2. Academic achievement analysis
2.3 Secondary moderator analysis on academic achievement
In the first stage moderator analysis, the result showed that some categories Qw are significant, meaning that the ES in those groups are not homogenous and further analysis is necessary. Therefore, the researcher set three criteria to elicit proper data:
firstly, the Qw is significant; secondly, k>20; thirdly, researcher’s concern. Eight groups fitted the criteria: nature and science, elementary, university, Asia, Europe, North America, puzzle games and simulation games. The research results are as follows:
2.3.1 Nature and science group on academic achievement (1) Nature and science and educational level
As Table 4.8 shows, different educational level is the reason to cause the heterogeneity of academic achievement in nature and science. Research results showed that except junior high, each weighted mean ES is positive and 95% CI is significantly different from zero, meaning that in elementary, senior high and university, the effect of applying DGBL on students’ nature and science academic achievement is significantly better than the effect of applying Non-DGBL on students’ nature and science academic achievement.
Because educational level (QB = 10.99, p<.05) achieved significant level, it meant that educational level is the moderator of nature and science. Post hoc showed that when applying DGBL on nature and science, the effect of senior high (ES=1.05) is significantly better than university (ES=.71) while junior high school is not having significant effect.
While first stage analysis showed that applying DGBL in each educational level is better than Non-DGBL, secondary moderator analysis showed that applying DGBL on nature andscience in junior high is not significantly better than Non-DGBL.
(2) Nature and science and sample location
As Table 4.8 shows, different sample location is the reason to cause the heterogeneity of academic achievement in nature and science. Research results showed that each weighted mean ES is positive and 95% CI is significantly different from zero, meaning that in Asia, Europe, and North America, the effect of applying DGBL on students’ nature and science academic achievement is significantly better than the effect of applying Non-DGBL on students’ nature and science academic achievement.
Sample location (QB =5.99, p<.05) achieved significant level, meaning that sample location is the moderator of nature and science. Post hoc showed that when applying DGBL on nature and science, the effect of North America (ES=.91) and Asia (ES=.88) is significantly better than Europe (ES=.57).
Compare secondary moderator analysis with first stage moderator analysis, the result is consistent.
(3) Nature and science and digital game categories
As Table 4.8 shows, different digital game category is the reason to cause the heterogeneity of academic achievement in nature and science. Research results showed that excluding role-playing games(k=0), each weighted mean ES is positive and 95% CI is significantly different from zero, meaning that in puzzle
games, simulation games, and strategy games, the effect of applying DGBL on students’ nature and science academic achievement is significantly better than the effect of applying Non-DGBL on students’ nature and science academic achievement.
Digital game categories (QB = 9.02, p<.05) achieved significant level, meaning that digital game categories is the moderator of nature and science. Post hoc showed that when applying DGBL on nature and science, the effect of simulation games (ES=1.18) is significantly higher than puzzle games (ES=.69).
While first stage analysis showed that role-playing games is the most effective game among digital game categories, secondary moderator analysis showed in nature and science subject matter, simulation games is the most effective game in digital game categories.
Table 4.8
Nature and Science Secondary Moderator Cross Analysis
Group Moderator
Q
B kd
+subject matter (Nature and science)
Educational level (3>4) 10.99* 20
1.Elementary 7 .99#
Digital game categories(6>4) 9.02* 18
4. Puzzle games 10 .69#
6. Simulation games 3 1.18#
8. Strategy games 5 .86#
Note. *: significant at the .05 level; d+:
weighted mean effect size; k: number of articles; #: d
+positive and 95% CI significantly different from zero; (): post hoc.
2.3.2 Elementary group on academic achievement (1) Elementary and subject matter
As Table 4.9 shows, different subject matter is the reason to cause the heterogeneity of academic achievement in elementary. Research results showed most elementary articles studied on language, math, and nature and science subject matters. Except others, each subject matter weighted mean ES is positive and 95% CI is significantly different from zero, meaning that in language, math, and nature and science, the effect of applying DGBL on elementary students’ academic achievement is significantly better than the effect of applying Non-DGBL on elementary students’ academic achievement.
Subject matter (QB = 79.67, p<.05) achieved significant level, meaning that subject matter is the moderator of elementary. Post hoc showed that when applying DGBL on elementary, the effect of language (ES=.95) and nature and science (ES=.99) is significantly higher than math (ES=.14). Compare secondary moderator analysis with first stage moderator analysis, the result is consistent.
(2) Elementary and sample location
As Table 4.9 shows, different sample location is the reason to cause the heterogeneity of academic achievement in elementary. Research results showed that except Europe, each sample location weighted mean ES is positive and 95%
CI is significantly different from zero, meaning that in Asia and North America, the effect of applying DGBL on elementary students’academic achievement is significantly better than the effect of applying Non-DGBL.
Sample location (QB = 137.09, p<.05) achieved significant level, meaning that sample location is the moderator of elementary. Post hoc showed that when applying DGBL on elementary, Asia (ES=1.28) is significantly higher than North America (ES=.26) while Europe doesn’t have significant effect.
While first stage analysis showed that applying DGBL on students’ academic achievement in each sample location is significantly better than the effect of applying Non-DGBL on students’ academic achievement in each sample location, secondary moderator analysis showed that applying DGBL on elementary students’ academic achievement in Europe is not significantly better than Non-DGBL on elementary students’ academic achievement in Europe.
(3) Elementary and digital game category
As Table 4.9 shows, the different digital game category is the reason to cause the heterogeneity of academic achievement in elementary. Research results showed that only puzzle games weighted mean ES is positive and 95% CI is significantly different from zero, meaning that only puzzle games showed the result that the effect of applying DGBL on elementary students’ academic achievement is significantly better than Non-DGBL.
While first stage analysis showed that applying DGBL on students’ academic achievement in each digital game category is significantly better than the effect of applying Non-DGBL on students’ academic achievement in each digital game category, secondary moderator analysis showed that applying DGBL by role-playing games or strategy games on elementary students’ academic achievement is not significantly better than Non-DGBL on elementary students’
academic achievement.
Table 4.9
Elementary Secondary Moderator Cross Analysis
Group Moderator
Q
B kd
+Educational level (Elementary)
subject matter (1,3>2) 79.67* 20
1. Language 4 .95#
Digital game categories 58.24* 21
4. Puzzle games 15 .35#
5. Role-playing games 2 1.39
8.Strategy games 4 .20
Note. *: significant at the .05 level; d+:
weighted mean effect size; k: number of articles; #: d
+positive and 95% CI significantly different from zero; (): post hoc.
2.3.3 University group on academic achievement (1) University and subject matter
As Table 4.10 shows, different subject matter is the reason to cause the heterogeneity of academic achievement in university. Research results showed that except math (k=2), each subject matter weighted mean ES is positive 95%
CI is significantly different from zero, meaning that in language, nature and science, health and PE and computer science, the effect of applying DGBL on university students’ academic achievement is significantly better than the effect of applying Non-DGBL on university students’academic achievement.
Subject matter (QB = 30.88, p<.05) achieved significant level, meaning that subject matter is the moderator of university. Post hoc showed that when
applying DGBL on university, language (ES=.94) is significantly higher than computer science (ES=.49) and health andPE (ES=.35) while math is not having significant effect.
While first stage analysis showed that applying DGBL on students’ academic achievement in each subject matter is significantly better than the effect of applying Non-DGBL on students’ academic achievement in each subject matter, secondary moderator analysis showed that applying DGBL on university students’ academic achievement in math is not significantly better than the effect of applying Non-DGBL on university students’academic achievement in math.
(2) University and sample location
As Table 4.10 shows, different sample location is the reason to cause the heterogeneity of academic achievement in university. Research results showed that each weighted mean ES is positive and 95% CI is significantly different from zero, meaning that in Asia, Europe, and North America, the effect of applying DGBL on university students’ academic achievement is significantly better than the effect of applying Non-DGBL on university students’academic achievement.
Sample location (QB = 16.63, p<.05) achieved significant level, meaning that sample location is the moderator of university. Post hoc showed that when applying DGBL on university, North America (ES=.80) is significantly higher than Europe (ES=.40).
While first stage analysis showed that applying DGBL on Asia and North America is significantly better than Europe, secondary moderator analysis showed that applying DGBL on North America university students is significantly better than Europe university students.
(3) University and digital game category
As Table 4.10 shows, different digital game category is the reason to cause the heterogeneity of academic achievement in university. Research results showed that each weighted mean ES is positive and 95% CI is significantly different from zero, meaning that in role-playing games, puzzle games, simulation games and strategy games, the effect of applying DGBL on university students’
academic achievement is significantly better than the effect of applying Non-DGBL on university students’ academic achievement.
Digital game category (QB = 14.52, p<.05) achieved significant level, meaning that digital game category is the moderator of university. Post hoc showed that strategy games (ES=1.04) is significantly higher than simulation games (ES=.72) puzzle games (ES=.54) and role-playing games (ES=.47); simulation games is significantly higher than puzzle games and role-playing games.
While first stage analysis showed that applying DGBL on students’ academic achievement in each digital game category is significantly better than the effect of applying Non-DGBL on students’ academic achievement, and role-playing games is the most effective game type, secondary moderator analysis showed that in university, the most effective game type is strategy games.
Table 4.10
5. Health & Physical Education 8 .35#
6. Computer Science 4 .49#
Digital categories(8>6,4,5 ;6>4,5) 14.52* 33
4. Puzzle games 9 .54#
5. Role-playing games 4 .47#
6. Simulation games 15 .72#
8. Strategy games 5 1.04#
Note. *: significant at the .05 level; d+:
weighted mean effect size; k: number of articles; #: d
+positive and 95% CI significantly different from zero; (): post hoc.
2.3.4 Asia group on academic achievement (1) Asia and subject matter
As Table 4.11 shows, different subject matter is the reason to cause the heterogeneity of academic achievement in Asia. Research results showed that both language and nature and science weighted mean ES are positive and 95%
CI are significantly different from zero, meaning that in language and nature and science, the effect of applying DGBL on Asia students’ academic achievement is significantly better than the effect of applying Non-DGBL on Asia students’academic achievement.
Subject matter (QB = 90.66, p<.05) achieved significant level, meaning that subject matter is the moderator of Asia. Post hoc showed that when applying DGBL on Asia, language (ES=1.40#) is significantly higher than nature and science (ES=.88#), while math and health and PE are not having significant effect.
While first stage analysis showed that applying DGBL on students’ academic achievement in each subject matter is significantly better than the effect of applying Non-DGBL on students’ academic achievement in each subject matter, secondary moderator analysis showed that applying DGBL on Asia students’
academic achievement in math and healthand PE is not significantly better than the effect of applying Non-DGBL.
(2) Asia and educational level
As Table 4.11 shows, different educational level is the reason to cause the heterogeneity of academic achievement in Asia. Research results showed that excluding junior high(k=0), except senior high, elementary and university weighted mean ES are positive and 95% CI are significantly different from zero, meaning that in elementary and university, the effect of applying DGBL on Asia students ’ academic achievement is significantly better than the effect of applying Non-DGBL on Asia students’academic achievement.
Educational level (QB = 72.49, p<.05) achieved significant level, meaning that educational level is the moderator of Asia. Post hoc showed that elementary (ES=1.28) is significantly higher than university (ES=.63), while senior high is not having significant effect.
While first stage analysis showed that applying DGBL on students’academic achievement in each educational level is significantly better than the effect of applying Non-DGBL on students’ academic achievement in each educational level, secondary moderator analysis showed that applying DGBL on Asia students’academic achievement in senior high is not significantly better than the effect of applying Non-DGBL on Asia students’academic achievement in senior high. Moreover, in first stage moderator analysis, elementary has the smallest effect among educational level while in secondary moderator analysis, elementary is the most effective educational level when applying DGBL in Asia.
(3) Asia and digital game categories
As Table 4.11 shows, different digital game category is the reason to cause the heterogeneity of academic achievement in Asia. Research results showed that excluding strategy games (k=0), each weighted mean ES is positive and 95% CI is significantly different from zero, meaning that in puzzle games, role-playing games and simulation games, the effect of applying DGBL on Asia students’
academic achievement is significantly better than the effect of applying Non-DGBL on Asia students’academic achievement.
Digital game category (QB = 24.91, p<.05) achieved significant level, meaning that digital game category is the moderator of Asia. Post hoc showed role-playing games (ES=1.24) is significantly higher than puzzle games (ES=.76) and simulation games (ES=.41); puzzle games is significantly higher than simulation games.
While first stage post hoc analysis showed that simulation games is significantly higher than puzzle games, secondary post hoc moderator analysis showed that puzzle games is significantly higher than simulation games when applying DGBL on Asia students’academic achievement.
Table 4.11
Asia Secondary Moderator Cross Analysis
Group Moderator
Q
B kd
+Sample location (Asia)
Subject matter(1>3) 90.66*
1. Language 5 1.40#
2. Math 3 .02
3. Nature & science 8 .88#
5. Health & PE 3 .20
Educational level(1>4) 72.49*
1. Elementary 11 1.28#
3. Senior High 4 .19
4. University 6 .63#
Digital game categories(5>4,6;4>6) 24.91*
4.Puzzle games 12 .76#
5.Role-playing games 3 1.24#
6.Simulation games 6 .41#
Note. *: significant at the .05 level; d+:
weighted mean effect size; k: number of articles; #: d
+positive
and 95% CI significantly different from zero; (): post hoc.
2.3.5 Europe group on academic achievement (1) Europe and subject matter
As table 4.12 shows, different subject matter is the reason to cause the heterogeneity of academic achievement in Europe. Research results showed that only nature and science weighted mean ES is positive and 95% CI is significantly different from zero, meaning that in nature and science, the effect of applying DGBL on Europe students’ academic achievement is significantly better than the effect of applying Non-DGBL on Europe students’academic achievement.
Subject matter (QB = 14.75, p<.05) achieved significant level, meaning that subject matter is the moderator of Europe. Only nature and science (ES=.57) showed positive effect while language, math, health and PE, and computer science are not having significant effect.
While first stage analysis showed that applying DGBL on students’ academic achievement in each subject matter is significantly better than the effect of applying Non-DGBL on students’ academic achievement in each subject matter, secondary moderator analysis showed that applying DGBL on Europe students’
academic achievement in language, math, health and PE, and computer science is not significantly better than the effect of applying Non-DGBL.
(2) Europe and educational level
As Table 4.12 shows, research results showed that different educational level is the reason to cause the heterogeneity of academic achievement in Europe. Both senior high and university weighted mean ES are positive and 95% CI are significantly different from zero, meaning that in senior high and university, the effect of applying DGBL on Europe students ’ academic achievement is significantly better than the effect of applying Non-DGBL on Europe students’
academic achievement.
Educational level (QB = 16.58, p<.05) achieved significant level, meaning that educational level is the moderator of Europe. Senior high (ES=.58) and university (ES=.40) have positive effect, while junior high and elementary is not having significant effect. There is no significant difference between senior high and university.
While first stage analysis showed that applying DGBL on students’ academic achievement in each educational level is significantly better than the effect of applying Non-DGBL on students’ academic achievement in each educational level, secondary moderator analysis showed that applying DGBL on Europe students ’ academic achievement in junior high and elementary is not significantly better than the effect of applying Non-DGBL on Europe students’
academic achievement in junior high and elementary.
(3) Europe and digital game category
As Table 4.12 shows, research results showed that excluding role-playing games (K=0), except strategy games, puzzle games and simulation games weighted
mean ES are positive and 95% CI are significantly different from zero, meaning that in puzzle games and simulation games, the effect of applying DGBL on Europe students’academic achievement is significantly better than the effect of applying Non-DGBL on Europe students’ academic achievement.
Digital game category (QB =4.88, p> .05) is insignificant, meaning that digital game category is not the moderator of Europe.
While first stage analysis showed that applying DGBL on students’ academic achievement in each digital game category is significantly better than the effect of applying Non-DGBL on students’ academic achievement, secondary moderator analysis showed that excluding role-playing games (k=0), applying DGBL by strategy games on Europe students’academic achievement is not significantly better than the effect of applying Non-DGBL on Europe students’
academic achievement.
Table 4.12
Europe Secondary Moderator Cross Analysis
Note. *: significant at the .05 level; d+:
weighted mean effect size; k: number of articles; #: d
+positive and 95% CI significantly different from zero; (): post hoc.
2.3.6 North America group on academic achievement (1) North America and subject matter
As Table 4.13 shows, research results showed that different subject matter is the reason to cause the heterogeneity of academic achievement in North America.
Math, nature and science, health and PE weighted mean ES are positive and 95% CI are significantly different from zero, meaning that in math, nature and science, health and physical education, the effect of applying DGBL on North America students’ academic achievement is significantly better than the effect of applying Non-DGBL on North America students’academic achievement.
Subject matter (QB = 35.55, p<.05) achieved significant level, meaning that
Digital game categories 4.88
174. Puzzle games 9 .16#
6. Simulation games 6 .34#
8. Strategy games 2 .43
subject matter is the moderator of North America. Post hoc showed that health and PE (ES=.90) and nature and science (ES=.91) are significantly higher than math (ES=.48).
While first stage analysis showed that language and computer science are significantly better than the effect of applying Non-DGBL on students’
academic achievement, secondary moderator analysis showed that in North America, applying DGBL in language and computer science are not significantly better than the effect of applying Non-DGBL.
(2) North America and educational level
As Table 4.13 shows, research results showed that different educational level is the reason to cause the heterogeneity of academic achievement in North America. Each educational level weighted mean ES is positive and 95% CI is significantly different from zero, meaning that in elementary, junior high, senior high and university, the effect of applying DGBL on North America students’
academic achievement is significantly better than the effect of applying Non-DGBL on North America students’ academic achievement.
Educational level (QB = 52.74, p<.05) achieved significant level, meaning that educational level is the moderator of North America. Post hoc showed that junior high (ES=.89), senior high (ES=.92) and university (ES=.80) are significantly higher than elementary (ES=.26). Compare secondary moderator analysis with first stage moderator analysis, the result is consistent.
(3) North America and digital game category
As Table 4.13 shows, research results showed that different digital game category is the reason to cause the heterogeneity of academic achievement in North America. Except role-playing games, each digital game category weighted mean ES is positive and 95% CI is significantly different from zero, meaning that in puzzle games, simulation games and strategy games, the effect of applying DGBL on North America students’academic achievement is significantly better than the effect of applying Non-DGBL on North America students’academic achievement.
Digital game category (QB = 42.38, p<.05) achieved significant level, meaning that digital game category is the moderator of North America. Post hoc showed that when applying DGBL on North America, simulation games (ES=.90) is significantly higher than puzzle games (ES=.59) and strategy games (ES=.46) while role-playing games is not having significant effect.
While first stage analysis showed that applying DGBL on students’ academic achievement in each digital game category is significantly better than the effect of applying Non-DGBL on students’ academic achievement, secondary moderator analysis showed that applying DGBL by role-playing games on North America students’academic achievement is not significantly better than the effect of applying Non-DGBL.
Table 4.13
5. Health & Physical Education 4 .90#
6. Computer Science 2 1.02
7. others 7 .82#
Educational level (2,3,4>1) 52.74* 29
1.Elementary 4 .26#
2.Junior High 4 .89#
3.Senior High 3 .92#
4.University 18 .80#
Digital game categories (6>4,8) 42.38* 30
4. Puzzle games 12 .59#
5. Role-playing games 2 .04
6. Simulation games 9 .90#
8. Strategy games 7 .46#
Note. *: significant at the .05 level; d+:
weighted mean effect size; k: number of articles; #: d
+positive and 95% CI significantly different from zero; (): post hoc.
2.3.7 Puzzle games group on academic achievement (1) Puzzle games and subject matter
As Table 4.14 shows, research results showed that different subject matter is the reason to cause the heterogeneity of academic achievement in puzzle games.
As Table 4.14 shows, research results showed that different subject matter is the reason to cause the heterogeneity of academic achievement in puzzle games.