Initial word tests contribute to effective human anatomy education

Abstract

Dissection‐based education should be improved in response to persistent concerns regarding its quality and the impact on student performance. We hypothesized that a word test implemented during the initial stage of a dissection course would be an effective intervention for dissection‐based education, especially for struggling students. From 2013 to 2017, 25 basic line drawings with English anatomical terms were distributed to 640 second‐year medical students at Akita University, Japan, for dissection preparation. The word test was designed with anatomical terms and images and was conducted in the second week of dissections; it contained 10 line drawings and asked students to provide 30 English anatomical terms. The scores were returned immediately. Practical examinations were administered at mid‐term and on the last day of the course. The word test scores were positively correlated and significantly predicted the practical examination scores. Friedman test rankings for students with low word test scores (< 60%) significantly improved through practical examinations. These positive effects may be attributable to several factors, including the design with anatomical terms and images, early implementation, providing students with appropriate information for dissection preparation, and immediate feedback on scores. The effects are discussed through the lens of adult learning theory. The word test could help learners understand how to study anatomy and motivate them for dissections, and help teachers support struggling students. The word test may be effective for anatomy education in the context of reduced curriculum hours.

INTRODUCTION

Anatomy education has long been an important and fundamental element of medical education, not only for developing anatomical knowledge of the human body but also for obtaining three‐dimensional and spatial abilities, learning about death and dying, and fostering both the temperament of young medical students and the professionalism of medical doctors. ¹ , ² , ³ Students' reflections about death can be initiated through in‐person education with real human body donors or human samples. ⁴ , ⁵ Additionally, research indicates that most students value dissection practice. ⁶ , ⁷

The number of curriculum hours for anatomy education, especially for anatomical dissection, was reduced considerably in many countries because of the paucity of qualified educators, increasing cost to keep human body donors, and expansion of modern academic fields. ⁸ , ⁹ , ¹⁰ , ¹¹ , ¹² , ¹³ Reduced curriculum hours raised concerns that the quality of anatomy education would decline, leading to worse student performance. In response, institutions introduced various innovative educational methods to maintain the quality of anatomy education, including introductory videos for dissection, ¹⁴ , ¹⁵ using computers at dissection tables, ¹⁶ online web tutorials, ¹⁷ group discussions, ¹⁸ integrations with other disciplines, ¹⁹ and collaborations with clinical medicine. ²⁰ , ²¹ Despite reduced curriculum hours, students completed their instruction in anatomy through a modified curriculum. Overall, student performance was maintained or improved because new educational methods were developed, including online education and flipped classroom, which are comparable to traditional education. ²² , ²³ , ²⁴ Student satisfaction decreased in the first year of curriculum reform but recovered several years later. ¹⁹ , ²⁰ Students could learn anatomy more effectively using various educational methods. ²⁵ , ²⁶ , ²⁷

Although institutions seemed to manage effectively by modifying their curriculum and anatomy content, there were concerns about reduced curriculum hours for dissection. ²¹ , ²⁸ According to Shin et al., ²⁸ 117 anatomical course directors in the United States (US) indicated that they perceived the weakest aspects of the anatomy course to be insufficient time dedicated to dissection (23.1%) and insufficient total course time (15.1%). With reduced curriculum hours, student performance suffers. ²¹ , ²⁵ , ²⁸ Attention should be paid to struggling students and additional assistance should be provided during the curriculum reform transition period, even if overall mean scores of the cohort do not change significantly. These studies also indicate that several interventions are necessary for dissection‐based anatomy education due to reduced curriculum hours. ²¹ , ²⁸

To effectively improve students' anatomical knowledge, we propose identifying the students at risk of failure as early as possible in the dissection laboratory through the administration of a simple word test. The objective of this study was to investigate the usefulness of the word test for struggling students in dissection‐based anatomy education. To achieve this goal, the authors investigated how much the initial word tests affected the practical examinations for effective dissection and then further analyzed the change in performance for low‐scoring students, as identified by the initial word test. We hypothesized that, for dissection‐based anatomy education, a word test would be an effective intervention during the initial stage of the dissection course.

MATERIALS AND METHODS

Composition of medical student cohorts

Most of the medical students from the 2013–2017 cohorts entered the university after high school. A small number of students entered medical school with a bachelor's degree. Most of them were Japanese, however, several permanent residents who graduated from Japanese high schools were also included. The number of students in the 2013–2017 cohorts was 127, 127, 126, 134, and 126, respectively, with a total of 640 students, all of whom completed the anatomical dissection course. The percentage of female students was 39.4%, 42.5%, 42.1%, 40.3%, and 38.1%, respectively, for the 2013–2017 cohorts, which were not significantly different as determined by the Chi‐square test of independence. A dissection group typically comprised four students who participated simultaneously. Therefore, the maximum number of groups per cohort was 33. Each dissection group was formed based on their rankings from the previous year and included a high‐ranking student, two middle‐ranking students, and one low‐ranking student. Each dissection group also included two or three males and one or two females.

Human anatomy curriculum at Akita University

The human anatomy course ran during the second year of medical school from early April until early July as a stand‐alone subject, for 9 weeks in total. The dissections were performed 4 days a week from 12:50 to 18:30. During the ninth week of the course, five dissections were added in the morning, for a total of 41 dissections for the entire body. The course consisted of 203 education hours, including classroom lectures for 39 h, dissection in the laboratory for 156 h, and two practical examinations for 8 h.

The dissection content was similar to that found in Grant's Dissector. ²⁹ The dissection began with the nuchal and back regions, and the flexor surface of the lower extremities was examined. After turning over the body donor, the superficial part of the cervix and trunk was dissected. Then, the thoracic and abdominal organs and the four extremities were dissected. The pelvis was opened by bisecting the pubic symphysis and cutting through the ilium. Next, the organs in the pelvic cavity were removed. ³⁰ In the latter half of the anatomy course, the face and head were dissected. In the anatomy of the orbit, the ciliary ganglion was observed laterally. ³¹ The inner and middle ear were dissected using hobby rooters to examine fine structures. ³²

Typically, each anatomy dissection overlapped on two different body regions. In each dissection group, the two areas were assigned two by two. At the end of each dissection, the students explained to each other the areas they dissected.

Students could independently ask teachers questions for self‐directed learning, and teachers valued students' autonomy regarding these questions. Teachers supported dissection groups to align their progress of dissection procedure, especially when body donors had problem areas to be dissected, for example, excessive subcutaneous fatty tissue, removal of organs with postoperative adhesions, or severe contracture of joints.

Comparison of dissection education between Akita University and other institutions

The dissection curriculum at Akita University was compared with that of other institutions. The 156 laboratory (dissection) hours at Akita University were greater than that for other anatomy programs: 81 h at Indiana University, 55 h at the University of Alabama, 104 h at Rush University, and 83 h at the Mayo Clinic College. ³³ , ³⁴ According to McBride and Drake, ¹⁰ the average laboratory and lecture hours for anatomy were 76 and 51, respectively, in the US. Additionally, the number of students in a dissection group at Akita University was four, compared to four to six in the aforementioned institutions.

Word test design

The word test was designed to test two factors: identify basic human structures and understanding of anatomical terms. For students' preparation, simple anatomical images with anatomical terms were printed and delivered to every student at the beginning of the dissection course or before spring break (Figure 2, Supplement 1). The printout contained 25 line drawings covering the entire human body with approximately 200 basic anatomical terms, with each pointing to the corresponding structure with arrows. These images and anatomical terms were originally selected from the content of a 1994 dissection course by the corresponding author. The TERMINOLOGIA ANATOMICA JAPONICA ³⁵ was used to determine the standard for the anatomical terms because it contains Japanese and English anatomical terms in accordance with Terminologia Anatomica (1998).

FIGURE 2.

A schematic diagram showing schedules of dissection, the word test, and two practical examinations. Printouts with 25 line drawings accompanied by English anatomical terms were distributed before the start of dissection. The word test was administered in the second week of the dissection course. After the word test was administered, dissection sessions continued for approximately 7 weeks.

To clearly indicate to students the goals of the two factors, the word test was conducted during the initial stage of each dissection course (Figures 1 and 2). The word test included appropriate profiles for supplying students with the anatomical terms in advance and giving them feedback on their test scores as an incentive to improve their dissection practice.

FIGURE 1.

Four samples of line drawings used in the word tests. Upper left: Aorta and related organs; Upper right: Female pelvic organs; Lower left: Eye ball and related structures; Lower right: Heart and blood vessels. The full set of 25 line drawings is presented in Supplement 1. In the word tests, 10 of the 25 line drawings were used to prompt 30 anatomical terms, which students filled in on the drawings. Correct anatomical terms with accurate English spelling were considered correct, and the total number of correct terms was the word test score.

Although students had typically begun dissecting the back and flexion side of the leg, the word test covered the entire human body. The corresponding author selected approximately 10 images from the original prints and 30 corresponding anatomical terms with pointing arrows for each cohort (Figure 1). Undergraduate medical students need to know English medical terms for future clinical learning. Students were asked to provide the exact anatomical terms with the correct spellings within 15 min. No validation by external institutions or a pilot study has been done previously. Four anatomy teachers double‐scored the answer sheets, which were returned to students as early as possible for review purposes and checking simple grading errors. Word test scores were used as a formative assessment and contributed a 10% weighting to students' summative assessments.

Practical examinations

A day before the practical examinations, students had the opportunity to review the anatomical structures by examining all body donors dissected by their cohort. The number of tables used as examination stations was adjusted to 33. On examination day, all anatomy instructors were involved in the question setting with a tag (or a pin) on each body donor. Questions were selected for structures that were without damage, indicating a proper topographical anatomy dissected carefully by students. Most were identification questions. However, the test occasionally included questions about functions of the structures and the names of nerves innervated to the muscles (Table 1). In each cohort, two practical examinations were performed: Practical Examination 1 was administered at mid‐term in June, and Practical Examination 2 was administered on the last day of the dissection course (Figure 2). The two practical examinations had 66 questions for each cohort. Practical Examination 1 could not cover all body regions.

TABLE 1.

Questions and item difficulties for practical examinations from 2013–2017.

(a‐1)
Practical Examination 1			Practical Examination 2
Date	May 20, 2013		Date	June 12, 2013
Station number	Question	Item difficulty	Station number	Question	Item difficulty
1	Thyrocervical trunk	0.402	1	Femoral nerve	0.622
2	Deep temporal artery	0.150	2	Posterior cricoarytenoid muscle	0.205
3	Rhomboid major muscle	0.315	3	Coracoacromial ligament	0.236
4	Serratus posterior inferior muscle	0.339	4	Vertebral artery	0.685
5	Auriculotemporal nerve	0.142	5	Tibial collateral ligament	0.512
6	Superior laryngeal nerve	0.157	6	Superior gluteal artery	0.740
7	Superior labial artery (branch)	0.354	7	Median sacral artery	0.465
8	Spermatic cord	0.417	8	Levator veli palatini muscle	0.449
9	Proper hepatic artery	0.457	9	Ampulla (of uterine tube)	0.071
10	Inferior vena cava	0.630	10	Obturator internus muscle	0.449
11	Frontal nerve	0.165	11	Thoracic duct	0.165
12	Vertebral artery	0.205	12	Superior mesenteric artery	0.622
13	Inferior (pharyngeal) constrictor muscle	0.213	13	Pronator teres muscle	0.693
14	Great (or long) saphenous vein	0.795	14	Subscapular nerve	0.181
15	Transversus thoracis muscle	0.126	15	Annular ligament of radius	0.213
16	Median nerve	0.173	16	Dorsalis pedis artery(dorsal artery of foot)	0.205
17	Right pulmonary artery	0.598	17	Fibularis (or peroneus) longus muscle	0.134
18	Superior gluteal nerve	0.394	18	Extensor pollicis longus muscle	0.638
19	Third cervical nerve	0.236	19	Lateral circumflex femoral artery	0.283
20	Latissimus dorsi muscle	0.528	20	Lumbosacral trunk	0.724
21	Hyoglossus muscle	0.197	21	Piriformis muscle	0.472
22	Biceps femoris muscle (short head)	0.417	22	Styloid process	0.575
23	Mitral valve	0.417	23	Opponens pollicis muscle	0.244
24	Transversus abdominis muscle	0.543	24	Salpingopharyngeal fold	0.142
25	Longus capitis muscle	0.118	25	Superior suprarenal artery	0.213
26	Sciatic nerve	0.827	26	Infraorbital nerve	0.346
27	Pyramidalis muscle	0.094	27	Tibialis posterior muscle	0.386
28	Left testicular vein	0.276	28	Seminal vesicle	0.480
29	Inferior epigastric artery	0.102	29	Renal column	0.094
30	Levator scapulae muscle	0.630	30	Hypophysis (or pituitary gland)	0.378
31	Peyer's patch	0.409	31	Ulnar nerve	0.835
32	Abducens nerve	0.441	32	Inferior epigastric artery	0.197
33	Middle lobar bronchus	0.142	33	Nerve of pterygoid canal	0.465

(a‐2)
Related table	Cohort	N	Mean	Standard deviation	Standard error	Significance level	Lower bound	Upper bound
(a‐1)	2013	66	0.372	0.209	0.026	0.05	0.321	0.423

(b‐1)
Practical Examination 1			Practical Examination 2
Date	June 10, 2014		Date	July 11, 2014
Station number	Question	Item difficulty	Station number	Question	Item difficulty
1	Inferior oblique muscle	0.559	1	Geniohyoid muscle	0.362
2	Lingual nerve	0.425	2	Lesser palatine nerve	0.331
3	Short gastric artery	0.252	3	Hemiazygos vein	0.638
4	Submandibular ganglion	0.063	4	Internal carotid artery	0.260
5	Superior cervical ganglion	0.425	5	Femoral nerve	0.724
6	Ileocolic artery	0.079	6	Testicular artery	0.803
7	Coronoid process	0.236	7	Anterior tibial recurrent artery	0.031
8	Phrenic nerve	0.646	8	Levator veli palatini muscle	0.535
9	Vertebral artery	0.339	9	Thoracic duct	0.244
10	Levator scapulae muscle	0.638	10	Adductor longus muscle	0.811
11	Rectouterine pouch	0.110	11	Uterine tube	0.307
12	Semitendinosus muscle	0.402	12	Suprascapular nerve	0.535
13	Pyloric sphincter muscle	0.520	13	Anterior cruciate ligament	0.646
14	Spinous process of axis	0.220	14	Renal sinus	0.157
15	Circular fold	0.606	15	Obturator externus muscle	0.323
16	Cephalic vein	0.307	16	Ulnar nerve	0.244
17	Right coronary artery	0.039	17	Peroneus (or fibularis) longus muscle	0.134
18	Serratus anterior muscle	0.228	18	Inferior epigastric artery	0.307
19	Inferior epigastric artery	0.488	19	Posterior cricoarytenoid muscle	0.291
20	Transverse cervical artery	0.512	20	Posterior cord	0.157
21	Apical segmental bronchus	0.000	21	Semilunal hiatus	0.291
22	Gluteus minimus muscle	0.394	22	Greater splanchnic nerve	0.402
23	External branch of superior laryngeal nerve	0.055	23	Extensor carpi radialis longus muscle	0.181
24	Internal oblique muscle	0.654	24	Corona mortis	0.606
25	External carotid artery	0.575	25	Deep peroneal (or fibular) nerve	0.465
26	Common peroneal nerve	0.189	26	Transversus abdominis muscle	0.654
27	Inferior laryngeal nerve	0.142	27	Lingual nerve	0.449
28	Multifidus muscle	0.236	28	Obturator internus muscle	0.480
29	Thyrohyoid membrane	0.614	29	Sacrospinous ligament	0.236
30	Trapezius muscle	0.370	30	Lumbosacral trunk	0.669
31	Fossa ovalis	0.528	31	Ganzer muscle	0.079
32	Lesser occipital nerve	0.425	32	Internal intercostal muscle	0.205
33	Zygomaticus major muscle	0.102	33	Posterior longitudinal ligament	0.220

(b‐2)
Related table	Cohort	N	Mean	Standard deviation	Standard error	Significance level	Lower bound	Upper bound
(b‐1)	2014	66	0.366	0.209	0.026	0.05	0.315	0.417

(c‐1)
Practical Examination 1			Practical Examination 2
Date	June 9, 2015		Date	July 10, 2015
Station number	Question	Item difficulty	Station number	Question	Item difficulty
1	Portal vein	0.762	1	Inferior epigastric artery	0.405
2	Transverse cervical artery	0.421	2	Thoracic duct	0.341
3	Ileocolic artery	0.333	3	Genitofemoral nerve	0.437
4	Middle lobar bronchus	0.452	4	Frontal sinus	0.405
5	Auriculotemporal nerve	0.278	5	Corona mortis	0.556
6	Anza cervicalis	0.643	6	Pudendal nerve	0.159
7	Pulmonary valve	0.548	7	Stellate ganglion	0.286
8	Anterior interventricular branch	0.397	8	Levator ani muscle	0.063
9	Cystic artery	0.341	9	Achilles tendon	0.738
10	Facial artery	0.754	10	Ductus (or vas) deference	0.603
11	(Left) hepatic vein	0.317	11	Iliacus muscle	0.119
12	Inferior rectus muscle	0.397	12	Palatopharyngeus muscle	0.095
13	Internal oblique muscle	0.500	13	Obuturator externus muscle	0.310
14	Suprascapular artery	0.095	14	Hypoglossal nerve	0.341
15	Common peroneal (or fibular) nerve	0.206	15	Axillary nerve	0.413
16	Pectoralis minor muscle	0.762	16	Malleus	0.579
17	Spinous process of axis	0.794	17	Carotid (sinus) branch	0.063
18	Sacrotuberous ligament	0.246	18	Inferior nasal concha	0.611
19	Vertebral artery	0.270	19	Ulnar nerve	0.246
20	Intercostobrachial nerve	0.111	20	Left testicular vein	0.532
21	Serratus anterior muscle	0.071	21	Posterior longitudinal ligament	0.643
22	Middle cervical ganglion	0.095	22	Levator veli palatini muscle	0.524
23	10th intercostal nerve	0.095	23	Radial nerve	0.444
24	Mylohyoid muscle	0.405	24	Ureter	0.563
25	Levator scapulae muscle	0.675	25	Tibialis anterior muscle	0.476
26	Ligamentum arteriosum	0.389	26	Ovary	0.222
27	3rd cervical nerve	0.341	27	Popliteal artery	0.198
28	Stylohyoid muscle	0.397	28	Thyrocervical trunk	0.667
29	Small (or short) saphenous vein	0.563	29	Obturator internus muscle	0.563
30	Mitral valve	0.373	30	Vestibular fold	0.579
31	Gluteus minimus muscle	0.659	31	Coracobrachialis muscle	0.595
32	Trochlear nerve	0.270	32	Maxillary nerve	0.151
33	Vagus nerve	0.452	33	Vertebral artery	0.667

(c‐2)
Related table	Cohort	N	Mean	Standard deviation	Standard error	Significance level	Lower bound	Upper bound
(c‐1)	2015	66	0.409	0.203	0.025	0.05	0.359	0.459

(d‐1)
Practical Examination 1			Practical Examination 2
Date	June 8, 2016		Date	July 8, 2016
Station number	Question	Item difficulty	Station number	Question	Item difficulty
1	Submandibular ganglion	0.112	1	Tensor veli palatini muscle	0.791
2	(Left) pulmonary vein	0.478	2	Corona mortis	0.731
3	Middle meningeal artery	0.709	3	Ampulla (of uterine tube)	0.627
4	Superior gluteal nerve	0.060	4	Corpus cavernosum penis	0.701
5	Vagus nerve	0.910	5	Genitofemoral nerve	0.522
6	Thoracic duct	0.507	6	Palatoglossus muscle	0.657
7	Thyrocervical trunk	0.634	7	Extensor digitorum brevis muscle	0.545
8	Nasociliary nerve	0.306	8	Hemiazygos vein	0.851
9	Femoral vein	0.642	9	Posterior cricoarytenoid muscle	0.530
10	Trochlear nerve	0.634	10	Lateral meniscus	0.201
11	Sternocleidomastoid muscle	0.970	11	Anterior interosseous artery	0.425
12	Sciatic nerve	0.963	12	Maxillary sinus	0.366
13	Root of second cervical nerve	0.127	13	Radial nerve	0.672
14	Common bile duct	0.493	14	Pituitary gland	0.948
15	Sacrotuberous ligament	0.724	15	Levator ani muscle	0.284
16	Aortic valve	0.716	16	Adductor pollicis muscle	0.828
17	Vertebral artery	0.313	17	Pronator teres muscle	0.761
18	Superficial inguinal ring	0.627	18	Anterior semicircular canal	0.597
19	Internal oblique muscle	0.746	19	1st lumbar nerve	0.179
20	Pectoralis minor muscle	0.925	20	Pectineus muscle	0.731
21	Pericardiacophrenic artery	0.015	21	Dorsalis pedis artery	0.634
22	Ileal orifice	0.157	22	Lumbosacral trunk	0.694
23	Gluteus minimus muscle	0.858	23	Superior gluteal artery	0.821
24	Serratus anterior muscle	0.201	24	Ethmoidal sinus (or cell)	0.448
25	Deep cervical artery	0.007	25	Vertebral artery	0.925
26	Gastroduodenal artery	0.567	26	Ulnar nerve	0.590
27	Common peroneal (or fibular) nerve	0.470	27	Innermost intercostal muscle	0.351
28	Left gastric artery	0.172	28	Sacrospinous ligament	0.530
29	Inferior epigastric artery	0.590	29	Superior suprarenal artery	0.672
30	Great (or long) saphenous vein	0.784	30	Popliteal artery	0.791
31	Sigmoid sinus	0.104	31	Cisterna chyli	0.769
32	Testicular artery	0.493	32	Axillary nerve	0.604
33	Facial nerve	0.060	33	Right ovarian vein	0.701

(d‐2)
Related table	Cohort	N	Mean	Standard deviation	Standard error	Significance level	Lower bound	Upper bound
(d‐1)	2016	66	0.554	0.262	0.032	0.05	0.489	0.618

(e‐1)
Practical Examination 1			Practical Examination 2
Date	June 14, 2017		Date	July 14, 2017
Station number	Question	Item difficulty	Station number	Question	Item difficulty
1	Auriculotemporal nerve	0.540	1	Coracoid process	0.857
2	Circular fold	0.960	2	Interosseous membrane of forearm	0.476
3	Superior thyroid artery	0.817	3	Inferior phrenic artery	0.794
4	Cystic artery	0.825	4	Transverse arytenoid muscle	0.143
5	Thoracic duct	0.056	5	Lateral femoral cutaneous nerve	0.770
6	Vertebral artery	0.722	6	Flexor halluces longus muscle	0.603
7	External oblique muscle	0.563	7	Renal pelvis	0.381
8	Sympathetic trunk	0.389	8	Lateral circumflex femoral artery	0.556
9	Nasociliary nerve	0.548	9	Inferior constrictor muscle	0.310
10	Digastric muscle	0.563	10	Dorsal interosseous muscle	0.302
11	Azygos vein	0.722	11	Interlobar artery	0.024
12	(Right) recurrent laryngeal nerve	0.857	12	Inferior gluteal artery	0.500
13	Inferior epigastric artery	0.841	13	Ulnar nerve (deep branch)	0.730
14	T11	0.032	14	Lumbosacral trunk	0.802
15	Great (or long) saphenous vein	0.849	15	Facial nerve	0.738
16	Inferior mesenteric artery	0.952	16	Radial nerve	0.817
17	Levator scapulae muscle	0.635	17	Median nerve	0.825
18	Superior gluteal nerve	0.794	18	Stellate ganglion	0.603
19	C3	0.492	19	Fibular (or peroneus) tertius muscle	0.365
20	Internal thoracic artery	0.865	20	Vertebral artery	0.937
21	Pectoralis major muscle	0.968	21	Obturator internus muscle	0.460
22	Small (or short) saphenous vein	0.698	22	Popliteal artery	0.690
23	Greater occipital nerve	0.333	23	Adductor brevis muscle	0.389
24	Great cardiac vein	0.429	24	Greater splanchnic nerve	0.548
25	Suprascapular artery	0.611	25	Palatine tonsil	0.016
26	Perforating artery	0.238	26	Anterior talofibular ligament	0.040
27	Superficial inguinal ring	0.770	27	Left ovarian vein	0.825
28	Sartorius muscle	0.381	28	Epiglottis	0.532
29	Pyloric sphincter muscle	0.762	29	(Posterior) intercostal artery	0.532
30	Gemellus inferior muscle	0.881	30	Lateral sacral artery	0.675
31	Rectouterine pouch	0.587	31	Medial circumflex femoral artery	0.444
32	Circumflex branch	0.437	32	Pterygopalatine ganglion	0.524
33	Ciliary ganglion	0.349	33	Middle rectal artery	0.611

(e‐2)
Related table	Cohort	N	Mean	Standard deviation	Standard error	Significance level	Lower bound	Upper bound
(e‐1)	2017	66	0.580	0.250	0.031	0.05	0.530	0.642

Note: Tables (a‐1)–(e‐1) show the questions and item difficulties for Practical Examinations 1 and 2 for each cohort from 2013 to 2017, respectively. Tables (a‐2)–(e‐2) show the statistical data for each cohort from 2013 to 2017, respectively, including 95% confidence intervals with lower and upper bounds. Station number; the number of the dissection table on which a cadaver was placed with a pin or tag as a question; N: the total number of questions for each cohort.

For each practical examination, participants were divided into four groups. In each cohort, the first group of 33 students entered the dissection laboratory, positioning themselves individually beside each examination station where a body donor was lying. They were allotted 1 min to answer each question, moving to the next dissection table after the sound of a bell. After answering all 33 questions, the students moved again in the same direction, according to the bell every 30 s, providing them the opportunity to reconfirm their answers. Afterward, they immediately left the dissection laboratory and moved into the next room where other body donors were stored. While the first group was taking the test, the other three groups were waiting in the lecture room. Immediately after the first group left the dissection laboratory, the second group entered it and took the practical examination in the same way. After the second group finished, the first and second groups left the storage room or the dissection laboratory through opposite doors. Then, the third and fourth groups entered the dissection laboratory for the test, following the same process as the first and second groups. After the fourth group finished the exam, all groups were invited to the dissection laboratory again to review the questions. Mobile devices were prohibited, as was communication among students during the examination. All authors continuously monitored students in the laboratory and storage room to prevent any communication among them.

Outcome analysis

All student score data were obtained from the cohorts of 2013 through 2017. The authors started planning this research in 2018, establishing the study hypothesis and purpose. In other words, when collecting data from 2013 to 2017, both teachers (observers) and students (objectives) were unaware of the study hypotheses and purpose. Therefore, this study could be regarded as double‐blinded research in terms of data collection. Data included the scores of the initial word test and two practical examinations in each cohort. They were recognized as objective data, with each term contained in the Terminologia Anatomica Japonica. ³⁵ All authors attended to questions regarding the practical examinations, and all double‐scored the answer sheets to avoid differences among observers. Multiple authors collaborated to avoid observer bias in data processing and statistical analyses.

For word test scoring, one point was given for each correctly spelled answer and zero points for each incorrect answer or incorrectly spelled answer (Supplement 2a‐e). For Practical Examination 1 and 2 scoring, each question with correct identification were recorded as “1” in an Excel file. Four anatomy teachers double‐scored the answer sheets. The answer sheets were then returned to students for confirmation. The item difficulty of each question is shown in Table 1. The statistics for student scores with correct English spellings, including means, standard deviations, standard errors, and 95% confidence intervals, are provided in Supplement 2. All second‐year medical students took part in the word test and two practical examinations, with all scores being used for the statistical analyses.

At each practical examination, participants were divided into four groups depending on their ID numbers and took the exams separately. Scores of these four groups in Supplement 2 were analyzed using analysis of variance (ANOVA), and multiple comparisons when necessary. The correlation indexes were obtained among three scores of the word test and Practical Examinations 1 and 2 (Supplement 2). Single and multiple regression analyses were performed among the three scores (Supplement 2). Students were divided into two groups with high word test scores (18 points, 60% or higher) and low word test scores (less than 18 points), respectively. For students with low scores, regression analyses were executed. The three scores of the word test and Practical Examinations 1 and 2 were standardized for each student group with low or high word test scores, respectively. A Friedman test of these three scores was performed for students with low or high word test scores in each cohort.

The appropriate statistical test was selected depending on the data structures. SPSS (Version 17.0) was used for most statistical analyses, including Cronbach's alphas, correlation coefficients, ANOVA, linear regression analyses, and Friedman tests. Excel (Microsoft) and “BellCurve for Excel” (Social Survey Research Information Co., Ltd., Tokyo) were also used for correlation coefficients and regression analysis. The EZR (Saitama Medical Center, Jichi Medical University, Japan) was used for making the regression planes for multiple regression analyses.

RESULTS

Word test

Correct English spellings were required for anatomical terms. The number of students with low word test scores, defined as less than 60% correct, was 70, 49, 41, and 26 for the 2013, 2014, 2016, and 2017 cohorts, respectively (Supplement 2a, b, d, and e). In 2015, the word test was not administered. The scores on the word test were a continuous variable.

Practical examinations

Internal consistency

Cronbach's alpha was obtained for the 66 questions from Practical Examinations 1 and 2 in each cohort: 0.886 for 2013; 0.862 for 2014; 0.841 for 2015; 0.809 for 2016; and 0.811 for 2017. All cohorts had coefficients greater than 0.8, demonstrating adequate reliability of the practical examination. ³³ , ³⁶

Differences in mean scores for the practical examinations

In each practical examination, participants were divided into four groups, and the test was repeated four times with the same questions on the same day. Anatomy teachers did not notice any communication among students. Additionally, to ensure that no communication occurred among students during each practical examination, the mean scores of the independent four groups were analyzed using ANOVA. All correct identifications, including English and Japanese answers, were used in this analysis. When a significant difference (p < 0.05) was obtained in the ANOVA, multiple comparisons were performed to identify the significant groups.

No significant differences were observed in the 2013, 2014, and 2016 cohorts, in Practical Examination 2 in 2015 and Practical Examination 1 in 2017 (Table 2a).

TABLE 2.

Analysis of variance (ANOVA) and multiple comparisons for mean scores among the four groups in each practical examination from 2013 to 2017.

(a) Summary of ANOVA at each practical examination from 2013 to 2017
Cohort	Practical Examination 1 or 2	Degree of freedom and F value	p value
2013	Practical Examination 1	F (3, 123) = 2.099	0.104
	Practical Examination 2	F (3, 123) = 1.087	0.357
2014	Practical Examination 1	F (3, 123) = 0.306	0.821
	Practical Examination 2	F (3, 123) = 0.473	0.702
2015	Practical Examination 1	F (3, 122) = 3.277	0.023
	Practical Examination 2	F (3, 122) = 2.276	0.083
2016	Practical Examination 1	F (3, 130) = 1.732	0.164
	Practical Examination 2	F (3, 130) = 1.754	0.159
2017	Practical Examination 1	F (3, 122) = 0.788	0.503
	Practical Examination 2	F (3, 122) = 2.952	0.035

(b‐1) Descriptive statistics in ANOVA in Practical Examination 1 in 2015
Groups	N	Mean	Standard deviation	Standard error	Lower bound	Upper bound	Minimum	Maximum
1st	32	12.22	5.735	1.014	10.15	14.29	3	26
2nd	32	15.63#**	4.492	0.794	14.01	17.24	6	24
3rd	31	12.65	4.309	0.774	11.06	14.23	3	24
4th	31	13.13	4.342	0.780	11.54	14.72	2	23
Total	126	13.41	4.891	0.436	12.55	14.28	2	26

(b‐2) Descriptive statistics in ANOVA in Practical Examination 2 in 2015
Groups	N	Mean	Standard deviation	Standard error	Lower bound	Upper bound	Minimum	Maximum
1st	32	12.84	4.833	0.854	11.10	14.59	5	25
2nd	32	15.41#	4.717	0.834	13.71	17.11	6	29
3rd	31	13.58	5.130	0.921	11.70	15.46	2	22
4th	31	12.52	4.618	0.829	10.82	14.21	3	19
Total	126	13.60	4.901	0.437	12.73	14.46	2	29

(b‐3) Multiple comparisons among four groups in Practical Examination 1 in 2015 by the Turkey method
		Mean difference	Standard error	Significance probability	Lower bound	Upper bound
	2nd group	−3.406	1.191	0.025**	−6.51	−0.30
1st group	3rd group	−0.426	1.200	0.985	−3.55	2.70
	4th group	−0.910	1.200	0.873	−4.04	2.22
	1st group	3.406	1.191	0.025**	0.300	6.51
2nd group	3rd group	2.980	1.200	0.068	−0.15	6.11
	4th group	2.496	1.200	0.166	−0.63	5.62

(c‐1) Descriptive statistics in ANOVA in Practical Examination 1 in 2017
Groups	N	Mean	Standard deviation	Standard error	Lower bound	Upper bound	Minimum	Maximum
1st	32	21.44#	4.295	0.759	19.89	22.99	12	29
2nd	32	20.13	4.626	0.818	18.46	21.79	7	28
3rd	31	20.42	4.209	0.756	18.88	21.96	12	29
4th	31	19.87	4.295	0.771	18.30	21.45	11	27
Total	126	20.47	4.350	0.388	19.70	21.24	7	29

(c‐2) Descriptive statistics in ANOVA in Practical Examination 2 in 2017
Groups	N	Mean	Standard deviation	Standard error	Lower bound	Upper bound	Minimum	Maximum
1st	32	19.75#**	3.943	0.697	18.33	21.17	13	28
2nd	32	17.03	4.177	0.738	15.53	18.54	10	24
3rd	31	17.74	5.341	0.959	15.78	19.70	7	26
4th	31	16.71	4.345	0.780	15.12	18.30	9	24
Total	126	17.82	4.581	0.408	17.01	18.63	7	28

(c‐3) Multiple comparisons among four groups in the practical examination 2 in 2017
		Mean difference	Standard error	Significance probability	Lower bound	Upper bound
	2nd group	2.719	1.119	0.077	−0.20	5.63
1st group	3rd group	2.008	1.128	0.288	−0.93	4.95
	4th group	3.040*	1.128	0.040**	0.10	5.98
	1st group	−3.040*	1.128	0.040**	−5.98	−0.10
4th group	2nd group	−0.322	1.128	0.992	−3.26	2.62
	3rd group	−1.032	1.137	0.801	−3.99	1.93

Note: (a) Summary of ANOVA at each practical examination from 2013 to 2017. A total of 120 or more students in each cohort were divided into four groups. They took the practical examinations in order and in the same question settings. Differences in the mean scores among the four groups were examined by ANOVA. Degree of freedom, F‐value, and p‐value are indicated for each practical examination. Significant differences were observed in Practical Examination 1 in 2015 (p = 0.023) and Practical Examination 2 in 2017 (p = 0.035). Table b‐1: Descriptive statistics in ANOVA in Practical Examination 1 in 2015. The Turkey method revealed that the second group was significantly higher than the first group (**p = 0.025) in Practical Examination 1 in 2015 (see Table b‐3). Note the second group obtained the highest mean scores in both practical examinations (# in Tables b‐1 and b‐2). Lower and Upper bound: 95% confidence interval for mean; Mean: mean number of correct identifications in the 33 questions of the practical examination; N: number of students in each group in Practical Examination 1. Table b‐2: Descriptive statistics in ANOVA in Practical Examination 2 in 2015. The student groups were the same as in Practical Examination 1 in 2015. The second group obtained the highest mean scores (^#). Table b‐3: Multiple comparisons among four groups in Practical Examination 1 in 2015 using the Turkey method. Table c‐1: Descriptive statistics in ANOVA in Practical Examination 1 in 2017. The first group obtained the highest mean scores (#). Table c‐2: Descriptive statistics in ANOVA in Practical Examination 2 in 2017. The student groups were the same as in Practical Examination 1 in 2017. The results of the Turkey method showed that the first group was significantly higher than the fourth group (**p = 0.040) for Practical Examination 2 in 2017 (see Table c‐3). Note that the first group obtained the highest mean scores in both practical examinations (# in Tables c‐1 and c‐2). Lower and Upper bound: 95% confidence interval for mean; Mean: mean number of correct identifications in the 33 questions of the practical examination; N: number of students in each group in Practical Examination 1. Table c‐3: Multiple comparisons among four groups in Practical Examination 2 in 2017.

A significant difference was observed in Practical Examination 1 in 2015 (F (3, 122) = 3.277, p = 0.023) (Table 2a). Multiple comparisons showed that the score of the second group was significantly higher than that of the first group (p = 0.025) (**in Table 2b‐1,b‐3). The mean score of each group did not increase as the practical examination proceeded. The mean score of the second group was the highest in both practical examinations in 2015 (# in Table 2b‐1,b‐2).

Similarly, a significant difference was observed in Practical Examination 2 in 2017 (F (3, 122) = 2.952, p = 0.035) (Table 2a). Multiple comparisons showed that the score of the first group was significantly higher than that of the fourth group (p = 0.040) (**in Table 2c‐2,c‐3). The mean score of each group did not increase as the practical examination proceeded. In the 2017 cohort, the mean score of the first group was highest in both practical examinations (# in Table 2c‐1,c‐2).

These results strongly suggest that the anatomical abilities of the second group in 2015 and the first group in 2017 were higher than those of the other groups in 2015 and 2017, respectively. If communication among students occurred during the practical examination, the mean scores would have increased for each subsequent group. However, the mean scores did not increase for the examinations conducted on the same day. Barring any communication among student groups during the practical examinations, there is no evidence to suggest that communication among students significantly impacted the group scores.

Correlation coefficient among the word test and practical examinations

For each cohort, correlation analyses were performed between the word test scores and scores of Practical Examinations 1 and 2, each of which was a continuous variable with a normal distribution (Figure 3). As no word test was conducted in 2015, the correlation analysis was not performed for that year.

FIGURE 3.

A schematic diagram showing correlation coefficients (Pearson) among scores of the word test and two practical examinations in 2013, 2014, 2016, and 2017. The coefficients between corresponding scores in each cohort are indicated beside each arrow. Every correlation coefficient was significant at the 0.001 level.

Positive correlation coefficients were obtained: between the word test scores and scores of Practical Examination 1 (from 0.506 to 0.650) and between the word test scores and scores of Practical Examination 2 (from 0.389 to 0.658) (Figure 3). Moderate to high correlation coefficients were obtained between the scores of Practical Examination 1 and those of Practical Examination 2 (from 0.593 to 0.814) (Figure 3).

Regression analysis of the word test and practical examinations

For each cohort, simple regression analyses were performed between the word test score as an independent variable and score of Practical Examination 1 as a dependent variable (Figure 4A,B, Supplements 3‐1(a) and 3‐2(a)). Additionally, multiple regression analyses were performed with the word test and Practical Examination 1 as two independent variables and score of Practical Examination 2 as a dependent variable (Figure 4A,B, Supplements 3‐1(b) and 3‐2(b)). As no word test was conducted in 2015, the regression analysis was not performed for that year.

FIGURE 4.

(A) A schematic diagram showing variables and the significance of regression analyses for all students in each cohort of 2013, 2014, 2016, and 2017. p‐values in the upper left represent the significance of regression equations in simple regression analyses with the word test as an independent variable and Practical Examination 1 as a dependent variable. Similarly, p‐values in the lower right represent the significance of regression equations in multiple regression analyses with both the word test and Practical Examination 1 as independent variables and Practical Examination 2 as a dependent variable. Details of statistical data are shown in Supplement 3‐1. (B) A schematic diagram showing variables and the significance of regression analyses for students with low word test scores in each cohort of 2013, 2014, 2016, and 2017. p‐values in the upper left represent the significance of regression equations in simple regression analyses with the word test as an independent variable and Practical Examination 1 as a dependent variable. Similarly, p‐values in the lower right represent the significance of regression equations in multiple regression analyses with both the word test and Practical Examination 1 as independent variables and Practical Examination 2 as a dependent variable. The equation in the multiple regression analysis was not significant in 2017 (p = 0.594). Details of the statistical data are shown in Supplement 3‐2.

All participants

The word test scores as independent variables significantly predicted the scores in Practical Examination 1 as dependent variables (p < 0.001). Both the word test and scores in Practical Examination 1 as independent variables significantly predicted the scores in Practical Examination 2 as the dependent variable (p < 0.001). The p‐values for each regression analysis were less than 0.001, indicating the significance of the regression equation (Figure 4A, Supplement 3‐1). A regression line and three‐dimensional regression plane with each scatter plot in 2017 were presented as examples of the simple and multiple regression analyses, respectively (upper graphs in Figure 5). Detailed regression analyses for all participants are presented in Supplements 3‐1, 4‐1, and 5‐1.

FIGURE 5.

Four graphs showing the results of regression analyses in 2017. Regression equations and coefficients of determination (R ²) are indicated. Upper left: A scatter plot with the regression line from the simple regression analysis for all students; Upper right: A 3D scatter plot with the regression plane from the multiple regression analysis for all students; Lower left: A scatter plot with the regression line from the simple regression analysis for students with low word test scores; Lower right: A 3D scatter plot with the regression plane from the multiple regression analysis for students with low word test scores; X‐axes: The word test scores; Y‐axes: The scores of Practice Examination 1 (Practical Exam 1); Z‐axes: The scores of Practical Examination 2 (Practical Exam 2). In the multiple regression equations, the small x and y represent two independent variables, and the small z represents dependent variables of the 3D regression planes. Details of the statistical data in 2017 are shown in Figure 4A and Supplement 3‐1 for all students, and in Figure 4B and Supplement 3‐2 for students with low word test scores.

Students with lower word test scores

In the 2013, 2014, 2016, and 2017 cohorts, students were divided into two groups: students with lower word test scores (below 60%) and those with higher word test scores (60% or higher). Similar regression analyses for students with lower word test scores (below 60%) were performed. The p‐values for each regression analysis were less than 0.05 for all four cohorts of students with lower word test scores, indicating statistical significance of the equations, except for the multiple regression analysis in 2017 (p = 0.594) (Figure 4B, Supplements 3‐2). A regression line and three‐dimensional regression plane with each scatter plot for 2017 were presented as examples of the simple or multiple regression analyses, respectively (lower graphs in Figure 5). Detailed regression analyses for students with lower word test scores are presented in the Supplements 3‐2, 4‐2, and 5‐2.

Thus, in the multiple regression analyses, the word test and Practical Examination 1 scores could not significantly predict the scores of Practical Examination 2 for the 2017 cohort of students with lower word test scores (Figure 4B, Supplement 3‐2(b)). However, the scores for the word test and Practical Examination 1 significantly predicted the scores of Practice Examination 2 in the remaining three cohorts.

Performance changes after the word test

As previously mentioned, the scores of the word tests and Practical Examination 1 did not predict the performance on Practical Examination 2 according to the multiple regression analyses for students with lower word test scores in 2017 (data in 2017 of Figure 4B). The performance of students with lower word test scores might have improved through dissection practice. Therefore, the scores of the word tests and two practical examinations were standardized, and non‐parametric Friedman tests were performed for the three tests for students with lower word test scores (< 60%) and those with higher word test scores (60% or higher), respectively. The change in ranks in the lower‐scoring student group was compared with that in the higher‐scoring student group.

Students with lower word test scores

The mean rankings of students with lower word test scores increased significantly from the word test score to the scores of Practical Examinations 1 and 2 (p < 0.01) in every cohort, suggesting that their low performance in the initial stage improved during the practical examinations (Figure 6A). These findings also demonstrate that the significant difference among four student groups of Practical Examination 2 in 2017 (Table 2a) did not affect the performance changes in students with lower word test scores.

FIGURE 6.

(A) A graph showing changes in mean rankings in Friedman tests in the word test and two practical examinations for students with low word test scores (below 60%) in each cohort of 2013, 2014, 2016, and 2017. The mean ranking significantly increased (p < 0.01) in the practical examinations of every cohort. (B) A graph showing changes in mean rankings in Friedman tests in the word test and two practical examinations for students with high word test scores (60% or higher) in each cohort of 2013, 2014, 2016, and 2017. The mean ranking significantly decreased in 2013 and 2014 (p < 0.01). In contrast, no significant difference was observed in 2016 and 2017 (p > 0.05).

Students with higher word test scores

In turn, the change in the performance of students with higher word test scores was analyzed using the Friedman test. Changes in their mean ranks and p‐values are summarized in Figure 6B. In the 2013 and 2014 cohorts, the mean rankings of the scores in the practical examinations decreased significantly from those of the word test scores (p < 0.01). In the 2016 and 2017 cohorts, no statistical significance was observed in the higher‐scoring students' mean rankings from the word test (p = 0.224, p = 0.065, respectively).

DISCUSSION

The study findings indicate that participation in human dissection improves the performance of students with low scores on an initial word test. That is, students who began the course with less knowledge benefited and had reduced chances of failing because of their exposure to dissection practice. Consequently, anatomy instructors could be less concerned about these students failing.

In contexts where there are reduced curriculum hours, innovative educational methods contribute to maintenance or improved student performance. ²² , ²³ , ²⁴ For example, Pei and Wu ²² demonstrated that online education was more effective than traditional in‐class education in undergraduate medical education in terms of knowledge and skill. Choi‐Lundberg et al. ²³ reported that students who accessed all dissection audio‐visual resources prior to dissection sessions as preparation obtained significantly higher scores in the practical examination than those who did not access the resources at all. Sadik and Abdulmonem ²⁴ demonstrated that there was no significant difference between the flipped classroom and traditional lectures in written anatomy examinations.

Nevertheless, there were concerns that student performance was maintained or improved at the cost of anatomy education quality, including dissection‐related knowledge. ²¹ , ²⁸ Zill ²¹ noted disadvantages regarding the lack of learning origins and insertions of skeletal muscles and arterial branches in clinical medicine. She also highlighted that the deleted content as a result of reduced curriculum hours was not reflected in student assessment and test scores. To address these concerns, dissection‐based anatomy education must be more effective than before the curriculum hours were reduced.

Research strengths

This study demonstrated that the word tests had a predictive function for middle and final practical examinations. In previous reports, Nwachukwu et al. ¹ revealed a significant correlation between the quality of dissection and scores of practical examinations. McNulty et al. ³⁷ showed that the practical quizzes performed 3 days prior to the summative examination were related to the final scores. Azzi et al. ³⁸ reported a strong correlation between midterm and final examinations, and a weak correlation between multiple formative quizzes and summative examinations.

This study demonstrated the value of students' initial performance. The performance of low‐scoring students identified in the initial word test significantly improved through the subsequent practical examinations. Specifically, the results highlight the importance of knowing initial performance and the roles of dissection practice afterward. Pizzimenti et al. ³⁹ reported that students in the bottom quartile demonstrated greater improvement than those in the higher three quartiles after a one‐ or two‐week dissection program. Similarly, Manyama et al. ⁴⁰ found that students with low performance benefited more from reciprocal peer teaching methods in anatomy dissection compared to those who had higher performance. Cale et al. ⁴¹ demonstrated that repeating students could effectively improve their study strategies and time management. Unlike these previous studies, the initial performance improved during the regular dissection curriculum in this study, not in the repeating or additional program.

We provided peer‐to‐peer teaching time after each dissection practice to encourage student self‐directed learning. In each dissection practice, two dissection issues were assigned to these students two‐by‐two, respectively. Each pair of students cooperated to solve the assigned issue, usually sitting across from each other at the dissection table. At the end of each dissection practice, the students explained the dissected results on each assigned issue. The effect of collaboration and that of peer‐to‐peer learning could be of more benefit to low‐scoring students than others. ⁴⁰

In this study, the authors did not plan to screen students at risk but instead examined the relationship between performance in the word test and those in the practical examinations. Ahmady et al.'s ⁴² review also reported the importance of identifying problem learners at risk of failing the academic course as early as possible. Typically, they are identified after finishing anatomy courses and knowing their final scores. If anatomy teachers know the partial scores of their students as early as possible and predict their later performance, rather than give them a second chance at dissection after unsuccessfully finishing the first course, the teachers could conduct more effective anatomy education.

Word test rationales and adult learning theory

Multiple aspects of the word test may have contributed to its positive effect. The fundamental design of the word test was its composition of two factors: anatomical structures and anatomical terms. Printouts for study preparation were supplied to students in advance and scores on the test were immediately returned to them. Students were aware of their initial performance and were motivated to practice anatomy dissection after the word test. These three rationales of the word test—its design, providing students with appropriate information, and early implementation—may have contributed to its positive effect, and each is discussed below.

The authors re‐examined the aforementioned three rationales of the word test from the perspective of adult learning theory, ⁴³ , ⁴⁴ because these rationales could be related to the active involvement of students during the dissection practice afterward. According to Knowles, ⁴³ andragogy is central to adult learning theory. There are six assumptions upon which andragogy is founded: (1) the need to know; (2) the learner's self‐concept; (3) the role of the learner's experience; (4) readiness to learn; (5) orientation to learning; and (6) motivation. ⁴³ Through the basic anatomical terms and simple drawings in the word test, students had an opportunity to discuss human structures combined with anatomical terms for the first time. They could attain readiness and internal motivation for learning by understanding the necessity of anatomical terms for the subsequent dissection course and clinical medicine. The word test results contributed a 10% weighting to the final grade, which could have increased students' readiness and internal motivation to take the word test, and provide an opportunity for students' self‐directed learning. ⁴⁵ Immediate feedback of word test scores may have enhanced the aforementioned self‐directed learning and intrinsic motivation. Learners could recognize the gap between their initial performance and their target scores. The immediate feedback and timing of the word test gave them an opportunity to become internally motivated and self‐disciplined learners.

Taylor and Hamdy ⁴⁴ proposed a multi‐theory model about how adults learn, positing that all learning starts with what learners already know. Their learning activities move with the dissonance phase when their prior knowledge is found to be incomplete, moving through the subsequent phases, including refinement, organization, feedback, consolidation phases, and reaching a new dissonance phase for another learning cycle to attain higher levels of knowledge. ⁴⁴ They defined several considerations by educators during the dissonance phase: the learner's prior knowledge, their learning styles, the stage of development of the learner, the learner's motivation, the resources (time), and the nature of the task. ⁴⁴

Regarding learning style, Hattie and Donoghue ⁴⁶ analyzed 400 learning strategies and created a learning model, in which the learners start with surface learning and move to deep learning to acquire higher levels of knowledge. They then transfer again to understand new knowledge. The authors claimed that both learning styles are important and that a sufficient volume of surface knowledge is required to advance into deep learning.

Using images could influence item difficulty, however, the effects of the images on students' scores were not simple and depended on various types of images used in the extended matching questions. ⁴⁷ , ⁴⁸ Students attained significantly higher scores on the anatomical questions with both terms and structures than those with terms but no images. ³⁶ In the current word test, line drawings with simple identification questions could help reduce the cognitive load related to anatomical terms and structures. ⁴⁹ It is reported that one of the most challenging aspects of dissection practice is the sheer volume of information, including anatomical terms and structures. ⁵⁰ Reducing cognitive load by automating anatomical terms and structures in the early stages of dissection practice should be useful for efficient dissection and for communicating with peers. Knowledge of anatomical terms and structures (word test) can serve as a scaffolding for students' early dissonance and help them overcome initial barriers to learning and advance to the next stage.

Ramsden ⁵¹ discriminated between surface learning and deep learning, however, their difference was blurred and both learning styles with appropriate balance are necessary during dissection practice. ⁵² , ⁵³ Pandey and Zimitat ⁵² argued that students used both a surface learning approach (memorization) and deep learning approach (understanding and visualization) for successful learning of anatomy. Smith et al. ⁵³ highlighted that adequate surface approach to learning may be important in anatomy during the initial stages before deep learning. In this study, studying anatomical terms in the word test could be regarded as surface learning. Accordingly, conducting the word test during the initial stages of anatomy education is important, as it represents adequate surface learning, consistent with the claim made by Smith et al. ⁵³ The word test in this study comprised basic anatomical terms and related simple images, which included both a surface learning and deep learning approach for students. The word test in this study was consistent with the learning styles proposed by Hattie and Donoghue. ⁴⁶

Word test scores were immediately returned to students as indicators of their initial performance. This enabled them to realize the learning gap, motivating them to engage in subsequent dissection practice. Feedback was effective for improving learning processes. ⁴⁴ Adult learning requires proper educational resources, for example, enough space, books, and digital materials. The most precious resource is the time for planning, performing, and evaluating for learning. ⁴⁴ According to Ahmady et al., ⁴² motivation, study strategies, and time‐management were positively correlated to the effect of learning. ⁴² In this study, students whose word test scores were less than 60% could understand their initial low performance in the second week of the dissection course. They were given enough time to address this in the following dissection practice for 7 weeks. During these periods, they had an opportunity to improve their learning strategies and time management ⁴¹ and to cooperate in each dissection group. ⁴⁰ The early implementation of the word test in this study could be a required resource of adult learning.

The researchers in this study usually accepted active and open‐ended questions by students as well as respected their remedy action. Since we did not restrict communication between students and teachers during dissection sessions, it may have influenced the examination scores. However, the possibility that the students applied remedy actions during dissection sessions with their peers and teachers could be the intended mechanism of the intervention. According to adult learning theory, the word test could inherently trigger such a type of remedial action.

Study Implications

Interpreting the current results through the lens of adult learning theory could re‐shape the context of how our results are used by students and faculty. By including the initial word test as part of the curriculum, faculty could provide the above‐mentioned opportunities that align with adult learning theory, in which readiness and internal motivation are improved for students' self‐directed learning. ⁴⁴ Faculty could provide evidence that the initial word test scores predict that of the final examination and that the performance of students at risk should improve. This would help students understand the gap more accurately and provide students at risk with the opportunity to more proactively improve their learning processes. By having faculty and students consciously operate the word test with the particular responsibilities and roles as indicated by adult learning theory, ⁴⁴ we believe that it could contribute to more effective anatomy education.

Limitations

This study was performed for multiple cohorts of second‐year medical students in a single institution. Differences among these cohorts were not measured or accounted for in the study. Most of the participants were Japanese, and the percentage of female students in each cohort was not significantly different. It is unknown whether the educational effects of the word test might be replicated in more diverse types of cohorts, such as in graduates from various universities, or cohorts with different gender proportions.

CONCLUSIONS

The positive effect of the word test was statistically demonstrated. The initial word tests predicted the scores of the practical examinations. The performance of students with low word test scores (<60%) was significantly improved through the practical examinations at the mid‐term and last day of the dissection course.

This positive effect of the word test could be attributable to three rationales of the word test: its design with anatomical structures and terms, providing students with appropriate information, and early implementation, which motivated students to attain better performance in alignment with adult learning theory.

AUTHOR CONTRIBUTIONS

Hiroshi Abe: Conceptualization; data curation; formal analysis; investigation; methodology; project administration; supervision; writing – original draft. Ryoji Suzuki: Data curation; investigation; project administration. Akimitsu Ishizawa: Data curation; investigation. Ming Zhou: Data curation; investigation. Yoshinori Kanatsu: Investigation; resources. Hideo Akashi: Data curation; formal analysis; investigation; writing – review and editing.

ETHICS STATEMENT

The Ethical Review Committee of the Akita University Graduate School of Medicine determined that this research was in accordance with relevant ethical standards.

Supporting information

Data S1: Supplements 1‐1(a‐k), 1‐2(l‐y): 25 line drawings in print form were distributed to students for preparation before the start of the dissection course. Approximately 200 English anatomical terms were attached to these images. Ten out of 25 images were selected for the word tests in each cohort that required students to fill in the blank for 30 English anatomical terms, which had already been specified in the print.

Data S2: Supplements 2a‐e: Number of correct answers for the word tests and Practical Examinations 1 and 2 with correct spellings correspond to each student in (a) 2013, (b) 2014, (d) 2016, and (e) 2017, respectively. Those for the Practical Examinations are indicated in (c) 2015, because the word test was not performed in 2015. N: number of students; n (< 18); number of students with a score on the word test less than 18 (60%); SD: standard deviation; SE: standard error; 95% CI: 95% confidence interval; LB: lower bound; UB: upper bound.

Data S3: Supplement 3‐1: Regression analyses between word tests and practical examinations for all students in each cohort. (a) Simple regression analyses. The significance of regression equation was surveyed using ANOVA. The independent variables (word test scores) significantly predicted the dependent variables (scores in Practical Examination 1) in simple regression analyses for all students (p < 0.001). R: multiple correlation coefficient; R ²: R square, coefficient of determination; adj‐R ²: adjusted R square; ANOVA: analysis of variance; Sig: significance probability; Constant: Intercept of regression equations; B: partial regression coefficient. (b) Multiple regression analyses. Two independent variables (word test scores and scores in Practical Examination 1) significantly predicted the dependent variable (scores in Practical Examination 2) in multiple regression analyses for all students (p < 0.001). B(X ₁): partial regression coefficient of word test scores; B(X₂): partial regression coefficient of Practical Examination 1. For other abbreviations, please refer to Supplement 3‐1(a).

Supplement 3‐2: Regression analyses between word tests and practical examinations for students with low word test scores (below 60%) in each cohort. (a) Simple regression analyses. The independent variables (word test scores) significantly predicted the dependent variables (scores in Practical Examination 1) in simple regression analyses for students with low word test scores (p < 0.05). For abbreviations, please refer to Supplement 3‐1(a). (b) Multiple regression analyses. Two independent variables (word test scores and scores in Practical Examination 1) significantly predicted the dependent variable (scores in Practical Examination 2) in multiple regression analyses for students with low word test scores in both 2013 and 2014 (p < 0.001), and 2016 (p < 0.05). In 2017, a significant regression was not demonstrated (p = 0.594). B(X ₁): partial regression coefficient of word test scores; B(X₂): partial regression coefficient of Practical Examination 1. For other abbreviations, please refer to Supplement 3‐1(a).

Data S4: Supplement 4‐1: Scatter plots with regression lines from the simple regression analyses for all students in each cohort of 2013, 2014, 2016, and 2017. Regression equations and coefficients of determination (R ²) are indicated. The X or Y axes show the scores of the word test or Practical Examination 1 (PE1), respectively. In the regression equations, the small x or y represents the independent or dependent variables, respectively, of the regression lines. Details of the statistical data are shown in Supplement 3‐1(a).

Data S5: Supplement 4‐2: Scatter plots with regression lines in the simple regression analyses for students with low word test scores in each cohort of 2013, 2014, 2016, and 2017. Regression equations and coefficients of determination (R ²) are indicated. The X or Y axes show the scores of the word test or the Practical Examination 1 (PE 1), respectively. In the regression equations, the small x or y represents the independent or dependent variables, respectively, of the regression lines. Details of the statistical data are shown in Supplement 3‐2(a).

Data S6: Supplement 5‐1: Scatter plots with 3D regression planes from the multiple regression analyses for all students in each cohort of 2013, 2014, 2016, and 2017. Regression equations and coefficients of determination (R ²) are indicated. X‐axes: the word test scores; Y‐axes: the scores of Practical Examination 1 (Practical Exam 1); Z axes: the scores of Practical Examination 2 (Practical Exam 2). In the regression equations, the small x and y represent two independent variables, and the small z represents the dependent variables of the regression planes. Details of the statistical data are shown in Supplement 3‐1 (b).

Data S7: Supplement 5‐2: Scatter plots with 3D regression planes from the multiple regression analyses for students with low word test scores in each cohort of 2013, 2014, 2016, and 2017. Regression equations and coefficients of determination (R ²) are indicated. X‐axes: the word test scores; Y‐axes: the scores of Practical Examination 1 (Practical Exam 1); Z axes: the scores of Practical Examination 2 (Practical Exam 2). In the regression equations, the small x and y represent two independent variables, and the small z represents the dependent variables of the regression planes. Details of the statistical data are shown in Supplement 3‐2 (b).

Biographies

Hiroshi Abe, M.D., Ph.D. is a Professor Emeritus at the Department of Anatomy, Akita University Graduate School of Medicine, Akita, Japan. He taught anatomy and neuroanatomy to second‐year medical students. His research interests are anatomy education and the localization and function of ATP‐sensitive K channels.

Ryoji Suzuki, M.D., Ph.D. is an associate professor at the Department of Anatomy, Akita University Graduate School of Medicine, Akita, Japan. He teaches anatomy and neuroanatomy to second‐year medical students. His current major research interest is the function of fatty acid binding protein in Peyer's patches.

Akimitsu Ishizawa, Ph.D. was an assistant professor at the Department of Anatomy, Akita University Graduate School of Medicine, Akita, Japan. He taught anatomy and neuroanatomy to second‐year medical students. His research interest lies in the variations in the human body, especially vascular variations in the heart.

Ming Zhou, M.D. was an assistant professor at the Department of Anatomy, Akita University Graduate School of Medicine, Akita, Japan. He taught anatomy and neuroanatomy to second‐year medical students. His research interests are the localization and function of ATP‐sensitive K channels.

Yoshinori Kanatsu is a technician at the Department of Anatomy, Akita University Graduate School of Medicine, Akita, Japan. He embalms donated human body donors. He teaches anatomical dissection to second‐year medical students. His research interest is maintaining body donors in good condition.

Hideo Akashi, Ph.D. is an assistant professor at the Department of Anatomy, Akita University Graduate School of Medicine, Akita, Japan. He teaches anatomy and neuroanatomy to second‐year medical students. His main research interest lies in the relations between variations and human genes, biotechnology, and bioinformatics.

DATA AVAILABILITY STATEMENT

The data that supports the findings of this study are available in the Supporting Information of this article.