A Study on the Effects of Response Time on Travel Package

Attributes

Usha Ananthakumar and Sagun Pai

Indian Institute of Technology Bombay, Mumbai, India

Keywords: Consumer Behavior, Conjoint Analysis, Demographic Profiling, Tourism Preferences, Willingness to Pay.

Abstract: The rapid growth of online surveys in the past decade has raised questions about the effects of response time

on the results. The focus of our current study is to discuss the impact of response time on various travel

package attributes, thereby understanding consumer cognitive process. This study makes use of a recently

conducted conjoint analysis experiment on travel package preferences in order to gain insights into the impact

of response time on attribute importance and willingness to pay (WTP). Accordingly, the respondents are

grouped as fast and slow depending on their response time and their differences in conjoint attribute

importance estimates are investigated. The study also examines the changes in consumer willingness to pay

for the two groups. Additionally, the distinctions in socioeconomic characteristics between the fast and slow

respondents are also analyzed. The results and conclusions obtained from this research will help tour operators

to scrutinize the time taken by consumers and thereby deploy appropriate marketing strategy based on the

respective importance values and WTP trends.

1 INTRODUCTION

Survey based research has been one of the most

prominent mechanisms to elicit social response

through both direct and indirect techniques. With the

advent of the internet, online surveys have become a

popular way to conduct survey research due to better

data management, cost-effectiveness, wider target

population and higher user interactivity (Benfield and

Szlemko, 2006; Couper, 2000; Malhotra, 2008; Van

Selm and Jankowski, 2006; Höhne et al, 2018). The

use of online surveys also provides respondents the

flexibility to take as much time to think as required

because of absence of external pressure (Cook et al,

2011). The response time taken to fill up the survey

can be easily measured in case of online surveys

without obstructing the respondent’s thinking

process. The time thus gathered, can be used to derive

valuable insights into consumer behavioral patterns

and decision making process (Baxter and Hinson,

2001; Skowronski and Carlston, 1987; Mayerl et al,

2019; Gibbons and Rammsayer, 1999; Hertel et al,

2000; Sheppard and Teasdale, 2000; Mayerl, 2013).

Although response time is well researched to have

impact on customers, not many studies analyze the

impact of response time on the importance given by

customers to various product attributes. One such

study was done by Holmes et al (1998) wherein they

observed willingness to pay trends according to

survey response time for a conjoint analysis setup.

Conjoint analysis is a popular indirect measurement

technique used to understand consumer preferences

and tradeoff depending upon product attributes

(Green and Wind, 1975; Ryan, 1999). Consumers are

asked to rate different product profiles with varying

attributes in order to understand changes in their

underlying importance (Hobbs, 1996; Phillips et al,

2002). These ratings are then modified to part worth

utilities and relative importance measures for product

attributes so as to evaluate consumer preferences. In

addition to determining product attribute utilities,

conjoint analysis has also been used to estimate

willingness to pay for changes in certain attributes

(Gensler et al, 2012; Palumbo, 2011; Breidert et al,

2006). However, the existing literature on conjoint

analysis with implications of response time is limited

and it fails to provide a detailed analysis of the effects

of time on importance values of product attributes.

The focus of our current study is to discuss the

impact of response time on conjoint importance

values, thereby understanding consumer cognitive

process. This study makes use of the survey results

carried out to find out the importance and willingness

Ananthakumar, U. and Pai, S.

A Study on the Effects of Response Time on Travel Package Attributes.

DOI: 10.5220/0010600600790087

In Proceedings of the 10th International Conference on Data Science, Technology and Applications (DATA 2021), pages 79-87

ISBN: 978-989-758-521-0

to pay for different attributes of tour packages on

analyzing tourist preferences (Pai & Ananthakumar,

2017). In order to examine the effects of response

time, we categorize the survey respondents into two

types, namely slow and fast, as per their survey

response times. We then move on to identify

differences between the two groups depending upon

their preference for certain attributes. We also aim to

examine the changes in consumer willingness to pay

for the two groups. The differences between the two

groups in terms of both attribute importance and

willingness to pay can be utilized while developing

effective marketing strategies.

The remainder of the paper is structured as

follows: Section 2 involves related literature review

for response time effects on consumer decisions;

Section 3 describes our data and methodology;

Section 4 provides the results of our analysis and

Section 5 presents the conclusions and marketing

implications of our study.

2 LITERATURE REVIEW

For answering a typical survey question, the

respondents are required to complete multiple

cognitive tasks such as question comprehension,

judgment, choice comparison, response formatting

(Tourangeau et al, 2000) and hence the entire process

of survey completion involves significant cognitive

efforts. Prior to the advent of internet-based surveys,

researchers explored the effect of respondents’ time

to think on survey responses (Svedsater 2007; Cook

et al. 2007, 2012). It was observed that time to think

reduces uncertainty about response as well as the

willingness to pay (WTP). Cook et al (2012) also

found a correlation between individual reflection time

and demand for the product to be valued under the

valuation task.

With the increase in the number of online and

computer assisted surveys, multiple studies have tried

to explore the impact of objectively measured

response time on the responses. Holmes et al (1998)

observed that response time systematically affects the

preference structure for rainforest protection in an

adaptive conjoint analysis experiment. They also

found that an increase in response time also results in

an increase in preference intensity, i.e. how strongly

respondents prefer one product over another. Haaijer

et al (2000) used a multinomial probit (MNP) model

to show that greater response time results in more

systematic responses due to decrease in the error

variance.

Rose and Black (2006) demonstrate that response

time not only influence heterogeneity within the mean

of parameter distributions, but also has a significant

influence upon variance heterogeneity. Brown et al

(2008) noticed that responses become more stable

with time and the response time falls with increase in

the number of comparison tasks, due to increasing

familiarity and experience. Bonsall and Lythgoe

(2009) examined the determinants of response time in

a choice experiment survey and showed that

demographic factors such as age, education level as

well as choice order and scenario complexity

influence the response time. Hess and Stathopoulos

(2013) deployed a response time model using survey

engagement as a latent variable and observed positive

correlation between response time and engagement.

Campbell et al (2018) extended the latent class time

model with different scales across classes to reveal

that response becomes more and more deterministic

with increase in response time. In a recent research,

Marquis (2021) utilized the response times to study

the problem of cheating in political knowledge tests

and clearly indicated response time analysis to be a

promising strategy for alleviating the problem of

cheating behaviour.

The causes for faster or slower response times can

differ for various respondents. Well informed,

opinionated individuals may respond quicker than

their counterparts (Krosnick 1989; Bassili 1993).

Alternatively, individuals with lesser motivation or

cognitive skills might rush through the survey and

hence report lower response times (Malhotra 2008).

The mode and format of survey administration, online

or offline, might also affect the response times.

Heerwegh (2002) compared the response times for

radio buttons and drop-down boxes and found that

radio buttons recorded faster response time.

Additionally, respondent involvement has also

been known to have a significant impact on

willingness-to-pay estimates (Berrens et al 2004).

Svedsater (2007) found that giving respondents time

to think decreased WTP for donations to an

environmental program among students. Macmillan

et al (2002) found that respondents who were not

given time to think had mean WTP 2–4 times greater

than those who were given significant time. Similarly,

Subade (2007) observed that WTP estimates reduced

without intervention of interviewers, who would

otherwise control the flow of the survey and thereby

increase the completion time. Cook et al (2012) argue

that time to think removes interviewer bias and helps

interviewees take well informed choices after

properly researching or consulting with their friends

and family. However, the literature lacks when it

DATA 2021 - 10th International Conference on Data Science, Technology and Applications

comes to studying the effects of response time

specifically on conjoint analysis estimates of WTP.

This motivates our study to answer how exactly

response time affects importance attached by

consumers to various attributes and WTP values.

3 DATA AND METHODOLOGY

This study uses data from a survey that investigates

the importance of various travel package attributes in

a conjoint analysis setup (Pai & Ananthakumar,

2017). The main aspects of this work are stated so that

it can be taken forward in the present study. Conjoint

analysis is a popular marketing analysis technique

used to estimate both the individual importance

values of product attributes as well as its combined

influence with other attributes on customers’ product

choices (Lewis, Ding and Geschke, 1991). To

conduct a conjoint analysis experiment, three main

steps must be followed: (1) identify the attributes, (2)

determine attribute levels, and (3) compile the

attribute profiles (Van der Pol and Ryan,1996). For

this study, a conjoint analysis experiment to explore

consumer preferences for travel packages was

conducted. Six different travel package attributes

including price, length of stay, hotel rating, season of

travel, destination and mode of transport were chosen.

Each of the six attributes was further expanded into

five levels as described in Table 1. By applying the

principles of orthogonal array design, the entire full

factorial design of all levels (5x5x5x5x5x5 i.e 15625

profiles) was reduced to 25 tour packages so that the

effects of attributes could be studied without any

interference (Green, 1974; Hair et al., 2006).

A comprehensive online questionnaire

comprising 33 questions split into two segments was

devised. The first segment included questions to

assess the socioeconomic background of the survey

participant. Details such as gender, age, occupation,

income and marital status were gathered in this

segment. In the second segment, the participants were

asked to rate the 25 tour packages on a scale of 1 to

10, with 10 representing most interested and 1

representing least interested. Instead of just providing

plain textual description, the 25 tour packages were

accompanied with pictures displaying features of the

corresponding package so that the participants would

get a better idea of the levels. A total of 168

individuals completed the survey, out of which 153

responses (15 incomplete responses removed) were

used for the analysis. Details of the respondents’

backgrounds are described in Table 2.

Table 1: Tour package attributes and levels.

ttributes

evels

Price < 20k, 20 - 35k, 35 - 50k, 50 - 75k, > 75

Len

th of sta

<3, 4, 5 - 7, 8 - 10, >11

Hotel ratin

1, 2, 3, 4, 5

Season Winter(December - February),

Spring(March - April), Summer(May -

June), Monsoon(July - September),

Autumn(Septembe

- November)

Destination Adventure & Activity, Beach, Hill

station, Herita

e & Wildlife, Pil

rima

Mode of

transport

Flight, Train, Bus, Car, Minibus

Table 2: Socioeconomic characteristics of the sample.

Characteristic %

Gender:

Male 48

Female 52

ears

)

Less than 25 27

25 - 40 18

–

60 49

60 o

above 6

Famil

Status:

Single 34

Marrie

Occupation:

Working 69

Non-working* 31

Income (in INR):

Less than 0.2 million 29

0.2 - 0.6 million 22

Greate

than 0.6 million 49

*Non-working refers to respondents who are students or

retired or unemployed

Conjoint analysis gives attribute importance and

part-worth utilities at both aggregate as well as

individual level (North, De Vos, & Kotze, 2003).

Part-worth utility values measure the consumer

preference for levels within attributes, with greater

values denoting higher consumer liking for the level.

The basic conjoint analysis model used in our study

is as follows:

𝑟=𝛽



+𝛽



𝑑



+𝛽



𝑑



+𝛽



𝑑



+𝛽



𝑑



+𝛽



𝑑



+𝛽



𝑑



+𝜀 (1)

where r denotes user rating for corresponding tour

package, d1 denotes price, d2 denotes length of stay,

d3 denotes hotel rating, d4 denotes season, d5 denotes

destination, d6 denotes mode of transport, βs the

corresponding coefficients and ε denotes the error

term. The six tour package attributes were chosen as

independent variables whereas respondent ratings

A Study on the Effects of Response Time on Travel Package Attributes

were chosen as the dependent variable. Using the

estimated beta values, relative importance of the

attribute is a measure to understand how important a

particular attribute is compared to the rest (Orme,

2010) and can be given as:

𝑅𝐼















 











∑















 















(2)

where βjkl indicates the corresponding conjoint

model weights for attribute k. Additionally, we also

estimate the marginal willingness to pay (MWTP) for

certain attribute level changes. MWTP can be viewed

as the marginal rate of substitution (MRS) between

price and non-price attributes. It provides a monetary

value to identify fluctuations in the utility associated

with product adjustments. Utility and price can be

used to estimate MWTP from level l to level h of

attribute k for an individual j as follows (Jedidi and

Jagpal, 2009):

𝑀𝑊𝑇𝑃





ℎ,𝑙



=











 





















 











 𝛽



−𝛽



 (3)

where pj denotes the levels of the price attribute and

the βj1l represents the corresponding part-worth

utilities. The minimum and maximum price levels used

for computing minimum MWTP values were 20,000

and 75,000 INR respectively. In order to compute

MWTP, classes were identified for the attributes. For

hotel ratings, the classes were - Budget hotels (1, 2 star)

and Luxurious hotels (3, 4, 5 star). Similarly, the length

of stay was divided into two categories - Short stay (4

or less), and Long stay (5 or more). Mode of transport

was divided into flight and land routes. Minimum

attribute MWTP values were found by considering

maximum difference in the change of class by using

minimum utility value from the higher class and the

maximum utility value from the lower class. It was

observed that under the given conditions, the best

representative based on the frequencies of budget hotel

to luxurious hotel was 2 star to 3 star, and for short stay

to long stay was less than 3 days to greater than 11

days. Trains were selected for representing land-based

mode of transport due to their maximum utility value.

Yan and Tourangeau (2008) found multiple item

specific features involving response time of survey

such as question length, toughness of the question and

position of the question within the survey. Total

completion time averages out the effects of such

various factors pertaining to each question and hence

in our study, the total survey completion time (in

minutes) was recorded for response time analysis.

Descriptive statistics for completion time are presented

in Table 3.

To analyze the survey response time, the

respondents are divided into two groups.

Respondents with response time less than 7 minutes

are classified as fast respondents whereas those with

response time greater than or equal to 7 minutes are

classified as slow respondents. The chosen time

boundary value is the greatest integer completion

time less than the median completion time. In order

to identify differences between the groups,

multivariate analysis of variance (MANOVA) test is

carried out on the attribute importance values derived

from the conjoint analysis. This is followed by

analysis of variance (ANOVA) on each of the six

attribute importance to recognize where the

difference stems from within the groups. Similarly,

the three MWTP values are also tested for differences

by using MANOVA, followed by ANOVA. We also

analyze the groups to determine which group

dominates for the individual willingness to pay.

Following this, we conduct a thorough profiling of the

socioeconomic data gathered from the survey and

attempt to figure out the slow and fast groups based

on their demographic features. Subsequent to this,

appropriate statistical tests are carried out to confirm

if the groups are different based on various

demographic characteristics.

Table 3: Descriptive Statistics.

Measure

Completion time (mins)

Mean 7.57

Median 7.27

Standar

deviation 3.13

Maximu

14.67

Minimu

1.77

ercentile 5.8

ercentile 8.45

4 RESULTS AND DISCUSSION

The following subsections contain a detailed

discussion of the most noteworthy findings of our

study.

4.1 Differences based on the Response

Time

The relative importance values estimated by the

conjoint analysis technique can be used to get an idea

of the desirability of each of the travel package

attributes. Figure 1 shows the aggregate average

relative importance values for all the attributes. The

attributes in descending order of importance are hotel

DATA 2021 - 10th International Conference on Data Science, Technology and Applications

rating (21.57%), price (19.58%), length of stay

(16.63%), mode of transport (14.98%), destination

(13.99%) and season (13.25%). Tables 4 and 5

respectively describe the MANOVA and ANOVA

results for the conjoint importance values for the two

groups. MANOVA reveals that there are significant

differences between the importance values for slow

and fast respondents (p < 0.05). ANOVA results show

that among the attributes, importance for length of

stay, hotel ratings, destination and mode of transport

differed between the two groups. Price and season are

found to have more or less the same importance for

all the respondents. This could be because, price of a

tour package is a factor which is somewhat pre-

decided by individuals depending upon their budget.

Likewise, season is another factor which is given

similar importance by all consumers probably due to

high preference for a pleasant travel weather which

matches their work schedule.

The group means for the four significantly

different attributes are compared to each other. It is

found that mean importance for duration (Mfast =

15.32, Mslow = 17.74) and hotel rating (Mfast =

19.31, Mslow = 23.46) are greater for slow

respondents compared to fast respondents. On the

other hand, mean importance for destination (Mfast =

15.39, Mslow = 12.79) and mode of transport (Mfast

= 14.09, Mslow = 12.54) are found to be higher for

fast respondents. Duration and hotel ratings are found

to have higher importance values in the aggregate

results whereas destination and mode of transport

feature in the bottom two attributes. It is also found

that the mean importance values for all the four

attributes are somewhat similar for fast respondents

whereas slow ones specifically show more

importance towards overall importance attributes i.e

duration and hotel ratings. This suggests that slow

respondents actually gave a more careful thought

while deciding the importance attributes and not so

Figure 1: Average importance of the attributes in

percentage in increasing order.

Figure 2: Average importance values in percentage for fast

and slow groups.

much for the relatively unimportant attributes. Figure

2 shows the comparisons between importance values

for the fast and slow respondents.

Table 4: MANOVA results for conjoint importance values.

Statistic Value F Num Df Den Df Pr (>F)

Wilk’s

Lambda

0.853 4.208 3 146 0.00062

Table 5: ANOVA results for conjoint importance values.

Attribute

used for the

test

Source Mean

square

F value Pr (>F)

Price Model

Residual

121.64

56.76

2.143 0.1453

Duration Model

Residual

350.86

52.90

6.632 0.0109

Hotel rating Model

Residual

620.99

74.04

8.839 0.0434

Destination Model

Residual

310.67

60.726

5.116 0.0251

Season Model

Residual

62.3

33.258

1.873 0.1731

Mode of

trans

ort

Model

Residual

850.20

47.79

17.79 4.2e-05

When it comes to marginal willingness to pay

(MWTP), MANOVA results suggest that there are

differences between the groups (p < 0.05). ANOVA

reveals that the source of these differences is the

MWTP for transport (train to flight) and duration

(short to long stay) to some extent (p < 0.10). The

remaining MWTP value i.e budget to luxurious hotel

is found to be consistent across the groups. Thus,

hotel rating, which was relatively an important

attribute, show similar willingness to pay in spite of

differences in importance. It is also found that MWTP

for train to flight is considerably larger for fast

respondents (Mfast = 18983 INR) as opposed to slow

ones (Mslow = 5448 INR). Similarly, MWTP for

A Study on the Effects of Response Time on Travel Package Attributes

short to long stay is higher for fast respondents (Mfast

= 8073 INR) as compared to their slow counterparts

who have negative MWTP values (Mslow = -4355

INR). This means that for a travel agent willing to

make economic benefits from tweaking the travel

package attributes, the best option is to change mode

of transport and duration rather than any other feature.

The huge difference in the MWTP values reveals that

fast respondents have a strong opinion both in terms

of importance as well as spending budget for

transport and duration. Tables 6 and 7 respectively

describe the MANOVA and ANOVA results and

Figure 3 depicts the MWTP values.

Table 6: MANOVA results for willingness to pay.

Statistic Value F

Num

Den

(

)

Wilk’s

Lambda

0.929 3.74 3 149 0.0125

Table 7: ANOVA results for willingness to pay.

Attribute

used for the

test

Source

Mean

square

F value Pr (>F)

Budget to

luxurious

hotel

Model

Residual

1.66e+09

1.03e+09

1.603 0.207

Short to

long sta

Model

Residual

5.86e+09

2.07e+09

2.83 0.0946

Train to

fli

Model

Residual

6.96e+09

1.60e+09

4.343 0.0388

Figure 3: Average MWTP values in INR for fast and slow

groups.

4.2 Demographic Analysis of the

Respondents

As suggested by past studies, it is found that slow and

fast respondents indeed show different preference

structures. Importance for length of stay and hotel

rating are observed to be high amongst slow

respondents. These two attributes have considerable

share of importance in the overall travel package

combination. Fast respondents show higher affinity for

destination and mode of transport, both of which had

relatively low contribution to the travel package. This

demonstrates that slow respondents give more careful

thought in relatively important variables. However,

fast respondents are found to be more strongly

opinionated about their choices, which can be seen in

the high value of their MWTP for both transport and

duration. These findings align well with previously

conducted studies (Bassili 1993, Malhotra 2008).

Since we have established that response time does

have effects on consumer preferences, it is of interest

to study the characteristics of a customer and thereby

find whether a particular customer would take less or

more time. Accordingly, we decide to analyse the

demographic characteristics of the respondents in fast

and slow groups. The demographic details of the fast

and slow groups are described in Table 8.

In terms of demographics, respondents in the slow

group (n = 83) are more than that of fast group (n =

70). The slow group has more middle-aged adults

over the age of 40 and senior citizens. This group also

consists mainly married, working individuals within

the higher income group (greater than 0.6 million

INR). The fast group has more of younger population

of the age group below 25 years. This group consists

Table 8: Demographic profiles of the respondents.

Demographic attribute Group 1

(slow)

Group 2

(fast)

Total

Gender Male 47 32 79

Female 36 38 74

Age Less than 25 11 30 41

25 - 40 11 16 27

–

60 56 20 76

60 o

above 5 4 9

Family

Status

Marrie

67 34 101

Single 16 36 52

Occupation Working 62 44 106

Non-working 21 26 47

Income < 0.2 m 21 24 45

0.2 - 0.6 m 18 15 33

> 0.6 m 44 31 75

Average time taken (mins) 9.87 4.85

Table 9: Chi-square test results.

Attribute Chi-square df p value

Gende

2.7773 1 0.2494

e 22.081 1 2.613e-06

Famil

Status 16.093 1 6.032e-05

Occu

ation 1.9765 1 0.1598

Income 0.8342 1 0.361

DATA 2021 - 10th International Conference on Data Science, Technology and Applications

of mostly unmarried individuals. This profiling

indicates that the response time of an older customer

who is married is higher than a younger, single

person. It is interesting to note that the two groups did

not show any significant difference with respect to

gender, occupation and income. On inspection of the

average time to respond for the two groups, the slow

respondents clocked 9.87 minutes while fast

respondents took approximately half the time.

In order to confirm the differences between the

groups in terms of different demographic

characteristics, hypothesis testing was carried out for

each of the attributes. For comparing gender, we

chose the null hypothesis (H0) that the percentage of

male members is same for both the groups. For

comparing age, the null hypothesis was the

percentage of people above 40 years is the same for

both the groups. For comparing family status, the null

hypothesis was the percentage of married people is

the same for both the groups. For comparing

occupation, the null hypothesis was the ratio of

working people in both the groups was the same.

Finally, for income, the null hypothesis was the

fraction of people with income higher than 0.6 million

was the same in both the groups. Results of the chi-

square test for these hypotheses are shown in Table 9.

Based on the p value, we can reject the null

hypotheses for age and family status. This confirms

our claims that the two groups have differences

mainly in case of age and family status.

5 CONCLUSIONS

The focus of this study was to comprehend the effects

of response time on preferences of Indian travellers

for different tour package attributes. The results of the

conducted study have direct implications on key

purchase decisions as it aids travel agents in

understanding the consumer preferences. The results

from this research can be applied in tourism

management to create effective marketing strategies

that attract potential clients.

Based on the preliminary analysis, price, hotel

rating and duration were found to be relatively crucial

attributes whereas destination, season and mode of

transport were less important. This is consistent with

previously conducted studies on tour package

attributes (Jacoby and Olson, 1977; Chan and Wong,

2006). The classification of respondents into fast and

slow groups also provided key insights into the

consumer decision making process. Price is more or

less determined by the spending power of the

individual and not so much by the time taken.

Similarly, season of travel is something for which

consumers have a clear preference. Apart from these

two attributes, other four had significant differences

depending on the response time of respondents. Slow

respondents demonstrated a more careful attitude by

showing more importance for overall important

attributes such as duration and hotel ratings and less

for others. On the other hand, fast respondents

showed similar importance for all the attributes and

hence, had relatively more importance for mode of

transport and destination as compared to slower

individuals. This suggests that slow respondents are

more careful in their choices and spend more effort in

determining the overall important attributes as

compared to fast ones.

When it comes to willingness to pay, both the

groups showed similarities for MWTP for budget to

luxurious hotels. Though hotel ratings attributes were

more important for slow respondents, no significant

differences within MWTP were observed. This

means that although slow respondents identify these

as important attributes, they are not willing to shell

out more money. Alternatively, fast respondents were

found to have significantly larger MWTP for

transport (train to flight) and duration (short to long

stay). Transport was initially found to have more

importance for fast individuals and they also showed

larger economic preference towards the same. This

can be interpreted as faster individuals having strong

opinions about their preferences. On the other hand,

in case of duration, even though slow respondents

showed higher importance, they were not willing to

pay as much as the fast ones. Travel agents can use

this information to change transport factors in their

travel packages so as to maximize their economic

profits. Future studies can use these results and

attempt to identify random outliers such as older,

married people with faster responses, so as to

effectively analyze survey findings.

The demographic profiling of our slow and fast

respondents revealed that older, married, working

individuals tend to take higher time. This can be

attributed to their eagerness to understand more about

the products which they plan to purchase. On the

other hand, the younger, single, non-working

population showed a significantly lower time to

respond. This brings into light the effects of advances

in technology and social media that has created a

more informed youth population having lower

attention span. The way of handling response time in

our study can also be used in case of online marketing

wherein travel websites can track user behavior to

classify the respondent as fast or slow and thereby

provide suitable tour packages.

A Study on the Effects of Response Time on Travel Package Attributes

Few recent developments with regard to attributes

are worth considering for future studies. Gonzalez

(2019) focuses on the aspect of variation in attribute

importance measures and highlights the need for

reporting how these measures are obtained for better

comparison and interpretation. Though previous

literature is used to select the attributes in our present

study, Webb etal. (2021) has used best-worst scaling

survey to come up with certain criteria to guide

attribute selection for discrete choice experiments.

There are some shortcomings to this research.

Convenience sampling method was deployed to

collect the data because of limited availability of

manpower and resources. Therefore, the results

obtained in the study may not represent the entire

tourism sector. Orthogonal array design method was

applied in converting the full factorial design into a

smaller one that was finally used for the survey. There

could be some initial biases in the respondents due to

the order of survey questions, which could have been

reduced by randomizing the order of questions.

Lastly, this study was conducted for the Indian

tourism market and hence, all results might not be

directly extended to other regions. However, there is

good scope for future research broadening the

findings from this study, which can improve the

quality of tour packages provided and thereby the

global tourism industry.

REFERENCES

Bassili, J.N. (1993). Procedural efficiency and the

spontaneity of trait inference. Personality and Social

Psychology Bulletin, 19(2): 200-205.

Baxter, B.W., Hinson, R.E. (2001). Is smoking automatic?

Demands of smoking behavior on attentional resources.

Journal of Abnormal Psychology, 110(1): 59-66.

Benfield, J.A., Szlemko, W.J. (2006). Internet-based data

collection: Promises and realities. Journal of Research

Practice, 2(2), Retrieved from http://jrp.icaap.org/

index.php/jrp/article/viewFile/30/122.

Berrens, R.P., List, J.A., Bohara, A.K., Joe Kerkvliet, J.

(2004). Examining the role of social isolation on stated

preferences. American Economic Review, 94(3): 741-

752.

Breidert, C., Hahsler, M., Reutterer, T. (2006). A review of

methods for measuring willingness-to-pay. Innovative

Marketing, 2(4): 8-32.

Brown, S.D., Heathcote, A. (2008). The simplest complete

model of choice response time: Linear ballistic

accumulation. Cognitive psychology, 57(3): 153-178.

Bonsall, P., Lythgoe, B. (2009). Factors affecting the

amount of effort expended in responding to questions

in behavioural choice experiments. Journal of Choice

Modelling, 2(2): 216-236

Campbell, D., Mørkbak, M.R., Olsen, S.B. (2018). The link

between response time and preference, variance and

processing heterogeneity in stated choice experiments.

Journal of Environmental Economics and

Management, 88: 18-34.

Chan, E.S.W., Wong, S.C.K. (2006). Hotel selection: when

price is not the issue. Journal of Vacation Marketing,

12 (2): 142–159.

Cook, D.A., Thompson, W.G., Thomas, K.G. (2011). The

motivated strategies for learning questionnaire: score

validity among medicine residents. Medical education,

45(12): 1230-1240.

Cook, J., Jeuland ,M,, Maskery, B., Whittington, D. (2012).

Giving stated preference respondents “time to think”:

results from four countries. Environmental and

Resource Economics, 51(4): 473-496.

Cook, J., Whittington, D., Canh, D.G., Johnson, F.R.,

Nyamete, A. (2007). Reliability of stated preferences

for cholera and typhoid vaccines with time to think in

Hue, Vietnam. Economic Inquiry, 45(1): 100-114.

Couper, M.P. (2000). Usability evaluation of computer-

assisted survey instruments. Social Science Computer

Review, 18(4): 384-396.

Gensler, S., Hinz, O., Skiera, B., Theysohn, S. (2012).

Willingness-to-pay estimation with choice-based

conjoint analysis: Addressing extreme response

behavior with individually adapted designs. European

Journal of Operational Research, 219(2): 368-378.

Gibbons, H., Rammsayer, T.H. (1999). Differential effects

of personality traits related to the P-ImpUSS dimension

on latent inhibition in healthy female subjects.

Personality and Individual Differences, 27(6): 1157-

1166.

Gonzalez, J.M. (2019). A Guide to Measuring and

Interpreting Attribute Importance. Patient 12: 287–295

Green, P.E. (1974). On the design of experiments involving

multifactor alternatives. Journal of Consumer

Research, 1: 61-68.

Green, P.E., Wind, Y. (1975). New way to measure

consumers’ judgments. Harvard Business Review,

53(4): 107-117.

Haaijer, R., Wagner, K., Wedel, M. (2000). Response

latencies in the analysis of conjoint choice experiments.

Journal of Marketing Research, 37(3): 376-382.

Hair, J., Black, W., Babin, B., Anderson, R.E. (2010).

Multivariate Data Analysis, 7th ed. Pearson Prentice

Hall, Upper Saddle River, NJ.

Heerwegh, D., Loosveldt, G. (2002). An evaluation of the

effect of response formats on data quality in web

surveys. Social science computer review, 20(4): 471-

484.

Hess, S., Stathopoulos, A. (2013). Linking response quality

to survey engagement: a combined random scale and

latent variable approach. Journal of choice modelling,

7: 1-12.

Hobbs, J.E. (1996). A transaction cost analysis of quality,

traceability and animal welfare issues in UK beef

retailing. British Food Journal, 98(6): 16-26.

Höhne, J.K., Schlosser, S. (2018). Investigating the

adequacy of response time outlier definitions in

DATA 2021 - 10th International Conference on Data Science, Technology and Applications

computer-based web surveys using paradata

SurveyFocus. Social Science Computer Review, 36(3):

369-378.

Holmes, T., Alger, K., Zinkhan, C, Mercer, E. (1998). The

effect of response time on conjoint analysis estimates of

rainforest protection values. Journal of Forestry

Economics, 4(1): 7-28.

Hertel, G., Neuhof, J., Theuer T., Kerr, N.L. (2000). Mood

effects on cooperation in small groups: Does positive

mood simply lead to more cooperation? Cognition &

emotion, 14(4): 441-472.

Jacoby, J., Olson, J.C. (1977). Consumer Response to Price:

an Attitudinal, Information Processing Perspective. In

Y, Wind & M, Greenberg (Ed.). Moving Ahead with

Attitude Research (pp. 73-86). Chicago: American

Marketing Association.

Krosnick, J.A., Alwin, D.F. (1989). Aging and

susceptibility to attitude change. Journal of personality

and social psychology, 57(3): 416-425.

Lewis, R., Ding, S., Geschke, U. (1991). Conjoint analysis

and its application in the hospitality industry. Retrieved

from https://scholar.lib.vt.edu/ejournals/JIAHR/issue2

Macmillan, D.C., Philip, L., Hanley, N., Alvarez- Farizo,

V. (2002). Valuing the non-market benefits of wild

goose conservation: a comparison of interview and

group-based approaches, Ecological Economics, 43:

49-59.

Malhotra, N. (2008). Completion time and response order

effects in web surveys. Public opinion quarterly, 72(5):

914-934.

Marquis, L. (2021) Using response times to enhance the

reliability of political knowledge items: An application

to the 2015 swiss post-election survey, Survey Research

Methods, 15: 79-100.

Mayerl, J. (2013). Response latency measurement in

surveys: Detecting strong attitudes and response

effects. Survey Methods: Insights from the Field,

Retrieved from https://surveyinsights.org/?p=1063.

Mayerl, J., Andersen, H., Giehl, C. (2019). Identification of

Measurement Problems of Survey Items and Scales

Using Paradata. In N, Menold & T, Wolbring (Ed.).

Qualitätssicherung sozialwissenschaftlicher

Erhebungsinstrumente, (pp. 9-35). Wiesbaden:

Springer VS.

North, E.J., De Vos R.B., Kotze, T. (2003). The importance

of apparel product attributes for female buyers. Journal

of Consumer Sciences, 31(1): 41-51.

Orme, B.K. (2010). Getting Started with Conjoint Analysis:

Strategies for Product and Pricing Research. Research

Publishers LLC.

Pai, S., Ananthakumar, U. (2017). Understanding tourist

preferences for travel packages: a conjoint analysis

approach. Asia Pacific Journal of Tourism Research,

22(12): 1238-1249.

Palumbo, A., De La Fuente, P., Rodríguez, M., Sanchez, F.,

Martinez-Salazar, J., Munoz, M., Marqueta, J.,

Hernandez, J., Espallardo, O., Polanco, C., Paz, S.,

Lizan,L. (2011). Willingness to pay and conjoint

analysis to determine women's preferences for ovarian

stimulating hormones in the treatment of infertility in

Spain. Human reproduction, 26(7): 1790-1798.

Phillips, K., Johnson, R., Maddala, T. (2002). Measuring

what people value: a comparison of attitude and

preference surveys. Health Services Research Journal,

37(6): 1659-1679.

Rose, J.M., Black, I.R. (2006). Means matter, but variance

matter too: Decomposing response latency influences

on variance heterogeneity in stated preference

experiments. Marketing Letters, 17(4): 295-310.

Ryan, M. (1999). Using conjoint analysis to take account of

patient preferences and go beyond health outcomes: an

application to in vitro fertilization. Social Science &

Medicine, 49: 535-546.

Sheppard, L.C., Teasdale, J.D. (2000). Dysfunctional

thinking in major depressive disorder: A deficit in

metacognitive monitoring? Journal of Abnormal

Psychology, 109(4): 768-776.

Skowronski, J.J., Carlston, D.E. (1987). Social judgment

and social memory: The role of cue diagnosticity in

negativity, positivity, and extremity biases. Journal of

personality and social psychology, 52(4): 689-699.

Svedsater, H. (2007). Ambivalent statements in contingent

valuation studies: inclusive response formats and

giving respondents time to think. Australian journal of

agricultural and resource economics, 51(1): 91-107.

Tourangeau, R., Rips, L.J., Rasinski, K. (2000). The

psychology of survey response. Cambridge University

Press.

Van Der Pol, M., Ryan, M. (1996). Using conjoint analysis

to establish consumer preferences for fruit and

vegetables. British Food Journal, 98(8): 5-12.

Van Selm M, Jankowski NW (2006) Conducting online

surveys. Quality and Quantity, 40(3): 435-456.

Webb, E.J.D., Meads, D., Lynch, Y., Judge, S., Randall, N.,

Goldbart, J., Meredith, S., Moulam, L., Hess, S.,

Murray, J. (2021). Attribute selection for a discrete

Choice experiment incorporating a best-worst scaling

survey. Value Health, 24(4): 575-584.

Yan, T., Tourangeau, R. (2008). Fast times and easy

questions: The effects of age, experience and question

complexity on web survey response times. Applied

Cognitive Psychology: The Official Journal of the

Society for Applied Research in Memory and

Cognition, 22(1): 51-68.

A Study on the Effects of Response Time on Travel Package Attributes