A Statewide Performance

Study of Academic Variables Affecting of Baccalaureate Nursing Graduates on Licensure Examination

ELEANOR MCCLELLAND, PHD, ORPHA

J.

GLICK,

The purpose of this study was to validate, using a statewide sample, findings from two previous smaller studies investigating the relationships between admission selection variables and subsequent achievement in baccalaureate nursing programs and performance on the National Council Licensure Examination for Registered Nurses (NCLEX-RN). Subjectsfor this study were 1,069 graduates of nlne Iowa basic baccalaureate nursing programs. The analyses included three major components. The first addressed the relationship between admission selection variables and subsequent achievement in the nursing programs and performance on a standardized nursing achievement test (AssessTest) and the NCLEXRN. The second component investigated the extent to which achievement In nursing courses predicted performance on the NCLEX-RN. In the third component a path analysis was used to formulate a causal model describing the relationships among the variables in the study. The results of the study suggest that students’ prenursing grade point average and American College Testing scores predict their performance on the NCLEX-RN. Knowledge of performance predictors can both facilitate admission selection and the use of educational resources to develop nursing competence and promote success in obtaining licensure to practice. (Index words: Baccalaureate; Education; NCLEX-RN performance; Nursing) J Prof Nurs 8:342350,1992. Copyright 0 1992 by W.B. Saunders Company

T

HE ROLE OF THE NURSE has changed considerably in the last 25 years. More attention is

being directed toward college and university programs that prepare graduates for professional nursing. The mandate is that nurses must be educated for *Associate Professor, College of Nursing, The University of Iowa, Iowa City, IA TAdjunct Assistant Professor, College of Nursing, The University of Iowa, Iowa City, IA. *Assistant Professor, College of Nursing, The University of Iowa, Iowa City, IA. Address correspondence and reprint requests to DC McClelland, PhD, RN, 101 L Nursing Bldg, College of Nursing, The University of Iowa, Iowa City, IA 52242. Copyright 0 1992 by W.B. Saunders Company 8755-7223/92/0806-0008$03.00/O

342

Jownal

JUNE C. YANG, PHD, RN,?

RN,*

of Professional Nursing,

PHD,

RN*

present

and future challenges

care in rapidly

changing

American

Association

addressed

these

edge, practice, cation.

AND

of providing

health

competent

care systems.

of Colleges of Nursing

issues by defining

(AACN)

essential

and values for professional

The final report of the project,

The

knowl-

nurse edu-

“Essentials

of

College

and University Education for Professional Nursing, ” was published in 1986. This report calls on nurse educators to design curricula that provide learning opportunities addressing these standards. Curricular implications include careful selection of clinical learning sites, provision of new areas of knowledge, and development

of an understanding

responsibilities

of the changed

of the nurse.

In addition, determination of the abilities of candidates for admission to baccalaureate nursing programs

has taken

on new meaning

nurse shortage and a reduction Although basic baccalaureate have increased

in the wake of a

in the applicant pool. nursing enrollments

for the first time

in 5 years,

the de-

mand-driven nurse shortage is predicted to continue and increase. The mandate that nurses be educated for future well as present health care delivery continues to be a challenge that nurse educators need to address (Naisbitt & Aberdene, 1990). Research designed to establish

predictive

criteria

results

has yielded

conflicting

admission (Allchnie

&

Bellucci, 1981; Hayes, 1981; Richards,1977; Seither, 1980). Other than past academic achievement, there are no clearly identified criteria that consistently predict a student’s success in the nursing major and the licensure examination that can be measured before admission. Grant (1983) notes that reading skill may be one exception. In contrast to early attempts to identify predictive admission criteria, recent studies of academic performance have focused on identifying students at risk for failure early in their program of study (Feldt & Donahue, 1989; Froman & Owen, 1989; Jenks, Selekman, Bross, & Paquet, 1989; McKinney, Small, VoI 8, No 6 (November-December)z

1992: pp 342-350

STATE STUDY OF PERFORMANCE

O’Dell,

& Coonrod,

1988; Payne

The aim of this approach support

for students

National

Council

tered

Nurses

McKinney

343

ON NCLEX-RN

& Duffey,

is to provide

educational

et al. concluded

Examination Payne

for Regis-

and Duffey

and

that early identification

of

students

who may be at risk for NCLEX-RN

provides

opportunities

for intervention

instruction,

small group

sistance

test-taking

strategies,

programs

to reduce

test anxiety.

strategies

seek to prevent

nomic losses associated by Grant

seminars, and

These

personal,

failure

such as com-

puter-assisted

as-

relaxation educational

social,

and performance are members

who may be at risk for failing the Licensure

(NCLEX-RN).

with

1986).

and eco-

with failure that are described

Nursing

These schools”

(IACN).

1. What

The research questions

is the relationship

selection

of Colleges

variables

a. achievement

of

were:

between

admission

and:

in baccalaureate

nursing

pro-

grams? b. performance

on the NCLEX-RN?

2. Does achievement NCLEX-RN

in nursing

courses predict

performance?

3. What are direct dictor variables

and indirect on

effects of pre-

NCLEX-RN

perfor-

mance?

( 1983).

Analyses the

on the NCLEX-RN.

of the Iowa Association

higher

the

included

two major components.

The first

component addressed the relationship between admission selection variables and subsequent achievement

student’s

Gbi, the more /ikely she or he would perform well on the NCLEX-RN examination.. .

in baccalaureate nursing programs and performance on a standardized nursing achievement test and the NCLEX-RN.

In the second component

an effort was

made to formulate and test a causal model describing relationships among the variables in the study. Since 1983 a series of studies was conducted

at The

Methodology

University of Iowa College of Nursing to examine relationships between admission selection variables and performance

in the nursing

sure examination

(NCLEX-RN).

DESIGN

major and the licenIn the pilot

study

The design of the study was exploratory.

Academic

Glick, McClelland, and Yang (1986) found that the grade point average (GPA) in the biologic science

performance before, during, and after completion of a baccalaureate nursing program was examined retro-

courses (BIO-GPA) and the overall prenursing grade point average (PN-GPA) were the best predictors of success in one nursing program. There were not sta-

spectively to identify predictors of achievement could be used in making admission decisions.

tistically

significant

relationships

tors and NCLEX-RN showed high correlation

between

rion variables, eg, nursing courses grades. In a followup study with a larger sample (n = 2 lo), the validity of using PN-GPA

to predict

achievement

ing program was also supported. the best predictor for performance

SAMPLE

the predic-

scores, even though they coefficients with other crite-

in the nurs-

On the other hand on the NCLEX-RN

was the American College Testing Programs, Inc, (ACT) social studies subscore. Additionally, when the relationships between stratified GPAs and mean NCLEX-RN scores were analyzed, the higher the student’s GPA, the more likely she or he would perform well on the NCLEX-RN examination (Yang, Glick, & McClelland, 1987). The present study, the third in the series, was conducted to investigate the relationship between admission selection variables and achievement in all basic baccalaureate nursing programs in the state of Iowa

that

Data for the study were obtained from the academic records of 1,070 graduates of the nine basic baccalaureate nursing

programs

in the state of Iowa. Only the

graduates who wrote the NCLEX-RN examination in the state of Iowa during the d-year period from 1985 through 1988 were included in the sample. This represents 80 per cent of the total basic baccalaureate graduates (N = 1,335) for this period (Iowa Board of Nursing, April 1990, personal communication). The number of subjects from each of the schools ranged from 17 to 457. Cases with missing data were excluded from each analysis; this accounts for differences in sample size for individual analyses. The project met the criteria for exemption from review as determined by The University of Iowa College of Nursing Human #Clarke, Coe, Luther, Marycrest,

Graceland, Grand View, Iowa Wesleyan, Mount Mercy. and The University of Iowa.

MCCLELLAND,

344

Subjects Review Committee involved

because this research only

the use of existing

educational

data. The data

were coded in a way such that individuals identified.

Findings

are reported

could not be

only

as grouped

data. The demographic similar

characteristics

to previous

graduates.

studies

The majority

of the sample were

of baccalaureate

were white

and female (93 per cent), ranging

(96.8

nurse per cent)

in age from 2 1 to 57

CLN-GPA

for five schools reflected

cal specialties

eg, maternity

dren,

health

mental

nursing,

etc,

course titles that implied

content,

such as Nursing

all nursing

nonclinical

research,

nursing

courses

nursing

cliniof chil-

and four schools integrated

I. N-GPA

courses with a clinical

required

traditional

nursing,

reported

YANG, AND GLICK

nursing

was comprised

practicum

of

as well as

such as pharmacology,

trends and issues, pathophysiology,

and nutrition.

years. The mean age was 24.6 years. The non-white ethnic

groups

(3 per cent) represented

were American

Indian

12), Hispanic

(n =

in the sample

l), Afro-American

(n = 2), and Asian (n =

(n =

11).

all predictors had statistically ’ significant relationships with NCLEX-RN scores. .

’

PROCEDURE

Coding

sheets designed

uted to the chairperson The chairpersons

for the study were distrib-

of each participating

program.

were asked to complete

a form for

each subject according to directions given and to list the courses making up each GPA. The directions included the procedure for computing an example of a completed form. Variables grade

included

point

composite

average

in the study (HS-GPA),

GPAs as well as

The NCLEX-RN was developed tion Committee of the National

Boards of Nursing and is designed with the nursing process and locus of decision-making as the conceptual framework. The NCLEX-RN was first administered in July 1982 and replaced the State Board Test Pool Examination.

were high school ACT

subtest

scores, grades from all required

and

prenursing

(PN-GPA) and nursing (N-GPA) courses, and scores from a standardized nursing achievement test AssessTest and the NCLEX-RN examination. Although the

The predictor

variables

ranging Program

from .83 to Inc., 1986).

were: (1) HS-GPA;

validity

for the NCLEX-

areas and the percentage of items in each category. Kuder-Richardson Formula 20 values have been reported in the high .80 range (J. Bosma, sonal communication).

1987, per-

Results RELATIONSHIPS AMONG ADMISSION SELECTION VARIABLES AND ACHIEVEMENT

(2) ACT

subtest and composite scores (ACT-COMP); and (3) GPAs for chemistry (CHEM-GPA), BIO-GPA, social sciences (SOC-GPA), GPA was comprised

Content

RN has been established from the study of nursing practice. This resulted in a plan specifying content

reliability of high school and prenursing course grades is not known, ACT subtest and composite scores have been shown to have reliabilities .96 (American College Testing

by the ExaminaCouncil of State

and all PN-GPA. The CHEMof organic and inorganic chem-

istry course grades. Course grades in biology, human anatomy and physiology, and microbiology comprised the BIO-GPA, and SOC-GPA included course grades from psychology, sociology, and anthropology. PNGPA incorporated all courses completed at the time of admission to the nursing program. The courses included for all schools were chemistry, biology, human anatomy and physiology, microbiology, psychology, and sociology. In addition several schools included English composition, developmental psychology, health, nutrition, and anthropology. The criterion variables included grades from required clinical nursing courses (CLN-GPA), cumulative N-GPA, and AssessTest and NCLEX-RN scores.

Correlation

coefficients

between

predictor

variables

used as selection criteria for admission and achievement in the nursing major, AssessTest, and the NCLEX-RN licensure examination were obtained. This information was used to determine whether success in a nursing program and the licensure examination could be predicted at the time admission decisions are made. The Pearson Product-Moment Correlation Coefficients obtained when CLN-GPA, N-GPA, CUM-GPA and AssessTest, NCLEX-RN scores were used as criterion variables are presented in Table 1. The predictor variables included in this analysis were HS-GPA, ACT scores, CHEM-GPA, BIOGPA, SOC-GPA, and PN-GPA. The correlation coefficients indicate that all predictors had statistically significant relationships (P d .OO 1) with NCLEX-RN scores. The highest predictor was the ACT composite score (t. = .48), followed by

STATE STUDY OF PERFORMANCE

TABLE

1.

Correlation

345

ON NCLEX-RN

Coefficients

Between Predictor Variables and Criterion Variables N-GPA

CLN-GPA

Predictor HS-GPA ACT Scores English Nat Sci sot ss Math Comp CHEM-GPA BIO-GPA SOC-GPA PN-GPA Abbreviations: lP s ,001. tP c .Ol.

CUM-GPA

AssessTest

NCLEX-RN

.-

--_______ r

n

I

n

r

”

r

”

r

n

36*

(409)

40*

(527)

44*

(519)

.22t

(141)

26*

(488)

36” .31’ 29* .27* .37* .44* .50* .49* 57*

(643) (644) (644) (644) (646)

.33* .30* 30’ 24* 36* .42’ .52* 50” 61’

(837) (838)

37* 33’ .32* .30k 40’ .61* .71* .67* 81’

(820) (821)

.44* 35 .43* 32’ 46’ 30* 33* 33’ 57*

(131) (131) (131)

42’ 41’ 44’ 29’ 48* 28” 34* 36* 41*

(793) (794) (794) (794) (795) (969) (973) (973) (807)

(750) (753) (755) (683)

Nat SCI, natural sciences,

the ACT social studies

subscore

(838) (838) (843) (1058) (1062) (1061) (850)

(821) (821) (826) (1033) (1037) (1036) (836)

Sot SS, social studies subscore,

(Y =

.44), the ACT

tion

(131) (138) (226) (228) (227) (154)

Comp, composite

of the regression

was examined

by scatterplots

English subscore (Y = .42), PN-GPA (Y = .41), the ACT natural science subscore (Y = .4 l), and the ACT

and considered to be met. Since the Ns were sufficiently large, the normality of the dependent variables

math subscore

was also assumed.

(Y

=

.29).

It is interesting

to note that

a similar pattern of prediction holds for performance on the AssessTest. These findings are consistent with

The results of the stepwise multiple regression tests when each predictor variable was entered last is pre-

the predictive

sented

in Table 2. When

variable

and HS-GPA,

relationship

the AssessTest

between

and the NCLEX-RN

performance

on

(r = .74; Mosby,

April 1990, personal communication). However, it should be noted that not all schools represented in the sample administer the AssessTest. This accounts for the smaller scores. Although formance courses nursing

NS in the analysis

using

the AssessTest

ACT scores were most predictive

on the NCLEX-RN,

grades

of per-

in prenursing

were most predictive of performance in the major (N-GPA). For example, note that the

correlation

coefficient

for N-GPA

and PN-GPA

was

BIO-GPA,

GPA contributed

tween the above predictors cant at P d .OOl.

and N-GPA

be-

the greatest

amount

of unique

vari-

ance (6 per cent) when AssessTest scores (P d .Ol) were the dependent variable and HS-GPA, ACTCOMP,

CHEM-GPA,

BIO-GPA,

SOC-GPA,

and

PN-GPA were independent variables. However, when NCLEX-RN scores were the dependent variable, ACT-COMP

coefficients

as indepen-

accounted for approximately 5 per cent of the variance (P d .OOl) over and above the other variables. PN-

variance

All correlation

CHEM-GPA,

were entered

dent variables, the largest R’ increments were observed with BIO-GPA and SOC-GPA, each of which

N-GPA

.40).

was the dependent

ACT-COMP,

and SOC-GPA

.61, for BIO-GPA .S2, for SOC-GPA .50, and for CHEM-GPA .42. Similarly, HS-GPA correlated with (Y =

N-GPA

contributed

the greatest

amount

of

(13 per cent; P < .OOl).

were signifi-

The statistically significant correlation coefficients among all the predictor variables as well as the similarities in strength of the correlation coefficients indicate that these variables predict achievement in a baccalaureate nursing program and the licensure examination. Therefore, stepwise multiple regression analyses using N-GPA, and AssessTest, and NCLEXRN scores as dependent variables were performed to determine the unique contribution of each predictor variable in the equation. In contrast to the forward selection procedure, a stepwise multiple regression analysis explores the effect of each new variable on the other variables entered earlier. The linearity assump-

RELATIONSHIP BETWEEN PERFORMANCE

The relationship

NURSING ACHIEVEMENT ON THE NCLEX-RN

between

achievement

AND

in nursing

courses and performance on the NCLEX-RN was examined with Pearson Product Moment Correlation Coefficient and regression analyses. Variables reflecting nursing achievement included a composite of grades in clinical nursing courses (CLN-GPA), composite grade for all nursing courses (N-GPA), and AssessTest scores. As shown in Table 3, the correlation coefficients for all nursing achievement variables and performance on the NCLEX-RN were highly significant (P c .OO l), with AssessTest scores having the

346 TABLE

MCCLELLAND,

2.

Results of Stepwise Entered Last

AssessTest (n = 100; df = 1 93)

NCLEX-RN (n = 426; df = 1 419)

N-GPA (n = 479; df = 1 473)

Abbreviation: ‘P c ,001. tP c .Ol.

SE, standard

Regression

When Each Predictor

R

RZ

R2 Change

PN-GPA N-GPA AssessTest HS-GPA ACT-COMP CHEM-GPA BIO-GPA SOC-GPA PN-GPA HS-GPA ACT-COMP CHEM-GPA BIO-GPA SOC-GPA PN-GPA HS-GPA ACT-COMP CHEM-GPA BIO-GPA SOC-GPA

,737 707 652 572 537 572 561 573 521 ,591 ,467 592 591 ,590 592 ,648 ,662 663 ,629 ,631

,543 ,450 425 328 ,288 328 315 ,329 ,271 ,349 ,218 ,350 .349 ,348 350 ,421 439 ,439 ,396 398

,003 046 121 001 040 001 014 ,000 ,058 ,003 134 002 003 004 002 022 004 003 ,047 045

Variable

Beta 075 314 ,450 -

041 ,276

-

050

-.197 - ,000 567 - ,066 438 ,059 098 103 106 174 ,071 072 288 275

Was

SE Beta

F

084 087 ,078 106 117 135 143 124 ,201 ,047 ,047 ,054 071 062 099 040 040 ,043 046 045

79 12 91% 33.70’ 15 5.60t 14 1.88 00 7 99t 1.95 86.71* 1 19 1 93 2 77 115 18.53* 310 2.84 39 49* 37 76*

error.

greatest association (Y = .66) followed by CLN-GPA (Y = .55) and N-GPA (Y = .53). It is also interesting to note that when PN-GPA, scores were entered

Analyses

Independent Variable

Dependent Variable

NCLEX-RN (n = 131, df = 1 127)

Multiple

YANG, AND GLICK

N-GPA,

as independent

and AssessTest variables

for the

NCLEX-RN in the regression analyses (Table 4), AssessTest scores accounted for 12 per cent of the vari-

ables. An original theoretical causal model was formulated from previous research findings and modified to be tested with path analysis.

Path analysis examines

the relevant path coefficients, providing insight about direct and/or indirect effects of model components on

ance (P 6 .OO l), and N-GPA accounted for 5 per cent (P d .OO 1). Although AssessTest performance is more

variable (NCLEX-RN). This that is, eliminating pathmay lead to “trimming,” ways of the hypothesized model found to be insignif-

predictive,

icant.

relationship

the findings between

show that there

students’

is a strong

achievement

the final

endogenous

in nurs-

ing courses, particularly clinical nursing courses, and their performance on the licensure examination.

THEORETICAL

The original

Causal Model: From PN-GPA to Performance on the NCLEX-RN In this phase of the analyses the researchers lated and tested a causal model relationships among the predictor TABLE

3.

Correlation

for explaining the and criterion vari-

Coefficients

Among

N-GPA

CUM-GPA AssessTest NCLEX-RN *P i

.OOOl.

Criterion

r

n

1 00

757

I

theoretical

studies reading subscore, CHEM-GPA, SOC-GPA,

(Fig

1) was devel-

ACT-COMP, BIO-GPA, PN-GPA, N-GPA, As-

sessTest scores, and the criterion variable (NCLEXRN scores). This model was postulated based on the Variables CUM-GPA

”

model

oped to illustrate hypothesized causal relationships among the predictor variables: HS-GPA, ACT social

N-GPA

CLN-WA

CLN-GPA

formu-

MODEL DEVELOPMENT

r

AssessTest n

r

NCLEX-RN ”

I

”

95’

755

.81*

737

.54*

105

.55*

717

1 .oo

1,066

80’

1,038

43

228

53’

973 952

1.00

1,041

37*

221

49*

1.00

229

66” 1 00

194 976

347

STATE STUDY OF PERFORMANCE ON NCLEX-RN

TABLE

4.

Correlation HS-GPA

HS-GPA

Coefficients

ACT-ENG

Among Predictor Variables

ACT-NAT SCI

ACT-SOC SS

ACT-MATH

ACT-COMP

CHEM-GPA

BIO-GPA

SOC-GPA

PN-GPA

1.00

.42*

.39*

.34*

.41*

.47*

.25*

.32*

.34*

.37*

(528)

(476) 1 00

(477) .54*

(477) .59*

(477) .55*

(483) .79*

(526) .17*

(526)

(526) .33*

(496) *

(846)

(840) 1 .oo

(640) *

(840) .47*

(836) .80’

(634) .23*

(& .24

(836) .27*

(73208) .28’

(841)

(8:) 1.00

(841) *

(837)

(835) .15*

(838) t

(637) .35*

(729) .27*

(841)

(8:) 1 00

(8:;)

(835) .20*

$8) .22*

(837) *

(729) .27*

(641)

(8;;) 1 00

(835) .23

(838) .26*

(iZ7) *

(729) .34*

(846)

(640) 1.00

(843) .56 (1,061) 1.00 (1.066)

$2,

(731) .73*

ACT-ENG ACT-NAT SCI ACT-SOC SS ACT-MATH

l

ACT-COMP

*

CHEM-GPA

(1,062) BIO-GPA

l

(1 ,,“:O, *

(646) .82*

(l.&

(851) .70’

1.00

SOC-GPA

(1,065)

(850) 1.00

PN-GPA

(853) Abbreviations. lP s .OOOl.

findings

ENG, English, NAT SCI, natural sciences;

of previous

studies

published

SOC SS, social studies subscore;

in the litera-

ture. For example, results of several researchers indicated that students’ academic performance during high school predicted academic success in college and/ or in performance on the licensure (Munro, 1980; Seither, 1980; Yang, Clelland,

1987). Therefore,

HS-GPA

examinations Glick, & Mcwas selected

as

COMP, composite.

ables for the original

theoretical

model were selected

according to the magnitude of correlation to other variables in the study and according to the logical sequence speculated ACT social studies correlation

by the researchers. For instance, reading test scores showed high

coefficients

and SOC-GPA,

with BIO-GPA,

CHEM-GPA,

which are major components

of PN-

the exogenous variable, the variable whose variance is unaccounted for by variables in the model (Pedhazur,

GPA. N-GPA and AssessTest and NCLEX-RN were added to complete the logical sequence

1982, p. 581). Endogenous variables

hypothesized pathways jective observation.

are the variables

whose vari-

ances are accounted for by either exogenous or endogenous variables in the model. The first in the sequence of endogenous variables was performed on standardized college entrance examinations (ACT tests). In the present study, ACT-COMP correlated with NCLEX-RN scores. This outcome pected

because

ACT-COMP

most highly could be ex-

is a compilation

four subtest scores, which results test-lengthening effect. Additional

of the

in an additive or endogenous vari-

The original

1.

Original theoretical model.

based on the researchers’

theoretical

model described

sub-

above was

subsequently modified for testing in this study and is shown in Fig 2. It graphically depicts the pattern of relationships among the exogenous and endogenous variables actually examined in the present path analysis.

As shown

in Fig 2, HS-GPA

scores, although highly scores, were eliminated

Figure 2, Figure

scores of the

Theoretical

and

ACT

test

correlated with NCLEX-RN from the original model be-

causal model for IACN study (n =

192). *P 6 .05; **P s .ol; **+P s ,001.

348

MCCLELLAND,

cause PN-GPA

is more recent and was found to be a

strong predictor

of GPA in baccalaureate

grams

as well as performance

This diagram

exhibits

a recursive

causal flow is unidirectional. tions

that

(Pedhazur,

underlie

nursing

pro-

on the NCLEX-RN. model in which the

The following

the application

assump-

of path

1982, p. 82) were examined

analysis

and consid-

ered satisfied: 1. Linearity

assumption:

examined

and the linearity

every other variable 2. Residuals

Scatter diagrams between

were

each and

be interval tenability

scales.

of a theoretical

model

MODEL TESTING

y5-3

=

reproduced

0.2 I. The magnitudes

the reproduced

the direct correlation

nal correlation

coefficients

predictor

variables

inappropriate.

0.26)

noted

to

coefficients were 0.14,

0.09,

and 0.15

scores may be

Based on the above findings

a trimmed

the

(Fig 2) is

model shows that both CHEM-GPA

and SOC-GPA

(P =

0.20)

scores significantly

On the other

showed

a statistically

hand,

BIO-GPA

significant

indirectly

via AssessTest (P =

direct

0.27)

effect

on

NCLEX-RN scores. All the pathways depicted in the trimmed model were statistically significant at the P d .OOl level.

FOR IACN DATA

Summary and Conclusions

and indirect effects, (2) reproduction of a portion of the correlation matrix to determine whether or not the

Data were obtained nine basic baccalaureate

relationships

of Iowa (1) to explore the relationships

in the data are consistent

tical significance (P d coefficients were obtained;

and

between

and the origi-

and NCLEX-RN

scores.

Model testing for this study included three steps: (1) calculation of path coefficient (P) to explore direct

ory, and (3) elimination

coeff-

ys2 = 0.25,

of discrepancies

correlation

and

This perhaps sugfor y51, ys2, and y5), respectively. gests that some hypothesized direct pathways between

(P =

tested by path analysis. THEORETICAL

The

affect NCLEX-RN

scales were considered As previously

effects.

by summing

cients were as follows: Y>, = 0.14,

3). The trimmed

uncorrelated.

3. The causal flow was unidirectional. 4. The measurement

indirect

matrix

model was drawn in the final step of path analysis (Fig

was confirmed.

were assumed

the correlation

YANG, AND GLICK

with the the-

of pathways with no statis.05). In the first step path they are presented

in Fig

2. As shown, the effects of CHEM-GPA (P = 0.16), BIO-GPA (P = 0. lo), and SOC-GPA (P = 0.22) on AssessTest scores were direct effects. The direct effects of CHEM-GPA and SOC-GPA on AssessTest scores were statistically significant (P d .05). In contrast, the effects of the variables of CHEMGPA, BIO-GPA, SOC-GPA, and AssessTest scores on NCLEX-RN scores were in part direct and in part indirect. Direct effects of CHEM-GPA and SOC-GPA on NCLEX-RN scores were not statistically significant (P c .05). However, direct effects of BIO-GPA and AssessTest scores on NCLEX-RN scores were statistically significant at P d .O 1 and P d .OO 1, respectively. This finding led the researchers to conclude that except for BIO-GPA, there were no significant direct effects of prenursing GPAs on NCLEXRN scores and that CHEM-GPA and SOC-GPA only affected NCLEX-RN scores indirectly through its correlation with AssessTest scores. Indirect effects of CHEM-GPA, BIO-GPA, and SOC-GPA on NCLEXRN scores were estimated to be 0.088, 0.055, and 0.120, respectively. In the second step the researchers examined the correlation components and reproduced a portion of

mission selection in baccalaureate

from 1,070 graduates of the nursing programs in the state between:

ad-

variables and academic performance nursing programs and on the

NCLEX-RN and between N-GPAs and NCLEX-RN scores and (2) to examine direct and indirect effects of predictor variables on NCLEX-RN performance. Pearson Product-Moment Correlation Coefficients between predictor variable and criterion variables were statistically significant at P d .OO 1. PN-GPA was the best predictor for achievement in the baccalaureate nursing programs, whereas ACT social studies reading and English test scores best predicted performance on the AssessTest.

The ACT composite

score alone

was the strongest predictor of success on the AssessTest and the NCLEX-RN. This is in contrast to the previous study, which showed that the ACT social studies reading most highly correlated with NCLEX-RN

Figure 3.

Trimmed model (n = 192). ***P

test was scores

c

.OO 1.

STATE STUDY OF PERFORMANCE

(Yang et al., there

1987).

It was speculated

are some similarities

studies

reading

social studies prehension

349

ON NCLEX-RN

between

that perhaps

the ACT social

test and the NCLEX-RN. reading

test purports

and analytical

The test requires new situations

and evaluative

the examinee

draw inferences

The ACT

to measure

reasoning.

to read passages and to

and conclusions by examining

and apply

importance

of ideas. Perhaps these abilities those required

tests in biology, grams.

to

are unique

and social science courses;

used in baccalaureate

nursing

pro-

why the ACT social studies

ing test was less predictive mance than ACT-COMP The high predictability

of NCLEX-RN

in the present of PN-GPA

readperfor-

study. on academic

achievement was not surprising because PN-GPA resents the GPA over a longer period of time

repthan

individual components of prenursing GPA. These results were similar to the previous findings of the researchers. This may have been influenced by the fact that 42.5 per cent of the sample was taken from the same institution Multiple tered last,

as in the earlier studies.

regression analyses showed that when enthree of the independent variables (HS-

GPA, BIO-GPA, and SOC-GPA) made statistically significant unique contributions to the total variance when N-GPA was the dependent variable. When AssessTest scores were the dependent variable, ACTCOMP

and

PN-GPA

unique

contributions

made

statistically

significant

over and above the other vari-

ables (HS-GPA, CHEM-GPA, BIO-GPA, and SOCGPA). Using NCLEX-RN scores as the dependent variable,

only ACT-COMP

to the total percent variables

(HS-GPA,

variance

contributed

significantly

over and above the other

CHEM-GPA,

BIO-GPA,

SOC-

GPA, and PN-GPA). This does not mean that other variables were unimportant predictors. In fact, they were excellent predictors as shown by high correlation coefficients (Table 4). However, because of multicollinearity among the independent variables, when one of the variables dictive value.

was entered

last it did not add pre-

A path analysis was carried out in an attempt to test a causal model. The exogenous variables examined were CHEM-GPA,

Socioeconomic

and Figure

for the other ACT subtests;

chemistry,

It is unclear

them

interrelationships

and unlike

or examinations

com-

BIO-GPA,

and SOC-GPA,

and

the

4.

Model for future study.

endogenous

NCLEX-RN

variables

were

AssessTest

scores. A path coefficient

way (Fig 2) was obtained. of direct and indirect

and

for each path-

After a careful examination

effects and reproduction

of the

correlation matrix, one pathway, from BIO-GPA to the AssessTest, was eliminated. In addition, pathways from CHEM-GPA and SOC-GPA to the NCLEX-RN were deleted. These changes resulted in the trimmed model

displayed

researchers GPA did

in Fig 3. This

trimming

to conclude that CHEM-GPA not have significant direct

NCLEX-RN

performance,

led the and SOCeffects on

but they had indirect

ef-

fects through the AssessTest. However, the BIO-GPA showed a significant direct effect on NCLEX-RN performance.

The theoretical

model tested

in this study

was a small segment of a potentially large theoretical model for future study as depicted in Fig 4. The results of this study suggest that PN-GPA and ACT test scores can predict mance

in baccalaureate

future

nursing

quently on the NCLEX-RN. tion based on valid predictors serves students’

educational

perfor-

and subse-

Careful admission selecof academic success con-

resources

acquisition

academic

programs

as well

of the knowledge

as facilitate and skills re-

quired to successfully enter professional practice. This knowledge and these skills provide the basis for clinical competence, which ultimately quality of patient care.

contributes

to the

Acknowledgment The authors would like to thank the following people for their contributions to this project: Thomas Kruckeberg, Systems Programmer Ill, The University of Iowa College of Nursing; Younghae Chung, PhD Candidate, Biostatistical Consulting Center, Department of Preventive Medicine, The University of Iowa; and the deans and department heads of the participating schools and colleges.

References Allchnie, M., & Bellucci, J. (1981). Prediction of freshman students’ success in a baccalaureate nursing program. Nursing

Research,

30,

49-53.

American Association of Colleges of Nursing.

(1986).

Essentials ing

of college and univwsity

(final report). Washington,

education for professional

nurs-

DC: Author.

The American College Testing Program, Inc. (1986). Content of the tests in the ACT assessment. Iowa City: Author.

350

MCCLELLAND,

Feldt, R. C., & Donahue, J. M. (1989). Predicting nursing GPA and National Council licensure examination for registered nurses (NCLEX-RN): A thorough analysis. Psychological Reports, 64, 4 15 -42 1. Froman, R. D., & Owen, S. V. (1989). Predicting performance on the National Council licensure examination. Westem Journal

of Nursing

Research,

11(3),

344-346.

Glick, 0. J., McClelland, E., & Yang, J. C. (1986). NCLEX-RN: Predicting performance of graduates of an integrated baccalaureate nursing program. Journal of Professional Nursing,

2, 98- 103.

Grant, R. E. (1983). Predicting academic access. In W. L. Holzemer (Ed.), Reviw of research in nursing education (pp. 89-90). Thorofare, NJ: Slack. Hayes, E. (198 1). Prediction baccalaureate nursing education ing Education,

20,

of academic success in a program. Journal of Nurs-

4-8.

Jenks, J., Selekman, J., Bross, T., & Paquet, M. (1989). Success in NCLEX-RN: Identifying predictors and optimal timing for intervention. Journal of Nursing Education, 28(3), 112-l 18. McKinney, J., Small, S., O’Dell, (1988). Identification of predictors

N., & Coonrod, B. of success for the

YANG, AND GLICK

NCLEX and students at risk for NCLEX failure in a baccalaureate nursing program. Journal of Professional Nursing,

4, 55-59. Munro, B. H. (1980). Dropouts from nursing education: Path analysis of a national sample. Nursing Research, 29(6),

37 l-377.

Naisbitt, J., & Aburdene, P. (1990). Megatrends 2000. New York: William Morris. Payne, M. A., & Duffey, M. A. (1986). An investigation of the predictability of NCLEX scores of BSN graduates using academic predictors. Journal of Professional Nursing, 2,

326-332.

Pedhazur,

E. J. (1982).

research: Explanation

Multiple

regressronin behavioral

and predictors (2nd

ed.). New York: Holt, Rinehart & Winston. Richards, M. A. (1977). One integrated curriculum: An empirical evaluation. Nursing Research, 26, 90-95. Seither, F. F. (1980). Prediction of achievemenr in baccalaureate nursing education. Journal of Nursing Education, 19,

28-36.

Yang, J. C., Glick, 0. J., & McClelland, E. (1987). Academic correlates of baccalaureate graduate performance on NCLEX-RN. Journal of Professional Nursing, 3, 298306.