The Relationship Between STEM School Achievement And STEM

The Relationship Between STEM School Achievement And STEM Out-Of-School Activities: Possible Role Of Family Social Background And Students Gender Ivan Dević1, Josip Burušić1, Toni Babarović1, Dubravka Glasnović Gracin 2, Marija Šakić Velić1 1 Institut društvenih znanosti Ivo Pilar 2 Učiteljski fakultet Sveučilišta u Zagrebu Funding source: Croatian Science Foundation, grant n° JOBSTEM HRZZ IP-09 -2014 -9250

JOBSTEM PROJECT: STEM career aspirations during primary schooling: A cohortsequential longitudinal study of relations between achievement, selfcompetence beliefs, and career interests o. Croatia ØIvo Pilar Institute of Social Sciences ØUniversity of Split, Faculty of Philosophy ØUniversity of Zagreb, Faculty of Electrical Engineering and Computing ØUniversity of Zagreb, Faculty of Teacher Education o. France ØUniversité de Poitiers o. Hungary ØCentral European University o. United States ØUniversity of California Irvine

The longitudinal-sequential design of the study with a two-group pre- and post-test randomized experiment

Participants • Three cohorts of primary school students in three waves during three consecutive years

The JOBSTEM project: Three main research goals 1. To examine how students’ general and specific STEM career aspirations form and how they change over time 2. To examine how students’ school achievement and self-competence beliefs relate to students’ general and specific STEM vocational preferences and the dynamics of these relations during primary school 3. To examine how students’ general and specific STEM career aspirations are shaped by their families and gender

Present study -goals §Within the third aim, we formulated specific research question for this study: To test the moderation effects of gender and SES on the relationship between out of school activities and school achievement.

Present study - Background - The results of Croatian students on PISA and TIMSS assessments show low or average results in science and mathematics compared to results in other countries in the EU (Braš Roth et al, 2008; NCVVO, 2012 a; 2012 b) Also, the achievement of Croatian students in science and mathematics is low in comparison to their achievement in other school subjects (Burušić, Babarović & Šakić, 2008). - Participation in out-of-school activities is positively associated to better school attendance, more positive attitudes towards school work, better work habits, better interpersonal skills, lower drop-out rates, less time spent in unhealthy behaviours, and improved grades (Clark, 1988; Hamilton & Klein, 1998; Mc. Laughlin, 2000) -clear positive correlation between student participation in STEM out-of-school activities and school achievement in the STEM area (Froschi et al. , 2003; Hansen, et. al. , 1995; Lee, 1997 - experimental studies conducted by James-Burdumy et al. (2005) and Gibson & Chase (2002) did not find a consistent positive relationship between participating in out-of-school activities and school achievement. - Recent studies in STEM school domain confirmed the strong influence of family SES in predicting children’s achievement in STEM school subjects. Students from families with lower SES are less likely to achieve high in STEM subjects (OECD, 2007), and consequently they have less opportunity to pursue careers in STEM field (Gorard & See, 2009)

Methods

Participants • 1798 primary school students from 16 primary schools in Zagreb and its surroundings attending grades 4 to 6 (N 4 th grade= 586, N 5 th grade=580, N 6 th grade=632) • equally distributed by gender (48. 6% girls). Assessment • Paper and pencil method • Group assessment, in the classes during the regular school activities • Data collection lasted 90 minutes

Measures - students’ grades in STEM domains In the Croatian school system and school curriculum the STEM area is covered by the following school subjects, depending on the grade: • Mathematics • Nature Subjects taught in participants’ grades • Informatics • Technical education • Geography • Biology • Chemistry • Physics - collected directly from the school administration. - average STEM achievement was calculated as the arithmetic means of the grades in STEM-related school subjects which were transformed into z-values and centered around the corresponding grade mean

Measures - students’ knowledge in STEM §Within the JOBSTEM project we developed tests of STEM school knowledge §Separate tests were developed through grades 4 th to 6 th §Current curricular documents related to STEM subjects in Croatia were used in the tests development How much water is in the tank? §A complete psychometric analysis was conducted for the tests for 4 th, 5 th and 6 th grade o Paper published (Dević, Babarović, Glasnović, Burušić, 2018) o All three test have unidimensional structure, acceptable reliability (α 4 th = . 78; α 5 th =. 70; α 6 th = . 79), good discriminating power and moderate to high correlations with achievement in STEM school subjects A. B. C. D. 205 l 201 l 202 l 210 l

Measures - STEM Out of school activities STEM out-of-school activity scale containing 12 most common activities within the STEM area in which children can be involved during leisure time. Items are adapted from the ROSE project (Relevance of Science Education) (Jenkins & Pell, 2006) taking into account the context and possible activities that children can participate. Likert scale: 1 - ‘almost never’, 5 – ‘very often’ Sample items: How often in leasure time do you do the folloving: You are visiting observatory, planetarium or zoo. You are programming on a computer You read books about the universe, Earth, animal and plant world. You are visiting natural or technical museums. You're thinking and solving math tasks. You are visiting fairs and science festivals.

Family characterstics – indicators of SES §parents’ education For educational level, we used semi – interval rating scale (response ranging from 1 = ‘Unfinished or completed elementary school’ to 6 = ‘Obtained Master’s or Ph. D level’) §family income Monthly family income perception was reported ‘as rating of per household member income’ on semi-interval 5 -point rating scale, ranging from ‘extremely below income of majority’ to ‘well above average of majority’. Average result on indicators – dichotomus variable: a) lower SES (42%) b) higher SES (58%)

Results

Does gender influece the relationship between out of school activities and STEM achievement

Gender as moderator of out of school activities on STEM achievement Gender STEM Out of school activities STEM school grades Model Summary R R-sq MSE F df 1 df 2 p , 18 , 032 , 94 20, 69 3 1755 , 00 Model coeff se t p constant -, 09 , 03 -2, 72 , 01 -> small effect of out of school activities Out of school , 19 , 03 6, 09 , 00 Gender , 19 , 04 4, 05 , 00 -> boys perform lower than girls Interaction -, 08 , 05 -1, 69 , 09

Gender as moderator of out of school activities on STEM knowledge test Gender STEM Out of school activities STEM knowledge Model Summary R R-sq MSE F df 1 df 2 p , 16 , 027 , 97 15, 13 3 1661 , 00 Model coeff se t p constant , 04 , 03 1, 05 , 30 Out of school , 16 , 03 4, 69 , 00 -> small effect of out of school activities Gender -, 09 , 05 -1, 82 , 05 -> girls perform slightly lower than boys Interaction -, 01 , 05 -, 15 , 89

Does SES influece the relationship between out of school activities and STEM achievement

SES as moderator of out of school activities on STEM achievement SES STEM Out of school activities STEM school grades Model Summary R R-sq MSE F df 1 df 2 p , 24 , 055 , 91 33, 33 3 1716 , 00 Model coeff se t p constant -, 18 , 04 -4, 9175 , 00 Out of school , 10 , 03 2, 58 , 01 -> Students with higher SES, perform better in STEM area SES , 35 , 05 7, 2846 , 00 Interaction -, 10 , 05 -2, 0931 , 04 -> Effect of Out of school activities on STEM achievement depends on SES!

SES as moderator of out of school activities on STEM achievement For low SES, out of school activities have bigger influence on STEM achievement than for high SES

SES as moderator of out of school activities on STEM knowledge test SES STEM Out of school activities STEM knowledge test Model Summary R R-sq MSE F df 1 df 2 p , 19 , 044 , 95 20, 09 3 1624 , 0000 Model coeff se t p constant -, 14 , 04 -3, 57 , 00 Out of school , 10 , 03 2, 58 , 01 -> Students with higher SES, perform better in STEM area SES , 22 , 05 4, 31 , 00 Interaction , 06 , 05 1, 12 , 26 -> Effect of Out of school activities on STEM achievement depends on SES!

Summary and Implications § Participating in out of school activities has a relative low effect on student school achievement § Students with lower SES perform lower in STEM area § Gender does not moderate the relationship between out of school activities and STEM achievment, regardless of indicator of school achievment used (school grades or STEM knowledge test) § SES moderates the relationship between out of school activities and school achievement but only when school grades are used as indicator of school achievement § Low-archieving students participate less in STEM-related out-of-school activities Fostering this involvement, especially in younger cohorts, can present a potential challange for parents and educators to increase later choices in STEM domain, since longitudinal research confirmed the links between early STEM involvement – positive STEM-related beliefs – and consequently STEM choices (Simpkins, Davis-Kean, & Eccles, 2006) § Further research is need…

Thank you for your attention! More information at: www. jobstem. eu