.06, 6.90) 0.95 p < 0.001 RCPM scores 0.58 (0.09, 1.07) 0.24 0.30 p = 0.021 D-type 0.12 (-0.02, 0.25) 0.07 0.21 p = 0.094 Step 3 Constant 2.87 (-0.02, 5.76) 1.44 p = 0.051 RCPM scores 0.59 (0.11, 1.07) 0.24 0.30 p = 0.017 D-type 0.02 (-0.15, 0.19) 0.09 0.03 p = 0.853 I-type 0.20 (-0.01, 0.42) 0.11 0.29 p = 0.062
Note. For the first model, R2 = 0.10 for Step 1; R2 = 0.10 for Step 2 (p < 0.05); R2 = 0.001 for Step 3 (p =
13 Psychometric properties of the CTD task
To assess the second hypothesis, the psychometric properties of the CTD task are acceptable or good, it was firstly tested whether the scores on the CTD task were normally distributed. Normality was tested by means of the Kolmogorov-Smirnov test. The scores on the CTD task, D(186) = 0.11, p < 0.001, deviated
significantly from normal. The CTD scores were therefore not normally distributed. To explore normality
further, a histogram graphing the distribution of CTD scores was created. Figure 2 shows that the distribution indeed deviates from normal and that there seems to be a ceiling effect, with a disproportional number of children obtaining a score of 14. In contrast to expectations, the CTD scores were not normaly distributed.
To test reliability, internal consistency was assessed by computing Cronbach’s alpha. It was expected that Cronbach’s alpha would be .7 or higher. Since the assumption of
normality was violated, first Pearson correlation coefficients and a non-parametric version, Kendall’s tau coefficients, were computed to see whether these coefficients would differ. There were no differences between these coefficients, so it is possible to interpret Cronbach’s alpha. The CTD task had high internal consistency, with Cronbach’s α = 0.90. Subsequently, item-total correlations were computed to test whether omitting specific items would increase Cronbach’s alpha. This was only the case for items 14 and 15. Omitting these items would increase Cronbach’s alpha to respectively 0.909 and 0.904. Consistent with expectations, the CTD task has a high reliability.
To test whether the CTD task is valid, both convergent and divergent validity were tested. Convergent validity was tested by assessing the strength of the relationship between CTD scores and the D-scale of the parental questionnaire and the curiosity ratings by the teachers. Firstly, it was expected that the correlation between scores on the CTD and scores on the D-scale would be significantly stronger than the correlation between scores on the
14
CTD and scores on the I-scale. As was previously mentioned, the scores on the CTD task were not normally distributed. However, since the sample was larger than 30, normality was assumed. The Kolmogorov-Smirnov test showed that both the scores on the D-scale, D(57) = 0.12, p = 0.037, and on the I-scale, D(57) = 0.17, p < 0.001, also differed significantly from normal. In this case normality was also assumed due to the sample size. Unexpectedly, there was no significant relationship between the scores on the CTD task and the scores on the D-scale from the parent questionnaire, r = - 0.08, p = 0.548, and between the scores on the CTD task and the scores on the I-scale from the parent questionnaire, r = 0.12, p = 0.357. Contrary to expectations, the relationship between scores on the CTD and on the D-scale was not significantly stronger than the relationship between scores on the CTD and the I-scale. In fact, the CTD scores did not correlate with any type of curiosity, as measured with the parent questionnaire. Convergent validity is therefore insufficient.
Subsequently, the relationship between teacher’s ratings of pupils curiosity and scores on the CTD was assessed by computing a Pearson correlation coefficient. It was expected that teacher’s ratings of curiosity would be significantly and positively correlated with CTD scores. Teacher’s ratings of curiosity were not distributed normally, D(180) = 0.30, p < 0.001, however since the sample was larger than 30 normality was assumed. There was a small but significant positive correlation between scores on the CTD and teacher’s ratings of curiosity, r = 0.21, p = 0.005. Scores on the CTD task seem to be positively related to the teacher’s
ratings of curiosity, in line with expectations.
Divergent validity was assessed by testing the strength of the relationship between CTD scores and age, sex and scores on the RCPM. The relationship between CTD scores and age was tested computing a Pearson correlation coefficient, which was expected to be non-significant. Age was not normally distributed, D(127) = 0.09, p = 0.011, however since the sample was larger than 30 normality was assumed. The relationship between age and CTD scores was non-significant, r = -0.07 , p = 0.443. Thus, consistent with expectations, age did not influence CTD scores.
Secondly, the relationship between sex and CTD scores was tested by means of an independent t-test. Sex was expected to have a non-significant effect on CTD scores. Levene’s test showed that for the CTD scores, variances were equal for boys and girls, F(1, 183) = 0.36, p = 0.552. On average, boys scored higher on the CTD task (M = 8.07, SE = 4.75) than girls did (M = 6.26, SE = 4.70). This difference of 1.81 was significant, t(183) = 2.60, p = 0.010. This means that contrary to expectations, boys obtained significantly higher CTD scores than girls. To test whether this difference was due to a few specific items a
chi-15
square test was conducted, separately for each item. χ2 ranged between 0.05 and 19.55. Boys
were significantly more likely to choose the ambiguous puzzle for items 1, 3, 6, 8, 9 and 13, but the gender ratio was equal for the other items.
Finally, to test the relationship between scores on the RCPM and scores on the CTD, a Pearson correlation coefficient was computed. The correlation between RCPM- and CTD scores was expected to be non-significant. As was previously mentioned, the RCPM-scores are not normally distributed, however since the sample was larger than 30, normality was assumed. There was a small but significant positive relationship between scores on the RCPM and scores on the CTD, r = 0.20, p = 0.005. Contrary to expectations, RCPM- and CTD scores were positively related. Since both sex and RCPM-scores influenced CTD scores, divergent validity was insufficient.
Discussion
In the current study, the relationship between I- and D-type curiosity and school performance was assessed. D-type curiosity was hypothesized to be associated with school performance, whereas I-type curiosity was not. However, the opposite was found: I-type curiosity was a good predictor of school performance, whereas D-type curiosity was not. The first hypothesis was therefore rejected. Secondly, psychometric properties of the CTD task were tested. Since the scores were not normally distributed, the first subhypothesis had to be rejected. The CTD task was very reliable, supporting the second subhypothesis. The CTD was not related to D-type curiosity, however was related to teacher’s ratings of curiosity. Finally, age was not related to the CTD task, however sex and intelligence were. Thus neither
convergent nor divergent validity proved to be sufficient, and the third subhypothesis was rejected.
The finding that D-type curiosity is not a predictor of school performance, is
surprising. The main arguments why D-type curiosity was thought to be better suited for an educational setting is that it is associated with seeking accurate and useful information (Lauriola et al., 2015) and with concerns about performance (Ryan & Deci, 2000). However, it is possible that these arguments do not yet apply to first and second graders. Seeking accurate and useful information might not be important for subjects in first and second grade, such as learning the basics of maths, reading and spelling, but more for learning specific information, for instance in geography or history. Also children in the first and second grade might not yet be concerned about their academic performance, thus D-type curiosity could become a better predictor of school performance as they grow older and grades become more
16
important. This would be in line with previous research that shows that intrinsic motivation declines with age, whereas extrinsic motivation, which is only associated with the D-type, is more stable (Lepper, Corpus & Iyengar, 2005). A longitudinal study could explore this possible explanation.
An alternative explanation is that the low response rate of the parental questionnaire measuring I- and D-type curiosity resulted in a systematic bias. Children who’se parents returned the questionnaire were not more intelligent or better school performers. However, it is still possible that higher educated parents are more likely to fill in this questionnaire. Those with a higher education level usually score higher on the I-type of curiosity (Litman et al., 2010), thus these parents may place more value on I-type curiosity. This systematic bias could have resulted in a stronger link between I-type curiosity and school performance for these children.
Whereas reliability of the CTD task was very high, both convergent and divergent validity were not acceptable. Firstly, the CTD task was not related to D-type curiosity ass rated by parents, but is was related to teacher’s ratings of curiosity. One possible explanation is that parents and teachers differ in the way they rate curiosity. Parents may rate their child as high on curiosity due to social desirability, whereas teachers do not have this bias. Parents might also differ in determining how frequent their child shows curious behaviours. For instance, if a child displays a certain behaviour once a week, some parents might rate this as displayed ‘sometimes’, whereas other parents might rate it as ‘often’. Teachers however are able to rate a child’s curiosity by comparing it to other pupils, which is perhaps a more objective measure but can also be biased. The same child could for instance be rated as either average or highly curious, depending on the pupils he or she is compared to. To determine whether the CTD task really measures D-type curiosity, it would be best if future research examined the relationship between the CTD task and other direct measures of D-type curiosity, such as the Underwater Exploration task (Jirout & Klahr, 2012), instead of ratings by parents or teachers.
Divergent validity was also insufficient. Although age was not related to CTD scores, both sex and intelligence were. Firstly, boys obtained higher scores on the CTD task than girls. For some items, boys were more likely than girls to choose the ambiguous puzzle. In the Appendix can be seen that these items were more ‘girly’, such as a heart, bunny or butterfly, so for these items boys might have chosen the ambiguous CTD puzzles not out of curiosity but because the unambiguous puzzle was unappealing to them, resulting in a higher score.To overcome this, more gender-neutral puzzles should be used in the CTD task.
17
Secondly, intelligence was related to CTD scores. This could mean that the CTD task measured intelligence as well as curiosity, which is undesirable. However, some researchers have found curiosity to be related to intelligence in adults (Von Stumm, Hell & Chamorro-Premuzic, 2011). If this is the case, the relationship between the CTD task and intelligence would be a matter of convergent validity, instead of a limitation of the task. However, other reseachers have found no relationship between curiosity and intelligence in young children (Jirout & Klahr, 2012, Henderson & Wilson, 1991). Until the relationship between
intelligence and curiosity is clarified, it remains uncertain whether the relationship between the CTD task and intelligence is indeed a limitation of the task.
This study has a few limitations. As was previously mentioned, the response rate of the parental questionnaire was low, which may have resulted in a systematic bias in the measurement of I- and D-type curiosity. Secondly, this study only consisted of one
measurement. Therefore, it is not possible to assess the causal relationship between I- and D-type curiosity and school performance. A longitudinal study could explore this, and also asses whether I-type curiosity is a better predictor for school performance in younger children, whereas D-type curiosity could become more important as children age.
This study has tried to shed clarity on the relationship between I- and D-type curiosity and learning. This is important since I-type children could have different educational needs than D-type children. A better understanding of how curiosity influences learning can therefore contribute to education and enable teachers to tailor education to children’s needs.
References
Arnone, M. P., Grabowski, B. L., & Rynd, C. P. (1994). Curiosity as a personality variable influencing learning in a learner controlled lesson with and without
advisement. Educational Technology Research and Development, 42(1), 5-20. Baesjou, B., Boonstra, Z., ten Nijenhuis, R., van Steijn, A., & van Trommel, H. (2013). Het
meten van deprivatie nieuwsgierigheid bij kinderen tussen de zeven en negen jaar oud. Onderzoeksverslag, Universiteit van Amsterdam.
Egberink, I. J. L., Janssen, N. A. M., & Vermeulen, C. S. M. (1985). COTAN documentatie: Coloured Progressive Matrices, CPM, 1947. Opgehaald van:
http://www.cotandocumentatie.nl/test_details.php?id=286
Elliot, A. J., & Church, M. A. (1997). A hierarchical model of approach and avoidance achievement motivation. Journal of Personality and Social Psychology, 72(1), 218-232.
18
Field, A. (2009). Discovering Statistics Using SPSS. Sage Publications.
Grossnickle, E. M. (2014). Disentangling curiosity: Dimensionality, definitions, and
distinctions from interest in educational contexts. Educational Psychology Review, 1-38.
Henderson, B. B., & Wilson, S. E. (1991). Intelligence and curiosity in preschool children. Journal of School Psychology, 29(2), 167-175.
Jirout, J., & Klahr, D. (2012). Children’s scientific curiosity: In search of an operational definition of an elusive concept. Developmental Review, 32(2), 125-160.
Lauriola, M., Litman, J. A., Mussel, P., De Santis, R., Crowson, H. M., & Hoffman, R. R. (2015). Epistemic curiosity and self-regulation. Personality and Individual
Differences, 83(2015), 202-207.
Lepper, M. R., Corpus, J. H., & Iyengar, S. S. (2005). Intrinsic and extrinsic motivational orientations in the classroom: Age differences and academic correlates. Journal of
Educational Psychology, 97(2), 184-196.
Litman, J.A. (2005). Curiosity and the pleasures of learning: Wanting and liking new information. Cognition & Emotion, 19(6), 793-814.
Litman, J. A. (2008). Interest and deprivation factors of epistemic curiosity. Personality and Individual Differences, 44(7), 1585-1595.
Litman, J. A. (2010). Relationships between measures of I-and D-type curiosity, ambiguity tolerance, and need for closure: An initial test of the wanting-liking model of information-seeking. Personality and Individual Differences, 48(4), 397-402. Litman, J. A., Crowson, H. M., & Kolinski, K. (2010). Validity of the interest-and
deprivation-type epistemic curiosity distinction in non-students. Personality and Individual Differences, 49(5), 531-536.
Litman, J., Hutchins, T., & Russon, R. (2005). Epistemic curiosity, feeling-of-knowing, and exploratory behaviour. Cognition & Emotion, 19(4), 559-582.
Litman, J. A., & Jimerson, T. L. (2004). The measurement of curiosity as a feeling of deprivation. Journal of Personality Assessment, 82(2), 147-157.
Loewenstein, G. (1994). The psychology of curiosity: A review and reinterpretation. Psychological Bulletin, 116(1), 75-98.
Piotrowski, J. T., Litman, J. A., & Valkenburg, P. (2014). Measuring epistemic curiosity in young children. Infant and Child Development, 23(5), 542-553.
19
Richards, J. B., Litman, J., & Roberts, D. H. (2013). Performance characteristics of measurement instruments of epistemic curiosity in third-year medical students. Medical Science Educator, 23(3), 355-363.
Ryan, R. M., & Deci, E. L. (2000). Intrinsic and extrinsic motivations: Classic definitions and new directions. Contemporary Educational Psychology, 25(1), 54-67.
Von Stumm, S., Hell, B., & Chamorro-Premuzic, T. (2011). The hungry mind intellectual curiosity is the third pillar of academic performance. Perspectives on Psychological Science, 6(6), 574-588.
Teacher questionnaire
Wij willen u vragen om voor elke leerling in te vullen of deze zwak, gemiddeld of sterk presteert op onderstaande vaardigheden
1 = zwak 2 =
gemiddeld 3 = sterk
Naam
leerling Taalvaardigheid Rekenvaardigheid Creativiteit
Mate van
nieuwsgierigheid
voorbeeld 3 2 2 3
Instructions CTD task
(Instructie begint als alle boekjes zijn uitgedeeld.)
Op jullie tafel hebben jullie nu als het goed is allemaal een boekje liggen, klopt dit? Pak nu allemaal een potlood of een pen.
(Zo nee; dan komt begeleider 2 of 3 het boekje brengen.)
Als je naar de voorkant van het boekje kijkt, zie je onder het plaatje een lijntje waar je je naam op kunt schrijven. Willen jullie daar nu allemaal je naam op schrijven?
(Begeleider 2 en begeleider 3 lopen door de klas om te kijken of dit lukt en helpen als nodig.) Is het bij iedereen gelukt?
In dit boekje staan allemaal puntjes net zoals je hier op dit papier ziet. (Begeleider 1 wijst naar de tekening op het A3 op het bord.)
Als je de puntjes met elkaar verbindt, krijg je een tekening. Wie heeft wel eens zo’n soort tekening gemaakt?
Zoals de kinderen die dit al gedaan hebben weten, trek je een lijn van 1 naar 2, naar 3, naar 4, (Begeleider 1 doet dit voor met haar vinger) en zo verder tot je bij het laatste stipje bent. Zo ontstaat er een figuur.
In dit boekje staan op elke bladzijde twee figuren van puntjes, net zoals op dit papier.
(Begeleider 1 wijst naar het A3.) Jullie mogen op elke bladzijde kiezen welk figuur jullie het liefst zouden willen tekenen. Je ziet op iedere bladzijde een figuur boven en een figuur onder. Tussen deze twee figuren kan je kiezen. Bij elk figuur staat een hokje (begeleider 1 wijst naar
20
het hokje). Bij het figuur wat je het liefst zou willen tekenen, zet je in het hokje een kruisje. Het andere hokje laat je leeg. Je mag het figuur in dit boekje niet echt tekenen!
Van deze twee figuren, zou ik het liefste het bovenste figuur willen tekenen. Wat moet ik nu doen? (Begeleider wijst eventueel iemand aan die antwoord mag geven.) Juist, ik zet bij dit figuur een kruisje in het hokje. Het onderste hokje laat ik leeg.
Nu mag je je boekje open doen. Kijk eens naar de eerste bladzijde. Welk van deze twee figuren zou jij het liefst willen tekenen, de onderste of de bovenste? Zet een kruisje in het hokje van dit figuur. Let op: je mag het figuur dus niet tekenen. Jij mag kiezen: er is geen fout antwoord.
Is dit bij iedereen gelukt?
Nu mogen jullie zelf alle opdrachten uit het boekje maken. Doe dit zonder met je buurman of buurvrouw te praten. Als je klaar bent met alle opdrachten doe je je boekje dicht en steek je je vinger op. Dan komen wij je boekje halen. (Begeleiders 1 en 2 halen de boekjes op als de kinderen hun vinger opsteken, controleren of de naam duidelijk leesbaar is en geven elk kind een figuur om zelf te tekenen. Hierbij wordt tegen de kinderen gezegd dat als ze klaar zijn met de tekening, dat ze dan hun boek mogen pakken en mogen lezen.)
Als je tussendoor een vraag hebt, of het niet meer snapt mag je je vinger opsteken en komt er iemand om je vraag te beantwoorden.
Heeft er iemand nog een vraag?
(Indien er vragen zijn deze beantwoorden.) Dan mogen jullie nu beginnen. Succes! Tijdslimiet: 10 minuten
Let er bij het ophalen op dat er op elke pagina iets ingevuld is en er geen twee antwoorden zijn gegeven. Als er niets is ingevuld: vragen toch nog iets in te vullen; als er twee
21 CTD task
