Influencer Credibility and Purchase Intention among College Students in Bengaluru (Survey Project, SPSS)

Influencer Credibility and Purchase Intention among College Students in Bengaluru is a Sem-5 BCU Survey Project (BBA 5.7) that tests which of an influencer's perceived traits — trustworthiness, expertise or attractiveness — actually moves students from watching a reel to buying what is in it.

About 250 undergraduate students from four Bengaluru degree colleges respond to a structured questionnaire built around Ohanian's classic tri-dimensional source-credibility scale (1990), adapted to 5-point Likert items and anchored to the influencer each respondent follows most closely in their self-chosen category — fashion, beauty, food or gadgets. The questionnaire also captures influencer tier (nano, micro, macro, mega), platform mix, disclosure awareness (whether respondents notice and understand #ad and 'paid partnership' labels, which ASCI's influencer guidelines make mandatory), and a purchase-intention battery.

The analysis is deliberately deeper than a typical BBA survey: after reliability checks (Cronbach's alpha in R and SPSS), the study runs chi-square (gender × following behaviour), independent-samples t-tests (micro vs mega followers on trust), one-way ANOVA (credibility across tiers), Pearson correlations of each credibility dimension with purchase intention, and a multiple regression of purchase intention on the three dimensions plus disclosure awareness and category.

The deliverable is a credibility-by-tier dashboard in Tableau and a set of evidence-based takeaways for the small D2C brands that dominate Bengaluru's campus marketing: which trait to screen for when shortlisting creators, and whether disclosure labels scare students away or signal honesty. Every statistic in the report can be rerun from the raw data file.

Bengaluru's D2C brands and campus-facing businesses spend a growing share of their marketing budgets on influencers, but creator selection is still driven by follower counts and engagement screenshots. What these metrics hide is credibility — whether students actually believe the creator's claim about a serum, a cafe or a gadget. When a creator students distrust promotes a product, money buys exposure but not intention; when a small nano creator with 8,000 followers recommends the same product to a tight community, the effect may be far larger per rupee spent.

At the same time, ASCI's influencer guidelines mandate disclosure labels on paid promotions, and brands quietly worry that visible '#ad' tags erode persuasion. Whether disclosure awareness reduces or increases trust among Indian students is exactly the kind of question local evidence can answer — yet most available studies are global, pre-disclosure-era, or not student-specific.

This project therefore measures, among Bengaluru undergraduates: how credible they rate the influencers they actually follow (trustworthiness, expertise, attractiveness), how credibility varies by creator tier and product category, whether disclosure awareness changes attitudes, and which combination of these factors best predicts purchase intention — giving campus-facing brands a defensible, statistics-backed basis for choosing creators.

Abstract

This survey project measures the credibility of social-media influencers as perceived by Bengaluru undergraduates and tests which credibility dimensions drive purchase intention. Using an adapted Ohanian (1990) scale, a disclosure-awareness block based on ASCI's influencer guidelines, and SPSS-based hypothesis testing on ~250 responses, the study compares credibility across influencer tiers and quantifies each trait's contribution to intention to buy.

Introduction

Influencer marketing is the default entry point for brands targeting students: college-going consumers discover cafes, skincare, fashion and gadgets through creators long before they see a brand's own advertisement. Bengaluru, with one of India's largest student populations and a dense D2C startup scene, is a natural laboratory. Brands, however, still select creators mostly by follower count — a metric that says nothing about whether students believe the person.

Source-credibility research offers a measurable alternative. Ohanian (1990) validated a 15-item scale capturing three dimensions — trustworthiness, expertise and attractiveness — that remains the standard starting point for endorser studies. India adds a regulatory layer: ASCI's Guidelines for Influencer Advertising on Digital Media require paid promotions to carry clear disclosure labels (#ad, #collab, 'paid partnership'), so credibility now coexists with disclosure cues students actually see.

Existing system / literature gap

• Global studies using Ohanian's scale are abundant, but tier-aware comparisons (nano/micro vs mega creators) and disclosure-awareness effects in the Indian student segment are rarely combined in one undergraduate-accessible design.
• College marketing practice in Bengaluru still equates reach with influence; there is little local, replicable evidence on which credibility dimension matters per product category.
• Student projects in this space often stop at percentages; this design adds tier comparisons, correlations and regression so findings survive examiner scrutiny.

The proposed study

A cross-sectional survey of ~250 undergraduates across four colleges, quota-sampled by gender and year of study. Each respondent anchors all scale items to one influencer they follow closely in a chosen category. Measured: the three credibility dimensions (5 items each), purchase intention (4 items), influencer tier, platform mix, and disclosure awareness (2 items). Hypotheses: H1 credibility means differ across tiers (ANOVA); H2 each dimension correlates positively with purchase intention (Pearson); H3 micro-vs-mega followers differ on trustworthiness (t-test); H4 purchase intention is predicted by trustworthiness, expertise, attractiveness and disclosure awareness (multiple regression). Outputs: a tier dashboard in Tableau and category-wise brand takeaways.

Feasibility

• Technical: the scale is published and adaptable; SPSS and Excel are curriculum tools; R (BBA 4.3) computes alphas; Tableau Public is free.
• Economic: QR-code distribution on campuses keeps cost under ₹500.
• Operational: 8–10 minutes per response; no personal identifiers beyond age band, gender and college code; informed consent is built into the form's first screen.

In scope

• Undergraduate students (18–24) of four Bengaluru degree colleges, quota-sampled by gender and year of study (~250 responses).
• Influencers followed on Instagram or YouTube in four categories: fashion, beauty, food/cafes and gadgets.
• Constructs: Ohanian's three credibility dimensions (adapted, 5 items each), purchase intention, influencer tier, platform mix and disclosure awareness of #ad / paid-partnership labels.
• Statistical analysis in Excel, SPSS and R of one survey wave: reliability, chi-square, t-test, ANOVA, correlation and multiple regression.

Out of scope

• Actual purchase tracking or brand sales data — intention, not behaviour, is measured.
• Celebrity endorsers who are not social-media creators, and athlete/film-star endorsements.
• Brand-specific experiments or any identification of individual influencers or respondents.
• Causal claims: the design is cross-sectional and reports association only.
• Collection of names, phone numbers or social-media handles of respondents.

Research design: descriptive, cross-sectional survey with hypothesis testing, using a stimulus-anchored questionnaire: every scale item refers to 'the influencer you follow most closely' in the respondent's chosen category, which keeps ratings concrete instead of hypothetical.

Sampling: quota sampling across four colleges (two each from north and south Bengaluru), with quotas on gender (≈ 50/50) and year of study; within quotas, respondents are recruited through a QR poster at canteens and common rooms plus paper forms for those who prefer offline. Target n ≈ 250. First screen of the form states purpose, voluntary nature and anonymity, and consent is required to proceed.

Instrument: Section A respondent profile (age band, gender, year, college code, category chosen, platforms used, hours on the platform); Section B anchor-influencer profile (tier by follower count band: nano <10k, micro 10k–100k, macro 100k–1M, mega >1M; content style; how long followed); Section C credibility: 15 items adapted from Ohanian (1990) — trustworthy/untrustworthy, reliable, sincere, expert, knowledgeable, skilled, attractive, classy, elegant and similar pairs on 5-point Likert; Section D purchase intention (4 items: seek more information, visit the product page, buy if budget permits, recommend to friends); Section E disclosure awareness (notices #ad / paid-partnership labels; understands what they mean) plus one open-ended 'what makes you unfollow an influencer?'.

Pilot: 20 students from a non-sample college; Cronbach's alpha computed in R for each 5-item dimension (target ≥ 0.7); item wordings with high deletion-impact revised.

Ethics: informed consent on the form's first screen; anonymity; no influencer or respondent names recorded; open-ended answers paraphrased in the report; aggregate-only reporting.

The study is a single pipeline from theory to brand takeaways: Ohanian's scale supplies the construct definitions, the questionnaire anchors them to each respondent's own followed influencer, and the analysis stack separates scale validation (R) from hypothesis testing (SPSS) and communication (Tableau).

flowchart TD
  A["Ohanian 1990 scale + ASCI disclosure guidelines"] --> B["Conceptual model and 4 hypotheses"]
  B --> C["Questionnaire: 15 credibility items, 4 intention items, disclosure block"]
  C --> D["Pilot 20 students; Cronbach alpha >= 0.7 in R"]
  D --> E["Quota sample ~250 students across 4 colleges"]
  E --> F["Excel: coding, tier bands, composite scores"]
  F --> G["SPSS: chi-square, t-test, ANOVA, correlation"]
  G --> H["Multiple regression: purchase intention"]
  H --> I["Tableau: credibility-by-tier dashboard"]
  I --> J["Category-wise creator-selection recommendations"]

Conceptual model

flowchart LR
  T["Trustworthiness (5 items)"] --> P["Purchase intention"]
  E["Expertise (5 items)"] --> P
  A["Attractiveness (5 items)"] --> P
  D["Disclosure awareness (#ad)"] --> P
  TIER["Influencer tier: nano to mega"] --> T
  TIER --> E
  TIER --> A
  CAT["Product category"] --> P

Design notes. Tier is banded by follower count so respondents classify their anchor influencer without handling exact numbers. Composite scores are means of their five items after reliability checks; the regression is examined for multicollinearity (VIF < 5) and residual normality before any claim is written. Because the anchor differs by respondent, category and tier enter the model as controls, and this design choice is stated openly in the limitations.

• Category deep-dive: repeat per category with larger samples to test whether attractiveness matters more in beauty than in gadgets, as the literature hints.
• Experimental upgrade: show the same testimonial with and without an #ad label to two randomised groups and measure the disclosure effect causally.
• Behaviour link: a follow-up wave eight weeks later checking whether stated intention became actual purchases, with receipts self-reported in bands.
• Regional language layer: a Kannada-language form for tier-2 colleges around Bengaluru to widen the sample beyond English-dominant campuses.
• Creator-side view: short interviews with five local nano/micro creators on how they choose brands, pairing the audience-side model with supply-side reality.