Indian Politics & Policy Volume 3, Number 1, Spring 2020 | Page 108

Indian Politics & Policy
tion with candidates from other parties.
To create the dependent variable at first,
we recoded PCs into two categories: 1
as constituencies where BJP candidates
ran against other candidates and 0 as
constituencies where candidates from
BJP allies or independent candidates
supported by the BJP ran against other
candidates. Then we selected a sample
of young voters and constituencies
where BJP candidates ran and used the
vote choice variable as the dependent
variable, where a vote for the BJP was
labeled as 1 and a vote for other parties
was labeled as 0. It was a binary variable
having two categories; therefore, a logit
regression model was used. The total
sample used in the regression analysis
was 2641 (refer to Table 15 for logic of
the sample selection).
Table 15: Sample Selection for Regression Model
Sample size
NES 2019 total sample 24235
PCs BJP candidates running against other candidates 19866
Young voter sample in our survey 3188
Young voter sample in selected PCs where BJP candidates ran against
other candidates
2641
The independent variables included
in the model were assessment
of the incumbent government’s performance,
closeness to a party, leadership
choice, election issues such as unemployment,
price increases, opinion on
demonetization, and awareness of the
Pulwama attack. These factors were also
controlled by socioeconomic factors
such as caste-communities, level of educational
attainment, locality, economic
class, and exposure to both social and
news media.
The model was statistically significant
as the goodness of fit test was
significant (the Hosmer and Lemeshow
Test sig value was .510). Socioeconomic
factors, such as economic class, locality,
level of educational attainment, level
of social and news media exposure, and
caste-communities that young voters
belong to, did not have an impact on
youth’s voting for the BJP, as these factors
were not found to be statistically
significant. On the other hand, attitudinal
variables, such as satisfaction with
the incumbent government’s performance,
closeness to a party, leadership
choice (where respondents were asked
to name a person they wanted to see as
the next PM of India in an open-ended
structure), and support for demonetization,
were found to be statistically significant.
Other election issues, such as
unemployment, price increases, and the
Pulwama attack, were not significant in
the presence of other factors. It was Modi’s
leadership as PM of the country that
emerged as the most significant factor
with a sig value .000; however, partisanship
(those who like the BJP as a party
over other parties) was also significant
(p value was .016), but not to the extent
104