Domuz Gribi (H1N1) Salgınıyla İlişkili Algıların, Kaygı ve Kaçınma Düzeyi Değişkenleri Bağlamında İncelenmesi

Participants

The sample of the present study was composed 

of 379 women and 318 men (N = 697) as two groups: 

university students and non-students. The mean age for 

the entire sample was 27.59 (SD = 1.46). 422 univer-

sity students from two private and two state universities 

in Ankara, Turkey participated in the study voluntarily. 

This sample included 281 (66.60 %) women and 141 

66     Turkish Journal of Psychology

(33.40 %) men with a mean age of  20.70 (SD = 1.33) 

ranging from 17 to 34. The non-student group consisted 

of 275 participants, 98 (35.60 %) women and 177 (64.40 

%) men with a mean age of 38.16 (SD = 1.64) ranging 

from 16 to 86. 

Materials

The measurement tool of the present study includ-

ed four main sections: (1) demographic information and 

descriptive questions regarding H1N1, (2) perception of 

H1N1, possible causes, control methods and attitudes to-

ward the vaccine, (3) avoidance behaviors, and (4) anxi-

ety measurement. In the demographic data sheet, partici-

pants were asked to indicate their age, gender, education 

and employment status. This section also included some 

global questions such as the perceived dangerousness of 

H1N1, the level of worry about catching the disease and 

the predicted duration of the pandemic. 

Perception Scales. The perception sections of the 

present study were designed to utilize a factor analytic 

strategy in order to elicit participants’ causal attribu-

tions. For the purpose of this section, a group of items 

was generated by the author in order to assess partici-

pants’ attributions for H1N1 as a disease, possible causes 

of the disease and the outbreak, control methods and at-

titudes toward the vaccine. Therefore, three perception 

scales were used in the study: (1) Perception of H1N1 

(2) Causes of H1N1, and (3) Control of H1N1. Attti-

tudes toward the vaccine were measured by a separate 

scale. For all scales, the participants were asked to use 

a fi ve-point Likert-type scale to indicate their agreement 

for a given statement (1 = not agree at all, 5 = certainly 

agree). Therefore, higher mean scores indicated higher 

agreement of participants on scales and components. 

Some items in the scales were written as negative state-

ments in order to prevent response biases, and they were 

reversed before statistical analyses. The psychometric 

properties of the perception scales are presented in the 

results section.

Avoidance Behaviors. Cognitive and behavioral 

avoidance was measured by 14 items generated by the 

author with the same methodology described above. Par-

ticipants were asked to indicate their frequency of avoid-

ance behaviors in the previous week on a given Likert-

type scale ranging from one to fi ve (1 = I never did this 

behavior, 5 = I frequently did this behavior). Therefore, 

the increasing mean scores in this scale indicated higher 

frequencies of avoidance behaviors engaged in by the 

participants during the previous week.

Trait Anxiety Inventory (TAI). Anxiety was mea-

sured by TAI form of State-Trait Anxiety Inventory 

(STAI) developed by Spielberger, Gorsuch and Lushene 

(1970) and adapted to Turkish by Öner and Le Compte 

(1975). TAI is a 20-item self-report scale which measures 

individuals’ general response tendency and individual 

differences in experiencing anxiety in the face of threat-

ening or stressful situations. Each item on TAI is rated 

between 1 to 4 (1 = Almost never, 4 = Almost always) 

and 7 items are reversed before scoring. Increasing total 

scores on the scale indicate higher level of trait anxiety. 

The internal consistency of the scale in this study was 

.84.

Procedure

The data of the study was collected in two ways de-

pending on the samples enrolled. The data from universi-

ty students was collected in class sessions on a voluntary 

basis. However, the data from the non-student group was 

collected by conventional methods. The administration 

of the questionnaires took approximately 15 minutes. 

All participants were informed about the purpose of the 

study and confi dentiality before the data collection. 

Results

Pre-analyses

Perception, Attitude, Avoidance, and Anxiety 

Measures. A 2 (gender) x 2 (status) MANOVA was per-

formed on general questions before the main analyses. 

MANOVA results indicated that gender (F

2,649

 = 16.24, 

p < .05, η

2

 = .04) and status (F

2,649

 = 5.01, p < .05, η

2

 = 

.01) had main effects on dependent variables. The gen-

eral worry of women about the disease (M = 3.15, SD = 

1.17) was signifi cantly higher than men (M = 2.74, SD = 

1.20), (F

1,650

 = 20.90, p < .05, η

2

 = .03). The predicted du-

rations of the outbreak also differed signifi cantly (F

1,650

 = 

9.11, p < .05, η

2

 = .01). Men predicted signifi cantly high 

duration (M = 7.12 months, SD = 5.60) than women (M 

= 6.23 months, SD = 4.60). Status had a main effect on 

only the predicted duration (F

1,650

 = 7.31, p < .05, η

2

 = 

.01) where students (M = 6.93, SD = 5.47) predicted du-

ration of the outbreak longer than the nonstudents (M = 

6.17, SD = 4.44). Principal Component Analysis (PCA) 

with Varimax rotation was used in analyzing all percep-

tion, attitude and and avoidance scales.

Perception of H1N1. The scale consisted of 8 items 

regarding the perception of H1N1. PCA with 8-item scale 

revealed 2 components. The fi rst component consisted of 

5 items that were related to participants’ perception of 

the  dangerousness of the disease. The second compo-

nent consisted of 3 items concerning how participants 

perceived the contagiousness of the disease. The two 

components explained 52.62 % of total variance with an 

acceptable internal consistency (.69).

Causes of H1N1. The fi rst component of the scale 

consisted of 6 items that were related to conspiracy be-

liefs presented in mass media and shared by the public. 

The second component, named as environment, con-

Perception of H1N1, Anxiety and Avoidance     67

sisted of 8 items pertaining to social and physical envi-

ronment as possible causes of the H1N1 outbreak. The 

last component, faith, consisted of four items examining 

participants’ religious and spiritual beliefs regarding the 

causes of the H1N1 pandemic. These three components 

explained 51.93 % of total variance. The internal consis-

tency for the total scale was found to be .80.

Control of H1N1. This scale consisted of 18 items 

aiming to explore participants’ beliefs about how to con-

trol the outbreak at different levels (personal, institu-

tional, global). The macro control component consisted 

of statements pertaining to participants’ beliefs about the 

effi cacy of measures implemented at institutional, na-

tional and global levels. The second component, named 

as personal control, consisted of items regarding beliefs 

about the effi cacy of measures taken personally. The last 

component consisted of four items and explored beliefs 

about the unavoidability of being infected by the virus. 

The component explained 14.08 % of the total variance 

and had an internal consistency of .63. The total variance 

explained by the three components of Control of H1N1 

was 56.46 %. The internal consistency of the total scale 

was .76.

Attitudes toward the Vaccine. This scale was de-

signed to explore participants’ attitudes toward the H1N1 

vaccine that was developed and implemented during the 

outbreak. The two components of the scale were named 

as positive attitude and negative attitude. The total vari-

ance explained by the two components was 55.93 %. 

Internal consistencies for positive attitude, negative at-

titude and total scale were .85, .67 and .82 respectively..

Avoidance Behaviors. The fi rst component of 

the scale consisted of 7 items that were mostly related 

to  cognitive avoidance behaviors, such as directing at-

tention to or thinking about other things. Public place 

avoidance, was composed of four items that aimed to 

measure the frequency of avoidance behaviors that 

the participants engaged in, such as not visiting shop-

ping centers or not using public transportation. The last 

component of avoidance behaviors scale was personal 

contact, which consisted of three items measuring the 

frequency of avoidance behaviors, such as not shak-

ing hands or touching other people while greeting, and 

avoiding public toilets. The internal consistency for the 

total scale was found to be .83. The three components of 

the scale explained 59.37 % of the total variance.

Multivariate Analysis of Variance (MANOVA) and Hi-

erarchical Regression Analysis

The effects of gender and status on depending 

variables were tested by MANOVA. For this purpose 

a 2x2 (gender x status) MANOVA was performed on 

the components of perception of H1N1(dangerousness, 

contagiousness), perception of causes of H1N1 (con-

spiracy, environment, faith), perception of control of 

H1N1 (macro control, personal control, unavoidability), 

attitudes toward the vaccine (positive attitude, negative 

attitude), avoidance behaviors (cognitive avoidance, 

public place avoidance, personal contact, mean avoid-

ance) and TAI scores as dependent variables. With the 

use of Wilks’ criterion, the main effects for both gender 

(F

15,668

 = 4.77, p < .05, η

2

 = .09) and status (F

15,668

 = 6.13, 

p < .05, η

2

 = .12) were obtained. 

The analyses revealed that gender had a signifi cant 

effect on the contagiousness component indicating that 

women (M = 4.09, SD = 0.82) perceived the disease as 

more contagious than men did (M = 3.92, SD = 0.96). 

Another signifi cant gender difference was observed 

on mean faith scores. Men had higher mean scores on 

this component (M = 1.84, SD = 0.93) than women (M 

= 1.62, SD = 0.75). Consistent with this fi nding,  men 

also perceived the disease as more unavoidable. Means 

and standard deviations of the unavoidability compo-

nent for men and women were M = 2.96, SD = 1.09 and 

M = 2.74, SD = 1.00, respectively. As for the attitudes 

toward the vaccine, men had signifi cantly higher mean 

scores on both positive attitude (M = 2.24, SD = 1.10) 

and negative attitude (M = 2.74, SD = 0.81) than women

(M

pos.att.

 = 1.83, SD = 0.95; M

neg.att.

 = 2.50, SD = 0.78). The 

analyses revealed two signifi cant gender differences in 

terms of avoidance behaviors. On the personal contact 

component women had higher mean scores (M = 2.51, 

SD = 1.20) than men (M = 2.17, SD = 1.13). The mean 

avoidance scores for women (M = 1.88, SD = 0.67) were 

also high compared to that of men (M = 1.79, SD = 0.66). 

The fi nal signifi cant gender difference was observed on 

TAI scores where women (M = 42.47, SD = 8.55) had 

higher total scores than men (M = 40.49, SD = 7.57).

The analyses also revealed signifi cant differences 

in the dependent variables due to the status of the par-

ticipants. The mean faith scores of the student group (M 

= 1.87, SD = 0.96) were signifi cantly higher than those 

of non-students (M = 1.63, SD = 0.74), indicating that 

students attributed the possible causes of the outbreak 

to religious and spiritual sources more than the non-stu-

dents. In the control of H1N1, two signifi cant differences 

were observed. The non-student group (M = 2.71, SD = 

0.91) placed more emphasis on macro control measures 

taken by global, national and institutional authorities in 

stopping the outbreak compared to the student group (M 

= 2.25, SD = 0.83). The unavoidability mean scores of 

the student group (M = 2.93, SD = 1.02) was found to be 

signifi cantly higher than those of the non-student group 

(M = 2.72, SD = 1.07). Regarding the attitudes toward 

the vaccine, the non-student group had higher scores on 

both positive attitude (M = 2.32, SD = 1.13) and negative 

attitude (M = 2.75, SD = 0.84) than students (M

pos.att.

 = 

1.82, SD = 0.92; M

neg.att.

 = 2.51, SD = 0.77). 

68     Turkish Journal of Psychology

As for avoidance behaviors, all mean scores dif-

fered signifi cantly between groups except for cognitive 

avoidance. The mean scores of the non-student group 

for public place avoidance (M = 1.71, SD = 0.85), per-

sonal contact (M = 2.46, SD = 1.20) and mean avoidance 

scores (M = 1.91, SD = 0.70) were signifi cantly higher 

than that of the student group (M = 1.46, SD = 0.70; M = 

2.29, SD = 1.16; M = 1.79, SD = 0.64, respectively).

The only gender x status interaction effect was ob-

tained for the unavoidability component (F

1,682

 = 3.93,

p < .05, η

2

 = .01). Post-hoc analyses using Tukey’s HSD 

revealed that the unavoidability scores of non-student 

women (M = 2.39, SD = 0.10) were signifi cantly lower 

than those of non-student men (M = 2.90, SD = 1.09), 

student women (M = 2.87, SD = 0.99) and student men 

(M = 3.04, SD = 1.08).

Hierarchical Regression Analysis was performed 

to assess the predictive role of the components of percep-

tion and attitude scales and anxiety measures on mean 

avoidance scores. Gender and status variables were en-

tered in the fi rst block and after their effects were con-

trolled, other variables (dangerousness, contagiousness, 

conspiracy, environment, faith, macro control, personal 

control, unavoidability, positive attitude, negative atti-

tude and TAI) were entered as the second block.   

The Hierarchical Regression Analysis revealed that 

both the fi rst (R = .13, F

2,683

 = 5.83, p < .001) and the 

second block (R = .31, F

11,672

 = 5.86, p < .05) variables 

predicted mean avoidance scores. While the fi rst block 

variables explained 1 % of the variance, the variables 

in the second block explained 10 % of the variance. 

The gender (β = -.09, p < .05) and status (β = -.12, p < 

.05) variables in the fi rst block had negative effects on 

avoidance behaviors. In other words, avoidance scores 

of the women and non-student participants were higher. 

Among the second block variables, after controlling the 

effects of gender and status, dangerousness (β = -.08, p < 

.05), faith (β = -.08, p < .05), personal control (β = -.11, 

p < .001) and TAI (β = -.12, p < .001) had signifi cant ef-

fects on avoidance behaviors.

Discussion

In the present study we investigated H1N1-related 

perceptions and their relationship with anxiety and avoid-

ance behaviors observed during the swine fl u pandemic 

in Turkey. The results of the study suggest that there exist 

some gender differences in perceiving the disease-related 

components. Women participants in our study perceived 

the disease as more contagious than men did, and their 

total anxiety scores were higher. In our case, women’s 

perceptions of high contagiousness may be interpreted 

as higher perceived risk of being infected. Therefore, for 

women, a relationship between higher level of anxiety 

and increased risk perception can be expected. A simi-

lar pattern was obtained during the initial phase of the 

pandemic in France (Raude & Setbon, 2009) and Turkey 

(Akan, 2010). 

Our  fi ndings with respect to women’s higher 

anxiety and perception of contagiousness become more 

meaningful if they are considered together with their 

personal contact avoidance and higher mean avoidance 

scores. Since women experienced greater anxiety and 

believed that the disease was contagious, they may have 

exhibited more frequent avoidance behaviors. One pos-

sibility to account for gender difference in the avoidance 

of personal contact is the combination of high risk per-

ception and high level of anxiety on the part of women. 

One important point that should be explored regarding 

avoidance behaviors is that some of the avoidance be-

haviors investigated in this study were also recommend-

ed personal measures by health authorities to prevent the 

disease from spreading. For this reason, personal contact 

avoidance behaviors can be accepted as non-pharma-

ceutical preventive behaviors as well. In the literature, 

there exists evidence that females have a tendency to 

initiate touching behaviors more than males, and same-

gender touching among females is more frequent than 

males (Stier & Hall, 1984). One may argue that although 

females touch more in social contexts, they have a ten-

dency to decrease their frequency of touching behaviors 

when needed such as during infectious pandemics. For 

these reasons, women’s tendency toward higher personal 

contact avoidance has several practical implications. 

First, it can be argued that since women are more sensi-

tive to (or anxious about) diseases and avoid more, pub-

lic campaigns during pandemics may target women pri-

marily to develop protective behaviors. This may lead to 

increased visibility of desirable behaviors in society such 

as washing hands, not touching others while greeting, 

and these behaviors may be modeled by others. Second, 

women in traditional families are the primary caregivers. 

Therefore, they may disseminate desirable behaviors not 

only to children but also to others such as the elderly. In 

sum, our study supports a high risk perception-anxiety-

avoidance linkage model for women. 

The differences on the part of males can also be 

interpreted consistently in themselves. Among the causal 

attributions made by men, faith had signifi cantly higher 

mean scores compared to women. That is, the men in the 

study placed more emphasis on religious and spiritual 

beliefs in explaining the causes of the disease. A similar 

attributional pattern in men was obtained in another study 

that utilized the same methodology in Turkey (Çırakoğlu, 

Kökdemir, & Demirutku, 2003). In this study men rated 

religious practices as more useful than women in the 

treatment of unipolar depression. The results of these 

studies propose that men tend to attribute causes of some 

Perception of H1N1, Anxiety and Avoidance     69

phenomena to uncontrollable factors such as faith. In ad-

dition, the present study indicated a higher perception 

of the unavoidability of the disease in men. This fi nding 

may be more meaningful if it is interpreted together with 

men’s ratings for faith attributions. Since men explain 

the causes of the pandemic by referring more to uncon-

trollable sources, such as religious and spiritual factors, 

it can be expected that they may perceive the disease as 

more unavoidable.

As for the attitudes toward the vaccine, men had 

signifi cantly higher scores on both positive and negative 

opinion components. In fact, confl icting or vague opin-

ions about vaccines for fl us are not new (Mangtani et al., 

2006; Wicker, Rabenau, Doerr, & Allwinn, 2009). In our 

study, men’s higher emphasis on faith and unavoidability 

beliefs may have resulted in making attributions with-

out questioning or without considering the available data 

in deciding about the effi cacy of the vaccine. In sum, 

when we consider the results in the light of desirable or 

functional health behaviors during pandemic diseases, it 

is worth proposing that, compared to men, women have 

some more favorable behaviors (e.g., decreasing the fre-

quency of physical contact while greeting). 

In addition to gender, the results of our study also 

indicated some signifi cant differences in terms of the sta-

tus of participants. The mean faith scores of the student 

group were signifi cantly higher than those of non-stu-

dents. When considered together with the higher educa-

tional level of the student group, this was an unexpect-

ed  fi nding. This difference may stem from the student 

group’s more limited experience of pandemics, which 

may have caused an unrealistic perception of the disease. 

Another possible explanation may be the increasing con-

servative tendencies in Turkey (Yılmaz, 2006).

The difference in the unavoidability components 

can also be explained through the daily routines of the 

sample groups. Although we did not collect information 

regarding working conditions (e.g., place, working hours 

etc.) in the study, it can be suggested that the student 

sample may have perceived their classrooms as a place 

where the probability of catching the disease is higher 

compared to the work places of non-students. 

In terms of the predictors of avoidance behaviors, 

our data proposed that being a woman, having a job, an 

increased perception of dangerousness, higher beliefs 

in faith, higher beliefs in personal measures and higher 

anxiety level predicted avoidance behaviors. 

This study has some limitations mostly due to cer-

tain practical reasons. Our study has a time-sensitive 

problem similar to that proposed by Leung et al. (2005). 

During the period of data collection the offi cial number 

of deaths announced by the national and global health 

authorities (WHO or the Turkish Ministry of Health) 

due to H1N1 was increasing. The number of death cases 

and increasing media coverage could have caused cer-

tain changes in the perception of lay people and, in turn, 

these possible changes may have had contaminating ef-

fects on the present data. Therefore, these results should 

be considered cautiously. Although the study revealed 

meaningful and interpretable results, the sample size 

may be considered “insuffi cient” for a lay theory study 

investigating the perception of a world-wide pandemic. 

Also, the relatively small sample size did not allow mak-

ing statistical analyses based on certain variables (e.g., 

age, working conditions, social contact with infected in-

dividuals, sources of information etc.), which may nor-

mally have effects on perceptions, attitudes, anxiety and 

avoidance levels.